can you help me to resurrect this valuable laboratory instrument?

Discussion in 'The Projects Forum' started by scary, Mar 11, 2012.

  1. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    Hello everyone, 'd really appreciate some help with this challenge I am confronting. I have a Shimadzu Diode Array Detector in the university lab I run which is rather old yet works absolutely fine. The instrument is used to detect the absorption spectra of a flow of fluid passing through a flow cell inside it for the purposes of characterisation and quantification of analytes. The detector is driven from a full length PCI board in an old pentium-driven PC running Windows 95. The problem is that the software & this PCI board are incompatible with the modern versions of Shimadzu's HPLC software.

    Here are some photos of the communication port on the instrument, the cable and its connector (I think this is a 40 pin Molex, I would like to be corrected) and the same connector port on the PCI as well as some shots of the PCI board inside the PC.

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    I would like to be able to use the detector in parallel with my more modern equipment from the same PC as that running the modern instrumentation. The challenge is to duplicate the control and communication functions of the PCI board and associated W95 software using something like a logic analyser to eavesdrop on communciations between the instrument & PC . I could then interpret this and use an Arduino or similarly programmable microprocessor to mimic the control and communication signals from the host PC. I am already familiar with datalogging with the Arduino and so the interception of the logic signals (I am presuming that communication via that Molex connector is going to be digital, right?) and the interpretation and separation of the control & data transfer functions are the challenge I face.

    Is this even vaguely possible? I can seek help from the school of engineering but they are a busy bunch (and if you need too much of their time, costly) and so I'd like to make as much progress on my own as I can before I go begging for their advice.

    I should mention that Shimadzu will provide no support for this issue beyond offering me a good price for a trade-in for my current instrument. That price contains five figures and is well down my list of priorities for the lab so it would be super-awesome if I could hack out a fix for it without the need to spend megabucks. I have several projects on the go at the moment which would benefit greatly from this instrument's capabilities!

    Many thanks in advance for advice & input.

    SC
     
    Last edited: Mar 11, 2012
  2. crutschow

    Expert

    Mar 14, 2008
    12,977
    3,220
    Do you have an old computer that you can use to talk to the instrument while you eavesdrop on the signals?

    Do you have a listing of the command structure for the instrument that lists the old digital signal protocol as well as the new?
     
  3. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    Hi Crutschow, thanks for your reply.

    Yes, I have the Windows 95 PC mentioned above in which the PCI board is installed. I'm not sure how much of the signal processing this board does so that function might need to be replicated in my solution.
    *edit* Sorry- I might have misunderstood you there. Do you mean a second to eavesdrop on the first? I assume that would require splitting the signal from the instrument two ways. Wouldn't this be done with a logic analyser or is that what you are suggesting?


    I will have a look in the manuals but I imagine not. Newer Shimadzu instruments use fibreoptic cables to communicate with a data module which then communicates via ethernet to the PC running the software interface.
     
  4. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    I have correctly identified the connector as a Parallel ATA connector, not Molex. Will be interesting to see if I can borrow a bus analyser for this format.
     
  5. Sparky49

    Active Member

    Jul 16, 2011
    834
    417
    Hi there. That looks like one cool piece of kit.

    If I've understood, then you are having problems connecting this to a more modern computer? Would running a virtual machine be a possible route?

    I can't wait to see this working. :)

    Sparky
     
  6. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    From the pictures. I would say that the card is a ISA bus card, and not PCI.
     
  7. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    Hi Sparky, thanks for your reply. The problem, if I can clarify, is that the instrument is dependent upon a full-length PCI card that can't be installed in a newer PC, preventing its integration into existing systems (I'd have to have 2 PCs on the bench and there's no room. Also, the software is old and unfriendly.)

    My goal is to duplicate the control & data processing functionality of the PCI card using an Arduino or similar.
     
    Sparky49 likes this.
  8. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    Cool, thanks for the input. Can you tell me how this difference might affect how I proceed with the project?
     
  9. crutschow

    Expert

    Mar 14, 2008
    12,977
    3,220
    I meant just using a logic analyzer while running the software on the old computer.

    If you can get the command structure/protocol for the old computer than you could likely design the interface from that. That would be easier then trying to determine it with a logic analyzer. But if it's all proprietary then it may not be available.
     
  10. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    You can still some motherboards with ISA bus. Like these http://www.adek.com/ATX-motherboards.html It may be other also. And I have not used any board from Adek ;)I would have taken that path. The job in order to reverse engineering the protocol may be quite hard. If you run the computer with XP or Windows 2000. The old software may still work. However this is just a tip. I can not say if this will work or not. But perhaps Adek can give some help. I was once running XP using a GPIB ISA card. And that worked nice
    Ps here Is a picture of the ISA bus. To big in size to post it
    http://upload.wikimedia.org/wikipedia/commons/0/0b/ISA_Bus_pins.png
     
  11. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    The Adek motherboard solves the electrical problem, that leaves the software. AFAIK Windows completely changed the driver model going forward from Windows 95. I could almost make a simple driver in 95, the learning curve to work in XP or such was too much for me to comprehend (also, I could do my tasks in USB and thus lost the need to do that).
     
  12. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    It looks like a parallel digital aquisition board, they were normally very simple.

    Most likely it is just sending parallel bytes as 8bits/8pins, and probably a strobe pin to latch each new data byte.

    And data is very likely one directional.

    The board is marked Shimadzu, you may still be able to get a datasheet or schematic for it, or at least a pinout for the main conenctor which would likely have been standard across many of their parallel data aqc boards.
     
    scary likes this.
  13. scary

    Thread Starter New Member

    Sep 14, 2011
    8
    1
    t06afre and ErnieM, I'd like to eliminate the second computer and its ISA card all together. I am hoping it is possible to work this thing directly from a microprocessor connected to the 'host computer' socket on the back of it.
    If the function of the card is as simple as THE_RB suggests this might be possible.

    I opened up the case of the instrument today to grab some shots of its guts if this is of use to contributors/ of interest to geeks & the curious.

    [​IMG]

    LH side showing in- and out-flow tubing leading to the flow cell. I believe the black box contain the light sources (there are two producing different wavelengths).

    [​IMG]

    RH side showing . . . well . . . some circuits and stuff. Relays, caps, I dunno what's going on here. I'm regretting not taking a close up of the device at the rear, which was made up of two thick copper windings.

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    I'd like to thank everyone for their input so far. I am simultaneously forming a plan of action and learning at the same time!
     
    Last edited: Mar 12, 2012
  14. t06afre

    AAC Fanatic!

    May 11, 2009
    5,939
    1,222
    This is a educational project right ;) Your pictures does not help very much. What would help a lot would be a good picture of the IO board on the PC. As then we indentfy the circuits. We may then identify the components. As a first step. Determine if the board has an onboard controller or not.
     
  15. wayneh

    Expert

    Sep 9, 2010
    12,087
    3,026
    As appealing as that goal is, is it really worth the effort just to eliminate a box? Just think of the old PC as an ugly, power-wasting interface box, and use it to communicate between your new stuff and the old instrument. Is that really so awful, compared to hours and hours spent salvaging old equipment, with no certain outcome? You might find a more modern instrument very inexpensively some where. I know it's hard to let go of the old good stuff - no one's a bigger pack rat than me - but sometimes it's the prudent choice.
     
Loading...