Matrix metering chips

Discussion in 'Nikon' started by ro_ro, Aug 5, 2004.

  1. Does anyone have detailed information about the lens protocol used in
    old Nikkor AF and AI-P lenses?

    I wonder if it should not be impossible to make custom chips by using
    a 1$ PIC12C508 or similar. Open source preferred.

    Why that? Nikon doesn't sell chips to everyone, and they are
    expensive. Chips from old lenses don't often fit the apertures of the
    manual lens to be upgraded. An MF lens with an AF chip makes the
    Camera try to AF, not so good. Even if you find a chip with suiting
    apertures, the focal length is different.

    Thats what I found out myself:

    The names of the lens contacts could be found in some old service
    manuals that can be download on the internet. Search for 'n90.pdf'
    with google.

    I recently dissassembled an old Nikkor 35-70/3.3-4.5 to see whats
    inside. The only important thing is a NEC uPD7554G577. The datasheet
    was hard to find, but I now have one, which I could mail to anyone
    who's interested.

    The NEC CPU is an old and slow 4-bit one, so the protocol can't be
    that complicated.

    I drew a sipmle schematic to show the connections.

    But I don't have a logic-analyzer to figure out what data is
    transmitted. Could anyone help?
  2. The UPD7554A seems to be compatible.

    A good datasheet can be downloaded here:

    A link to a do it yourself logic analyzer that uses a PC parallel port.
  4. Hi Ro Ro,

    I can't help you on this, but whatever result you get (doesn't matter how long it takes), please let us know the final version by posting it here/in a new thread. Thanks,

  5. This analyzer software works great ;) I had to use the allowio driver with w2k ('allowio DIGITRACE.EXE' in the command prompt). The buffer was increased to 255000.

    I build a simple buffer with a 74LS14 to protect the camera (Nikon N4004s)

    lens --|>o--|>o-- lpt

    Many byte flying on the lines but slow enough to get cought. The screenshot shows the beginning of transfer.

    I'm busy the next days but will try to figure more out next week. The datasheet schuld tell about the bitorder.

    Perhaps someone else who likes experiments could try the same.
  6. Sorry, I forgot the picture
  7. it sounds like a very good idea, long overdue :)

    Roro - how did you "listen"? did you disassemble an old lens? a spacer tube like an AF macro ring might be great for this. this way different lenses could be tested.

    I would not be surprized if the protocol is really simple. even if not understood the only data we need to know is the max opening and the current opening? perhaps all needed is to fill in the correct data replacing the once found in an test lens. (in case one finds a matching lens it would even be simpler.)
  8. Wonderful, long overdue idea! I just hope this doesn't end with Nikon USA smiting you for your efforts.
  9. to Walter: I disassembled a mechanically defective 35-70 and soldered wires direct to the 180R protection resistors. Extension tubes could maybe also do the job but are more expesnsive than a N4004/N50/N55 with broken shutter oder defective back. I try to get an working AF Nikon with defective shutter and solder wires direct to the inner lens connections.

    Some on the lens data is decribed in the N90s memory map. (LENS BANK)
    I think the bytes are the same.

    To Mike: I should contiue work on old Sigma lenses, the Nikon's databytes may be covered by US copyright. The protocol is standard I think.
  10. Its greek to me, but if it leads to a way to hack a D70 to meter w/ AI lenses my check will
    be in the mail. Unfortunately the lens I truly would like to use (35 1.4) can't be chipped.
  11. Ro Ro,

    How are the efforts going? To me it looks like that 1) is an interupt signalling end of a transfer. 2) is the clock signal and 3) is the actual data.

    If you need some help converting this serial data to parrallel format, send me some data you sampled and I will write a translator (I have done things like that before)

  12. It is written in the winds that great enlightenment will grace the eyes of one who looks upon US patent 4,896,181.
  13. Many thanks for your help. I got further now.

    I made a first test with a PIC (containing my special emulation software ;) connected to a N4004. The timing diagrams shown in the patent were very helpful, but the datebyte described are different from what I discovered. There are 26byte transferred instead to the 16byte in the patent. It works, the camera tries to AF (I took databytes from an AF lens) with only my PIC attatched.

    The bad news: Newer Nikons use a high speed clock, so the PIC cant't follow and no real data is transmitted :(

    The serial port in the NEC chip could handle clock rates of serveral 100kHz. My software serial port could only do 20kHz.

