I have an HP 3458A that seem to have open inputs:
Display shows "O.VLD" in all resistance ranges in spite of shorted inputs.
Display shows random numbers (floating slowly up or down) at all voltage ranges [edited: up to 10V range].
Front-rear switch position doesn't matter.
Selftest is OK, no errors (ERR 0).
Front-rear switch is checked, OK (I measured every single contact of the switch).
Connections from front panel inputs through the F/R switch to the P8 and P9 inputs on the PCB are measured and they are OK.
TEMP? shows ~29°C (open case at 22°C ambient temp), but I don't know if the main A/D converter is used for temp measurement.
What strategy would you recommend to find the root cause?
This is not a simple unit. I think the first step would be to try to understand its method of operation. My 3456A works well and my 3455A has died. I suspect yours is at least somewhat similar; did you walk through the troubleshooting procedure? I believe there is a step by step approach to isolating problems.
While it doesn't solve everything it does give you a shot at finding the area that needs work.
In the case of my 3455A it pointed to an HP specific IC which I can't buy so readily. I am hoping to find someone with a broken unit to sell me that piece. Spending a lot on this one when I have a working 3456A woouldn't make sense but maybe I can find one at a reasonable price. It's A10U26 and the part number is 03455-62501.
Hello Carl_Lab,
yes the temperature sensor is also fed to the main A/D, like V/A/Ohm
I strongly assume, that DCI , ACI are ok, but ACV is not working?
Have a look in the CLIP, analog PCB A1, 3458-66501, sheet 1/5.
There you see 8 relays, which feed signals of different modes into the DCV main input MUX, that is K3 (to R17). Input to this relay K3, are relay K5 for ACAL, the T- sensor from label W3A over K6, to ACV circuit over K9, and Ohm test current over K2. These should all work, and latter Ohm test current gives an Overload. Self test will not include input signals coming from the jacks, so there's the problem.
The signal from Volts Hi Input do not arrive at this relays MUX, so K1, K7 are not operating, either defect (both???), or they do not get proper steering signals from inverter driver U2, and from 'Defeat' FFs U22A,B. Check clear line of these FFs, pin1, or pin13, should be +5V normally, but maybe they are permanently shut off by the processor. overload detection, U10A and related circuit.
You could check first , if you get continuity from input of both relays pins K7-1, K1-3, 'Volts Hi Input', to the front jacks.
Maybe something is broken in between there.. see also input switching diagram
Frank
PS: I assume, that 100V and 1kV DC ranges give static zero reading, disregarding input voltage?
Thanks for you replies.
yes the temperature sensor is also fed to the main A/D, like V/A/Ohm
OK, then the A/D is working.
I strongly assume, that DCI , ACI are ok, but ACV is not working?
I assume, that 100V and 1kV DC ranges give static zero reading, disregarding input voltage?
Yes, exactly.
Just some additional thoughts, what could have happened:
If these Flip Flops are really held RESET (NOT R=0V) permanently, there might have been an overload in DCV mode, damaging the DEFEAT circuit, consisting of Q26 /Q27, CR6 / CR8 (labels are barely readable), U10A.
That might also have affected MUX FET Q13, so check bias current afterwards.
At first, these logic ICs (e.g. U22) also might simply fail after years (was the case initially with my instrument), as well as the comparators (U10).
Damage depends on the application, where DMM was used at last.
Any knowledge about that?
Frank
You could check first , if you get continuity from input of both relays pins K7-1, K1-3, 'Volts Hi Input', to the front jacks.
I've checked before:
Connections from front panel inputs through the F/R switch to the P8 and P9 inputs on the PCB are measured and they are OK.
The signal from Volts Hi Input do not arrive at this relays MUX, so K1, K7 are not operating, either defect (both???), or they do not get proper steering signals from inverter driver U2, and from 'Defeat' FFs U22A,B. Check clear line of these FFs, pin1, or pin13, should be +5V normally, but maybe they are permanently shut off by the processor. overload detection, U10A and related circuit.
I've bypassed from P8 (Volts Hi Input) to R17, and my applied 100mV DC input test signal was displayed correctly.
So at least one of the relays K3, K7+K1 does not switch on.
Damage depends on the application, where DMM was used at last.
Any knowledge about that?
Sorry, no.
Check clear line of these FFs, pin1, or pin13, should be +5V normally...
Pin 1 and Pin 13 are at +5V level.
