Tuesday, October 05, 2010

Interview with Foveon Executives

DPReview published an interview with Sigma COO Kazuto Yamaki, Foveon VP for Technology and Operations, Shri Ramaswami and its VP for Strategic Marketing, Rudy Guttosch talking about challenges on new 15.4x3 MP sensor design. The new sensor will be at the heart of the Sigma new SD1 DSLR, which is due in early 2011.

Other than talking about the new sensor design, the article also mentions Foveon's past project of a sensor for mobile phone.

14 comments:

  1. Where is Caver Mean now ? Is he still with Foveon ?

    ReplyDelete
  2. Carver Mead left Foveon about the same time as he started his next venture, Impinj.

    FYI - contrary to the claims in George Gilder's book, Dr. Mead was not involved in the current Foveon technology. The original point of starting Foveon was to build cameras with three-chip prisms and Foveon's first products were monochrome chips for these. Only when this did not work out marketwise did Foveon switch to the layered technology, which had been invented by Dick Merrill at National Semi, a large original Foveon investor and the fabricatior of the early Foveon sensors both monochrome and layered.

    ReplyDelete
  3. With all respect for the late Dick Merrill, he was a great engineer and a nice colleague in the imaging community, but he did not invent the layered photodiode structure to perform color imaging. This invention was done by Wolffenbuttel. I think it was in the '80s and I do not recall which one of the two Wolffenbuttel brothers made the invention, but I am 100 % sure that it was one of them. A.T.

    ReplyDelete
  4. I know of these contributions. Albert, can you email me the Wolffenbuttel reference at your convenience?

    C.F. Gay and R.D. Wieting, "Vertically Integrated Solid State Color Imager," US patent 4,581,625, 1986. (Filed 1983)
    H. Nozaki and T. Adachi. "Color Sensor," US patent 4,677,289, 1987. (Filed 1985)

    ReplyDelete
  5. Unless you mean this later reference:
    R.F.Wolfenbuttel, “Operation of the silicon colour filtering element”, Sensors and Actuators, 1989, Vol.
    16, pp. 13-23

    ReplyDelete
  6. How about "layered photodiode structure used by Foveon". That will save everyone unnecessary work.

    ReplyDelete
  7. "A simple integrated color indicator", Reinoud F WolfenButtel. IEEE SSC Vol sc-22, No.3, June 1987

    "active interference filters using silicon compatible materials" D.P. Peonar, P.J. French and R.F.Wolfenbuttel; 1995 journal unknown

    "an integrated silicon colour sensor using selective epitaxial growth" Bartek, Gennissen, Sarro, French, Wolfenbuttel. Sensors and actuators A, 41-42 (1994)

    A 1996 thesis from delft, "thin film colour sensors" D.P.Peonar promoters Middelhoek and Wolfenbuttel

    and

    Wolfenbuttel R.F. thesis, 1988, "integrated silicon colour sensors" also from delft promoter Middelhoek

    And in the "glass house and rock throwing" category

    "Photosensitivity and scanning of Silicon Image detector arrays" Savvas G. Chamberlain, IEEE journal of SSC vol. sc-4, No. 6, December 1969 See fig 1 in particular.

    Since Merrill developed X3 at Foveon AFTER the initial B&W sensors and well after Foveon had broken away from National. Foveon started out at National but was spun out as an independent company. The development of X3 was a number of years after.

    You can question whether or not Dick knew about Wolfenbuttel and if so when at when at National or later. But the development work clearly happened after.

    The wolfenbuttel approach was clearly not manufacturable.

    Also:
    Look in S.M.Sze book "semiconductor sensors"page 305, section 6.4 visible-light color sensors. Wolfenbuttels thesis mentioned ref 8.

