Fluke 8860A : Error 17 & 16 , need help and U6 EPROM data

[Elger]

Hello.

When switching on, my Fluke 8860A is showing “Error 17” most often but it can also happen that “Error 16” is shown.

Although the 8860A is well documented, I was not able to fix the errors and need your help.

I read the hints § 4-63 to 4-67 and followed the steps in Table 4.8 Testing Guard-Crossing Circuity in the Troubleshooting Section of the Service Manual Chapter 4.

When exercising the tests of Table 4.8, I found that the Controller Board is transmitting the Send Clock and Send Data Signals.
But the Data Signal is only valid when pin 15 of U6 is not connected otherwise it is held low.
The Send Data Signal is also valid when U6 / pin 15 is connected AND the EPROM to the Piggyback Socket of U6 is NOT fitted.

How to interpret this behavior ?
When the EPROM is not in its socket, then the U6 which is a Mostek MK38P70 receives its RESET Signal when the 8860A is powered on.
I guess by default all of the Ports of the MK38P70 are switched to INPUTS or similar e.g. ISOLATED.
When the EPROM is fitted, the therein stored program is executed and switches U6 / pin 15 to be an OUTPUT.

But what is causing this ?
Is it a faulty MK38P70 (it is a CMOS from the late 70’) ?
Or is it a faulty EPROM with some changed bits sending the MK38P70 false instructions ?

Because the MK38P70 is a rare to find device in these days here in Germany, I would like to test for a faulty EPROM at first.

Does anybody already have an image of the U6 Piggyback EPROM or could anybody be so kind and read out his EPROM ?

Regards,
Erwin

[m k]

Do some piggypack connection and signal measurements.
Are all data and address lines equally resistive and active?
If you can read the ROM upload it and we can check if it includes anything sane.

Since you still can't find the ROM image, after all these years, your copy can be special.
3780 internal ROM can be read using a special test mode but then TEST pin must be put to sort of an over voltage state and for casual owner that is obviously not very easy thing to do, even with spare.

Maybe we should start a 3780 invalid ROM image collection project.
Finally we will have so many of them that a correct one can be combined and compiled.
8860A is also not the only model using that chip.

[Elger]

Sorry for some delays.
I was on a short vacation over the weekend.

Yes, I already have checked the data and address lines of the EPROM 2716.
Because I do not own an In Circuit Debugger or similar but an Tek 11302 Analog Scope only,
my tests are on only one data or address line at a time.
The timing on the address lines is following the rules for binary address spaces.
The low and high levels are valid.
For the data lines I can say, that low and high levels of each data line are also ok.
But I am not able to determine the value of the data byte.

Regarding the EPROM Image Project, you have the same idea that I was thinking about.
I have already made an image file of my EPROM.

Are we allowed by Fluke to make the EPROM code public ?

[m k]

Take your time.

When you file a patent application you are revealing your invention.
So it is known but not available to use.

ROM code is sort of the same.
You can duplicate it as much as you like but you are not allowed to profit from it.
Technically putting it available for public is a bit different, but not much and preventing it is practical impossibility, at least for now and with these computer powers.
There are also fair use things and EU has more, if manufacturer is not providing you are allowed to start hacking for repair purposes.
What providing and allowed are exactly is probably still without high court precedent.

So you are almost free to put it up.
You can also send it to me and I will put it up.

One example.
You can freely resell your device but selling only its ROM chips is tricky.
If you sell only one set you're fine but if you put up a shop and start selling en masse Fluke will eventually come after you and then they don't want only your profits but also enormous amounts of extra compensations for their pain and misery.

[m k]

Just in case the MCU must be changed.
Manual says that there are client selectable i/o-port output latch style possibilities, pin by pin.

Mask ROM version is easy, you can't use the wrong one.
For P-version ports 0 & 1 the manual I have has nothing to say, ports 4 & 5 have dropped 2 FET and inverter direct drive output so only open drain or 6k pulled up is used.
MK97400 is pulled up and MK97410 is open drain, no idea how that is visible on the package.

[Elger]

Regarding the IO-Port Configuration, Figure 4 is explaining and showing the possible realizations.
Ports 0 and 1 are always standard configuration with TTL inputs and outputs as shown in the schematics.

The Output Options are described in the second picture.

For the MK38P70/02H one must know that Mostek is using alternating device names in their data books.
When Mostek is talking about the EPROM version of the MK3870 the MK38P70, they like to use the name spaces starting with MK97xxx.
That is the MK38P70/02H is called MK97400R which is designated to Standard IO-Pins.
The Suffix R designates the PiggyBack EPROM version.
In this case the MK38P70/02H has nothing to do with the MK97410 which has Open Drain Ports.
So there should be no confusion when looking for spare parts, you must choose a MK38P70/02H.
By the way, I could not find any explanation to the suffix H in MK38P70/02H.

As can be seen on the picture I have posted before, the exact device description is printed on the upper side of the device.

[m k]

My bad, I was already happily forgotten that order number and was mixing individual pins to the equation.
It's definitely indicating pullups.

Anyway, I can only see that port pins are bidirectional and instructions are affecting the whole port.
Included schematics has 6k2 pullups for pins 14 and 15 but not for pins 12 and 13, all those pins are for port 4.
Pins 3, 16 and 17 are for port 0 and have 2k7 pullups but there are special connections.
Then AC/DC connector J1 and A/D Ohms connector J2 have no pullups and their counterparts have solid 2.5V references.
So pins 14 and 15 can be voted to have extra pullups and MK97400 is the correct order code.

Fast read of '78 databook couldn't see those order codes.
'81 databook has them on page III-84, page III-82 has emulation possibilities, page III-10 has old part numbers and external ROM version MK97401.
MK9750x has open drain port 5, so no good but their generic codes are also different, '82/'83 databook page VIII-11, page VIII-1 has more selections and MK97401 is missing.
'81 databook page III-99 has more numbering stuff.

[Elger]

I'm surprised no one else is taking part in this discussion.

Is the Fluke 8860A so little used or is it so rarely defective?

In the latter case, the repair would be all the more worthwhile.
So please let's work together and help bring these almost nostalgic devices back to life.

The original label on my EPROM had come off and was lying around in the device.
It is labeled as follows:
800312
(8860)

[m k]

I think the situation is pretty clear and has been the same for long.
After certain unsuccessful waiting period devices are dismantled for parts.

Situation can change but ROM image is needed first.
Reasonable MCU cloning can't really start without it, or can but who would do it.

[m k]

Just in case,
what Sean did.
http://www.seanriddle.com/f8.html

Few ROM images are also available.
So at least some code examples.

[m k]

Are you still without a solution?