From SCHILD at aiub.unibe.ch Fri Dec 22 16:36:25 1995 From: SCHILD at aiub.unibe.ch (Schildknecht Thomas) Date: Thu Jul 29 11:54:07 2004 Subject: Thomson THX7897M Message-ID: <0099B40B.B7A6C9F6.26@ubeclu.unibe.ch> Ciao, Does anyone have some experience with the 2kx2k four strip THX 7897M CCD? We are currently purchasing a Camera from Astrocam using this chip. As this was the first time they integrated this CCD some test with an engeneering grade version were performed. The results were not at all promizing... We need a 2kx2k device with 4 outputs because the readout time is critical for our application (astrometry of artificial satellites). High readout rates (with of course increased noise) for certain applications will be used. The normal operation is slow scan however. Backside illumination is not mandatory. Would there be an alternative to this Tompson device (apart from the SIte device; e.g. from EEV?)? Thanks Thomas **************************************************************************** Thomas Schildknecht Internet: SCHILD@AIUB.UNIBE.CH Astronomical Institute SCHILD@130.92.4.11 University of Berne HEPNET/SPAN: 20579::49202::SCHILD Sidlerstrasse 5 Phone: +41-31-6318594 (91) CH-3012 Bern, Switzerland Fax: +41-31-6313869 **************************************************************************** From prj at mail.ast.cam.ac.uk Fri Dec 22 09:34:25 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:07 2004 Subject: info and idea request for PI CCDs Message-ID: Greetings Qian Song & all, > I bought a LN/CCD 1752PB from Princeton Instruments which is a thinned, > back illuminated and pixel size 15um, 1752X532 spectrometric detector. I don't > know why its full well capacity is only 5900 e and readout noise 8.6 e. PI > response me it's the correct CCD. I am curioused if anyone got a better one > from PI for the ad. I saw before I bought says the FWC not less than 260,000 e > and readout noise 4-6 e (only it's an ad. titled with 'preliminary'). > This may not be very helpful, but it seems to me you should complain or return the detector to PI. The electronics may not be set up properly with the CCD. If it is an MPP chip the full well would be less than 260K but more than 6K! Try contacting John Geary (John any comments?)- he has tested these chips, and reported results at the recent ICSOI conference at Grand Cayman. > I got trouble too when I use LN/CCD 1024TKB also from PI. When > it works under -100C, water vapor comes up on the quartz window ahead of CCD > chip and behind of the shutter (the vacuum condition of the space between the > chip and the window is good) as the exposure time reaches several tens of > seconds (the enviomental humidity about 60% and temperature 10C to 15C). The > lower the working temperature, the quiker vapor comes up. when I uninstalled > the shutter (there's another quartz window in front of shutter) with the outer > window together, the vapor no longer came. I cannot do without the shutter in > my future work. So, I need your help if you can give some ideas to solve this > problem. > This sounds like 'normal' dewar window condensation. The cold detector can cause radiative (& maybe some conductive) cooling of the window. The usual tricks are to install some heater wire or resistors in a ring around the window or to blow dry air across its surface. Good Luck, Paul Jorden ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From prj at mail.ast.cam.ac.uk Mon Dec 18 09:43:22 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:07 2004 Subject: zero force insertion sockets Message-ID: Hi Andre & others, > I remember remotely a discussion concerning the "zero force insertion > sockets". Some people was against using them with cryogenic CCD's, but I > do not remember what were the arguments. Any comment (pros and cons) ? > I have not looked up a previous discussion, but I would say we have used them with no real problems. We originally used various forms of simple 'dil' sockets or custom made 'sil' strips for early small CCDs. However, the larger CCDs, with lots of pins tend to be hard to remove from multiple sockets (especially the turned-pin sort). We started using ZIF sockets for EEV-05 series, & then Tek1024 chips. The advantage is that they allow very easy removal & insertion. They need some degree of machining for our central cooling block of course. The disadvantage is that they are quite chunky- and add to the mass of material to be cooled; this need not matter too much. Now we are using the butted style of chips (Lesser 2k3ebs, EEV CCD42) and these come with a connector as part of the package which is mated to asimilar one on a lead- hence no chip 'IC' socket is needed. Great! I would also say that we have come to the conclusion that a chip like the Tek1024 (in a Kovar package) in a ZIF socket alone may not make really good thermal contact with a cold block. For this design, I would favour an indium foil behind it (to allow for non-flat KOVAR package), coupled with a clamp on top of the chip to help press it down. Seasons Greetings to all, Paul Jorden ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From ynaogh at public.bta.net.cn Fri Dec 22 14:21:59 1995 From: ynaogh at public.bta.net.cn (user 1156) Date: Thu Jul 29 11:54:07 2004 Subject: info and idea request for PI CCDs Message-ID: <199512220521.NAA25546@public.bta.net.cn> Dear colleagues: I bought a LN/CCD 1752PB from Princeton Instruments which is a thinned, back illuminated and pixel size 15um, 1752X532 spectrometric detector. I don't know why its full well capacity is only 5900 e and readout noise 8.6 e. PI response me it's the correct CCD. I am curioused if anyone got a better one from PI for the ad. I saw before I bought says the FWC not less than 260,000 e and readout noise 4-6 e (only it's an ad. titled with 'preliminary'). I got trouble too when I use LN/CCD 1024TKB also from PI. When it works under -100C, water vapor comes up on the quartz window ahead of CCD chip and behind of the shutter (the vacuum condition of the space between the chip and the window is good) as the exposure time reaches several tens of seconds (the enviomental humidity about 60% and temperature 10C to 15C). The lower the working temperature, the quiker vapor comes up. when I uninstalled the shutter (there's another quartz window in front of shutter) with the outer window together, the vapor no longer came. I cannot do without the shutter in my future work. So, I need your help if you can give some ideas to solve this problem. Consult me with the address: ynaogh@public.bta.net.cn Many thanx. Qian Song. From ynaogh at public.bta.net.cn Fri Dec 22 14:21:59 1995 From: ynaogh at public.bta.net.cn (user 1156) Date: Thu Jul 29 11:54:07 2004 Subject: info and idea request for PI CCDs Message-ID: <199512220521.NAA25546@public.bta.net.cn> Dear colleagues: I bought a LN/CCD 1752PB from Princeton Instruments which is a thinned, back illuminated and pixel size 15um, 1752X532 spectrometric detector. I don't know why its full well capacity is only 5900 e and readout noise 8.6 e. PI response me it's the correct CCD. I am curioused if anyone got a better one from PI for the ad. I saw before I bought says the FWC not less than 260,000 e and readout noise 4-6 e (only it's an ad. titled with 'preliminary'). I got trouble too when I use LN/CCD 1024TKB also from PI. When it works under -100C, water vapor comes up on the quartz window ahead of CCD chip and behind of the shutter (the vacuum condition of the space between the chip and the window is good) as the exposure time reaches several tens of seconds (the enviomental humidity about 60% and temperature 10C to 15C). The lower the working temperature, the quiker vapor comes up. when I uninstalled the shutter (there's another quartz window in front of shutter) with the outer window together, the vapor no longer came. I cannot do without the shutter in my future work. So, I need your help if you can give some ideas to solve this problem. Consult me with the address: ynaogh@public.bta.net.cn Many thanx. Qian Song. From roger at ctiol3.ctio.noao.edu Wed Dec 20 16:54:29 1995 From: roger at ctiol3.ctio.noao.edu (roger smith x294) Date: Thu Jul 29 11:54:07 2004 Subject: MIcrotemp thermal fuses Message-ID: <9512201854.AA03684@ctiol3> Olaf, I quite agree about the need to be able to warm the mount fairly quickly for maintenance purposes, and it doesn't take much heat to get to high temperature. I (we) use Microtemp thermal fuses in series with the heater for all our detector mounts. These satisfy all of your criteria for thermal runnaway protection except that it is not self resetting. ie: a.) small b.) reliable c.) suitable for vacuum d.) still works at and withstands the N2 temperatures e.) obtainable for different temperature switch points The particular attraction of these devices is that their setpoint does not drift as a result of thermal cycling. We adopted them after finding that bimetal "button" type thermostats suffer inelastic deformation at cryogenic temperatures and change their setpoint unacceptably. The Microtemp fuses are an axial leaded, two pin component, 18.5mm long including minimum length leads and 4.0 mm in diameter and are rated to 15 amps (!). On resistance = 1.5 milliohm. Breakdown voltage = 1200V They contain a single pole sigle throw switch opened by a spring, but "potted" in wax inn the closed position. At the specified temperature the wax melts and the swicth opens. They are sealed so that the hot wax does not escape, and we have not detected any problems with outgassing into the vacuum. We have used them for about 11-12 years in both IR and optical detector mounts. They are very simple and reliable and are used in consumer appliances in great volume. Micro devices A division of Emerson Electric Co. 1881 Southtown Blvd Dayton, Ohio 45439 (We obtained our last batch from Arcom Sales Inc, 7300 E. Stetson Drive,Scottsdale AZ 85251, tel: (602) 945-5745) Part# Temperature +0 -4C tolerance 4158 72 C (lower temperatures used to be available) 4168 77 4178 84 4191 91 4194 93 4204 98 4208 100 4218 104 4227 109 etc 117, 121, 128, 141, 152, 188, 184, 216, 228, 240 We first adopted the Microtemp Fuses when building a failsafe 200W heater for a cryogenic IR spectrograph which had a substantial thermal inertia. After use in about 20 cryogenic applications we have never seen one fail to work, though we have had several trip prematurely (?) when we have been baking detector mounts, even though we thought the oven was set to below the nominal trip point. We now mount them in a socket for easy removel prior to baking (...a nuisance to be sure). We suspect that the wax begins to soften allowing the switch to open prematurely if kept near the thermal trip point for many hours. At present we use 72 degrees C and simply choosing a higher trip point maight solve this problem without jeopardizing the detector. Roger Smith, CTIO From richard at ucolick.org Tue Dec 19 21:46:00 1995 From: richard at ucolick.org (Richard Stover) Date: Thu Jul 29 11:54:07 2004 Subject: SITE 2K bow Message-ID: <9512200446.AA07778@gardiner.ucolick.org> Hello all, As we go to faster and faster cameras the CCD shape is becoming ever more important. In addition, the very tight constraints on overall flatness of our large CCD mosaics requires us to know the shape of our CCDs to rather high accuracy. Therefore at Lick Observatory we have been developing a system to measure CCD surface shape. The concept behind the system is similar to a Hartmann test. We bounce a laser off the CCD surface and measure where the reflected beam goes. The system will allow us to measure the shape of the CCD in a dewar under cryogenic conditions. We have so far looked at a number of Loral 2kx2k CCDs. We're using the system to help us develop better methods for CCD packaging. We also measured a SITe 2Kx2K CCD at room temperature. Unfortunately we have not measured the same one cold yet. We are expecting delivery of a new SITe 2kx2K any day now, and we will be measuring that device both warm and cold. We should have accurate measurements of this device before the end of January. I'll report back to this forum on our results. Our surface measuring system is undergoing significant modifications to make it automated. Once this is done it should be relatively easy and quick for us to measure CCD shapes. At that point we may make the facility available to others. Is there much interest in having such a facility available? Richard Stover UCO/Lick Observatory University of California Santa Cruz, CA 95064 408-459-2139 From oiwert at eso.org Tue Dec 19 21:51:30 1995 From: oiwert at eso.org (Olaf Iwert) Date: Thu Jul 29 11:54:07 2004 Subject: Overheat temp. switches Message-ID: <9512191951.AA18868@mc14.hq.eso.org> Dear all, Here it goes with yet another request : As many others we would like to increase the heating power inside the cryostat head to heat the CCD. This proves to be useful for maintenance, faster warm-up and clean cryo conditions as needed e.g. for UV flooding. However one problem can be to provide a lot of heating power without a hardware safety device against potential overheating of the CCD inside the dewar. (I know we can have precautions in the software such as temperature sensor contact loss detection etc., but an additional hardware protection seems more than useful to me in view of the charcaterization time, budget and replacement constraints for CCD devices). My question is whether anyone knows and has tested such a device which is : a.) small b.) reliable c.) suitable for vacuum d.) still works at and withstands the N2 temperature e.) obtainable for different temperature switch points (e.g. +70,+90,+120 C) f.) self repairing (not like a fuse one has to replace, breaking the vacuum) ???????????? In case you would even have a data sheet, I'd be more than happy to receive a copy of that via fax, as this would enable me to evaluate all the mechanical things asap..... Thanks a lot =============================================================================== Olaf Iwert E-mail : oiwert@eso.org European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 Optical Detector Lab FAX : +49-89-320-23-62 Karl-Schwarzschild-Str 2 85748 Garching GERMANY =============================================================================== From oiwert at eso.org Tue Dec 19 21:35:28 1995 From: oiwert at eso.org (Olaf Iwert) Date: Thu Jul 29 11:54:07 2004 Subject: SITE 2K bow Message-ID: <9512191935.AA18799@mc14.hq.eso.org> Dear all, I believe this is a VERY old, but nevertheless important issue. You are right Alistair, the plot from SITE is definitely "warm" data. We have also measured the bows of our TEK2K's here at ESO and found fairly good agreement with the plots we got from them. I've been discussing the issue about bow variation under cryogenic conditions with Site already years ago, however without a chance to really follow it up by measurement under cryogenic conditions (sometimes there is just too much interface talk, but not enough