HP 34401A DMM Randomly Switching Functions

[wn1fju]

An HP 34401A DMM, date code ~1991.  Passes all self-tests and seems perfectly fine and accurate on all functions.

Today, I hooked up 10VDC to it and let it sit on VDC for a while.  I came back only to find that the 34401A had switched itself to VAC.  I thought maybe I inadvertently hit the VAC button, so I put it back to VDC and came back in one hour.  Sure enough, the thing had switched to VAC again.

Any clue to what's going on?  I guess I should at least scope out the power rails and see if there are spikes.

[TimFox]

I don't have a 34401A, but my 6.5-digit Keithley bench meter will switch to AC volts if there is a short line-power blackout while it is powered on.

[pcwrangler]

Interesting Tim. Thanks for that info!

[floobydust]

Not sure if it's rebooting every hour or it's some phantom pushbutton presses? What firmware version is the Front Panel and Main board?

We did not get far in this thread: https://www.eevblog.com/forum/testgear/agilent-34401a-switches-from-acv-to-resistance-measurement/msg1938730/#msg1938730

[wn1fju]

The firmware revision is 03-01-01, so it is a really old model.

I tried it again and after about one hour, it was fine.  At about two hours I checked and it had switched itself again to VAC.

I wouldn't think there would be any chance of the thing switching exactly at one hour points.  I mean is there even any sense of timing in the circuitry?  I suppose I could sit there starting at 59 minutes and see exactly what happens.

Other than that, I guess it is one of the mysteries of life.  Unless you guys have a better explanation!

[TimFox]

Again guessing from my Keithley experience:  is it possible that you have a loose or flakey AC line connection (connectors, line cord, outlet strip) or a slightly faulty power switch or fuse?  If you turn on the meter, switch to DC volts, wait a bit, and then quickly cycle the switch on the outlet strip, what is the result?

[wn1fju]

Tried it with a different power cord and different outlet.  Same thing happened.   I flicked the power strip off/on as fast as I could and the 34401A simply rebooted normally and went to the default DCV.

Perhaps I will hook up a scope to the 5V rail and set it to trigger on a downward edge at about 4V.  If the 34401A switches into ACV without triggering the scope, it's probably not a power transient to blame.

[mawyatt]

Could this be due to RFI? These older HP34401As likely weren't designed to suppress the RFI that we subject them to today due to the close proximity of Smart Phones.

Best,

[Kleinstein]

ACV may be the default mode after a reset. There are several possible reasons to get a glitch an random reset in such an older computer.
Checking the supplies could be a point, e.g. there may be a failing capacitor starting to show ripple and mabe dropouts to trigger reset. It could as well be a failing ROM causing the CPU to crash and maybe a watchdog than trigger a reset instead if just hang.

[chekhov]

I think first thing to do here is to confirm that meter was really rebooted - like setting something that goes to default after reset (slow DC mode for example), then verifying that it still there after switching to AC or not. Maybe I'm missing something, but haven't seen that 'reset' theory was proven.

Maybe not likely, but what about possible dirt/moisture on front panel PCB ?

[Dr. Frank]

An HP 34401A DMM, date code ~1991.  Passes all self-tests and seems perfectly fine and accurate on all functions.

Today, I hooked up 10VDC to it and let it sit on VDC for a while.  I came back only to find that the 34401A had switched itself to VAC.  I thought maybe I inadvertently hit the VAC button, so I put it back to VDC and came back in one hour.  Sure enough, the thing had switched to VAC again.

Any clue to what's going on?  I guess I should at least scope out the power rails and see if there are spikes.

Hello,
I bought my 34401A in 1990 or 1991, REV 03-01-01, still working fine, so I also got a lot of experience on this DMM.
I find some of the hints given here pretty misleading.
Neither does the 34401A switch its mode, if different signals are applied, especially not from DCV to ACV, nor does it have ACV as boot-up mode. One cannot even store any other boot-up mode than DCV, 5 1/2 digits, R = 10MOhm, that's it. Therefore, a reset would also not bring it into ACV.