    ReplyDelete
  8. See also US4613895 by Bruce Burkey, filed 1978

    ReplyDelete
  9. Reinvention occurs often. People are generally thinking about the same problems at the same time and come up with the same solutions. One of the problems with the Foveon layered photodiode is that the structure does not fully deplete. The other problem is that the color separation is not significant enough to render equivalent colors at equilavent noise levels compared to more standard approaches. Being able to see the significance of a concept and doing something with it may be more important than invention. Although the CMOS APS was arguably not an invention of JPL, they did see some significance in it and deserve that recognition for pursuing it. The layered diode has not been as significant. Notice that people here say that Dick Merrill did not invent the layered diode. He might argue that point since he had an incremental arrangement.

    ReplyDelete
  10. Some days I feel like a lightning rod.

    "And in the "glass house and rock throwing" category "Photosensitivity and scanning of Silicon Image detector arrays" Savvas G. Chamberlain..."
    and
    "Although the CMOS APS was arguably not an invention of JPL"

    You guys crack me up. As if this is new news. Why not check out my 1997 review paper that includes all of this stuff and more.

    IEEE TRANSACTIONS ON ELECTRON DEVICES, VOL. 44, NO. 10, OCTOBER 1997

    such as:
    [3] S. Morrison, “A new type of photosensitive junction device,” Solid-State
    Electron., vol. 5, pp. 485–494, 1963.
    [4] J. Horton, R. Mazza, and H. Dym, “The scanistor—A solid-state image
    scanner,” in Proc. IEEE, vol. 52, pp. 1513–1528, 1964.
    [5] M. A. Schuster and G. Strull, “A monolithic mosaic of photon sensors
    for solid state imaging applications,” IEEE Trans. Electron Devices, vol.
    ED-13, pp. 907–912, 1966.
    [6] G. P.Weckler, “Operation of p-n junction photodetectors in a photon flux
    integration mode,” IEEE J. Solid-State Circuits, vol. SC-2, pp. 65–73,
    1967.
    [7] R. Dyck and G. Weckler, “Integrated arrays of silicon photodetectors for
    image sensing,” IEEE Trans. Electron Devices, vol. ED-15, pp. 196–201,
    Apr. 1968.
    [8] P. K. Weimer, G. Sadasiv, J. Meyer, L. Meray-Horvath, and W. Pike,
    “A self-scanned solid-state image sensor,” in Proc. IEEE, vol. 55, pp.
    1591–1602, 1967.
    [9] P. Noble, “Self-scanned silicon image detector arrays,” IEEE Trans.
    Electron Devices, vol. ED-15, pp. 202–209, Apr. 1968.
    [10] S. G. Chamberlain, “Photosensitivity and scanning of silicon image
    detector arrays,” IEEE J. Solid-State Circuits, vol. SC-4, pp. 333–342,
    June 1969.

    The JPL inventions are the issued patents. Read them and then argue the actual claims if you want. Obviously I cannot respond to any claim discussion but at least read the literature and get your facts straight. And, pay attention to EXACT wording of claims and not what you think they might say.

    If you find things missing from the review paper that should be included, I am happy to hear about and include it in future publications if I ever get around to writing an updated review paper.

    ReplyDelete
  11. Regarding Foveon and Dick Merrill. A good man and well-funded company that pursued an interesting idea and put it into practice. It would have been better if they had an image sensor expert as part of their team, I think, so that they would better understand the noise implications of both the 3T readout and color processing.

    I think their latest chip sounds like a good accomplishment and hope it does not suffer from too many things (like color dependence on ray angle, kTC noise, etc.).

    I think the Foveon story points to the barriers facing new technology that does not have compelling advantages over existing technology. CMOS image sensors have a compelling power and integration advantage over CCDs, despite initial image quality disadvantage. Foveon's image quality (aliasing) advantage gave it a niche but not a compelling advantage. Time will tell for Invisage and other newcomers.

    ReplyDelete
  12. Eric, you are so modest - pleasure to read

    ReplyDelete
  13. Eric - Thanks for posting these references and your thoughts on barriers facing new technologies.

    ReplyDelete
  14. I just ran across this thread. According to a former member of the Foveon patent committee, neither Dick Merrill nor anyone else at Foveon was aware of Wolfenbuttel's work, which they would have cited if they had known of it at the time.

    ReplyDelete

All comments are moderated to avoid spam and personal attacks.