action). However at that time the information I got was that the bow would not vary more than 30% compared to the warm values, according to the SITe data. By the way, I digged out the very first hint on this issue, being a memo from Steve Vogt (Lick Observatory), dated 4/17/92 in the context of TEK2K's with HIRES in which it says : "Bev also tried measurements with the CCD warm and at LN2 temperatures, and found no difference in the CCD shape." May I assume from your Email, Alistair, that you got another 2 TEK2K's from the SITe special offer period with 50 % discount ? Where do you plan to use them ? Season's greetings to all folks in the CCD world and many more thin high speed pixels yet to come..... =============================================================================== Olaf Iwert E-mail : oiwert@eso.org European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 Optical Detector Lab FAX : +49-89-320-23-62 Karl-Schwarzschild-Str 2 85748 Garching GERMANY =============================================================================== From walker at ctiow3.ctio.noao.edu Mon Dec 18 19:21:11 1995 From: walker at ctiow3.ctio.noao.edu (a.walker x305) Date: Thu Jul 29 11:54:08 2004 Subject: CCD flatness Message-ID: <9512182121.AA29745@ctiow3> Dear CCD-people, When SITe deliver a thinned 2048 CCD they now include a plot showing the "bow" of the surface, measured corner-corner in two orthogonal directions. Two recent CCDs have amplitudes of 210 and 230 microns. These are presumably measured with the CCD warm, although they don't say so explicitly. The only independent set of figures I have seen originate with David Vaughan at KPNO. For one device he found that the bow amplitude increased by 34 percent when the CCD is cooled to -100C. The 2048 we use at the 4m prime focus shows focus changes amounting to around 200 microns center-corner (the optical field is flat) and I am about to use a "field de-flattener" as dewar window to compensate. Operating temp. is -105C. I would be very interested to hear if anyone has measured the bow for thinned SITe (Tek) 2048's, and whether the increase in amplitude upon cooling is confirmed. Alistair Walker From apo at ast.cam.ac.uk Mon Dec 18 16:28:19 1995 From: apo at ast.cam.ac.uk (Paddy Oates) Date: Thu Jul 29 11:54:08 2004 Subject: vacuum grease on window Message-ID: David, Thanks for the information. We used 3 materials to clean the inside of the camera - Cleaning fluid to start, detergent then isoprop. That did a good job. We wouldnt use this procedure with the chip! We will think carefully and try a couple of test runs 1st to see what happens- I will let you know!! Cheers Paddy -- E-mail: apo@ast.cam.ac.uk Machine: Soulianis.ast.cam.ac.uk Web Home Page - http://www.ast.cam.ac.uk/~apo CCD Home Page - http://www.ast.cam.ac.uk/~apo/docs/ccds.html Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 direct... -374836 Fax... -374700 Merry Christmas to one and all x =o= | o/^\o < . > o/ \o < . > o/ . \o < . > o/ \o < . . > o/ \o < . . > o/ . . \o < . > o/=============\o ) ( ) ( |___| -- From ouellett at as.arizona.edu Mon Dec 18 09:14:51 1995 From: ouellett at as.arizona.edu (ouellett@as.arizona.edu) Date: Thu Jul 29 11:54:08 2004 Subject: vacuum grease on window Message-ID: <9512181514.332E80@aspc77.as.arizona.edu> For whatever it may be worth, the best way to clean vacuum grease and heavy index oils I found was to use Xylene first, then to remove the oily Xylene residue with an acetone wash followed by methanol. I have not tried this on a CCD surface, however. Best wishes to all on the list for a Happy Holiday Season! David Ouellette University of Arizona (520) 621-5228 voice Steward Observatory (520) 626-4330 FAX 933 N. Cherry Ave Building #65 Tucson, AZ 85721 From apo at ast.cam.ac.uk Mon Dec 18 13:52:27 1995 From: apo at ast.cam.ac.uk (Paddy Oates) Date: Thu Jul 29 11:54:08 2004 Subject: vacuum grease on window Message-ID: Roger Hi, I am envious of the sandcastles - I think I could live with the kind of weather we had in Australia when we were there one Christmas a few years ago - 45C at mid-day but we still downed the roast pork, Yorkshire pudding et al. Though we jumped straight in John Atkins swimming Pool afterwards. It all had to be cooked in-doors too as there was a total fire ban due to the lack of H2O... Blisss eh. To answer your questions- The heaters werent on at ambient - yeah the stuff migrated from behind the chip i.e. facing away from the Field flattener and went everywhere (it turned out, though we could only actually SEE it on the FF). We looked at the chip a few times but didnt see anything obvious BUT - last week I had the chip out an put it in one of our lab dewars to check the temp performance of the clamp, Indium and mica solution and I got the chance to have a closer look and well - you guessed its on the chip as well - looks like small blobs on one half. I ma going to try and clean it in the New Year. I will use some ISO-P alcohol on an enginerring device 1st and see what happens. I am worried that although the grease may not show gross effects it could have damaged the AR coating - this would obviously be bad news! The mechanism by which it got from the back of the chip and out is a mystery. I assume it vaporised when at room temp then settled onto all the warm surfaces including the FF which is only a few degrees below ambient when the chip is at operating temp. Thats all for now folks Paddy -- E-mail: apo@ast.cam.ac.uk Machine: Soulianis.ast.cam.ac.uk Web Home Page - http://www.ast.cam.ac.uk/~apo CCD Home Page - http://www.ast.cam.ac.uk/~apo/docs/ccds.html Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 direct... -374836 Fax... -374700 -- From walker at ctiow3.ctio.noao.edu Mon Dec 18 10:04:19 1995 From: walker at ctiow3.ctio.noao.edu (a.walker x305) Date: Thu Jul 29 11:54:08 2004 Subject: 2kx2k displays Message-ID: <9512181204.AA27164@ctiow3> >From George Jacoby, KPNO. A quick Net search located a page from Peritek whom I called just now. They have a VME board which they anticipate upgrading from 1600x1200 to 2kx2k very soon (1-3 months). It already has all the pixels, but needs a faster DAC which they are currently negotiating with IBM to purchase. Thus, only minor mods are needed to their existing board. Anticipated price is $5K; several boards can co-exist in a single system running X-windows. It would be possible to have one board run a full-view 8kx8k compressed view and another running a 2Kx2K window into the full frame. From roger at ctiol3.ctio.noao.edu Mon Dec 18 09:56:01 1995 From: roger at ctiol3.ctio.noao.edu (roger smith x294) Date: Thu Jul 29 11:54:08 2004 Subject: vacuum grease on window Message-ID: <9512181156.AA09825@ctiol3> Paddy, It's grey here too. In fact its drizzzling but I dodn't think it will register on the rain gauge. The problem is that we rarely get enough at once to get the droplets to run down into the bottom where they get measured. So we are still stuck at 6mm for the year. Bummer. The sun will probably be out thsi afternoon and I'm thinking more along the lines of sandcastles and surfing for Christmas. Our family has never had Christmas in winter. Are we envious? No! Thanks for the warning about vacuum grease. How did it migrate from the back of the detector though? Did you somehow run your detector heater while at ambient temperature. Its already low vapour pressure must be zilch when at crogenic temperatures. Presumably it got on the CCD too. How did you get it off? Roger From apo at ast.cam.ac.uk Mon Dec 18 10:06:40 1995 From: apo at ast.cam.ac.uk (Paddy Oates) Date: Thu Jul 29 11:54:08 2004 Subject: ZIFs, grease and Spiders Webs Message-ID: Blecha, Hi, we have used ZIF sockets for a number of years with our large EEV devices (800x1180, 1280x1180 and 2200x1180) and Tek1024s. The smaller 2 of the EEV devices and the Tek worked fine, we saw very low dark currents, about 1e/pix/hr (EEV) and around 6 for the TEk, BUT the large EEV device, because it was designed with 2-edge buttability in mind was diffincult to clamp down to the cold surface and hence we had a lot of problems with high dark current. The EEV devices alsways performed okay. However just recently we tried clamping the Tek device down onto the cold surface, still using a ZIF socket to mount and this resulted in far superior dark current performance i.e. un-measurable over an hour! The ZIF socket for the large EEV device was modified, the socket being cut in half and separated, joined by aluminium levers to close the pins around the chip legs when the lever was operated. I would guess this is about as big as you would want ot go with this arrangement. I dont recall seeing any adverse problems using the sockets in a vauum or cold environment. On this last point what we have just recently had trouble with is vacuum grease. We have had to return an instrument from La Palma , WYFFOS (a wide field fibre optic spectrographic camera) because of a build-ip of a contaminant on the systems field-flattening lens, just in front of the chip. What happened was the material accumulated on the field flattener over a day or two with the result that we saw a horrendous increase in scattered light, making the system un-useable. We had the material analysed in a mass spectrometer and it instantly showed the characteristic pattern of, what is described as Ultra High Vacuum Grease. It turned out this stuff was all over the inside of the camera - we have cleaned everything and taken out any trace of this grease. The scattered light problem has gone. We used a very small amount of the greas to provide good thermal contact between the chip package (a Tek1024) and the cold surface. We are now using a thin mica sheet (for insulation as the tek package is at substrate potential and we operate with SS at -7V) and Indiulm foil to provide even thermal contact with the cold surface. I have just tested the cold performance of this, clamping the chip in its ZIF socket, and thats what produce d the v.low dark current performance I mentioned above. LESSON to be learned DONT USE ANYTHING BUT THE TINIEST AMOUNT AF U.H.V GREASE on the SEALS and NON anywhere else. We have bought in some Super U.H.V grease which is good to 10-12 mbar and we are using this on the seals, none inside the camera.... This whole process has taken since September to sort out as the camera has a large volume and cold tests take around a working week to do... I should be putting the EEV data from the chips we measured before the Caymen conference on the Web page today. http://www.ast.cam.ac.uk/~apo/docs/caymen/caymen.html We are still progressing with testing though the VCTE even with the smaller 0.5kx2k (VxH) devices still appears non too good. I plan to measure the dark current and will put all this up on the Web then. Its cold and grey here in the UK at the moment - looks like a portent of Snow for Christmas, lets hope.... Merry Christmas to one and all Paddy -- E-mail: apo@ast.cam.ac.uk Machine: Soulianis.ast.cam.ac.uk Web Home Page - http://www.ast.cam.ac.uk/~apo CCD Home Page - http://www.ast.cam.ac.uk/~apo/docs/ccds.html Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 direct... -374836 Fax... -374700 -- From prj at mail.ast.cam.ac.uk Mon Dec 18 09:43:22 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:08 2004 Subject: zero force insertion sockets Message-ID: Hi Andre & others, > I remember remotely a discussion concerning the "zero force insertion > sockets". Some people was against using them with cryogenic CCD's, but I > do not remember what were the arguments. Any comment (pros and cons) ? > I have not looked up a previous discussion, but I would say we have used them with no real problems. We originally used various forms of simple 'dil' sockets or custom made 'sil' strips for early small CCDs. However, the larger CCDs, with lots of pins tend to be hard to remove from multiple sockets (especially the turned-pin sort). We started using ZIF sockets for EEV-05 series, & then Tek1024 chips. The advantage is that they allow very easy removal & insertion. They need some degree of machining for our central cooling block of course. The disadvantage is that they are quite chunky- and add to the mass of material to be cooled; this need not matter too much. Now we are using the butted style of chips (Lesser 2k3ebs, EEV CCD42) and these come with a connector as part of the package which is mated to asimilar one on a lead- hence no chip 'IC' socket is needed. Great! I would also say that we have come to the conclusion that a chip like the Tek1024 (in a Kovar package) in a ZIF socket alone may not make really good thermal contact with a cold block. For this design, I would favour an indium foil behind it (to allow for non-flat KOVAR package), coupled with a clamp on top of the chip to help press it down. Seasons Greetings to all, Paul Jorden ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From blecha at scsun.unige.ch Wed Dec 13 12:29:08 1995 From: blecha at scsun.unige.ch (Blecha Andre) Date: Thu Jul 29 11:54:08 2004 Subject: zero force insertion sockets Message-ID: <1673*/S=blecha/OU=scsun/O=unige/PRMD=switch/ADMD=400net/C=ch/@MHS> Ahoj, I remember remotely a discussion concerning the "zero force insertion sockets". Some people was against using them with cryogenic CCD's, but I do not remember what were the arguments. Any comment (pros and cons) ? Thanks Andre Blecha ---------------------------------------------- | tel +41 (22) 755 26 11 | | please call at hours (winter time = GMT+1) | | (8h15-9h30;9h50-12h;13h-15h30;15h50-17h15) | | fax +41 (22) 755 39 83 | | e-mail blecha@scsun.unige.ch | | Anonymous ftp obsftp.unige.ch | | use directory "contrib" | | Observatoire de Geneve | | 51, ch. des Maillettes | | 1290 Sauverny | | Switzerland | ---------------------------------------------- From blecha at scsun.unige.ch Wed Dec 13 12:15:34 1995 From: blecha at scsun.unige.ch (Blecha Andre) Date: Thu Jul 29 11:54:08 2004 Subject: 2k2k displays (Paddy Oates) Message-ID: <1672*/S=blecha/OU=scsun/O=unige/PRMD=switch/ADMD=400net/C=ch/@MHS> Ahoj, Sun has now in standard 1600 x 1280 system (that is the display with Turbo GX+ graphics) for S-bus workstation. Above that resolution it will be sort of large factory if you want a flicker-free display. I remember some applications running with as much as 4k4k displays; I will try to dig out commercial informations about it (but it may be more expensive that the CCD chip). You will need the VME bus ... In mean time think about a more "soft" solution. I would suggest one of: 1. multiple display on a single machine (standard at SUN) 2. software showing simultaneously: under-sampled full image 1 by 1 sampled large portion (say 1k1k) in a large window or possibly dedicated display (multiple display HW) + usual zooms We adopted the solution 2) for our balloon experiments 3k3k images and it is perfectly convenient providing the graphic system is fast enough and you have the memory. Our system works on SUN/Solaris and you may have the code if you are interested (It works with several image formats or as a server receiving data from another program). Andre Blecha ---------------------------------------------- | tel +41 (22) 755 26 11 | | please call at hours (winter time = GMT+1) | | (8h15-9h30;9h50-12h;13h-15h30;15h50-17h15) | | fax +41 (22) 755 39 83 | | e-mail blecha@scsun.unige.ch | | Anonymous ftp obsftp.unige.ch | | use directory "contrib" | | Observatoire de Geneve | | 51, ch. des Maillettes | | 1290 Sauverny | | Switzerland | ---------------------------------------------- From oiwert at eso.org Thu Dec 14 10:31:43 1995 From: oiwert at eso.org (Olaf Iwert) Date: Thu Jul 29 11:54:08 2004 Subject: SUSI Message-ID: <9512140831.AA16258@mc14.hq.eso.org> Dear all, My understanding is based entirely upon the last telephone and video communiction to La Silla, where it was siad that the problem with SUSI could be fixed by refilling it twice a day. I had explicitly asked Jorge and Nicolas whether theuy would need any spare parts from Garching or further support, however the answer was that the filling tube must have been damaged somehow so the hold time would be much less than expected. As there are no other signs of misbehaviour what concerns the CCD itself, I would suggest to maintain a refilling twice per day as a workaround. It could then be fixed either by sending the SUSI cryostat back to Garching during a period of NON use or during Sebastian's stay for the B&C upgrade. I'm just worried because Sebastian is also needed here for FORS.... =============================================================================== Olaf Iwert E-mail : oiwert@eso.org European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 Optical Detector Lab FAX : +49-89-320-23-62 Karl-Schwarzschild-Str 2 85748 Garching GERMANY =============================================================================== From apo at ast.cam.ac.uk Tue Dec 12 17:32:40 1995 From: apo at ast.cam.ac.uk (Paddy Oates) Date: Thu Jul 29 11:54:08 2004 Subject: Highly Tecnical Stuff on WWW Message-ID: Hi Folks, Well back home - to cold and frost. I have started a page devoted to the Caymen meeting, its all highly er technical at the moment, its is a state of preparation so dont expect a high information data rate on CCDs. You might find one or 2 pictures however of various folk in compromising positions in the bar - that bit at the back of the meeting room in case anyone missed it. For the CCD page, point yer browsers at http://www.ast.cam.ac.uk/~apo/docs/ccds.html for the strong of heart, point at http://www.ast.cam.ac.uk/~apo/docs/caymen/caymen.html One question, someone mentioned on the last night - I think it may have been one of the Loral divers er sorry people, that they knew of a source for 2kx2k DISPLAY MONITORS not CCDs - I think I know where to get a few of those. Does anyone (from Loral) know of such a display device - we want it for a Sparc 20 NOT a PC.... Thanks and Merry thingumies Payde -- E-mail: apo@ast.cam.ac.uk Machine: Soulianis.ast.cam.ac.uk Web Home Page - http://www.ast.cam.ac.uk/~apo CCD Home Page - http://www.ast.cam.ac.uk/~apo/docs/ccds.html Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 direct... -374836 Fax... -374700 -- From jbeletic at eso.org Mon Dec 11 09:50:55 1995 From: jbeletic at eso.org (James Beletic) Date: Thu Jul 29 11:54:08 2004 Subject: ESO "CCD" Workshop - October 8-10, 1996 Message-ID: 11 December 1995 Dear Colleagues, The dates for the next ESO "CCD" workshop have been finalized. We have settled on October 8-10, 1996. We moved the time from 2nd half of October to the first half of the month in order to have a better chance of good weather. We have reserved nearly every available hotel room in Garching for period of Friday, October 4 through October 11. We will need to finalize the arrangements around April. With everyone staying in the same town, it will be easier to arrange dinners and the 2 km walk to ESO is a pleasant way to start and end the day (we will provide a chartered bus also). For flights it is important for you to book early; nearby Munich has a little party (attended by millions) called the Oktoberfest, which has its final day on Sunday, October 6. Please plan to arrive by the afternoon of October 7 since we will have a welcoming reception that evening. We have had a very good response from the community and we expect that every major observatory and every CCD manufacturer will be represented. Also, in listening to comments, we plan to minimize the formal presentation aspect of the workshop and instead to concentrate on roundtable discussions and other forums conducive to free flow of information (e.g. biergarten). We will have more details on the agenda as the ideas crystallize. Please continue to send comments and suggestions. Please send us your information (address, E-mail, phone, FAX) if you wish to attend. We will be sending announcements by mail in February-March, 1996. Please do not actually purchase any plane tickets until you get confirmation from us - we may need to limit attendance. Best wishes, Jim Beletic --------------------------------------------------------------------------- ANNOUNCEMENT OF "CCD" WORKSHOP --------------------------------------------------------------------------- Title: Workshop on Optical Detectors for Astronomy Date: October 8-10, 1996 Duration: 3 days Location: European Southern Observatory (ESO) headquarters Garching, Germany # of participants: 100 maximum external participants Papers on: Optical detectors - primarily CCDs new designs manufacturer products performance specifications high speed, low noise amplifiers skipper amplifiers adaptive optics CCDs Pushing into the UV (improvements in backside passivation and AR coatings) Pushing into the IR (thick, high res. silicon) Avalanche photodiodes Wavelength sensitive detectors Detector controllers Interfacing to scientific instruments Effect of CCD parameters on astronomical science Detector calibration / testbench design Cryostat design - bath, continuous flow and TEC Status reports from the detector groups at the major telescope projects Proceedings: Will be published from camera ready copy provided 4 weeks after conclusion of the workshop. We strongly encourage hardware demonstrations and industry participation. We expect to have representatives of the optical detector groups from every major astronomical observatory in the world. Members of the local organizing committee: James Beletic Olaf Iwert Sandro D'Odorico Elena Zuffanelli, conference secretary ------------------------------------------------------------------------------ ========================================================================= James W. Beletic E-mail: jbeletic@eso.org European Southern Observatory Phone: +49-89-320-06520 Karl-Sch-Str 2, D-85748 Garching, GER FAX: +49-89-320-2362 ========================================================================= From tmca at cfht.hawaii.edu Tue Dec 5 15:29:53 1995 From: tmca at cfht.hawaii.edu (Tim Abbott) Date: Thu Jul 29 11:54:08 2004 Subject: Grand Cayman Message-ID: <9512060029.AA28391@malama.cfht.hawaii.edu> International Conference on Scientific Optical Imaging Grand Cayman, Nov 19 - Dec 2, 1995 Well, I'm fresh back at the tropical island in one ocean from another tropical island in another ocean. Following is the list of presentations made at the Grand Cayman conference last week, together with a few short notes on content. Please note that these are taken from my own notes and are therefore second-hand. For first hand info, please contact the people concerned, especially as I cannot possibly do justice to the authors in one short email. (If I have made any mistakes, the same disclaimer applies - these days I can barely read my own writing...) All-in-all, the conference was very successful and we all owe our thanks to M. Bonner Denton for all his efforts in getting us together. ---- *** "High Speed Scientific CCDs" - Jim Janesick, PixelVision. Jim defines some new acronyms: EPR = Equivalent pixel rate (pixel rate per port x no. of ports) HSS CCD = High Speed Scientific CCD, EPR > 1 MPixel/sec Slow scan CCD has EPR < 1 MPixel/sec. Runs over performance problem areas in HSS CCDs and how optimisation is acheived. HSS CCDs typically require smaller pixels, degrading the QE. MPP CCDs do not do well at high speed. Described many new CCDs (manufactured in "Sandbox" runs, with several different designs per wafer), including: "CINEMA" - a 4096x4096x9micron, frame transfer, w/ 32 output ports. intended for 30 frames/sec (EPR = 251 Mpixel/sec) "PLUTO" - 32 frames/sec, EPR = 32 Mpix/sec, skipper amplifier on each column "PHILLIPS" - 7kx9k, 4 ports, 31.5 sec/frame. *** "Fiber-Optic Array Sensors" - David R. Walt, Tufts University For medical and industrial applications, observing chemical reactions at the ends of fiber-optic probes. *** "Imaging Spectroscopy for Infra-Red Astronomy" - Craig Mackay, AstroCam Ltd. Describes a system based on COHSI at UH, for IR spectroscopy in H & J, by dispersing the beam at high resolution, masking out the atmospheric lines, recombining the beam and resdispersing at lower resolution before detection on HAWAII 1024^2 IR detectors from Rockwell. *** "Recent Spectroscopic Applications of Array Detector Technology" - M. Bonner Denton, University of Arizona. Review of Lab-oriented echelle spectrographs, Raman FTS with CCDs. Use of CCDs to analyse solvent-diffused chromatographs and gel electrophoresis. Described "Spectrograph-on-a-chip" - an integrated grating spectrograph and linear array. *** "Atomic Absorption Spectrometry Using a Continuum Source and a Segmented CCD" - James Harnly, USDA. Detection of carcinogens in your hamburgers. *** "Thinned CCDs for Soft X-ray and Electron Bombarded Apllications" - Morley Blouke, SITe. Described UV-coatings yielding QE ~ 50% @ 200 nm. Peak absorption at 400 nm for 10 keV electrons, but this is damaging to the CCD and 1.8 keV electrons used instead, yielding a lifetime of ~1000 hrs. *** "A New Large-Format, Very low noise Scientific CCD" - Peter Pool, EEV. EEV now makes a 2kx4kx13.5micron, 3-side buttable CCD. Thinned, boron implant. Some (thick devices) sent to RGO, initial test results presented by Paul Jorden: r.o.n. = 1.6 e- @ 4usec sample time, signal & reset. *** "Status of Large Area CCD Imagers at Loral." - Richard Bredthauer, Loral Fairchild Imaging Sensors Loral is alive and well. Demonstrated a 9kx9kx8.75micron CCD (162 MByte/image!) no bad columns, all 4 outputs work! Not run slow scan yet. *** Spectrophotometers from Spectral Instruments - Gary Sims, Spectral Instruments *** "A Review of Imaging Optica for the Ultraviolet" - Donald K. Wilson, Optics for Reseacrh, Inc. *** "Advanced Imaging Systems from Xedar Corp." - Gregg Herbison, Xedar Corp. Fiber optic faceplates for CCDs. *** "Application Specific Imaging Device Fabrication at Sarnoff" - John R. Tower, David Sarnoff Research Center. Segmented CCD for Echelle spectroscopy - 224 custom-placed linear arrays. PtSi, 320x244 pixel array for Vis-SNIR-MWIR imaging, 40% fill factor. Backside, high speed CCD - 1kx2kx18micron, frame xfer 32 output ports & on-chip CDS. QE 60% @ 400 nm. noise floor 30-50 e-. p-channel device for UV imaging. *** "A Thinned, Three-side Buttable 2kx4k CCD for Low Level Light Applications" - Morley Blouke, SITe. 2kx4kx15micron, frame xfer. r.o.n. = 3.6 e- @ 50 kpixel/sec Dark current = < 5e/pixel/hr full well ~ 75-80 ke-/pixel QE > 85% @ 650nm Bow: 80-180 micron peak-to-valley. *** "Flourescence Imaging of While Microorganisms with Scientific Grade CCDs." - Kenneth D. Gughes, Georgia Tech. *** "THX7398ML A 2kx2k Thinned CCD, 3 sides buttable." - Alain Jutant, Thomson Components and Tubes Corp. *** "Practical Applications of Hybrid Neural Nets" - Jeremy Lerner, Physical Optics Corp. *** "A New Low-Noise CCD for Spectroscopy" - John Geary, SAO SITe 1752x532, backside illum, available in limited quantity. r.o.n. = < 3e- for integ time > 12 usec. Evidence for residual image. *** "CCD Detectors after 7 years in Interplanetary Space." - Horst Uwe Keller, MPIA. After 7 yrs, Giotto CCDs have 2x the number *and* amplitude of hot pixels. *** "CIDs for Atomic Emmission Spectroscopy: The State of Current Performance and Future Directions". Michael, J, Pilon, Thermo Jarrel Ash CIDs now approaching the performance of CCDs: CID38 - 512x512 pixels, r.o.n. = 25e-, dark current < 0.5 e-/sec @ -80 C QE = 42% @ 250 nm, 82% @ 550 nm, full well ~ 1.26 Me-. *** "Confocal Microscopy", Brakenhof. *** "A Progressive Scan CCD Image Sensor for High Speed Photography" - Rusty Winzenread, EG&G Reticon. 1024x1024 pixels, 10^3 frame/sec, interline xfer. *** "Low-noise CCD Imaging at Kilohertz Frame Rates" - Craig Mackay, Astrocam Loral 64x64 pixel CCD for adaptive optics. up to 8 MHz w/ 12-bit ADC 5 MHz w/ 14-bit 9 Gigabyte per night. run CCD at 865 frames/sec, w/ full resolution => r.o.n. = 100 e- Also, studies of mammalian neurons: EEV CCD37 - 512x512 frame xfer CCD, 15 Hz frame rate. *** "Optical Imaging in Highly Scattering Media Using Time-Resolved Measurements" - David H. Burns, McGill Univ. Looking into the first ~1cm layer of human body using IR. *** "A Circular CCD for application with a Fabry-Perot." Wayne Frame, Ball Aerospace and Technology. A neat design, w/ pixels arranged in circles. Amplifiers: 1.5 uV/e- r.o.n. < 10e- Well > 200 ke-/pix Advantages over rectilinear device: very low r.o.n. in large diameter circles CTE in one direction only simple post-processing *** "Echelle Spectroscopy and CCDs, and Ideal Union for Fiber Optic Raman Systems." - Michael Carrabba. low-risk ID of nasty compounds made by the army, then forgotten about. Advice to the cautious: if you find any suspicious looking vials of liquid in your back yard, dump them on your neighbour "the army'll camp in your yard for weeks"... *** "Optical Detector Systems at the European Southern Observatory." - Jim Beletic, ESO. Plans for 8kx8k, maybe 16kx16k imager. Jim's making waves at ESO, watch that space... *** "High Resolution X-Ray Microtomography System using a Directly-Coupled Structured Scintillator and CCD Detector." - Nigel Allinson, UMIST. See the inside of a bee's head. *** "A Hybrid Neural Net Used for Hyper Spectral Image Analysis System Optimisation and Calibration.", Jeremy M. Lerner. *** "Unique Advantages of APS Technology for Scientific Applications." - Eric Fossum, JPL Really neat stuff with CMOS technology. APS = Active Pixel Sensors, several transistors/pixel. Building a micro-CCD in each pixel. 