HP3456A Weirdness. Power cycle calibration change

[Insatman]

Hi...I have an HP3456A which I bought a few months ago.  When I first powered it on and after warmup, it agreed with my HP34401A within a few 10's of uV.   Then we had a power outage, not uncommon where I live.  When the unit came back on and again after warmup, the calibration had changed significantly...100's of microvolts on the 10V scale against a stable reference.   I waited several days and no change, so I recalibrated the unit to the HP34401A.   It worked quite well, even through a couple of short power cycles.  It tended to be within 20uV of the HP34401A.   Then I went on a 1 week trip and turned off the HP3456A before leaving.  Upon my return I powered up the unit and let it warmup overnight.  Now I have about 500uV difference between the 34401A and the HP3456A.  I believe the 34401A because it's measurement of several references hasn't changed much while the 3456A is low by ~500uV.   Has anyone see this behaviour before?   Given how the unit is built, i'm guessing the reference is changing due to some kind of power cycle hysteresis.   

Another side note which may or may not be related.  When I first got the unit I put it flat on my bench.  The unit lacks feet and has all it's vents on the bottom so it overheated and the display went to all "------".  After cooldown it worked fine and passed all self-tests.   This happened one more time after I put some spacers under the unit.  I ended up putting a heatsink on the 40pin microprocessor on the top board as it seemed the hottest chip I could see.   Since then the problem never returned.   I don't think the problems are related but I include this info for completeness.

The history of the unit as far as I know is that it was used in a semi-conductor factory here in Philippines before being sold to an ebay seller.  I think it sat in a warehouse for some period of time before I bought it. 

Comments, suggestions, thoughts and questions are solicited.

[Conrad Hoffman]

Don't know the 56, but if it's anything like my 55, that behavior is not at all normal. I'd start by making sure the fan is OK, then the power supplies. Regulator failure is not unheard of. Check grease and mounting. Then I'd monitor the reference directly and see if that's stable. If it is, the problem is downstream. This shouldn't take long as the unit shouldn't be changing by more than 20 ppm or so from cold to warmed up.

[SeanB]

Backup battery is dead, so the unit loses calibration info after being powered off for a while and the SRAM capacitors discharge. time to either do the Dallas replacement, or do the surgery to use an external coin cell on those potted units, or to do the replacement of the soldered on lithium battery on the board. There are a few threads on this forum about that.  Depends on the age of the unit if the battery backed SRAM is a potted unit, or if it is a version with a battery that simply clips onto the package, or if it uses a Tadiran or SAFT battery soldered to the board. The potted ones need a dremel and a careful hand, and then recalibration, while the clip on ones just need the battery popped off and replaced, and the soldered ones have instructions on how to do it without losing cal, though as you have lost cal you will just replace it with power removed and redo the cal afterwards on all ranges.

[alm]

Since when do mechanical trimmers need a backup battery? I think you're confusing the 3456A with later models like the 3457A, that used digital closed case calibration. The 3456A is all mechanical trimmers.

My guess would be something like dried out power supply caps that work slightly better or worse, depending on ambient temperature or other factors. So I would start checking power rails for ripple and noise.

[kj7e]

Since the issue started with a power surge, I would also check the power supply caps first, may as well just replace them if they are original.

Index   Quantity   Part Number   Manufacturer Part Number   Description   Customer Reference   Backorder   Unit Price   Extended Price
1   3   1189-3012-ND   50ZLH1000MEFC16X25   CAP ALUM 1000UF 20% 50V RADIAL      0   1.62000   $4.86
2   3   1189-1422-ND   50YXJ2200M16X35.5   CAP ALUM 2200UF 20% 50V RADIAL      0   1.93000   $5.79
3   2   1189-4026-ND   63PX220MEFC10X16   CAP ALUM RAD      0   0.73000   $1.46
4   2   1189-4028-ND   63YXJ100M10X12.5   CAP ALUM 100UF 20% 63V RADIAL      0   0.52000   $1.04
5   1   1189-3906-ND   35MXG5600MEFCSN22X35   CAP ALUM 5600UF 20% 35V SNAP      0   4.23000   $4.23
                     Subtotal   $17.38

Doing this to mine now, here are some part numbers from Digikey for reference if you like.  I would do this first before getting into too much troubleshooting.

