HP 5334A Counter with no display (guilty chip = MOSTEK GPIB uP stucking bus...)

[bleuos1613]

Hello to all...and happy new year    (that was what we already said in 2020  ).

I did buy on an auction website a HP 5334A counter.

Unfortunately when i arrived, there was display on the front panel. I could hear the buzz of the transformer.
Openning of the box, test of the voltages on the test points. Everything was OK

After multiple power on/off I noticed that there some display for "fraction" of a second during the power on... that kept
my spirit for trying on.

After getting an online copy if the service manual of HP 5334B ( a sibling), and some study... I noticed
- display is controlled through an 8 line I/O Port bus from the main cpu...
- With the scope and checking some of the direct lines (STR1 & STR2) from the EXEC CPU to the from panel board that this CPU was OK.   (of use here, we can put the unit in diag mode by pressing the reset button while powering up the unit. In this mode the unit keeps doing some diag tests going on step to another and cycling when pressing the reset button).

Unfortunately, some of the Data lines to the 7 segment drivers kept high impedance state.

Checking what other component was on the lines I found that this IO Ports of the EXEC CPU was also shared with the GPIB uC.
Removing this last CPU from its socket and starting the unit solved the display problem of the unit.   

I have now an error 9.2 (GPIB CPU problem) when I power on the unit and after validating the error the units goes to "counting"...

Still trying to solve the GPIB CPU problem, how can I check if the problem is the CPU or some other component.

Note : 4MHz clock and reset lines is common to both EXEC CPU and GPIB CPU. This CPU is MOSTEK MK3870/42... this is
a Factory MASK programmed CPU with fw MK16255N-15. There seems some way to read the code using the test pin of the CPU and doing a "serial dump"...and after that to program a MK38P70 chip with some external UV-PROM.. But this version of the CPU cost "an arm"...

Should I try to mount the CPU on a breadboard and do some testing ?

Thanks in advance for any hint

Olivier

[m k]

Seems that your options are quite limited.
What can you do if its ROM is masked and pins are not stuck.

Your three 3870 chips are all 42 versions, means full ROM and only that verion is suitable.
For replacements I would go to external ROM direction.
Seems that STM made them after Mostek.

Replicating 3870 shouldn't be overly complicated.
Is it done already?

[bleuos1613]

In fact  on the 3 mostek CPU 3870, 1 is a 38P70, this one has an external EEPROM.

38P70 is pin compatible with 3070, so, if I can find the prom code of the GPIB CPU, I can program a UV PROM for another
38P70. But those are quite expensive.

I have seen on some places some guy who has used the MOSTEK TEST pin to do the serial read of its PROM... this assumes
the CPU is still alive...

Just some fun ;-)... maybe not worth the value of the counter which is rather "old" gear...

I can also find on ebay some guy with a damaged 5334A and re-use its chip...

[m k]

Put test pin to half of the operation voltage.
Then port 4 is output and port 5 is inverted input of internal data bus.

When test pin is high voltage ROM is bypassed and intsructions can be fed through port 5.
No idea what high voltage actually is.

In both above cases strobe will be clock.

[bleuos1613]

I will give it a try, mounting the MOSTEK on breadboard. PDIP 40 pins adapter will be handy for that also trying to understand how the 3 CPU do their handshake with the main CPU.

Keep you in touch.

[m k]

https://archive.org/details/1981_3870_F8_Microcomputer_Data_Book

III-44, 69/586 of pdf has a description of the test process, input is described a bit differently.
Absolute high voltage is still unknown but I read that Vcc+2V is enough.
Oscillator stuff continues from there.

I read that the chip was used quite widely, maybe you can find a cheap P-version donor.
I'm sure that finally the software will be available somehow, if needed.