256x256x20micron 1kx1kx10micron (with AT&T) Add an ADC/column (per pixel?!) (1 clock in => digital images out) very low power, but poorish fill factor. Showed a camera not much bigger than a nickel (big brother is watching...) *** "An Application of Raman Spectroscopy in the Petroleum Industry." Phillip Marteau, Universite' Paris Nord. Continuous montioring of distillates *** "Back Illuminated CCD Processing" - Mike Lesser, Steward Obs. Mike has significantly upgraded his lab, is exploring MTF vs thinness considerations and is now generating full mosaics... -- Timothy M. C. Abbott, Resident Astronomer, tmca@cfht.hawaii.edu http://www.cfht.hawaii.edu/~tmca/tmca.html Canada-France-Hawaii Telescope, P.O. Box 1597, Kamuela, HI 96743 From reif at astro.uni-bonn.de Tue Dec 5 16:10:18 1995 From: reif at astro.uni-bonn.de (Klaus Reif) Date: Thu Jul 29 11:54:08 2004 Subject: spurious charge vs degree inversion Message-ID: <9512051410.AA02558@aibn55.astro.uni-bonn.de> Roger, we didn't try to only barely invert the phases. We were mainly interested in getting a high antiblooming efficiency and a low spurious charge generation at the same time. One would expect that with only barely inverted phases the AB will go down, as it must depend directly on the number of holes attracted from the channel stops during inversion. For the parallel transport the requirements are of course different. One could try to vary the low potential (still keeping the phase in inversion) and see what happens. Klaus ---------------------------------------------------------- Klaus Reif e-mail: reif@astro.uni-bonn.de Astronomische Institute Tel: +49-228-737834 der Universitaet Bonn FAX: +49-228-733672 Auf dem Huegel 71 D-53121 Bonn Germany ---------------------------------------------------------- From roger at ctiol3.ctio.noao.edu Tue Dec 5 10:27:09 1995 From: roger at ctiol3.ctio.noao.edu (roger smith x294) Date: Thu Jul 29 11:54:08 2004 Subject: spurious charge vs degree of inversion Message-ID: <9512051227.AA15120@ctiol3> Klaus, Have you explored the possibility of only barely inverting the clocks? ie. Do you know whether the spurious charge per cycle is proportional to the number of holes at the surface during inversion? Roger From reif at astro.uni-bonn.de Tue Dec 5 12:46:00 1995 From: reif at astro.uni-bonn.de (Klaus Reif) Date: Thu Jul 29 11:54:08 2004 Subject: AB clocking and spurious charges Message-ID: <9512051046.AA00699@aibn55.astro.uni-bonn.de> Hi Simon and Roger, we got some experience with antiblooming clocking with the LORAL 2k CCD's (not with TEK Chips) but the generation of spurious charge when clocking CCD phases between the inverted and the non-inverted state is a common effect. We experimented with various waveforms over a rising time of 128us when switching from inverted (-8V) to non-inverted (+2.4V). We finally introduced a midstep potential at -2.8V which is kept for 128us. This reduced the spurious charge generation by a factor of 2. Our impression is that this step should be best slightly above the voltage where the phase goes out of inversion: the increase of the rising time before the inversion reversal does not change very much while increasing it above this point makes a significant effect. This seems to agree with one explanation for the spurious charge generation that I know of: driving a phase from the inverted to the non-inverted state repels the holes toward the surface. Trapped holes can be released with enough energy to produce spurious charge over impact ionisation at the surface. We find that the antiblooming efficiency and the spurious charge generation increase linearly over a wide range of antiblooming cycle frequencies (30Hz to 700Hz). The slopes are: antiblooming efficiency: 400e-/pixel per cycle spurious charge generation: 0.0004 e-/pixel per cycle You may find more on this topic in a paper that was published in the SPIE Vol. 2415 (Charge-Coupled Devices and Solid State Optical Sensors V): R. Kohley et al.: "Operating a large area CCD with antiblooming" Klaus Reif ---------------------------------------------------------- Klaus Reif e-mail: reif@astro.uni-bonn.de Astronomische Institute Tel: +49-228-737834 der Universitaet Bonn FAX: +49-228-733672 Auf dem Huegel 71 D-53121 Bonn Germany ---------------------------------------------------------- From roger at ctiol3.ctio.noao.edu Mon Dec 4 16:51:45 1995 From: roger at ctiol3.ctio.noao.edu (roger smith x294) Date: Thu Jul 29 11:54:08 2004 Subject: Spurious charge generation. Message-ID: <9512041851.AA12628@ctiol3> Simon, I have not explored anti-blooming clocking but have investigated spurious charge generation by Tek CCDs to some extent. I presume this is what you are referring to when speaking of "LED events". In one of our Tek 2K's, for example, I found that spurious charge depended mainly on the P3_high clock level, viz: _______________________________________ Spurious_charge Full P1,P2 = -8V to +3.5V P3_hi per per Well P3_lo = -5.5V readout parallel_shift 90% risde time > ~ 50us (V) (e-) (e-) (e-) ______________________________________ 2.03 0.2 0.0001 3.56 0.4 0.0002 60 5.00 0.84 0.00042 140 6.00 1.7 0.00085 7.00 3.7 0.0019 260 The spurious charge was measured here as the height of the overscan transition in a zero length dark frame. Making P3_lo more positive also decreased spurious charge but with a steeper loss in full well. All parallel clocks had very slow rise (and fall times). My impression from operating similar CCDs with a range of slope control filter time constants is that the spurious charge does *not* depend on the rise time so much as it does on the clock levels when the clocks have very slow edges. This does not conform to the conventional wisdom that spurious charge is dependent on the positive going edge **rate**. I have to admit that I have not explored it quantitatively. An aside: At first I wondered why the Tek 2K spurious charge images (ie bias frames) are completely flat given that you see bright spots with your anti-blooming clocks. Before thinking about it I assumed, we would see bright columns in our bias frames but then I realized that the bright spots you describe generate only 6(e-/s) / 300(Hz) = 0.02 (e-/shift). Although this is much greater than the 0.0004 e-/shift which accounts for the spurious charge I see, it would be undetectable in a single frame since it occurred only in isolated pixels. I will be very interested to hear of other experiences with charge injection since it also has serious implications for very low noise operation of CCDs. AB Clocking sounds like a great tool for exploring spurious charge! Roger Smith, CTIO From smt at ast.cam.ac.uk Mon Dec 4 17:51:09 1995 From: smt at ast.cam.ac.uk (Simon Tulloch) Date: Thu Jul 29 11:54:09 2004 Subject: Anti-blooming clocks Message-ID: Hi all We are currently trying to implement anti-blooming clocks with our 1024 Tek CCDs. We are wobbling charge between phases 1 and 2 at about 300Hz. The rising clock edges are 3us , the falling edges 10us. The anti-blooming works well but has the unfortunate side-effect of producing several hundred LED events across the chip. The brightest of these generates about 6 electrons/second. These LEDs disappear if the voltage swing on the vertical clocks is less than 8V, but blooming re-appears if this is done. Increasing the rise and fall times of the clocks to 20 us had no effect on the LED brightness. Has anybody else experienced this kind of problem? Regards Simon Tulloch smt@ast.cam.ac.uk Royal Greenwich Observatory Madingley Road Cambridge CB3 OEZ UK From walker at ctiow3.ctio.noao.edu Wed Nov 15 13:06:49 1995 From: walker at ctiow3.ctio.noao.edu (a.walker x295) Date: Thu Jul 29 11:54:09 2004 Subject: SITe CCDs Message-ID: <9511151506.AA05305@ctiow3> SITe CCD year-end sale. What they had on Monday (2048's listed only) model front/back AR? grade number 424 front N/A 1 2 3 1 eng 3 elec 1 424 back usual 0 1 1 2 2 5 3 1 eng 10 elec 5 mech 2 424 back UV 1 1 eng 3 price structure really is 1/2 normal, eg: back illuminated grade --> 0 1 2 3 eng 4 outputs 80K 65K 45K 2 outputs 55K 35K 1 output 45K 27.5K 20K 7.5K additional charges: 10% for selesction if required 15% for UV AR coating prices apply until 31 Dec, but there aren't likely to be many left by then,,, Alistair Walker From oiwert at eso.org Wed Nov 15 11:55:38 1995 From: oiwert at eso.org (Olaf Iwert) Date: Thu Jul 29 11:54:09 2004 Subject: Separate GNDx connection Message-ID: <9511150955.AA06722@mc14.hq.eso.org> Dear CCD afficionados, Thanks a lot for the recent help on my request concerning "Sharing of bias phases". In summary the answers point into the direction that practical experience in this matter is not always in agreement with the theory - quite a common phenomenon related to CCD behaviour ----------------------------------------------------------------------------- Here it goes with another request : Looking carefully at the SITe CCD data sheets, one finds a separate GND(a/b/c/d) connection to the bulk of each output transistor on-chip. While we would really like to minimize the cross-talk between the different readout channels, the following questions occur : 1.) Is the output amplifier really isolated from the rest of the substrate this way ? Clocks and Bias signals are generated versus a certain ground potential, also the video chains have a certain reference potential. 2.) Has anyone of you done something different than connecting the individual output transistor GND(a/b/c/d) pins to the SUB / PKG GND ? 3.) Is anyone running this ground completely separated from the rest of the other grounds - and, if yes : HOW ? Thanks in advance =============================================================================== Olaf Iwert E-mail : oiwert@eso.org European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 Optical Detector Lab FAX : +49-89-320-23-62 Karl-Schwarzschild-Str 2 85748 Garching GERMANY =============================================================================== From prj at mail.ast.cam.ac.uk Mon Nov 13 09:50:26 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:09 2004 Subject: Site deal & wrong number! Message-ID: To Tim, Rick & all, > > No, It's true! That really is Sue's fax number! (Her comment > to this was "Aha! That's why I haven't received any faxes for > the last day or two!") So I guess there is a phone screwup of > some kind. > Until they get the phones strightened out there, please use the > SITe main fax number 503 644 0798. (and good luck!) > I think there may have been some confusion! Originally Rick quoted the wrong phone number, and the right fax number! This is what I believe they are: SITe (Sue Gordon) phone: 503-671-7140 (not 6140) fax: 503-671-7110 However, I've not had good responses from her in the past. There is also Barry Rahiman, the newish Director of Worldwide Sales: Phone 503-671-7100, fax 503-644-0798 Good luck! Paul ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From mski at dns.dao.nrc.ca Sat Nov 11 21:04:31 1995 From: mski at dns.dao.nrc.ca (Rick Murowinski) Date: Thu Jul 29 11:54:09 2004 Subject: Site deal & wrong number! Message-ID: <9511120404.AA09541@tsartlip.nrc.ca> No, It's true! That really is Sue's fax number! (Her comment to this was "Aha! That's why I haven't received any faxes for the last day or two!") So I guess there is a phone screwup of some kind. Until they get the phones strightened out there, please use the SITe main fax number 503 644 0798. (and good luck!) Rick > From tmca@cfht.hawaii.edu Fri Nov 10 12:08 PST 1995 > From: tmca@cfht.hawaii.edu (Tim Abbott) > Subject: Re: Site deal & wrong number! > To: ccd-world@cfht.hawaii.edu > Mime-Version: 1.0 > Content-Transfer-Encoding: 7bit > > > Rick Murowinski writes: > > > It's true. I'm visiting Site at the moment and found out > > that anything which is now in inventory and you can take > > delivery before Dec 31 is 50% off the sticker price. > > > > Call Sue at Site for all the gory details, > > > > 503 671 6140 (fax 503 671 7110) > > This is the wong number and the woman at Nike corporation in Oregon sounded > a little peeved when I spoke to her...! > > Rick, could you put us right on this one? > > Cheers, > > Tim > > -- > Timothy M. C. Abbott, Resident Astronomer, tmca@cfht.hawaii.edu > http://www.cfht.hawaii.edu/~tmca/tmca.html > Canada-France-Hawaii Telescope, P.O. Box 1597, Kamuela, HI 96743 > ----- End Included Message ----- From tmca at cfht.hawaii.edu Fri Nov 10 10:49:06 1995 From: tmca at cfht.hawaii.edu (Tim Abbott) Date: Thu Jul 29 11:54:09 2004 Subject: Site deal & wrong number! Message-ID: <9511101949.AA08674@malama.cfht.hawaii.edu> Rick Murowinski writes: > It's true. I'm visiting Site at the moment and found out > that anything which is now in inventory and you can take > delivery before Dec 31 is 50% off the sticker price. > > Call Sue at Site for all the gory details, > > 503 671 6140 (fax 503 671 7110) This is the wong number and the woman at Nike corporation in Oregon sounded a little peeved when I spoke to her...! Rick, could you put us right on this one? Cheers, Tim -- Timothy M. C. Abbott, Resident Astronomer, tmca@cfht.hawaii.edu http://www.cfht.hawaii.edu/~tmca/tmca.html Canada-France-Hawaii Telescope, P.O. Box 1597, Kamuela, HI 96743 From mski at dao.nrc.ca Fri Nov 10 08:56:54 1995 From: mski at dao.nrc.ca (Rick Murowinski) Date: Thu Jul 29 11:54:09 2004 Subject: Site/Tektronix 2048 thinned AR coated chip Message-ID: <9511101556.AA08072@dao.nrc.ca> Blecha, It's true. I'm visiting Site at the moment and found out that anything which is now in inventory and you can take delivery before Dec 31 is 50% off the sticker price. Call Sue at Site for all the gory details, 503 671 6140 (fax 503 671 7110) Rick Murowinski {Mail}& 18 [HMessage 18: >From blecha@scsun.unige.ch Fri Nov 10 03:32 PST 1995 Content-Identifier: 1633 Alternate-Recipient: Allowed From: Blecha Andre To: ccd@eso.org Subject: Site/Tektronix 2048 thinned AR coated chip Hello, SITE announced a "Year end product offering" for large Tektronix chips including 2048 THINNED AR coated chip (ref: SI 424A). Any information both on this program and on experience with recent version of this specific chip is welcomed. Thanks Andre Blecha [J{Mail}& From blecha at scsun.unige.ch Fri Nov 10 12:31:37 1995 From: blecha at scsun.unige.ch (Blecha Andre) Date: Thu Jul 29 11:54:09 2004 Subject: Site/Tektronix 2048 thinned AR coated chip Message-ID: <1633*/S=blecha/OU=scsun/O=unige/PRMD=switch/ADMD=400net/C=ch/@MHS> Hello, SITE announced a "Year end product offering" for large Tektronix chips including 2048 THINNED AR coated chip (ref: SI 424A). Any information both on this program and on experience with recent version of this specific chip is welcomed. Thanks Andre Blecha From Richard_Bredthauer at ccmail.lfist.loral.com Thu Nov 9 09:43:07 1995 From: Richard_Bredthauer at ccmail.lfist.loral.com (Richard_Bredthauer@ccmail.lfist.loral.com) Date: Thu Jul 29 11:54:09 2004 Subject: Sharing of bias phases / Jean-Charles Message-ID: <9510088158.AA815893103@ccmail.aer.loral.com> Jean-Charles Regarding your glowing phenomena on the CCD. Grounding VRD and VOD is a classic error. An unused VRD should never be grounded at 0V. It will then act as a current source and dump charge into the buried channel. To eliminate the injection VRD should be biased positively at a voltage greater than the VRD used for setting the channel potential. In this way no excess charge can be injected. It is not a short. Dick From prj at mail.ast.cam.ac.uk Mon Oct 30 15:13:01 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:09 2004 Subject: Junk mail- not these groups Message-ID: Hi everyone, >From various responses of others, it seems that my junk mail has not come from the CCD groups- since others do not get them. Cheers, Paul ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From prj at mail.ast.cam.ac.uk Mon Oct 30 09:25:31 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:09 2004 Subject: Junk mail!!! Message-ID: Hi Tim, Alain, & everyone, I seem to be getting unsolicited junk e-mail, at an increasing rate! I assume that it is coming through these CCD-groups. Does anyone know how to stop it?? Can you arrange that only e-mail from those that subscribe to these groups can get through? If I'm wrong about the origin/path of these nuisance messages, I'm sorry, but would still rather not have them! Cheers, Paul ps. Here's an example that I just got: Date: Sat, 28 Oct 1995 23:15:10 -0400 To: (Recipient list suppressed) From: exd03642@interramp.com (POWER NETO) Subject: FULL INTERNET ACCESS NOW! IT PAY$ TO BE ON LINE!! ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From lesser at as.arizona.edu Fri Oct 20 15:14:57 1995 From: lesser at as.arizona.edu (Michael P. Lesser) Date: Thu Jul 29 11:54:09 2004 Subject: The CCD that wouldn't Message-ID: <9510202114.AA22574@frodo.as.arizona.edu> Hi Tim, I can't agree there is really anything strange with the CCD other than contamination. The images I sent with the CCD show complete and uniform pinning with 15 min of flood. Those at some point in your tests show these condesation patterns. I agree the patterns are not just straight water drops, but that is because the pattern you see is a combination of the condensation effects (opacity, reflection differences, junk on the surface, etc) AND any surface marks produced here before oxidation. These latter marks are now seen because the CCD is not fully charged due to the damage done to the oxide. Even if you uniformly discharged a CCD you would see these marks since the surface has charge variations due to surface irregularities. These all go away in the fully charged condition as the lower charge regions match the fully charged regions. These patterns can also be seen on an undamaged device simply by not charging the surface. I do not understand the dark current effect you see, but since I have so little experience with Orbit devices, I probably am missing something. The usual effect is high dark current in the unpinned condition. I noramlly see higher DC after power cycling which goes away after a few readouts. I have always thought this to be due to charge trapping and eventual discharge. -Mike From bogus@does.not.exist.com Fri Oct 20 11:30:45 1995 From: bogus@does.not.exist.com () Date: Thu Jul 29 11:54:09 2004 Subject: The CCD that wouldn't In-Reply-To: <9510201504.AA20779@frodo.as.arizona.edu> from "Michael P. Lesser" at Oct 20, 95 08:04:20 am Message-ID: Hi Mike, Many thanks for your comments on our woes. In general, I would agree with the picture that you paint, but there are a number of contradictory details that still worry us. > All the flat field and QE problems are associated with condensation which > occurred on the device at CFHT and contaminated the surface. The flats > clearly show water-like streaks and droplets. This most likely happened > during a vacuum leak which occurred in the CFHT dewar during early tests. This was one of our theories too, but the actual structure of the contamination is difficult to explain. Judging by the shape of the trails, we would have had to rotate the dewar back and forth by about 90 degrees around its long axis while the contaminant ran down the surface (and, of course, the trails are quite short so this rocking would have had to have been at fairly high frequency). I expect that this is possible, but it is difficult to visualise. Another point in our favour is that these patterns were present on the images taken in February, but at considerably lower amplitude. (I guess that it is possible that these just represent areas of the CCD which would be more prone to the effects of a contaminant and some condensation event just finished the job.) We are not sure if the CCD was pinned at that time, but the observers did report a decrease in sensitivity over the course of their run (I was still at ESO at the time). The flat field that I have seen is dominated by the brushstrokes which lead me to believe that the CCD was unpinned by that time. > No amount of UV > charging can overcome the positive charge which results from this. Those > areas of the device not contaminated are uniform and have high QE because > they have these bonds tied up by the oxide. And yet the ends of the trails, where one might expect the contamination to be to be the worst, pin quite well. Moreover, we *have* been able to almost completely pin the CCD, once, after the first 3 hour flood. Also, we still do not understand how the dark current could be so high after the system power was cycled, rather than after a UV flood, but only when the CCD is near-pinned (the dark current doesn't appear to occur in the unpinned CCD). A very puzzling CCD... Thanks again, Tim -- Timothy M. C. Abbott, Resident Astronomer, tmca@cfht.hawaii.edu http://www.cfht.hawaii.edu/~tmca/tmca.html Canada-France-Hawaii Telescope, P.O. Box 1597, Kamuela, HI 96743 From lesser at as.arizona.edu Fri Oct 20 09:04:20 1995 From: lesser at as.arizona.edu (Michael P. Lesser) Date: Thu Jul 29 11:54:09 2004 Subject: The CCD that wouldn't Message-ID: <9510201504.AA20779@frodo.as.arizona.edu> A few comments regarding Tim's note about the CFHT CCD. All the flat field and QE problems are associated with condensation which occurred on the device at CFHT and contaminated the surface. The flats clearly show water-like streaks and droplets. This most likely happened during a vacuum leak which occurred in the CFHT dewar during early tests. The effect of the contamination is to break the surface bonds at the Si-SiO2 interface. No amount of UV charging can overcome the positive charge which results from this. Those areas of the device not contaminated are uniform and have high QE because they have these bonds tied up by the oxide. If the contamination can be removed (with a solvent spray as CFHT tried) sometimes the devices will recover. Often the contaminate attacks the oxide and the CCD must be stripped. We have seen this with water condensation (which often can be fixed) and with oil or oil/water condensation (which usually cannot). The "brush" strokes are another indication of the same problem - an uncharged and unoxidized UV floodable CCD will show all type of flat field variations, dependent somewhat on the history of the thinning process. Brush strokes are an artifact of cleaning the device before oxidation. These features all disappear after charging. This again indicates the oxide is damaged and not charged fully (or pinned). A 3 hours UV flood is too long and almost always causes increased dark current. As in my home page, the recommended flood is 15 minutes. The photogenerated electrons produced by the intense UV get trapped in the channel stops, at the oxide interface, and at the AR coating interface. I hope these comments help... -Michael Lesser From lesser at as.arizona.edu Fri Oct 20 09:04:20 1995 From: lesser at as.arizona.edu (Michael P. Lesser) Date: Thu Jul 29 11:54:09 2004 Subject: The CCD that wouldn't Message-ID: <9510201504.AA20779@frodo.as.arizona.edu> A few comments regarding Tim's note about the CFHT CCD. All the flat field and QE problems are associated with condensation which occurred on the device at CFHT and contaminated the surface. The flats clearly show water-like streaks and droplets. This most likely happened during a vacuum leak which occurred in the CFHT dewar during early tests. The effect of the contamination is to break the surface bonds at the Si-SiO2 interface. No amount of UV charging can overcome the positive charge which results from this. Those areas of the device not contaminated are uniform and have high QE because they have these bonds tied up by the oxide. If the contamination can be removed (with a solvent spray as CFHT tried) sometimes the devices will recover. Often the contaminate attacks the oxide and the CCD must be stripped. We have seen this with water condensation (which often can be fixed) and with oil or oil/water condensation (which usually cannot). The "brush" strokes are another indication of the same problem - an uncharged and unoxidized UV floodable CCD will show all type of flat field variations, dependent somewhat on the history of the thinning process. Brush strokes are an artifact of cleaning the device before oxidation. These features all disappear after charging. This again indicates the oxide is damaged and not charged fully (or pinned). A 3 hours UV flood is too long and almost always causes increased dark current. As in my home page, the recommended flood is 15 minutes. The photogenerated electrons produced by the intense UV get trapped in the channel stops, at the oxide interface, and at the AR coating interface. I hope these comments help... -Michael Lesser From tmca at cfht.hawaii.edu Thu Oct 19 18:38:08 1995 From: tmca at cfht.hawaii.edu (Tim Abbott) Date: Thu Jul 29 11:54:09 2004 Subject: The CCD that wouldn't Message-ID: <9510200338.AA06819@malama.cfht.hawaii.edu> Hello all, At CFHT, we're currently seeing a range of bizarre behaviour from our newest CCD and I thought I might pass it around a bit to see if anyone has any insight on what is going on. This is a 2048^2 Orbit CCD, thinned by Mike Lesser. There was a problem with the initial AR coat and it was stripped and recoated before being shipped. Mike was apparently able to pin it successfully. The CCD was initially run at CFHT in February this year and a successful, but uncalibrated, pin appeared to degrade over the course of about a week. The flat-field had the general appearance of horizontal and vertical "brushstrokes". The device was then shelved for a few months while we brought a new (Leach) controller on-line. We are now trying to bring the CCD on-line again to become our workhorse device. Unfortunately, we are having a hard time acheiving a uniform pin. We are using a Cadmium lamp as UV source. These are the observed phenomena: 1) After a 45 min UV flood in medical O2 atmosphere, there is a large amount of structure in the flat field, typical of an unpinned or poorly pinned thin CCD, but not like the data from February. This flat field appeared stable over the course of a week. (a sample flat-field can be seen in http://www.cfht.hawaii.edu/~tmca/orbit1.html, click on "image 1"). 2) We were unable to remove this partial pinning - even after the CCD was kept warm (but in vacuum) for a week and after soaking in H2 gas for 1.5 hrs, it was still only slightly degraded. 3) Since the flat field had the appearance that one might expect from a contaminant deposited by a condensation event on loss of vacuum, we washed the CCD in Isopropyl and rinsed with ethanol. This had no significant effect. (The same patterns are visible, but only very faintly, in the data from February). The dewar was thoroughly cleaned at the same time. 4) After a 3 hour UV flood, the flat-field was significantly smoother with only about 1/3 of the imaging area compromised. However, there was significant (1-0.1 electrons/pixel/sec) dark current which took several hours to dissipate. 4) The same dark current reappeared after the control system was power cycled. The initial form of the dark current was donut-shaped - ie. a bright ring centered on the CCD with a dark center and corners (possibly representing the temperature profile of the CCD?). As it decayed, over the course of 5 hours, it did so non-uniformly, with a bright blob appearing at upper right after about 2 hours, which had disappeared within an hour. 5) A small dark current appeared when the system was turned off for ~6 minutes, prompting the hypothesis that it is dependent on the system being powered down. 6) Therefore, we attempted a further 3-hr UV flood, this time with the control system turned on and idling such that the CCD was continually flushed. This appeared to severely degrade the pinning, but not to remove it entirely. 7) On the theory that cycling the CCD clocks during the warmup and UV flood caused the loss of pin, we tried another UV flood, for 4 hours with the system turned off. This appeared to make little or no difference. (a sample flat-field can be seen in http://www.cfht.hawaii.edu/~tmca/orbit1.html, click on "image 2"). Could we have "overpinned" the CCD...? So our questions are: Does any of this sound familiar to others with experience of pinning CCDS? Does anyone have any suggestions of how we might pin the CCD? The dark current that appears on cycling system power is completely novel to us. Might anyone be able to suggest a mechanism? Finally, is "overpinning" a realistic concept and how might it fit in with the current theories of the pinning mechanism? Many thanks for any replies, Tim Abbott On behalf of the CFHT detector group. -- Timothy M. C. Abbott, Resident Astronomer, tmca@cfht.hawaii.edu http://www.cfht.hawaii.edu/~tmca/tmca.html Canada-France-Hawaii Telescope, P.O. Box 1597, Kamuela, HI 96743 From stubbs at welkin.astro.washington.edu Tue Oct 17 02:56:17 1995 From: stubbs at welkin.astro.washington.edu (Chris Stubbs) Date: Thu Jul 29 11:54:09 2004 Subject: Sharing of bias phases Message-ID: <9510170856.AA12413@welkin.astro.washington.edu> greetings to all. Regarding the issue of crosstalk on a CCD when running with multiple outputs, the MACHO camera system runs with dual mosaics of 4 2k chips in each focal plane (32 Mpix. total). We run two outputs per chip and have acquired a fair bit of experience in dealing with the crosstalk. An early paper describing our approach is SPIE Vol 1900 pg. 192 (1993). The PhD thesis of one of the project students will be avail. soon. cheers, chris stubbs. From cuilland at srvdec.obs-mip.fr Mon Oct 16 19:10:42 1995 From: cuilland at srvdec.obs-mip.fr (Jean-Charles Cuillandre) Date: Thu Jul 29 11:54:09 2004 Subject: Sharing of bias phases Message-ID: <9510161610.AA00430@srvdec.obs-mip.fr> Hello, My experience is all about the loral 2k chips we used for MOCAM@CFHT(4x2k2k). Since the beginning it was clear for us that we had to run separately the two outputs, we even had to customize the SDSU analog board to get 2 VOD as high as 25V. During the optimization, it appeared indeed that both outputs didn't have the same optimal voltages. However we use only one output per chip as the analog board has only one video channel: we "shut down" the unused output by setting VOD=VRD. The only hard link between the two outputs is for the reset gate. The crosstalking in our mosaic was limited to capacitive coupling in the wiring of the four independent channels, and was eliminated by setting a delay (3us) after the video dump before starting the integration on the signal; But I have a question for those who use the same voltage for