The electrolytic caps are for sure not the best any more, but ripple on the power lines probably will not lead to such a specific mode change, as if it was a key press, but to randomly switching to any mode.


There's probably a problem with the display / keyboard PCB.
Most obviously, there might be contamination on the keyboard, especially on the golden fingers of the ACV switch.
The keys of such rubber membranes usually consist of conductive plastics, maybe with imprinted carbon, which may have disintegrated a bit after 30 years, and particles got stuck on the PCB.
Other similar failure mode could apply from general plastic outgassing, or external contamination, depending in which environment the DMM has been operated in the past.
 
So I would start by carefully disassembling the front panel and cleaning the PCB and each individual switch of the rubber membrane with ISOPROP.

If that does not cure the problem, the µP on the PCB might have a problem on one of its I/O lines.
Others have reported here, that the µP fails fatally after these years, but this would be a total loss, let's not assume that.

 Frank

[floobydust]

I would also suspect a dirty pushbutton for ACV. You can press it just a little or wiggle it and see if it triggers too easily.
One 34401a I took apart had wire clippings, an actual copper wire strand inside a key switch 

First generation Front Panel board 34401-66502 uses NEC uPD7527 MCU and the keyboard scan through a 'HC194 in rows/columns so I believe an old weak MCU I/O line would not misread one key, it would be a group of keys malfunctioning.

If you're going to take it apart, maybe check here first because there are a few traps, from the ribbon cable latch to the front plastics, and not breaking the glass VFD.
The keyboard membrane is also fragile and tears if you are rough with it.

[wn1fju]

Thanks to all the people who commented.

I did try the experiment with my scope on the 5V rail to see if it ever captured a drop-out.  Unfortunately (or maybe now fortunately), the 34401A ran for
four hours just fine, staying in DCV mode, and the scope never triggered.

Of course, I had it out of the cabinet and the shields removed, so I buttoned it all back up again, and after letting it sit a while, turned it back on.  It ran flawlessly for 2+ hours on DCV mode. 

I have checked the 5Va, 5Vd, 18V and -18V supplies and they all have minimal ripple.

At this point, I will just keep an eye on it.  If the switching to ACV starts happening again, I will dissect/clean the front panel.

[Kwakerman]

If memory serves me right there isn't an actual ACV button to select ACV isn't it a matter of pressing 'shift' (or 'function') then the DCV button to select the ACV option? 

[Kwakerman]

Disregard my last, was thinking about DCV to DCI selection - muppet ! 

[wn1fju]

I've been watching my 34401A for a couple of months now.  It definitely switches from DCV to ACV on its own sometimes after an hour or two, and once about 15 minutes after turn on.  I have verified, by supplying a 1 kHz, 1 Vrms signal, that I am getting an accurate ACV measurement after the random switch, so it isn't like just the display switched.  I have also replaced the electrolytics in the power supply and have disassembled/cleaned the front panel switches (they looked brand new).

But here is one more curiosity:  After the unintentional switch to ACV, when I press the front panel DCV button, I hear two relays clicking in fast succession.  When it is in normal operation and I toggle between ACV and DCV, I hear only one relay click.

Any insight would be helpful.  All I can really do now is wait until I get a total failure and something becomes obvious.  Other than that, the meter appears to work flawlessly.

[m k]

Can two relays mean that the machine is still partially in DCV mode and when you press the button it goes fully ACV and back?

Earlier opening it up did something, maybe the thing is just heat and very small connection problem.

A machine from the other thread does 2W Ohms from ACV and DCV.
My guess is that the situation is not just random dirt.

uPD7527 board has main functions connecting the same sense line, is it special, are there special components, not type but placement.
Other version doesn't have any support components, its ACV and DCV are connecting the same sense line but after that there's only a controller.

Are both problem boards uPD type?
Maybe it's the pull down(+13V) or next diode, that would explain those two clicks.
Is floating logic up?