Quick clues re faulty HP 5342A? (Edit Solved)

[VK5RC]

Hello,
Recently I switched on my old 5342a and it behaved oddly, bizarre display initially then after a minute or two, the display shows a fixed display (looks like a series of segments are lit) on the LEDs. See photo below.
A quick look at the downloaded manual ( http://literature.cdn.keysight.com/litweb/pdf/05342-90057.pdf ) the fault finding tables, shows no response to keypad test then putting an oscilloscope (high z) on the HDSPWRT pin (supposed to pulse positive logic level on start up) shows no sig signal (only .2V 20kHz positive and negative spike, looks like switchPSU noise).
 I am not an EE, and do not have the service extension cards so I have not spent a lot of time fault tracing; power rails test OK but it is looking like the microprocessor has failed on quick look. I see on ebay a few 'for parts' units with similar faults so assume it is a common failure.

If anyone has had a similar failure I would like to learn from their experience, I really dislike not maintaining good old gear but am getting a bad feeling about this. It is quite old gear and time spent fault finding may not be well spent.

Any comments/suggestions appreciated.
Rob Culver VK5RC

[Richard Head]

Rob

I've also got one of those counters. Mine is still working ok, luckily.
I can't offer any suggestions without studying the manual but if you require a comparison of voltages or waveforms I can open mine up and poke around.

Dick

[VK5RC]

Dick,  thanks for the offer,  if yours is working well I would leave it be,  a comment I got on the Agilent repair yahoo group is to ALWAYS remove the feet before removing the bottom case,  as they can foul some of the cables on the bottom,  fortunately I don't have the feet!
Regards Robert

[poot36]

I would look at the micro clock line or oscillator to check that it is working and in spec.  Also check the micro reset line to see if it is being held in a high or low state or pulsing.  If the unit has eproms check that they are not corrupted.  Also check for power supply ripple.

[VK5RC]

@poot36, thanks for the suggestions, ripple, rest line and micro oscillator I know how to check but eprom corruption will involve quite a bit of reading for an analogue guy!
Thanks Rob

[VK5RC]

Reset is going high (as it is supposed to I think) on switch on,  the clock is a good 1MHz, I even get some digital like signals at the vma pins on the micro board a14, voltage looks stable,  the display drivers and memory are next in line I think. Rob

[poot36]

Use the same approach as the reset for the data and address lines on the memory except that the lines should not stay at one value unless the equipment is not doing something.  If it always stops at the same data and address line values then that is its normal resting state or that could be a clue to what it is getting stuck on testing (try pressing the buttons on the front panel to see if this state changes or not (we are trying to generate an interupt for the micro here) and if it does then it is not totally dead).  Think of this as a POST card tester for a computer but you are going to have to manually test and decode the binary output.  If you have a logic analyzer use it (arduino perhaps for this low clock speed)?

[MarkL]

While looking at the address and data lines on the CPU pins as poot36 suggests, I would also try putting the CPU in free run mode as described on page 8-91.  This should allow the CPU to loop through the full address space, and you should see your highest frequency on A0, half of that on A1, half again on A2, etc.  You might not see any activity on the data lines in free run.

And I say "should" because the procedure describes putting the CPU on the extender card, so it's possible free run mode may have some dependency on the extender.

I'm assuming when you say the CPU clock is ok, you checked both clock pins (it's a two phase clock).  It's a good sign you see activity on VMA.

This doesn't look like the easiest unit to service with all the cards inserted vertically.  I suppose that's why they have such an extensive set of extender cards.  A signature analyzer would help narrow down the possibilities more quickly if you can get your hands on one.

[VK5RC]

poot36 and MarkL,  thanks for the suggestions, I appreciate your comments as my digital knowledge is expanding infinitely (1/0).
 I have an original Salae,  last night I soldered some extension wires onto 2 data lines of the RAM (u12 I recall,  not near the unit now)   and 2 address  lines,  both were active,  I will post a picture of the line voltages when I get back home.
Thanks Rob

[VK5RC]

A screen shot of the Saleae data screen on MCU board A14
Ch1= Inverted CS4 of RAM
Ch2= CS3 of RAM
Ch3= D6 of RAM
Ch4= D7 of RAM
Ch6=VMA of MCU
Ch7=VMA phi2 of MCU
The control and data lines of RAMwere just chosen at random (easiest to solder to as well)
PS I have found a HP Signal Analyzer, and have it coming, it will help quite a bit I think. hoping the extension card (or lack of it) doesn't alter the MCU too much in the test algorithms

[MarkL]

The traces indicate the clock is there and CPU is at least doing something.

Great that you found a signature analyzer!  The service manual for this unit is one of the most detailed ones I've ever seen from HP.  Hopefully the diagnostic steps will pinpoint the problem, or if not the analyzer will at least verify what's working (like the ROMs) to start eliminating suspects.

[VK5RC]

Problem solved. A tri-state hex inverter (an old DM8096N or Sn74366AN) U16 on board A14 had broken, replacing it fixed the problem. Part sourced from http://www.ebay.com.au/itm/230407534307?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1497.l2649.
The biggest problem with fault tracing this unit is the manual (very good really) uses extender cards that are unobtanium now. I used extension wires or accessed buses from the bottom of the unit.
The front display board is written to when HDSPWRT (high display write) goes high,  in my unit it was not moving. The whole address bus is accessable from the bottom of the unit, and a basic Saleae unit showed that the address line LA5 (on socket A14A (pin *8*) was also not moving. The rest is history. The part met its maker with some Widlarisation!
Thanks for the help.

[SeanB]

What is the name of the hammer.

[GreyWoolfe]

What is the name of the hammer.

Maxwell S. Hammer?

[free_electron]

excellent job on the Widlarization. Bob would approve !

[VK5RC]

The hammer didn't have a name,  I inherited it from my late father,  a civil engineer with a professional interest in instrumentation,  he was also a Bob,  so I think it is fitting I name  the hammer Bob, after both my father and Widlar.
As an aside if something really p****d my father off he would hurl it over the edge of a nearby steep scrub slope.

[tautech]

As an aside if something really p****d my father off he would hurl it over the edge of a nearby steep scrub slope.

Yes, done that......its called "giving it a flying lesson". 

[lowimpedance]

Successfully shown who is the BOSS 
Good job
and an old timer back in service (both hp and hammer)

[VK5RC]

I think something is missing,  any good ritual also has a chant,  the soundtrack that accompanies sending  the errant part to its maker must be obscene, 
All I can recall hearing is 'f**k you little b*****d'.  I am pretty sure most would naturally recite this after the tracking down of the fault.

[SeanB]

I seem to think something from ACDC works for me....... Thunderstruck.

[CristianCapra]

I have a HP5342A frequency counter and a meet the same behavior, but after few days of research I realize the problem was from rear panel switch, it was in EXTernal position. I switched in INTernal position, performed a  restart and all works fine now.