    ToDo: Get a faster micro or one with a hardware serial port. Case should be SO8 or smaller to be put in lenses with almost no space.

    I won't post any software or schematics in public to get not in trouble with Nikon. I only tell what can be done to get a working chip. More in next week.
    I still don't know what most databyte exacly do, but I gussed aperture and focal length.
  14. "don't know what most databyte exacly do, but I gussed aperture and focal length."

    well somewhere must also the "D" info come in. but you will see this by comparison of "D" with "non D" lenses once you got the hardware running.

    by the way where are you located on the globe. perhaps someone sharing the interest is close to you.
  15. 'D' chips are different to the old AF ones. Not only the distance information is transmitted, but also values for AF and other compensation values change.

    This works even on old cameras that use protocol 0x22. Newer cameras that include the 'D' feature use 0x25, 0x27 and 0x28 protocols for the 'D' chip lenses. These newer protocols provide addition information.

    My opinion is that even on older cameras a 'D' lens can provide better AF and metering.

    But for an upgrade of old lenses the 'D' isn't important. So only a basic 0x22 protocol used with the first AI-P lenes is needed.

    I got further using a fast AVR instead of a PIC. Clocked at 8MHz the AVR can handle clock speeds of 100kHz, but current consumption is too high (8mA). My old N4004 has problems with mirror and shutter if a chip takes more than 5mA.
    The old NEC chips take 1mA and the new AF-D ones take 5mA. Another problem is the missing capacitor on the power pins of the micro. AVRs don't work fine without this capacitor. The chips Nikon uses do.
    The old NEC 4-bit line is now obsolete and could not be used for a self-made chip. The new 'D' chip have 48pins are clock at 8MHz and have a Nikon logo, I don't know what's inside.

    I still haven't found a suiting micro and my time is limeted. For now I stop research on make compatible chips.
    If anyone wants to continue my work, please write me an email.
  16. Hello,

    I'm interrested in this topic.
    Does anybody knows a way to reach "Ro Ro" ?

  17. I am also interested in this. Has any progress been made?
  18. Definitively. Customised matrix chips are used in my lenses.
  19. Hi Bjoern,
    can you be more precise. Where are these chips available? Price? Protocol details?
  20. I'd love to know more about the chips too.
  21. The latest chip model allowed me finally to CPU-modify 35mm f/1.4 Nikkor. Runs are still experimental and availability limited.
  22. Who is doing the runs? Will they produce/sell them for us? Price? Contact info? Any other detail?

  23. Hello, To people who understand French, 18 pages of interresting discutions here
  24. Since the French site prevents the Google translation tools from accessing the whole thread, can anyone summarize?
  25. Hello,<br />
    <br />
    1° They managed to program a chip with any lens data.<br />
    2° It works on an old F8008 but too they're too shy to test on F5,D200 and so on.<br />
    3° They didn't manage to make new contact blocks ( hardware part of the project ).<br />
  26. Tested today on F5 it works.
  27. Hi Marc,
    is it possible to provide more informations? Schematics, code....?
    There are many out there who are interested in this.
  28. I used the 16F690 chip from microchip. You can get it in a 4x4 mm format. Now, i'm trying to post an image.
  29. Hi Marc,
    Is there some code available?
  30. For those which are interested in the french thread: Babelfish of yahoo is able to translate the thread. The translation is far from perfect but better than nothing.
  31. I've decided to put up a *very limited* number of customised CPU chips for sale. These can be programmed to exact specifications for any Nikkor lens (including AI/AIS, PC-Nikkor, Reflex-Nikkor etc.). Zoom range can be programmed as well. A modified lens will meter on all cameras with "F" mount and can use colour matrix metering and all operational modes (A, M, P, and S). Sorry no 3D matrix because the "D" info isn't available for a manual lens. I'm offering two basic types, one that can be fit to most lenses such as the 28/2, 28/2.8, 105/2.5, 300/2.8 etc., and a special small type that can - with some patience and braveness - go into 35/1.4 and 85/1.4 or similar lenses which have very limited internal space available for a chip. The chip will be programmed to match your specified lens before shipping. Contact me for prices and availability. This will be an one-time offer only. Some lenses are not currently possible to modify or the risk of damaging the lens in the process is too high (e.g. Noct-Nikkor and 50/1.2). (Shun: if the above violates guidelines, feel free to remove the post).
  32. Hello.
    Does anybody have device programmer's schematics to read or write D7554G firmware?

Share This Page