34401A problems on 4w resistance readings

[masterx81]

Hi!
I have a 34401A multimeter with a problem on the 4w resistance reading. I find the resistance a bit unstable, it takes a lot to stabilize, and when i select the front terminals, also without anything connected, i get some random readings. I've tried to check the bottom side of the board for flux residue, but it's really clean. I've read that also the front/rear switch itself can cause problems.

Any suggestion on what check for first?

Thanks!

[coppercone2]

push the switch a bunch of times and see if it does anything

[masterx81]

Already tried, not seen any improvement.

[bdunham7]

What resistances are you reading that are unstable and how are the connections made.  Can you set up a good 4-wire short?

[masterx81]

What resistances are you reading that are unstable and how are the connections made.  Can you set up a good 4-wire short?

For example, shorting the terminals it start at 40mohm, and slowly goes to almost 0 in near 5 minutes. This after proper unit warmup. If i open the short, and redo it immediately, it start back at 40. I not use the meter often, but last time i've used it few months ago was working fine.
I use true kelvin 4 wire clamps.
If i leave the back connection empty, and set the rear measurement, i see overload, as it should. With same conditions, selecting the front terminals i see random resistances from 100ohm up to 100's of kohms, i'm pretty sure that in the past it wasn't doing this. With 2w mode all seem working fine.

[bdunham7]

For example, shorting the terminals it start at 40mohm, and slowly goes to almost 0 in near 5 minutes. This after proper unit warmup. If i open the short, and redo it immediately, it start back at 40. I not use the meter often, but last time i've used it few months ago was working fine.
I use true kelvin 4 wire clamps.
If i leave the back connection empty, and set the rear measurement, i see overload, as it should. With same conditions, selecting the front terminals i see random resistances from 100ohm up to 100's of kohms, i'm pretty sure that in the past it wasn't doing this. With 2w mode all seem working fine.

The overload vs random readings with no connections in 4W mode isn't a big deal, but the 40mR thing is interesting--that's 10X normal and it is taking much, much longer to settle.  I would try leaving it on for 48-72 hours and see if that changes at all.  As a reference point, my 34401A has been on for several days in a 73F +/-5 room and I just tried it--it starts at 000.004 and settles to 000.001 flickering to 000.000 within 30-60 seconds.

[masterx81]

Last Time i've used it in 4w mode was last year, searching for a short using a set of Custom made Kelvin Probes. It was settling correctly in few seconds. I'll try to leave it warming up more as you suggested, but usually i was able to take decent readings after only 1h of warmup

[masterx81]

Interesting behavior: if i put a 4w short on both front and rear connectors, and allow the reading to stabilize, then i switch from front/rear, the reading reamain quite well at around 0. While, if i open the short on front (while reading the front terminals), and redo the short, it goes to 0 quite fast (as it should). Instead, if i open the rear short (while reading the rear terminals), it start from above 50mohm going slowing down, both on front and rear terminals. But to me not seem that there is a cross talk between front and rear terminals. This after around 12h of power on.

Note to add: while reading the front terminals it settle to 0 quite fast only if i leave the contacts open for small times. Maybe this behaviour is due that the front terminals read "phantom values".
Can change something switching from 10Gohm input impedance to 10Mohm?
In any case, nothing seem have sense to me on what it do 🤔

[Dr. Frank]

If you manually select different ranges, i.e. 100 Ohm, 1k, 10k, .. do you see different behaviors?

Does the 34401A measure correctly, when you apply different resistors, and do these measurements also show such a drift?

Does the 2W mode work properly, in comparison?

Frank

[masterx81]

After almost 24hours from power on, seem to work as it should, on both front and rear connectors 
No lag, fast settling.
Now i let it cool and see what it do, but seem to remember that the rear connection (where i leave a set of kelvin clamps always connected) had the same lag measuring an 1k vishay 0.005% 2ppm reference resistor.

[masterx81]

This morning i've powered it on, immediately tested the front and rear connectors and... it works! No settle time, precise reading I've heared that electrolytics can change they behaviour if leaved without tension for long times, maybe this is the case?
I can't explain this behaviour 

[JohanH]

Mine had erratic behavior in 4W mode, front input, due to cracked solder on input resistors (the large 2512 resistors near the terminal). Front panel was cracked and terminal latch was missing entirely. This caused bending of the PCB every time someone inserted leads into the front terminal jacks. Soldering the resistors again helped (and of course fixing the cracked front panel and getting a new terminal latch to secure everything together).

[Dr. Frank]

Mine had erratic behavior in 4W mode, front input, due to cracked solder on input resistors (the large 2512 resistors near the terminal). Front panel was cracked and terminal latch was missing entirely. This caused bending of the PCB every time someone inserted leads into the front terminal jacks. Soldering the resistors again helped (and of course fixing the cracked front panel and getting a new terminal latch to secure everything together).

Yes, that is another commonly known failure, and I initially thought of this one.
Obviously the failure of the T.O. vanished, so this case can be closed.