different outputs, what is the real amount of on-chip crosstalking and is there a way to cancel it? Finally, if anyone could help me to give a name to the following phenomena: on one loral chip, there is a strong glowing coming from, it seems, the corner; but this is not related to the output as I was able to set both VRD and VOD to 0V (and OG and SW) and it does not depend on the exposure time. It appeared that the amount of glowing depends on the time S2 is set high, and the high level of S2. I optimized this chip by setting S2+ as low as possible and changing the serial clocking. This effect was not so obvious at the beginning but this chip got worst and worst and finally stabilized. It is very well removed with a bias frame (glowing from this is visible on the 3 other chips) and the level is low (peak #70ADUs). Any idea of what happened, some kind of short? Regards, Jean-Charles. ----------------------------------------------------------- Jean-Charles Cuillandre Tel: (33) 61 33 28 29 Observatoire Midi-Pyrenees 14, Avenue Edouard Belin Fax: (33) 61 53 67 22 31400 Toulouse France E-mail: cuilland@obs-mip.fr jcc@cfht.hawaii.edu ____________________________________________________________________________ From roger at ctiol3.ctio.noao.edu Mon Oct 16 12:33:33 1995 From: roger at ctiol3.ctio.noao.edu (roger smith x294) Date: Thu Jul 29 11:54:09 2004 Subject: Sharing of bias phases Message-ID: <9510161433.AA19349@ctiol3> Reply to ..... > To: ccd-world@ocar01.obs-azur.fr, ccd_plumbers-l@ocar01.obs-azur.fr > Subject: Sharing of bias phases > Date: Mon, 16 Oct 1995 10:43:48 +0100 > From: Olaf Iwert > > I would like to call upon some comments for the following points : > > Most CCD controllers do NOT provide for the possibility to supply each bias > gate (e.g. of a Tektronix 2K CCD) with an independent bias voltage. > > As we know, some voltages especially on those devices do require individual > optimisation of their voltage (e.g. Last Gate per dedicated output), > my question is whether someone sees from the design and semiconductor > physics point of view any problems to drive the VRD voltages with individual > voltage ? > The same question for VDD ? > The same question for VLG ? > > To already prevent some answers I'd like to add that we have done so with > mostly good success (as probably most of you), however my question is more > whether it is against the theory to do so : > VRDA and VRDB are the end points of the same serial register. > In case of simultaneous clocking/resetting through/of output A and B a > potential gradient across the register channel exists which leads to a small > cross current in case of different biasing for VRDA and VRDB ? > > Furthermore I'd be interested to know whether there are observations from the > experimental or theoretical point of view to opt for certain preferred output > combinations in case one combines the voltages of specific output related bias > gates. > > E.g: On a TEK2K device : > Which combination is better ? > > Bias 1 :VDDA, VDDB > Bias 2 :VDDC, VDDD > > OR > > Bias 1: VDDA, VDDD > Bias 2: VDDB, VDDC > > OR any other combination ? > > The same question for combinations of VDRx ? > The same question for VLGx ? > > > I'm interested in ALL comments for all possible CCD types, however do look > forward specifically to comments on TEK2K devices and their users. > > > > =============================================================================== > Olaf Iwert E-mail : oiwert@eso.org > European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 > Optical Detector Lab FAX : +49-89-320-23-62 > Karl-Schwarzschild-Str 2 > 85748 Garching > GERMANY > =============================================================================== Olaf and everyone, We use fully independent bias voltages for each amplifier in the Arcons. Most of our experience is with Tek/SITe quad readout devices (1K, 2K square) and Lorals with 2 outputs. I was grateful for John Geary pointing out to me about a year ago that we might be driving a current through the channel due to VRD differences, since VRD is connected to the channel during reset. This left me puzzled that we do not actually see any effect which can be clearly attributed to VRD differences. I didn't see any change in performance when I connected all VRD's together in one system as an experiment. So I was interested to read David Burt's explanation which was relayed by Paul Jorden: (thanks Paul) > In general one could assume that outputs are independent, and should > operate with different voltages perfectly well. This is especially true > of relatively unsophisticated designs. Usually RD & OD are > reverse-biased with respect to the substrate, and so current cannot > flow between different outputs easily. We have found that having independent voltages is a plus for optimizing linearity and to a lesser extent noise. Although the differences in optimum voltages is usually small for Tek CCDs, we think it is worth having the option for the independent adjustment capability on VDD, VLG and VDD. Even if you don't end up with different settings it is very nice to be able to compare the amplifier you are adjusting with one which remains untouched. The other thing to consider is crosstalk, primarily for VDD, as we discussed in a recent mailstorm. If I was connecting biases in common, I would start by connecting a-b and c-d on the Teks so that the one serial register is driven by the same voltages. Are you interested in our clocks setup too? Each clock pin on the CCD has a separate analog switch so that timing is independent, but some levels are connected to common voltage buffers. We drive all RG's from the same pair of levels. Recently I have taken to running SW from the serial levels. I split the serials into left/right where possible to reduce the capacitive load on a given driver, but usually run them at the same levels. I could use this to adjust SW's independently if need be, since the other serial levels are not that critical. We have buffers fpr each level for each parallel phase and TG for each half of the chip... 16 in total. Again, this reduces the loading on a given driver and allows for one to play around with improving cosmetics, charge injection and full well, and dark current, if you can devise a way of navigating the parameter space efficiently! Roger Senior Electronics Engineer / Manager - Array Controller Projects Cerro Tololo Inter-American Observatory (CTIO), Casilla 603, La Serena, Chile Internet: rsmith@noao.edu Coordinates: 29.54 South, 71.16 West Phone: 56 (51) 225415 Bilingual receptionist (08:30-21:00) Fax: 56 (51) 205342/205212 Direct, or ask receptionist for fax machine Fax(USA): 1 (520) 318-8259 Fax & mail automaticly forwarded from Tucson Nat. Optical Astronomy Observatories (NOAO),PO Box 26732,Tucson AZ 85726-6732 From cdm at ast.cam.ac.uk Mon Oct 16 15:56:00 1995 From: cdm at ast.cam.ac.uk (Craig MacKay) Date: Thu Jul 29 11:54:10 2004 Subject: Sharing of Bias supplies Message-ID: Dear Olaf I am always concerned about cross-talk between outputs in CCDs such as the SITe 2048 x 2048 devices which are intended for low noise, high dynamic range operation. We (at AstroCam with our Antares Quad or 4400 controller) have all bias supplies for each amplifier separate (VOD, VRD, VOG/VLG). You will also want to allow of the possibility that the optimal voltage setting for each amplifier is slightly different for each amplifier. We find that the bigger the CCD the bigger are the differences between each channel. Hope this helps. Craig D Mackay. From prj at mail.ast.cam.ac.uk Mon Oct 16 13:37:59 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:10 2004 Subject: Sharing of bias phases Message-ID: Greetings Olaf & others, > As we know, some voltages especially on those devices do require individual > optimisation of their voltage (e.g. Last Gate per dedicated output), > my question is whether someone sees from the design and semiconductor > physics point of view any problems to drive the VRD voltages with individual > voltage ? > The same question for VDD ? > The same question for VLG ? > > I'm interested in ALL comments for all possible CCD types, however do look > forward specifically to comments on TEK2K devices and their users. > I've spoken to David Burt (the GEC/EEV CCD design expert); here is his information: In general one could assume that outputs are independent, and should operate with different voltages perfectly well. This is especially true of relatively unsophisticated designs. Usually RD & OD are reverse-biased with respect to the substrate, and so current cannot flow between different outputs easily. EEV, in some chips, provide guard rings which can be locally connected to OD; this provides an internal link between ODs (and should be shown on the chip schematic). Such a link would provide an internal path between different corners; it need not cause a problem, although it could cause a small unexpected current drain from/to the OD supplies. So, just a warning, in case some designs have internal links. Here at RGO we have only operated multiple outputs with the same supplies. We have connected unused outputs (of Tek chips) to an alternative 'pull-up' bias rail. I'll be interested to see any other comments. Regards, Paul ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From prj at mail.ast.cam.ac.uk Mon Oct 16 13:41:10 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:10 2004 Subject: IR sensors?? Message-ID: Hi Guys, I know this is basically a CCD domain. But, does anyone have any info on the competition? ie. Anyone know about low-noise IR sensors- in the 5 e- noise region? (Single, or small-pixel arrays). I'd be interested to hear. Thanks. Paul ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374811 \RGO Fax- 374700 From oiwert at eso.org Mon Oct 16 11:54:03 1995 From: oiwert at eso.org (Olaf Iwert) Date: Thu Jul 29 11:54:10 2004 Subject: Sharing of bias phases Message-ID: <9510160954.AA12467@mc14.hq.eso.org> Dear CCD gurus, I would like to call upon some comments for the following points : Most CCD controllers do NOT provide for the possibility to supply each bias gate (e.g. of a Tektronix 2K CCD) with an independent bias voltage. As we know, some voltages especially on those devices do require individual optimisation of their voltage (e.g. Last Gate per dedicated output), my question is whether someone sees from the design and semiconductor physics point of view any problems to drive the VRD voltages with individual voltage ? The same question for VDD ? The same question for VLG ? To already prevent some answers I'd like to add that we have done so with mostly good success (as probably most of you), however my question is more whether it is against the theory to do so : VRDA and VRDB are the end points of the same serial register. In case of simultaneous clocking/resetting through/of output A and B a potential gradient across the register channel exists which leads to a small cross current in case of different biasing for VRDA and VRDB ? Furthermore I'd be interested to know whether there are observations from the experimental or theoretical point of view to opt for certain preferred output combinations in case one combines the voltages of specific output related bias gates. E.g: On a TEK2K device : Which combination is better ? Bias 1 :VDDA, VDDB Bias 2 :VDDC, VDDD OR Bias 1: VDDA, VDDD Bias 2: VDDB, VDDC OR any other combination ? The same question for combinations of VDRx ? The same question for VLGx ? I'm interested in ALL comments for all possible CCD types, however do look forward specifically to comments on TEK2K devices and their users. =============================================================================== Olaf Iwert E-mail : oiwert@eso.org European Southern Observatory Phone & Voice Mail : +49-89-320-06-353 Optical Detector Lab FAX : +49-89-320-23-62 Karl-Schwarzschild-Str 2 85748 Garching GERMANY =============================================================================== From jbeletic at eso.org Mon Oct 9 05:35:18 1995 From: jbeletic at eso.org (J. Beletic) Date: Thu Jul 29 11:54:10 2004 Subject: ESO Optical Detector Workshop - October, 1996 Message-ID: <9510090435.AA07674@ns1.hq.eso.org> --------------------------------------------------------------------------- PRELIMINARY ANNOUNCEMENT OF "CCD" WORKSHOP --------------------------------------------------------------------------- Title: Workshop on Optical Detectors for Astronomy Date: 2nd half of October, 1996 (proposed) Duration: 3 days (proposed) # of participants: 125 maximum Papers on: Optical detectors - primarily CCDs new designs manufacturer products performance specifications high speed, low noise amplifiers skipper amplifiers adaptive optics CCDs Pushing into the UV (improvements in backside passivation and AR coatings) Pushing into the IR (thick, high res. silicon) Avalanche photodiodes Wavelength sensitive detectors Detector controllers Interfacing to scientific instruments Effect of CCD parameters on astronomical science Detector calibration / testbench design Cryostat design - bath, continuous flow and TEC Status reports from the detector groups at the major telescope projects Proceedings: Will be published from camera ready copy provided 2 weeks after conclusion of the workshop. We strongly encourage hardware demonstrations and industry participation. We expect to have representatives of the detector groups from every major observatory in the world. ------------------------------------------------------------------------------ 9 October 1995 Dear colleagues, ESO management has approved our proposal to host another ESO "CCD Workshop". You may have participated in one of the previous two workshops that were held in June, 1991 and October, 1993. Those workshops were valuable venues of information exchange between many of the people working in the field of astronomical CCD detectors. The format of a small workshop limited to 100-125 participants, with no parallel sessions, is a good environment for catching up with the efforts of other groups. Obviously, we have a bias toward astronomy, but fundamentally we want to hear about new developments in any field that could enhance the performance of optical detectors at telescopes. We have revised the workshop title to "Workshop on Optical Detectors for Astronomy" so that we make sure to include some of the new technologies that may be very important in our future (such as the wavelength sensitive detectors). I send out this preliminary announcement for several reasons: 1) To see if the proposed date of latter half of October, 1996 is good for most participant's schedules. We choose this date to avoid the airline / hotel saturation that occurs during the Oktoberfest, which ends early October. Please inform me if you are aware of conflicting conferences or meetings. The only two CCD conferences that I know of in the next year are the Cayman Island one in November, 1995 (which not everyone can justify) and the IS&T/SPIE symposium chaired by Blouke, Williams, Anagnostopoulos and Lesser in Jan-Feb, 1996. If there is another CCD conference in second half of 1996 that would conflict, please let me know. (Or let me know if there is an important meeting date, e.g. Gemini telescope meeting, that could swallow up a lot of the potential participants.) 