Note, this does not include the 12,000uF 15v filter cap for the outguard digital 5v supply.  I found some NOS 15,000uF 35v axial Sprague's on eBay that will fit nicely.  The ---- does seem to indicate there is an issue on the digital side of the unit, I would bet the temp sensitivity is just exposing the issue.

[saturation]

I have 3 working simultaneously without issue and in very close agreement, many of those early posts are on the forum.

3456a have no batteries, the constants are maintained by the setting of its pots.  However, once you tweak it after being untouched for decades, it will drift quite badly in the uV ranges, and need to be retweeked until it settles back down; I suggest if you did adjusted it, wipe the adjusts screw up and down multiple times on a range as these old hermetically sealed pots could still have a lot of deposits on them, only then start to calibrate it to match the 34401a.  Note, the mechanical adjusts affect the linearity on other ranges so do a performance check on all the ranges each time you move the pots.

The main disease in working 3456a units is the PSU, particularly bad old electrolytics, if they do not provide stable rail power, the readings will drift.  Check the power rails for AC leakage, as well as incorrect output DC.  If your disagreement is in the uV chances are its the PSU.

I agree fully with alm.  Its also documented on the HP yahoo users group. 

FWIW I found that semi-bad electrolytics in high quality old T&M gear can be 'reformed' by simply leaving the old device on 24/7 for 1 week or more after a PSU cap reform technique on first power up.  The sign this works is the drifting stops and the output stabilizes, not just on the 3456a, but many other T&M left in storage somewhere.  Ultimately its better to replace the caps, but 90% of my lab devices were salvaged that way. 

https://www.nmr.mgh.harvard.edu/~reese/electrolytics/#reform

[Insatman]

Backup battery is dead, so the unit loses calibration info after being powered off for a while and the SRAM capacitors discharge. time to either do the Dallas replacement, or do the surgery to use an external coin cell on those potted units, or to do the replacement of the soldered on lithium battery on the board. There are a few threads on this forum about that.  Depends on the age of the unit if the battery backed SRAM is a potted unit, or if it is a version with a battery that simply clips onto the package, or if it uses a Tadiran or SAFT battery soldered to the board. The potted ones need a dremel and a careful hand, and then recalibration, while the clip on ones just need the battery popped off and replaced, and the soldered ones have instructions on how to do it without losing cal, though as you have lost cal you will just replace it with power removed and redo the cal afterwards on all ranges.

As stated by others on this thread, the HP3456A does not have battery backup.  My 3457A does have but it works fine and I've never opened the case since buying it.

[Insatman]

FWIW I found that semi-bad electrolytics in high quality old T&M gear can be 'reformed' by simply leaving the old device on 24/7 for 1 week or more after a PSU cap reform technique on first power up.  The sign this works is the drifting stops and the output stabilizes, not just on the 3456a, but many other T&M left in storage somewhere.  Ultimately its better to replace the caps, but 90% of my lab devices were salvaged that way. 

I've had the unit for maybe four months now and it's been on 24/7 except for a few power outages and the one week mentioned in the origional post.   I think it's time to open up the unit and have a look at the PS caps and whatever else catches my eye.   

[saturation]

Yes, can't hurt.  Please report back.  From an older post:

https://www.febo.com/pipermail/volt-nuts/2010-November/000626.html

FWIW I found that semi-bad electrolytics in high quality old T&M gear can be 'reformed' by simply leaving the old device on 24/7 for 1 week or more after a PSU cap reform technique on first power up.  The sign this works is the drifting stops and the output stabilizes, not just on the 3456a, but many other T&M left in storage somewhere.  Ultimately its better to replace the caps, but 90% of my lab devices were salvaged that way. 

I've had the unit for maybe four months now and it's been on 24/7 except for a few power outages and the one week mentioned in the origional post.   I think it's time to open up the unit and have a look at the PS caps and whatever else catches my eye.   

[Insatman]

Yes, can't hurt.  Please report back.  From an older post:

https://www.febo.com/pipermail/volt-nuts/2010-November/000626.html

FWIW I found that semi-bad electrolytics in high quality old T&M gear can be 'reformed' by simply leaving the old device on 24/7 for 1 week or more after a PSU cap reform technique on first power up.  The sign this works is the drifting stops and the output stabilizes, not just on the 3456a, but many other T&M left in storage somewhere.  Ultimately its better to replace the caps, but 90% of my lab devices were salvaged that way. 

