HP53310A Modulation Domain Analyzer: failed Main Board DTACK test

[poppniete]

Hi folks!

I recently obtained a HP53310A Modulation Domain Analyzer which fails the power on self test with the message "failed Main Board DTACK test".
I bought the manuals and the component level identification package document including schematics of the unit from Artek Media.
According to the service manual, I isolated the CPU board from the mainboard and powered it separately. The built-in CPU board test routines show no errors, which means, the fault has to be on the mainboard. The mainboard is a huge PCB

Actually I do not have any experience with MC68K systems. I know what /DTACK is used for though, but I don't have a good idea on how to test it any further, because I don't know what test results are expected in the self test routine.

Does anyone here have any experience with this fault in a HP53310A? Any hints would be greatly appreciated, because the unit is in nice condition and is maxed-out option-wise...

Regards,

Jan Weber

[SoundTech-LG]

Now if you had a spare pare parts unit, you could start by swapping the board. Do you have the schematic? Artek manuals has the CLIP which includes a schematic.

[syau]

Suggest to check the voltage on the psu and the condition of the capacitors (especially those 1000uF)

[SoundTech-LG]

He didn't do that yet???  Bad assumption on my part, but with all the current posts on bad PS...

[poppniete]

I should have told you, but I did the usual basic checks.

Artek manuals has the CLIP which includes a schematic.

As I said, I bought the CLIP document.

I do not have a parts unit.

Suggest to check the voltage on the psu and the condition of the capacitors (especially those 1000uF)

I already did that. The power supply is in a like-new condition, all voltages are within spec with no ripple, and the caps have negligible ESR. There is no sign of electrolyte leakage.

Regards,

Jan

[SoundTech-LG]

Ok, well another consideration is the NVRAM (Dallas) contains a backup battery, is getting on in years, and could be getting low in voltage by now???  Not sure of the effects of low voltage on the NVRAM. Contains the CAL constants, so might be good to see if you can pull the data off, and put into the SOCKETED new one, if you decide to replace it. Assuming you have a programmer to do that...

[SoundTech-LG]

Any idea on the age of the unit? You can check the NVRAM chip's date code. First 4 digits of the S/N? Country of MFG. ?

[SoundTech-LG]

Have you actually checked the DTACK signal on the processor (or the main board if you can trace it) to see if it's being held high, or low, or it's not a real low, or a real high?  Tack a small wire on the pin, or trace, and reassemble board back into the unit.

[poppniete]

The instrument was manufactured in late 1999.
The self test procedures are quite extensive and include checks of the NVRAM checksum and all memories. The CPU board itself passes all selftests, so I don't think, there is a problem with the NVRAM.

The main board of the instrument has a whole bunch of /DTACK sources. It's hard to differentiate between the possible causes of failure.
I guess it might be a dead bus transceiver between the CPU board and the main board or something like that, which leads to a wrong initialization and a missing DTACK answer.

I guess I have to probe the bus every signal at a time to find out.

The documentation is lacking any functional description of the building blocks in general or the self tests in special. It is further complicated by the fact, that there are PALs thrown in everywhere, which are black boxes.

Regards,
Jan

[SoundTech-LG]

Not sure this can help, but it is full of DTACK references. Also searchable using adobe reader...
https://www.nxp.com/docs/en/reference-manual/MC68000UM.pdf

[Tomorokoshi]

In 53310A-CLIP.pdf from Artek on page 78, "A1-63", at coordinates H2 - G5 there are a bunch of "BUS DTACK JUMPERS". Are all of those in place? Are the connections good? They are around H5 to H6 on the A1 Main PCB.

[SoundTech-LG]

Excellent suggestion.

[SoundTech-LG]

"the unit is in nice condition and is maxed-out option-wise..."

I think a possible additional troubleshooting effort would be to "minimize" the options, removing them entirely...  then see how it reacts.

[mark03]

These analyzers are getting so cheap now on the used market that the most sensible course of action may be to buy another broken unit as a parts spare.

You didn't say where you live.  EU?  US?

[poppniete]

The options do not affect the core system. It's a high stability timebase and an optional 2,5 GHz input.

In Germany, where I live, HP53310A's are still quite expensive. Buying a parts unit is out of question for me, simply because I couldn't find any locally. Buying these overseas and paying for shipping is not a n economically viable option, I'd buy a working unit with some kind of warranty instead...

The unit I got was very cheap and in a very nice cosmetical condition. If my repair attempts do not work out, I guess I can sell the working bits with profit.

Regards

Jan

[poppniete]

Hi folks,

I just wanted to tell you for the sake of completeness that I was able to fix the unit.

In the end it was just a broken /LDS trace on the CPU PCB.

Regards,

Jan

[SoundTech-LG]

Yes...  nice to know you got it working! So a broken trace? Damaged? Corroded? Why was it broken?