2) To get feedback on content and length of the workshop. We propose a 3 day workshop this time, instead of the 2 days for previous workshop. Basic topics are listed above. 3) To get an estimate of interest level. (Please also inform us if families or significant others may accompany you and you wish for us to arrange tours / activities for them.) 4) To get participants to reserve their conference slot and travel money. We will have a number of invited talks from the major projects and the manufacturers and we will publish the workshop papers, both oral and poster presentations. We strongly encourage hardware demonstrations if you can bring your system and we will try to support a full range of video or computer demonstrations (we can play NTSC). With time set aside for informal interaction in cozy environments, we hope that everyone will find this to be a very useful workshop. Since we will hold the workshop at ESO Headquarters in Garching and we wish to create an intimate atmosphere, we will limit participation to a maximum of 125 persons. A more complete E-mail announcement will be sent in the coming months. The formal written announcement will be mailed in March, 1996. Please provide us with your mailing address if you wish to receive an announcement. Please send comments to me at jbeletic@eso.org. Ignore the vacation message that you will get this week, as I am gone on travel in a few hours time. I hope that most of you can make it. I look forward to helping to host you with an enjoyable and productive workshop. Best wishes, Jim Beletic Members of the local organizing committee: James Beletic Olaf Iwert Sandro D'Odorico Elena Zuffanelli, conference secretary . ========================================================================= James W. Beletic E-mail: jbeletic@eso.org European Southern Observatory Phone: +49-89-320-06520 Karl-Sch-Str 2, D-85748 Garching, GER FAX: +49-89-320-2362 ========================================================================= From tinbergen at ksw.rug.nl Fri Sep 29 11:41:14 1995 From: tinbergen at ksw.rug.nl (tinbergen@ksw.rug.nl) Date: Thu Jul 29 11:54:10 2004 Subject: comments requested Message-ID: <199509291057.AA23361@uwila.cfht.hawaii.edu> ++++++++++++++++++++++++++++++++++++ + From: Jaap Tinbergen + + Kapteyn Observatory + + Mensingheweg 20 + + NL-9301 KA Roden + + Netherlands + ++++++++++++++++++++++++++++++++++++ OUR OUTWARD INTERNET CONNECTIONS ARE COMPLICATED AND SOMETIMES GENERATE DISINFORMATION. PLEASE IGNORE THE "FROM" OR "REPLY TO" ENTRIES AND USE ONE OF THE ACCESS OPTIONS LISTED AT THE END (PREFERABLY THE INTERNET ONE). ===================================================================== (file: fast-ccd.txt) Sept 27 1995 Dear CCD-world, I am trying to assess the practical feasibility with present and near-future CCDs of a mode of operation that is very different from most of CCD-astronomy. If any of you have any views on the matter or can point my way to fruitful approaches, please email to TINBERGEN@KSW.RUG.NL. Any form of cooperation will be welcome; our small optical instrumentation group is being merged into the Dwingeloo radio establishment and it is not clear how much optical development will be possible in the near future, but it IS clear that small is beautiful, so any big-brother backing or sympathy would be at a premium. I shall explain the basic idea in a few sentences, after which I refer you to my paper in the Hague IAU Symposium 167 (p. 197), where I speculate about a few of the important details. If desired, I can send you a paper reprint, or I can email you the text (ASCII) and figure (Latex) files. ------------------------------------------------------------------------ The intended regime is high photon-flux, i.e. from a few times noise-equivalent signal to close to saturation (both within individual frame exposures of millisecs to a few secs at most). By rapid modulation between 2 states of a polarization modulator and synchronized readout of the entire CCD frame, imaging polarimetry should be possible to very high precision, limited only by the SHORT-TERM stability of the entire opto-electronic arrangement. The crux of the proposal is that THE SAME PIXEL records the fluxes in both states of the optical modulator, while everything else is common and constant EXCEPT the optical modulation of the polarization (for which there are very `pure' solutions); photomultiplier polarimeters have worked to a precision of 10^-5 easily, 10^-6 in some cases, and photomultipliers are notoriously unstable devices. With such an imaging polarimeter, robust very-high-precision SPATIALLY DIFFERENTIATING photometry (of images, spectra, interferograms, or any other `object') can be implemented by adding suitable, passive and very stable, polarization fore-optics. Large telescopes and long total integrations (co-addition of millions of the two alternate frames that correspond to the two phases of the modulator) on comparatively bright `objects' would allow `objects' of very low contrast to be photometered in the spatially-differential sense. For worked-out examples, see the paper. The required optics are fairly complex but feasible, the queries are with the detectors (CCD and PtSi?). -------------------------------------------------------------------------- The CCD requirements are: frame rates of several tens to several hundreds per second, readout noise of less than about 20 electrons, (short-term) stability of the `gain' of 10^-5 (ball park), linearity unimportant (i.e. even 10% acceptable if stable), dark current similarly unimportant unless huge and (short-term) unstable. Number of pixels, QE, spectral range and system dynamic range will have to follow from the above requirements and whatever is available, but my impression from manufacturers' leaflets is that useful combinations are on the market. Stability of the ratio of integration times of the alternate frames is of course essential, as is absence of parasitic (modulator frequency) variations of supply voltages leading to varying CTE etc at very low levels; anything that adversely affects the equality of the `gains' for the alternate frames is poison. Though possibly the same general class of CCDs will have to be used as are needed for adaptive optics, the emphasis in the requirements is different. The first experiment to perform on any candidate CCD and its electronics would, I suppose, be to examine the light from a well-stabilized lamp totally devoid of ripple and find out to what precision the long-term average of the difference of alternate short-exposure frames is indeed zero. But to even think about trying that, I must organize some moral support. Hence this circular. All comments are welcome, even if destructive (would save me a lot of work): have I overlooked something vital? has it all been done before? any secret knowledge or insights as to why it can't possibly work? Alternatively, any encouragement you feel is in order (in The Hague, d'Odorico in his summing-up said: "Thank you for trying". I need something more specific than that). N.B. -- I know about drift scanning; `my' method could go to even lower contrast, I expect. It could perhaps also be crossed with drift scanning; it is also arguably more robust, not depending on any synchronized mechanical and electronic scanning or sky transparency variations . And it is basically an imaging polarimeter, which is nice. -- Flatfielding to 10^-5 as an alternative seems pie in the sky. Needs absolute accuracy of that order, rather than differential precision. -- Ensemble photometry and similar relative calibrations by overlapping frames only relax the flatfielding requirements slightly, as far as I can see. -- I know about charge-sloshing as an alternative to complete readout of alternate frames and computer co-addition. Charge-sloshing is too sensitive to CTE to be a robust method for very low contrast and it is for much lower photon fluxes (single-frame saturation within the total exposure; I am aiming for precision of 10^-5 or thereabouts, so 10^10 photons per pixel in the total exposure time; so of order 10^5 saturation frames are needed for one --- polarimetric or differential-photometric --- `image'). THAT'S IT; THANK YOU FOR LISTENING. JAAP TINBERGEN ===================================================================== Internet : tinbergen@ksw.rug.nl --------------------------------------------------------------------- | Until October 10 1995 | After October 10 1995 | Phone : 31 (0)5908 28815 my extension 31 (0)50 5028815 28888 secretariat 31 (0)50 5028888 | TeleFAX : 31 (0)5908 28800 | 31 (0)50 5028800 | ========================================================================== From brucet at pcnet.com Tue Sep 26 09:31:19 1995 From: brucet at pcnet.com (Bruce E. Truax) Date: Thu Jul 29 11:54:10 2004 Subject: Still More Loral Devices Available! Message-ID: Dear CCD Aficionados, A couple of months ago I ran a notice concerning some surplus Loral CCD devices which the EOST group had from its recent wafer run. We had some interest in these devices but we still have a number of Good and Excellent quality devices remaining. If you are interested please contact me at brucet@pcnet.com. The text from the original posting follows. The Edgar Smith Telescope group has had a very successful run of Loral 2048x2048 3-edge buttable CCDs. We have a number of excellent and good devices (John Geary's grading system) and we may even have 2 or three flawless devices remaining after we satisfy our needs. It would be a shame to see these devices not be used, plus we would like to recover some of our costs for the wafer run so we would like to offer our surplus devices to the CCD community. The prices and quality of the devices are given below. If you have any questions please contact me at brucet@pcnet.com. Front Illuminated Grade E+ E G N/A $35,000 $20,000 Back Illuminated (Thinned) Grade E+ E G $120,000 $85,000 $55,000 - Bruce Truax From psinclai at eso.org Fri Sep 22 17:57:46 1995 From: psinclai at eso.org (Peter Sinclaire) Date: Thu Jul 29 11:54:10 2004 Subject: ccd contamination In-Reply-To: ; from "Geoff Evans" at Sep 15, 95 6:06 pm Message-ID: <9509222153.AA09188@mc3.hq.eso.org> > Two of our thinned , back-illuminated ccds have changed colour on cooling - > instead of the usual grey of the anti-reflection coating this has > become visibly green. On warmup the chip reverts to its usual colour. > One of these chips is back in service after cleaning the cryostat and > vacuum backing the chip. > Does anyone have any ideas as to what could cause this problem? > > Geoff Evans > SAAO Dear Geoff, could you provide also some more detailed information on the effect on the CCD performace after this colour change? And, what type of CCD, UV flooded or not, any suspected contaminants. Did the colour change happen gradually with time? I guess after the baking the problem disappeared.. (moisture?) cheers, Peter -- ============================== E.S.O. ================================= Peter Sinclaire e-mail: psinclai@eso.org (Internet) Optical Detectors, TRS, La Silla European Southern Observatory Tel: +56 (2) 699-3425 P.O.Box 19001, Santiago 19, CHILE Fax: +56 (2) 695-4263 ++++++++ Over 19 million pixels serving science every night! ++++++++++ From apo at ast.cam.ac.uk Tue Sep 19 15:12:51 1995 From: apo at ast.cam.ac.uk (Paddy Oates) Date: Thu Jul 29 11:54:10 2004 Subject: More Loral/EEV Qe data Message-ID: Folks, Hi again, in case you thought I had gone AWOL. First thanks to one and all - Roger(CTIO), Craig(IoA, AstroCam), Stuart(UCSB), Peter(ESO), Jim(ESO), Roland(ESO), Mike(Steward) for the response to my Xtalk question. I now have the chip reading out to both FETs but have not quantified performance yet... We will be taking delivery of our 4kx2k EEV devices later in the year so the problem will become gradually more important for these larger devices. For the present it is planned to get the Loral chip out to the WHT for November with single channel operation, going as fast as good, low noise performance will let us - hopefully around 90 secs for the full frame. Will keep the Xtalk responses handy as I am working through our system. I have put our most recent loral (W10-1) and EEV 15/11 response data on the Web. as usual its at - http://www.ast.cam.ac.uk/~apo/docs/loral-1.html W7-4 http://www.ast.cam.ac.uk/~apo/docs/loral-2.html W10-1 http://www.ast.cam.ac.uk/~apo/docs/eev-1.html 1st 15/11 http://www.ast.cam.ac.uk/~apo/docs/eev-2.html 2nd 15/11 You might also find one or 2 pics of me in the warm Antiguan waters, diving after getting my PADI ticket earlier this Summer. Now all we need is for some-one to organise a conference, somewhere pleasant at the end of the year so it can happen again - not much chance of that I guess..... Paddy -- Imagine that Cray computer decides to make a personal computer. It has a 150 MHz processor, 200 megabytes of RAM, 1500 megabytes of disk storage, a screen resolution of 1024 x 1024 pixels, relies entirely on voice recognition for input, fits in your shirt pocket and costs $300. What's the first question that the computer community asks? "Is it PC compatible?" -- E-mail: apo@ast.cam.ac.uk Machine: Soulianis.ast.cam.ac.uk Web Home Page - http://www.ast.cam.ac.uk/~apo CCD Home Page - http://www.ast.cam.ac.uk/~apo/docs/ccds.html Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 direct... -374836 Fax... -374700 Phone: +44 (0) 1954- 261019 Home -- From gpe at da.saao.ac.za Fri Sep 15 18:06:19 1995 From: gpe at da.saao.ac.za (Geoff Evans) Date: Thu Jul 29 11:54:10 2004 Subject: ccd contamination Message-ID: Two of our thinned , back-illuminated ccds have changed colour on cooling - instead of the usual grey of the anti-reflection coating this has become visibly green. On warmup the chip reverts to its usual colour. One of these chips is back in service after cleaning the cryostat and vacuum backing the chip. Does anyone have any ideas as to what could cause this problem? Geoff Evans SAAO From prj at mail.ast.cam.ac.uk Thu Sep 14 15:30:21 1995 From: prj at mail.ast.cam.ac.uk (Paul Jorden) Date: Thu Jul 29 11:54:10 2004 Subject: Electronics (CCD) job at RGO Message-ID: Here are details of a job that has recently been advertised in the press. If anyone is interested, then get in touch. Please distribute around as you wish. If anyone objects to having (relevant) job adverts here, then say so. Regards, Paul ------------- THE ROYAL GREENWICH OBSERVATORY, Cambridge ELECTRONICS ENGINEER The Royal Greenwich Observatory (RGO) has a vacancy for an Electronics Engineer to play an important role in the design and operation of the RGO's low- light CCD imaging systems. The RGO Technology Division is responsible for the development of