Tuesday, October 26, 2010

CMOS Sensors for Astronomy Challenges and Opportunities

Richard Crisp published an undated "CMOS Imagers for Astronomy?" presentation talking about challenges in large area cooled CMOS sensor design. There are few things that I do not agree with in this presentation, but it's interesting read for sure. Here is one teaser slide:


  1. I suppose this is the same Richard Crisp who is Director of Applications and Semiconductor Technology at Tessera, according to his LinkedIn bio.

    APS = Amplified Pixel Sensor?? Actually no.

    Seems all the slides are either lifted from Bedabrata Pain or Jim Janesick, as noted on each one. I am just wondering if he recognizes that the CMOS APS technology was invented at JPL for space applications, also the former stomping grounds of Pain and Janesick and myself.

    I am personally glad to see that after 15+ years, the CMOS APS is being applied to space-related scientific imaging applications. Interestingly enough, this is about how long it took for CCDs to reach the same stage of "space maturity".

  2. @ "this is the same Richard Crisp who is Director of Applications and Semiconductor Technology at Tessera"

    Yes, he is. However, I believe this presentation does not belong to Tessera. It does not carry any Tessera logo and it's located at Richard's private web site.

  3. yes same guy: guilty as charged.

    this was from a talk I gave to some amateur astronomers in mid 2008

    they wanted to hear about cmos sensors and whether they would ever be useful for astronomy

    yes the slides largely came from Pain and Janesick, as I had taken a short course they taught at UCLA back in 2007. Given the short time I had to prepare the talk I found their material accessible and dealt with the points I wished to make.

    If I had to redo this talk today, it would be significantly different.

    I am fully aware that it was developed at JPL where all of you gentlemen have or currently work.

    You folks did some pioneering work and if you feel slighted in any way for not noting the history and motivation for the development, it was not my intent: the focus of the talk was where it could go and not so much the history. I would love to give a two hour talk, but I got something like 15 minutes :-)

    I am actually a bit surprised to see this linked here as it was posted on some amateur astronomy chat rooms last week due to several people asking me about it: some had seen the talk when I gave it in mid 2008.

    I know Eric is credited with the inventorship of the original concepts. I hope to meet Eric someday!

    I have heard it called both Amplified Pixel Sensor and Active Pixel Sensor. Perhaps you'd have liked it better were it to have been called Active Pixel sensor?

    Best regards

  4. ISW is right: this was a talk done to a group of astronomy hobbyists as part of my hobby of astronomical imaging. It had nothing to do with Tessera.

    you may enjoy some of my images that are found here:


    there are many more (~2GB) on my website:


  5. I don't feel slighted at all and just amused with JJ's current endeavors, which I support completely. I left JPL in 1996, and Jim left a year or two before me. Pain left a few years ago to pursue Bollywood ambitions (and co-produced an award winning film with his wife).

    APS stands for active pixel sensor, a term coined by Tsutomu Nakamura of Olympus in relationship to their CMD device, but I am often credited with popularizing that term. There was an unfortunate collision with the APS film standard, and now there are APS-sized APS devices.

    There was an NHK sensor termed Amplified MOS imager (AMI). I have never seen amplified pixel sensor.

    Best regards to Sigler and Maloney if you bump into them. (I am also responsible for naming Siimpel Siimpel, even tho it is now Tessera).


  6. I see both Eric and Shawn on a daily basis if I am not traveling in the orient.

    I met Tony Tang and Roman Gutierrez last week at a company meeting. Enjoyed meeting them both.

    I hope we cross paths some day. I think you were involved in the ISSCC back when I was (memory subcommittee chair followed by program committee chair in 2000) but never had the pleasure to meet you in those days...


  7. Dear Richard,
    Your paper on residual bulk image retention/ghost images in CCDs is very interesting. Never thought that carrier trapped in the p-epi/p-sub interface may be released later forming ghost images. Nice!

  8. Well that info came from Janesick's CCD book. I can't claim credit for the theory and there's plenty of evidence that there are other trapping sites within the device layer, so I don't think that's the full story.

    However having said all that, the method for characterizing the RBI and the tie-in to amplifier luminescence are important results from my paper in my opinion.

    For the KAF series sensors such as the KAF09000, the one characterized in the cited work, it is clear that one *must* deal with the RBI for long exposures needed in astronomy and in so doing, it necessitates *far* more cooling than management of dark shot and dark fixed pattern noise alone.

    To me that is another important result. None of the amateur astrocams that I have seen provide sufficient cooling to keep the managed sensor's noise contribution limited to read noise for exposures much longer than 5 minutes.

    I regularly take 15 to 30 minute exposures for example....

    -Richard Crisp

  9. "I don't feel slighted at all and just amused with JJ's current endeavors"

    What are JJ's current endeavors and what is amusing about them? Bollywood as well?

  10. "For the KAF series sensors such as the KAF09000, ......"
    Has this effect (residual bulk image) noticed in CMOS image sensors (in low temp)? In CMOS as well there is the p-epi/p-substrate interface where charges could occlude.
    How about the STI bulk interface? Maybe the trap characteristics is different here...they may recombine and thus have a short life-time, leaving no room for re-emission...

  11. I have no data on that. I have plenty of experience using the KAF series of full frame CCDs

    I have heard that the backside illuminated E2V CCDs have the problem as well and that says that it cannot be solely due to trapping at the epi interface.

    I have not had the opportunity to use any scientific grade CMOS sensors in a cooled long-exposure environment so I cannot add anything to the discussion in that regard specifically. I do understand that image lag is sometimes a problem with charge transfer type CMOS pixels but that is a different mechanism than RBI.

    You may want to refer to Janesick's CCD book to read about what he says about RBI and RSI. That, and some private communications with him and B. Pain, were helpful when I was doing my investigation.

  12. JJ's current endeavors are just a continuation of his CCD and CMOS image sensor work that he started at JPL.

  13. I am pretty sure he didn't do any CMOS image sensor work at JPL.


All comments are moderated to avoid spam and personal attacks.