Fluke 8840A/AF GPIB board

[PaulAm]

I just got a pretty nice 8840A/AF the other day but it didn't have the GPIB option.  I've seen a couple posts that include the dump of the firmware but I wonder if anyone has cloned that board?  So far, the only schematic I've found uses the NEC UPD7210 controller IC, but at least one of the board versions is supposed to just bit-bang the bus and it would seem not too difficult to clone that board.  The NEC IC is still available from the grey market, not sure how much I'd trust that.

To make matters worse, I think the A and A/AF GPIB firmware is slightly different.  So I need the right schematic + right firmware + right model.

Any pointers?

[coromonadalix]

better buy a  defective or  ''for parts meter''   normally an gpib board should be fine

Just have to check the option marks on the rear panel  when a seller show it

You may face many problems to clone or redo a board

[PaulAm]

Possibly, but what's the fun in that?

[oPossum]

There where many ISA GPIB cards that used the uPD7210. Buy one of those and you would have a uPD7210, GPIB drivers and GPIB connector. Getting the Z8 and proper firmware for it may be a bit more difficult.

I did GPIB for the Fluke 45 using schematics in the manual and it turned out great. Got the TMS9914A, GPIB drivers, and GPIB connector from a HP GPIB module for a DSO.

[robert.rozee]

the interface between the 8840A's processor and the GPIB board appears to just be two-wire serial (of unknown format), passed through a couple of isolating transformers to give two balanced pairs. pins 4 and 5 of U202 (8842A schematic) seem to be the TTL level RxD and TxD. when i bought my 8842A there was no GPIB board fitted, and instead the cable that would normally run up to that board was brought out to an 7x2 berg connector sitting in place of the GPIB socket. see:https://xdevs.com/doc/Fluke/8842A/img/rear_1.jpg. i obtained a GPIB board on ebay and fitted it a number of years back, but have never actually tested it.

i'm a little surprised no one has got in and snooped on pins 4 and 5 in order to figure out the protocol and what data is exchanged. you could then replace the GPIB board with a small micro to translate to/from RS-232 serial. much simpler to work with these days.


cheers,
rob   :-)

[Kleinstein]

Translate to RS232 or directly use a cheap UART to USB converter module instead of drivers for RS232 levels.

[PaulAm]

The manual has a note that the serial connection runs at 65 kbit.

It's pretty clear that z80 on the gpib board isn't doing much (at least in the version with the uPD7210).  Probably just does the setup and passes data back and forth to the main processor.  All of that has to be in the 4K internal ROM.  There's probably an initialization protocol to let the main processor know the option is present.  When I get some odd moments, I'll poke around with this. 

If a board shows up, it would be trivial to put an LA on the serial lines and capture the protocol.   I guess that would be too easy 

[Burner_357]

Hi,

I'm working on a replacement for the GPIB-Board. The plan is to replace it with an Ethernet-Interface.
I already have the hardware in the correct form-factor to retrofit the 8840, but I'm still struggling to reverse engineer the protocol.

I have asked for some help and/or coop on this topic in the following thread:
https://eevblog.com/forum/projects/fluke8840a-ieee-488-redesign-(ethernet-replacement-for-gpib-option-05)/

I know, thread is 2 years old, but I was able to make some progress since then. parts are decoded by now.
If there is enough interest here, i can translate the progress i made by now to english and post it in my thread.

So far I'm pretty confidet that the following parts of the protocol (getting info from the device) are correctly decoded:
- Displayed measuremt value (sign, value, range, function)
- readout some settings (sample-rate, off-set, auto-range, F/R-input, ext-trigger)
- internal trigger event

I also have decoded some of the "set parameters of the device" part of the protocol.
But they are not tested yet.
- set the "Listen", "Talk", "Remote" LEDs in the front-panel
- request the current configuration (data for "G0" and "G5" requests)
- set the current configuration (F, R, S, T, D, B)

I haven't poked around with the calibration procedure. I don't want to lose/brick the cal-data.
And there are still many gaps in the understanding of some parts of the protocol. Especially error handling and error-code transmission.

So if there is inerest in this project, let me know.

Greetings, Alex

[Kjo]

@PaulAM
When dealing with this family of Fluke meters you need to keep in mind that they were produced for about 20 years.
You can expect, and there were many variations of the GPIB board in that time. I have personally verified about 9
different configurations between the three meter (8840A, AF, 8842A).

First, they fall into 2 catagories:
1) 4K ROM size using  two forms of the Z86, (Z8611, Z8613).
    The 8840A and 8840AF can use the same board & ROM and any of the several 4K ROM code versions for 8840As.
    The 8842A requires a different 4K ROM version, but the boards are interchangable. This is because the GPIB
    commands are tied to the physical locations of the front panel buttons, and on the 8842A they added a 20mV
    range to the left of the bottom row. Also AUTO moved to the top row.
    It is interesting to note that the display and keyboard are identical for ALL models. The 8842A just has an extra hole.


2) 8K ROM size using two forms of the Z86E21 (plastic OTP and ceramic windowed EPROM).
    These versions were introduced around 1990. I think the Z86 is doing quite a bit more than on the older boards
    in spite of the U7210. Looking into the 8K code it is clear that about 25% of the upper 4K is now used.
    This version (8K) board will work in ANY 884X meter, as the code supports an extended GPIB command set.
    Im not sure exactly the method used, but there are a number of GPIB command differences that Fluke could
    be using to distinguish an 8840A from an 8842A. A simple test with P8 is sufficient. P8 on an 8840A/AF generates
    an error 71, while on an 8842A it does not.

    I even have 1 8K GPIB board from 1990 with about 2-dozen white wires, and it has a completely different chip layout
    from all other 8K PCBs I have. It also has ROM version 0.3. (That is not a typo!)


Keep in mind that the manuals that are circulating the internet are a snapshot of production from the early 1990s
I think. The GPIB programming instructions therein are for the 8K version. If you try a G8 command on a 4K board, you
will get a buss error 71 on the front panel. I had to scrounge to find schematics for both the 4K & 8K PCBs.

Also, to be clear, to my knowledge ALL 4K PCBs use only a 4K Z86. ALL 8K PCBs use a Z86E21 & a UPD7210.

Kevin - KO3Y

[Kjo]

One thing I would ask readers to look out for is a 488 PCB using the UPD7210 but having a Z8611 or Z8613 rather than a Z86E21 (8K).
I have one schematic “Fig 9-7 IEEE-488 Interface PCA” that shows a UPD7210 and a Z8611 annotated on the drawing. Does this mean that initially Fluke expected to fit firmware into 4K space for this revision? Was this rev to be universal working with both 8840a & 8842a and that that didn’t work out. Maybe they found universal code took 5K so switched to Z86E21.
Or maybe someone just copied the schematic symbol that was available.
I do see that the original schematics were “hand drawn” and the latter was “CAD generated”.

Kevin KO3Y

[Kjo]

I have a thread (Fluke 8840A/8842A GPIB/IEEE488 Bus Card Clone) specifically on a 8840A/42A replica HPIB 488 bus card.
This thread here talks about this and other possible bus access approaches. I will keep the other thread up to date on testing the replica PCB
but noting my success here seemed to make sense.

So far the card is 100% operational.

Kevin (KO3Y)