advanced telescope, instrument and detector systems for the UK telescopes on La Palma. It is also carrying out a programme of development and construction of systems for the new 8 metre Gemini telescopes on Hawaii and in Chile. Applicants must be professional electronics engineers, have an Honours degree in an appropriate discipline and should have a minimum of 2 years relevant experience. They must have experience of the design and programming of microprocessor based systems and also very low noise analogue signal electronics. Experience of the design and operation of CCD imaging systems is desirable. Experience of high-performance instrumentation including optical and electronic systems would be valuable. The ability to work in a team and liaise with other scientists and engineers is important. The appointment is in the grade of Higher Professional Technology Officer (HPTO) and is fixed term for five years. Starting salary will be in the range of #13,025pa to #18,911pa. Further performance related increases to a maximum of #23,009pa are possible. Assistance with relocation to the Cambridge area may be offered. For further information and an application form please contact, Miss J Adams, The Royal Greenwich Observatory, Madingley Road, Cambridge CB3 0EZ Tel. 01223 374000 ext. 4865 (jadams@ast.cam.ac.uk). Please quote job reference number 24/95 Closing date 28 September 1995 The PPARC has an equal opportunities policy. ------------------------------------------------------------------- E-mail: prj@ast.cam.ac.uk Machine: Kria.ast.cam.ac.uk Royal Greenwich Observatory, Madingley Road, Cambridge, CB3 0EZ, UK. Phone: +44 (0) 1223- 374000 \direct phone- 374812 \RGO Fax- 374700 From rreiss at eso.org Thu Sep 14 08:42:59 1995 From: rreiss at eso.org (Roland Reiss) Date: Thu Jul 29 11:54:10 2004 Subject: crosstalk Message-ID: <9509140642.AA04439@te15.hq.eso.org> > Jim, That is an impressive result! What changes were necessary to get > this 25 times improvement? I am eager to hear what Roland measures with > a CCD. Well, the results obtained with a 4-output CCD (TK1024) gave crosstalk results between 1:10.000 and 1:20.000. The 1:10000 result came from two channels which have their unbuffered video lines running in parallel for several ten mm along a flexible PCB. We expect results in the order of 20.000 or better by improving the layout of this PCB. > > Roland, how exactly did you do this test? ...Did you inject the stimulus > at the CCD socket and thus measure the effects of coupling in > between the video signals? I injected a signal generated by the CCD clock driver with an impedance of 1000 Ohms and amplitudes between 1..100mV into one of the preamp inputs at a time. All other channels were terminated to ground with 1000 Ohms. The signal has 'half pixel' frequency (it is in fact generated by a DSP opcode 'Bit Change' which toggles a bit) and yields black/white/black/white... columns across the entire chip. By averaging lines together in the quiet channels one can see crosstalk below the noise level of the channel (amp. noise, quantization noise etc.) By shorting/interrupting the video signals at different locations in the video chain I was more or less able to find out where most of the crosstalk came from. As Jim said it was initially much higher. Surprisingly the most dominant source of CT were the Crystal ADCs! Why? The ADCs shared a common voltage supply (+/-5V) and the inputs are protected by diodes because the other stages were supplied by +/-12V. I guess that the crosstalk was caused by capacitive coupling through the diodes which shared common rails of finite impedance. crosstalk became stronger when the integrator outputs went over 5.5V and the diodes started to 'protect', probably through slight modulation of the supplies. The effect could be drastically minimized by lowering the supply voltage of the integrators to +/-8 V. I have further improved the decoupling of the ADC supply voltages but I haven't had a chance to test it yet, the revised board is currently being produced. I haven't made tests yet with injection directly to the CCD socket, but it is clear that we have problems with the PCB layout between CCD and preamps. One of the next steps could be to try shielded and properly separated cables instead of a PCB to assess the real contribution of the CCD. -- Roland Reiss rreiss@eso.org From stuart at lensing.physics.ucsb.edu Wed Sep 13 10:12:17 1995 From: stuart at lensing.physics.ucsb.edu (Stuart Marshall) Date: Thu Jul 29 11:54:10 2004 Subject: multi channels and crosstalk Message-ID: <9509131712.AA04734@lensing.physics.ucsb.edu> Hi Paddy and all, We run 8 2k x 2k Loral 2EB chips at Mt. Stromlo on the macho microlensing search. Each chip has a split serial register so we have 16 outputs which we read out in parallel in about 70 seconds with 100 overscan in each direction at ~30us per pixel. Roger Smith's comments about the clock sequences are the place to start. If possible, use an engineering chip (or no chip) and go over the clocks with an oscilloscope and the autocad map or the equivalent. If you need it I can send our serial timing info. I am writing because we have considerable experience with cross talk. A dubious distinction. Our system is configured as 2 seperate cameras at prime focus each with 8 channels. There is no crosstalk between cameras. Within each camera we find that the cross talk is nearly linear and very stable. We represent the crosstalk with an 8x8 matrix M for each camera with 0's on the diagonal and coupling coefficients off diagonal. We try to recover the true signal vector p from the measured vector p' (of 8 pixels) using p = (1 - M)p'. We find the coupling coef's have a median (mean) magnitude of 1.16 (2.05) in units of parts per 10000. So a signal of 10000 generates a mean blip of about 2 counts. This is a little more than reported by Roger Smith. Also we have both positive and negative crosstalk. We have a couple of large coefficients at about 9 parts per 10000. The coefficients which are larger than 2 or 3 really do need correction. This is a single user, dedicated telescope/camera system that is in use every clear part of every night. On occasional misty, soupy nights we measure crosstalk and do domeflats etc. To measure the crosstalk we stop the 50 inch aperature down to about 4 inches and point at Sirius. A couple second exposure puts a single saturated source in the focal plane. The response excursions can clearly be seen and measured in the other channels. Moving the source around in the focal plan probes all the couplings. We have been doing this every 6 months or so for almost 3 years and the values are stable to 1 part in 10000 (1 unit). In March 1995 we replaced 2 CCDs in one of the cameras and when we measured the crosstalk we found it unchanged so it is not just the individual CCDs that are the source. Also the matrices for the two cameras are very similar with the largest coefficients in the same positions. During our first few months (1992) of taking data the crosstalk was about a factor of 5 larger and really unacceptable. We redesigned our preamps and cabling scheme to arrive at the current situation at the start of 1993. We use differential output preamps to drive (7m) twisted pairs to the photometrics A/D units mounted near the mirror cell. The remaining crosstalk is probably somewhat avoidable with some changes in the wiring inside the cameras. We are using cryogenic stainless steel coax between the CCD outputs and the dewar wall. I would much rather have shielded twisted pairs in there but we were able to keep taking data by correcting the remaining crosstalk as part of our flatfielding process. We rolled our own software to make calibration images and to do the flatfielding, dark subtraction, and crosstalk removal. The image calibration is done on each incoming image as it is written to disk and before archiving. We also have an un-flatfielder to undo and redo images taken when a new set of flats is needed but not available. Anything to keep the data flowing into the datbase. Our system is now a bit old but if it is any help we could send more comments (and horror stories). Stuart From cdm at ast.cam.ac.uk Wed Sep 13 18:00:51 1995 From: cdm at ast.cam.ac.uk (Craig MacKay) Date: Thu Jul 29 11:54:10 2004 Subject: Crosstalk Message-ID: Dear All I have been interested to read what has been said about crosstalk in multi-output CCDs. I think crosstalk is potentially a problem since it is possible to have bright stars in one quadrant with essentially full wells that one does not want to contaminate other quadrants down to the detection limit which may be very small. A star image may well reach 10e6 electrons with a detection limit against the sky of 10e2 electrons, so factors in excess of 10e4 for crosstalk limits are in order, for stars of even a few pixels. Simulations without a CCD attached are likely to give optimistic results. The sort of problem that arises is if some of the output amplifier supplies are commoned up. This is an increasing problem with 3-edge buttable devices where there is limited space for pins, and AO sensors where there are vast numbers of output channels being developed. The sort of problem that arises is that although the capacitances are very low in output amplifiers, a large current pulse on one output can feed into other channels if the CCD design allows that. Talking to some CCD designers they never thought of this as a problem. Bias supplies that are designed for low-noise low-frequency operation can show high dynamic impedances when they see fast pulses. If you think as another example of a CCD system where the output drain supplies are common, then the critical ratio is the output source impedance divided by the dynamic impedance of the output drain supply to the fast step that arises from a large charge being clocked onto the output node. We all design very low noise output drain supplies but the dynamic impedance is really hard to measure. Using constant current sources in the output source helps, again because the dynamic impedance is high, but resistive loads can be troublesome especially when the new high-speed CCDs are used since they need really low impedanmces to allow the fast pixel rates needed for AO (so I guess I go not agree with Jim Beletic's belief that things are easier in AO since a SNR of at least 10:1 and 100:1 being ideal per star is needed for a Shack-Hartmann sensor, and any crosstalk will translate into low-level centroid offsets). What we do in the AstroCam (formerly Astromed) controllers is to use dummy output stages that simulate the actual CCD impedances and run these as a differential analog stage using the same output drain and analog ground contacts and paying obsessive attention to the grounding, even in single-channel controllers such as the 4200 series. In the 4-channel 4400 we have completely separate supplies for ALL the output amplifier bias, reset clock etc lines so that there is really nowhere that the crosstalk could occur. This may be overkill but we really cannot detect any crosstalk at all in these systems. This system runs at up to 250 KHz per channel, and all four channels are exactly synchronised. In the high-speed 4100 system that goes up to 8 MHz pixel rate the potential problems with crosstalk are harder since the output impedances can be as low as a few hundred ohms and the dynamic impedances with 25 nanosec clock transitions are interesting to say the least. No sign of running out of interesting and challenging problems to solve. The only good crosstalk is that on ccd-world (and in the Caymans!) Craig D Mackay Institute of Astronomy, University of Cambridge, England. tel:+44-1223-337543, fax:+44-1223-337523 and AstroCam Ltd, Cambridge Science Park, Cambridge, England. tel:+44-1223-420705, fax:+44-1223-423021 From cdm at ast.cam.ac.uk Wed Sep 13 18:00:51 1995 From: cdm at ast.cam.ac.uk (Craig MacKay) Date: Thu Jul 29 11:54:10 2004 Subject: Crosstalk Message-ID: Dear All I have been interested to read what has been said about crosstalk in multi-output CCDs. I think crosstalk is potentially a problem since it is possible to have bright stars in one quadrant with essentially full wells that one does not want to contaminate other quadrants down to the detection limit which may be very small. A star image may well reach 10e6 electrons with a detection limit against the sky of 10e2 electrons, so factors in excess of 10e4 for crosstalk limits are in order, for stars of even a few pixels. Simulations without a CCD attached are likely to give optimistic results. The sort of problem that arises is if some of the output amplifier supplies are commoned up. This is an increasing problem with 3-edge buttable devices where there is limited space for pins, and AO sensors where there are vast numbers of output channels being developed. The sort of problem that arises is that although the capacitances are very low in output amplifiers, a large current pulse on one output can feed into other channels if the CCD design allows that. Talking to some CCD designers they never thought of this as a problem. Bias supplies that are designed for low-noise low-frequency operation can show high dynamic impedances when they see fast pulses. If you think as another example of a CCD system where the output drain supplies are common, then the critical ratio is the output source impedance divided by the dynamic impedance of the output drain supply to the fast step that arises from a large charge being clocked onto the output node. We all design very low noise output drain supplies but the dynamic impedance is really hard to measure. Using constant current sources in the output source helps, again because the dynamic impedance is high, but resistive loads can be troublesome especially when the new high-speed CCDs are used since they need really low impedanmces to allow the fast pixel rates needed for AO (so I guess I go not agree with Jim Beletic's belief that things are easier in AO since a SNR of at least 10:1 and 100:1 being ideal per star is needed for a Shack-Hartmann sensor, and any crosstalk will translate into low-level centroid offsets). What we do in the AstroCam (formerly Astromed) controllers is to use dummy output stages that simulate the actual CCD impedances and run these as a differential analog stage using the same output drain and analog ground contacts and paying obsessive attention to the grounding, even in single-channel controllers such as the 4200 series. In the 4-channel 4400 we have completely separate supplies for ALL the output amplifier bias, reset clock etc lines so that there is really nowhere that the crosstalk could occur. This may be overkill but we really cannot detect any crosstalk at all in these systems. This system runs at up to 250 KHz per channel, and all four channels are exactly synchronised. In the high-speed 4100 system that goes up to 8 MHz pixel rate the potential problems with crosstalk are harder since the output impedances can be as low as a fe