I've had the unit for maybe four months now and it's been on 24/7 except for a few power outages and the one week mentioned in the origional post.   I think it's time to open up the unit and have a look at the PS caps and whatever else catches my eye.   

I have a "2201" series HP3456A, so no fan.

I looked around and found no obvious problem with capacitor leakage.  I measured voltages and the 33V rail is at 44V.  All other voltages on the header were close to nominal.  Time to look at the schematic and dive deeper.

[Insatman]

Ok the 33V rail is unregulated so 44V is actually within spec.  I set the input to 240VAC rather than 220 and this reduced it a bit.  I noticed a lot of 1uF axial caps on the various boards.  None appear to be leaking but I think I'll order some to have around.   The existing caps on the PS board are all Nichicon except the largest one which is Sprague.   I'm thinking the problem is likely oxide on the pots and/or front/rear terminal switch.   So I plan on working on that next.    Overall the unit is pretty clean inside.   Anyone have advice on if I should retrofit a small 2" fan?   There is an easy place to mount it.

[kj7e]

Most if not all of the 1uF axial caps are tantalum, I don't believe many have had issues with these on this meter.

[Insatman]

Another clue.  When  the meter is attached to a stable voltage reference it tends to count down slowly.  If you remove the source and re-attach it starts over.  Hmmmm.

[khs]

You may take a look at the K103 input relay.

If there is a problem, cleaning the relay contacts is a little bit tricky:

Remove the relay and drive the relay with a function generator. Take care not to ruin the coil.

Additional generate a small current  through the relay contacts and observe the resistance of the contacts with an oscilloscope by observing the voltage drop of a resistor connected to the relay contact.

The cleaning effect is made by the mechanical stress the two relay contacts bounce together.

Sure, there is a risk to damage the relay and therefore ruin your 3456.

But without risk there is no fun.



 

[Insatman]

One of my three HP3456A started acting the way you described.  I kept my units on all the time.  About 3 weeks later it failed, I had dashes in the display.  When I opened it up, it had a fine powder everywhere inside of the unit.  A capacitor had failed in the main power supply board. 

I opened a good 3456At and measured the power supply voltages on the power supply capacitors, as a lot of older equipment was design in the days of 110 volt power and most power now days seems to be about 125 volts.  I ordered new capacitors with the bumped up voltage ratings and replaced them.  The unit came back up.  I cross checked the unit against the other two and it appeared it was spot on for the calibration.

I had a 4th HP3456A which is a parts unit and I pulled the power supply board and rebuilt it and then replaced it in the second unit.  Then rebuilt the second unit to replace in the third unit, etc.

I used panasonic 105C rated caps for replacements from digikey.

I've ordered a set of replacement caps.  When they come in I will replace them all and see if the stability improves.   

[mikeg]

I can report the same. After replacing all the caps in the supply the meter came back up and worked perfectly.

[Insatman]

Weirdness Continues...  I replace all the electrolytic capacitors on the power supply board and the large one on the Outguard Logic board.  Thankyou KJ7e for the component list.  All of KJ7e's recommended capacitors were approx x2 in capacitance to the original values.   I measured all the removed caps and found no "bad" ones, in that nothing was shorted and the ESR looked within reason.  Measuring the new capacitors the ESR was usually about 1/2 of the old capacitor values.   I did find something else worth noting.   One of the wires between the transformer and the power supply board was not properly seated in it's plastic header.  I could pull it out easily.  I spent some time reforming it and pulling out the little barb that locks it in place.  After some fiddling it finally snapped in.  So I powered up the unit after the cap retrofit and it seemed to work just fine.   I let it warmup overnight as it was already quitting time for my hobby.  The next morning (today) I started doing the calibration procedure which starts with the self-test.   Bang...it fails test #7.   It has never failed a self-test before!?!   Sigh....so I take the covers off the unit and begin to follow the recommended trouble-shooting procedures.  A few steps in, after removing and replacing some headers and testing it appears the A/D is at fault.  So replace the jumpers and run self-test again.   No faults.   I run it many many times...no fault.   Hugh....

