HP8647A RF generator - "504 modulation cal restore failure"?

[carl_lab]

This HP8647A had a power supply problem (repeatedly blown fuse).
I replaced the SMPS module and the instrument started with "504 modulation cal restore failure" message.

Is there a simple trick (starting seft test/cal or something else) to get rid of that annoying message?

Service manual:



Seem to be a more serious problem... 

PS:
After running the instrument for a while, switching off and on again, the message changed:

"502 update cal file"
Description: The RAM calibration file has been updated from the calibration ROM (OK?)
Cause: This is normal when a module is replaced (but I changed the SMPS only, how do you know?)
What to do: No action required Nice (but true?)

[carl_lab]

Update:
Error 504 is back.
(couldn't post earlier because forum was offline...)

[edpalmer42]

I have an 8647A where the battery backup has failed.  Instead of drawing ua of current, it's drawing ma.  I gave up and took the battery out.  Every time I turn it on, it gives me 503 errors (cal corrupt) for each card and a 627 error (battery fail).  The only effect is that everything returns to a default condition and the calibration values are copied from ROM to RAM.  After that, the unit operates normally.  Calibration is unaffected.

Since there's no service info for this model, I'm guessing, but it seems that the battery is supposed to power RAM on each card.  In your case, the copy from ROM to RAM is failing on the A3 card.  The 502 error is considered okay on module replacement because changing the module would wipe the onboard RAM.  Remember that's all guesswork on my part.

Have you checked the battery in your unit?  Maybe it's dying or your circuit is failing like mine.

Ed

[carl_lab]

Hi Ed,
no I did not check the battery (I probably will do that today).
Battery is OK (3,15V).
If battery was empty I would expect same errors as on your unit - 503 cal corrupt and 627 error (battery fail).

I also do not understand neither how a replacement of SMPS could be recognised (I did not see some intelligence on the module like ROM with serial number, data bus or something like that) nor what the replaced SMPS have to do with calibration.

BTW: Does anybody can provide circuit diagram of that SMPS?



It is also used in the following instruments:

    71452B
    83437A
    83438A
    84300A-E30
    84300A-E34
    84300A-E45
    84300A-E47
    84300A-E52
    84300A-E53
    84300A-E56
    84300A-E57
    84300A-E58
    84300A-E59
    84300A-E63
    84300A-E64
    84300A-E66
    84300A-E67
    84300A-E74
    84300A-S06
    86037A
    86037B
    86037C
    86120B
    86120C
    86121A
    86130A
    8643A
    8647A
    8648A
    8648B
    8648C
    8648D
    87030A
    8703B
    87075B
    E2507B
    E2508A
    E5501A
    E5501B
    E5502B
    E5503A
    E5503B
    E5504A
    E5504B

   

[alm]

I find it hard to believe that this is directly related to replacing the SMPS. If anything, it might be related to the failure of the old SMPS. Did it put out excessive voltage before it died? Did it die due to a short inside the instrument?

My guess is what the error message states: damaged RAM or corrupt ROM. Could be a checksum failure in either reading the ROM or reading back the RAM.

[carl_lab]

Next steps:

I reseated all 3 RF modules and the ROM module - no success.

I swapped the ROM module with the one of my working reference unit - no success.

Error message 502 appers for a few milliseconds after power-on-self-test, mostly overwritten by error 504 .
I could not find out how to scroll thu the error display (tries al arrow buttons, jog-dials ...).

[alm]

With ROM assembly you mean the entire A3 board? Any history on that donor unit (two broken A3 assemblies?). Are the rails powering the digital logic (the +5.2 V rail?) clean of noise?

[carl_lab]

I swapped the small PCB with 4 populated EPROMS (it is locked by a small aluminum angle plate that is bolted on from the  instrument's back side).
I did not swap one of the 3 RF modules (just released them, blow the dust off and reseated them).

I have no history of the donor/reference but at least it seem to work (output signal comes out) without error messages.

I checked all PS control LEDs (+5V, +5V(F1), +38V, 12V, +12V): all LEDs are on.

I checked PS output voltages at connector J6 (against metal chassis):
Pin1: +5,22V (connected to pin 2)
Pin2: +5,22V
Pin3: 0V (chassis GND)
Pin4: 0V (chassis GND)
Pin5: +14,97V
Pin6: 0V (chassis GND)
Pin7: -14,95V
Pin8: +37,66V
(doesn't look too bad)

I found another point the defective unit behaves different to my reference unit:
After turning the instruments on while pressing "memory DEL" key to perform a clear memory, the GPIB adress is not reset to 19 an no "ERROR 627: battery RAM failure, memory lost" (as the manual states to be normal, and what is confirmed by my reference unit).

What the hell is going on here?

(Edit: I moved part of text of my last post to this post because I added after you wrote your reply)

[alm]

Defect on the digital (A3) board other than EPROMs? Can you swap it between units?

[carl_lab]

A3 board is the motherboard?
Swapping the motherboard is a lot of work to do, it seem to be much easier to swap all 3 RF modules (or one after the other).

BUT: I don't know, if that action will corrupt calibration data on the working unit.

To be honest: I don't want to risk the possibility of a needed new calibration (it's not cheap)...

[alm]

To me this does not sound like an issue with the RF modules. More like a digital problem (or digital power supply). Especially since you report different behavior between the defect and known good unit. But it makes sense to first exclude any issues that would be easy to diagnose/fix.

[carl_lab]

Yes, of course.
But I don't see much difference between your suggestion (swapping A3 board) and swapping the whole unit and swapping all 3 RF modules.
The defective unit needs a working PS (so I have to swap it too).  The only difference I can see is front panel assembly and the (hidden) output module.
I don't think the front panel assembly have anything to do with the problem.

Maybe replacing the static RAM-ICs (HM62256BLFP) is a way to go?

Datasheet:
http://ee-classes.usc.edu/ee459/library/datasheets/hm62256.pdf

[alm]

The SRAM is certainly be one of the suspects for me. That, or possibly a bad connection on the PCB (e.g. open address pin) or bad power supply to the digital circuit (noise, shorted tantalum pulling it down). Measuring the voltage and noise at the power pins of the SRAM (if accessible) could be helpful. If you have a spare SRAM IC and the tools to swap it, then it would be worth trying.

[carl_lab]

I have the tools (hot air station) but not the ICs (don't want to cannibalize the working unit)...

Original HM62256BLFP is hard to find...

I don't think that I could use these:
https://www.reichelt.de/DRAM-FRAM-SRAM/IS62C256AL-45UL/3/index.html?ACTION=3&GROUPID=2954&ARTICLE=146564&OFFSET=16&SID=96WXXZEqwQATYAAEhTkO806e2755dec47fd937a3de76290ec42a6&LANGUAGE=EN
Standby-power is 150uW (so-called "low-power") vs. 1uW (original), so 1,2Ah battery will last only 16000h (2Y) for 4 SRAM ICs...

This one looks better (standby current 1uA @3V -> 3uW), but not available in SO-IC case from that supplier:
https://www.reichelt.de/DRAM-FRAM-SRAM/AS-6C62256-55PCN/3/index.html?ACTION=3&GROUPID=2954&ARTICLE=146551

Do anybody know an available replacement for HM62256BLFP?

[TheDefpom]

My recommendation is to replace all caps, even if they test ok, at least then you have covered the simplest fault, then if it continues to be a problem you know for certain it's not due to a bad cap, and digging deeper is a requirement.