I've measured the coil driver voltages at U2:
U2 Pin 16, K7 Pin 2 = 5.0V -> K7 contact should be open
U2 Pin 15, K1 Pin 2 = 0.5V -> K1 contact should be closed
U2 Pin 13, K3 Pin 1 = 0.7V -> K3 contact should be closed
So K1 or/and K3 contact(s) are defective.
Can I check these without removing the PCB from the chassis? (I could not find any test points at the top layer of the PCB...)
Only indirectly, by probing between input high Jack and e.g. R17.
Common Star point of all 8 relays is not accessible from top PCB side
I found the root cause:
K1coil is broken.
I replaced K1 by a working COTO relay from a donor 34401A.
(Does anybody know, where to buy those COTO relays?)
Thank you, my friend!
Voila!
(Does anybody know, where to buy those COTO relays?)
I tried to get them from the manufacturer but at the end it did not work because they are made exclusively for Agilent / Keysight.
But you can order them from Keysight.
Hack one together yourself. Did so on a dodgy HP3456A input relay.
It has more space in there than in 34401A or 3458A, so I was able to strip down a 5kV reed relay down so it would fit. Extracted glass piece and coil from case, into which it was casted with hard wax and botch wired it into my 3456A.
Basically, find some cheap high voltage reed relays and botch wire it into broken DMM.
A 3458A is worthy of the exact, proper replacement relay.
me too for what its worth and the precision it gives, only genuine parts would go in
Assuming the part number for K1 is 0490-1659, they seem to be available from Keysight in limited supply for USD 40.43 - well worth it for a meter of that quality.
TonyG
Sorry for my late reply, I was offline over the holydays...
Thank you all.
@tppc
Are you sure the relays from a 44701A (5 1/2
or maybe 6 1/2 digits) meets the 3458A specs?
The relays from 44701A have different P/N...
@Tony_G
Assuming the part number for K1 is 0490-1659, they seem to be available from Keysight in limited supply for USD 40.43 - well worth it for a meter of that quality.
Yes you are right!
This is the part no. listed at the replacement parts list, but my 3458A has two 0490-1748 (K1 + K3)...
P/N: 0490-1659 (used in 3458A parts list)
Mod and Ser Nbr Reqd Relay, 1uV Reed, Protected Supply
price: EUR 33.31Edit: Spec from COTO: 1000V, 0.5A/1.0A, 10Watts, thermal EMF of
1uV and a 5V, 350 Ohm coil
Very old part, replaced by:
P/N: 0490-1748 (used in my 3458A)
Relay-reed 0,5/1A 5VDC 1A 480-Ohm THT
price: EUR 51.73
not on stock, but can be orderedEdit:
Spec from COTO: 1000V, 0.5A/1.0A, 10Watts,
1uV reed relay with a 5V, 480 Ohm coil
P/N: 0490-1914 (used in 34401A (K101), E1412A)
Mod and Ser Nbr Reqd Rly-Reed 2A,5VDC Coil, Protected Supply (from parts list: Relay-Reed 2A 500 Vdc 5 Vdc-Coil, 400 OHM coil)
price: EUR 46.69500V in a 1000V spec'd 34401A? Can't be true...Edit:
Spec from COTO: 1000V, 0.5A/1.5A, 10Watts, thermal EMF of
2uV and a 5V, 480 Ohm coil
P/N: 0490-1556 (used in 44701A, 44732A, 44733A, 70110A, E1326B, E1410A, E1411B, E5515A, E5515B, E5515C)
Relay-Reed
price: EUR 25.86Does anybody know the actual differences? EDIT: Added specs from COTO
Better contact material or coating (for low thermal EMF)?
Well, it looks like HP used different references over the years, or depending on their stocks...
I have a 44701A which uses yet another reference from COTO. (Can't tell you now).
Hopefully, the difference between a 5 1/2 digits DMM and a 7 1/2 one is not (only) on its input relays specs
Hopefully, the difference between a 5 1/2 digits DMM and a 7 1/2 one is not (only) on its input relays specs
Correct, but in 8 1/2 digit mode of the 3458A thermal EMF error of the Relais contacts will be much more visible.
If the 44701A can provide 6 1/2 digits, the factor is 100, compared to 8 1/2 digits .
Hi,
just for information, here are the references used in my 44701A
COTO 5180-8233
COTO 5180-8234
Regards.
Thanks, for all your replies!