I put the covers back on and calibrate all the voltage ranges...working fine.   

Thermal issue?    We shall see what happens down the road.

[montemcguire]

I remember reading an HP service note that blamed IC sockets for some problems of this nature. The fix was to remove the chips from the sockets, remove the sockets, and solder the ICs directly to the PCB. If your 3456A has socketed ICs, the act of removing and re-installing the jumpers may have wiggled the sockets enough to make them fail and then work again. Not all 3456A have socketed ICs, but if your unit has them, this is worth considering.

[Insatman]

I remember reading an HP service note that blamed IC sockets for some problems of this nature. The fix was to remove the chips from the sockets, remove the sockets, and solder the ICs directly to the PCB. If your 3456A has socketed ICs, the act of removing and re-installing the jumpers may have wiggled the sockets enough to make them fail and then work again. Not all 3456A have socketed ICs, but if your unit has them, this is worth considering.

I do have socketed ICs.  Another type of wierdness has also cropped up.  Sometimes the unit goes into a mode where all the lights are on and the display will be some random value not changing.  Pushing the reset button changes the display but to a blank numbers and a + sign only.  Subsequent resets have no effect.  Only a power cycle fixes the problem and the unit may work correctly for days at a time. 

Only once since changing the capacitors have I had a low reading value like before.   The unit's calibration has drifted a bit however...maybe 30uV since calibration a week ago.  This is reading my MV-106 at 10V.   The MV-106 is also being monitored by another 6.5 digit meter and it has not changed.

I'm guessing the digital outguard board is to blame and I know it's large chips are socketed. 

[kj7e]

You should read over the 3456A service notes found here;

https://www.eevblog.com/forum/testgear/hp3456a-service-notes-for-owners-info!/msg144820/#msg144820

Sound like some of your issues are described there.

[Insatman]

You should read over the 3456A service notes found here;

https://www.eevblog.com/forum/testgear/hp3456a-service-notes-for-owners-info!/msg144820/#msg144820

Sound like some of your issues are described there.

Thanks again for the information.  Today the 40pin precision sockets came in and I replaced all three in the instrument.  The rivets were also replaced with M2 bolts on the power supply connector.   The unit is up and running again.  Passed self-test and now I'm just running it 24/7 to see if it is stable.   

[Insatman]

Just a (hopefully) final followup on this thread.   I applied the fixes recommended; re-socketing the 40 pin ICs, replacing PS caps and replacing grounding rivets with machine screws.   All this work was completed on Feb 19th.   The HP3456A was powered up 24/7 and connected to my MV-106 reference set at 10.00000 volts.  It drifted up several 10's of uV over the first couple of weeks but then stabilized.  Perhaps this was due to my cleaning the area around the trim pots with electronics cleaner?   Anyway after a couple of weeks the drift stopped and remained stable until today.   Today I did a voltage calibration and now my meter is spot-on against my standard.   Hopefully my power-cycle problems have ended.   Note we had several power black-outs in the past month and the HP3456A seems to return to correct readings after an hour warmup.

Thanks again for the advice I received.

[BradC]

Thanks again for the advice I received.

Thanks for coming back to finish it off. For someone dealing with similar issues in the future, your final post is as important as everything else in the thread.
Always nice to have a good result too.

[MisterDiodes]

Insatman's experience is a good one and similar to what we've seen after recapping,  a good cleaning around the pots and general spruce-up (especially when the area around the pots has been cleaned).   It'll take several days for the unit to re-stabilize but once it gets there if should be a pretty solid & reliable unit.  We use them every day along with 3458a's.  Great bang for the buck!!

Good Job!

[Insatman]

Insatman's experience is a good one and similar to what we've seen after recapping,  a good cleaning around the pots and general spruce-up (especially when the area around the pots has been cleaned).   It'll take several days for the unit to re-stabilize but once it gets there if should be a pretty solid & reliable unit.  We use them every day along with 3458a's.  Great bang for the buck!!

Good Job!

Yeah...Great bang for the buck.  I have roughly $200 invested including purchase, shipping and all parts.   Of course the labor was free.   It's a bit of a boat anchor unit, but as long as you have the space a nice addition to almost any bench.