[Solved] HP 3457A MOhm Range Problem

[Kugelsicher]

Hello everyone, I recently purchased an HP 3457A on ebay (for $167). Everything seems to work except the high ohms ranges (problems start at the 30M range). The first issue I noticed was that on the 2-wire ohms setting, with nothing connected to the inputs, the auto range selects the 300M range and displays around 090.0840 MOhm (it drifts, but I hadn't allowed the meter to warm up). So naturally, I opened it up (Chips are dated 1993 and 1994).

I found several dime sized circles of spray paint on the underside of both main boards. Apparently, someone spray painted it without disassembling it and paint went right though the holes on the bottom where I think feet should go. I also found a little copper dust (from the screw inserts, I think). I tested the paint for conductivity and found it to be on the order of 10 MOhms per mm. So, I sprayed off everything with an air duster and cleaned the paint off with IPA and a copper brush. It didn't come off easily; so, a little still remains. Pictures below.

I really thought that this would solve my problem; so, I reassembled the meter, powered it up, and nothing had changed, at all. Still around 90 MOhms on the 2 wire setting and all of the rest of tests and checks came up the same.

Now for the tests/checks and results:

Auto Cal: AC -> No Errors
Auto Cal: Ohms -> Error Denunciator; AUXERR? = 0 (Isolation error during auto cal)
Complete Self-Test -> Pass

2-Wire Open Test
Range          -> Reading
Auto (300M) -> 090.1640 MOhm
30      -> OVLD
300    -> OVLD
3k      -> OVLD
30k    -> OVLD
300k  -> OVLD
3M     -> OVLD
30M   -> OVLD
300M -> 090.1640 MOhm
3G     -> Automatically changes to 300M range

2-Wire Short Test
Range      -> Reading
Auto (30) -> 00.00211 Ohms (jumps around +/- .0005)
30        -> 00.00241 Ohms (jumps around +/- .0005)
300      -> 000.0031 Ohms (jumps around +/- .0003)
3k        -> 0.000003 kOhms (jumps around +/- .000001)
30k      -> 00.00000 kOhms (jumps around +/- .00001)
300k    -> 000.0000 kOhms (jumps around +/- .0001)
3M       -> 0.000000 MOhms (jumps around +/- .000001)
30M     -> -01.00061 MOhms (jumps around +/- .00001)
300M   -> Automatically changes to 3G range
3G       -> 0.000000 GOhms (stable)

4-Wire Open Test (No connections between any inputs)
Range -> Reading
Auto    -> reading jumps around rapidly, changes ranges
30       -> OVLD
300     -> OVLD
3k       -> reading jumps around including OVLD
30k     -> reading jumps around including OVLD
300k   -> reading jumps around including OVLD
3M      -> reading jumps around including OVLD
30M    -> reading drifts slower but still seemingly without bounds
300M  -> reading drifts slower but still seemingly without bounds
3G      -> reading drifts slower but still seemingly without bounds

4-Wire Open Test (Input Hi connected to Sense Hi; Input Low connect to Sense Low)
Range      -> Reading
Auto (3G) -> OVLD
30      -> OVLD
300    -> OVLD
3k      -> OVLD
30k    -> OVLD
300k  -> OVLD
3M     -> OVLD
30M   -> OVLD
300M -> Automatically changes to 3G range
3G     -> OVLD

4-Wire Short Test (3-feet long leads connected to inputs and then all connected together at other end)
Range      -> Reading
Auto (30) -> 00.008~~ Ohms (~ indicates very unstable digit)
30        -> 00.008~~ Ohms (~ indicates very unstable digit)
300      -> 000.008~ Ohms (~ indicates very unstable digit)
3k        -> 0.00000~ kOhms (~ indicates very unstable digit)
30k      -> 00.00001 kOhms (jumps around +/- .00001)
300k    -> 000.0001 kOhms (jumps around +/- .0001)
3M       -> 0.000001 MOhms (jumps around +/- .000001)
30M     -> 00.00001 MOhms (jumps around +/- .00001)
300M   -> 000.0000 MOhms (stable)
3G       -> Automatically changes to 300M range

I measured a 10MOhm resistor with the 2 wire and 4 wire settings (Results below). 4-wire seems to not suffer from the 2-wire's problem. This is probably a big hint to someone smarter than I.

2-Wire 10.07 MOhm Resistor Test
Range         -> Reading
Auto (30M) -> 09.079~~ MOhm
30      -> OVLD
300    -> OVLD
3k      -> OVLD
30k    -> OVLD
300k  -> OVLD
3M     -> OVLD
30M   -> 09.079~~ MOhm
300M -> 006.72~~ MOhm
3G     -> Automatically changes to 300M range

4-Wire 10.07 MOhm Resistor Test
Range          -> Reading
Auto (30M) -> 10.07~~~ MOhm
30      -> OVLD
300    -> OVLD
3k      -> OVLD
30k    -> OVLD
300k  -> OVLD
3M     -> OVLD
30M   -> 10.07~~~ MOhm
300M -> 010.07~~ MOhm
3G     -> Automatically changes to 300M range

Checked the test current for every 2 wire ohm range with a UNI-T UT61E (has a uA range) and they were all within one count.

To verify that the test current was correct even during a faulty 2-wire reading, I put the meter on the 30MOhm 2-wire range and connected a 1MOhm resistor. The meter read -00.0048 MOhm. I then attached another multimeter (with 10MOhm sample resistor) in parallel with the 1MOhm resistor and measured the voltage across the 900kOhm resistance, 1MOhm + 10MOhm in parallel. The 3457A now read -00.0861 MOhm. The second meter read 0.0911 V, which when divided by the 3457A's 30M range current of 100nA produces the correct 900kOhm result.

I'm a senior physics major so I understand the basics but have little EE experience; so, any advice on troubleshooting would be appreciated.

Thank you all,
Matthew

[plesa]

Hi Mattew,

Probably you have schematic at the end of this document
http://cp.literature.agilent.com/litweb/pdf/03457-90011.pdf
and service manuals
http://cp.literature.agilent.com/litweb/pdf/03457-90003.pdf
http://cp.literature.agilent.com/litweb/pdf/03457-90012sg.pdf


Because 4W seems to be OK I will also check rear terminals if problems persists ( it will use different relays ).
But some 3457A does not have rear terminals and you will needs to connect it to scanner connector behind relay board cover.

BTW check the battery voltage and replace if needed.

[Jwalling]

Hello everyone, I recently purchased an HP 3457A on ebay (for $167). Everything seems to work except the high ohms ranges (problems start at the 30M range). The first issue I noticed was that on the 2-wire ohms setting, with nothing connected to the inputs, the auto range selects the 300M range and displays around 090.0840 MOhm (it drifts, but I hadn't allowed the meter to warm up). So naturally, I opened it up (Chips are dated 1993 and 1994).
[snip]

Matthew

Hi Matthew,

If you don't get any resolution here, have a go at posting your question in the Yahoo HP/Agilent equipment group.
https://beta.groups.yahoo.com/neo/groups/hp_agilent_equipment/info?referrer=Manual_Exchange

Jay

[Kugelsicher]

Thank both of you for replying.
plesa, you were right on the mark. I disconnected both inputs (via TERM OPEN command) and the 90MOhm reading went away (shows OVLD on the GOhm range). So according to the schematic, that narrows it down exactly one part, relay K101. I'll report back once I replace it. UPDATE: It turned out not to be that relay (somehow). Still great long distance diagnostics; it got me a step further.

Thanks again.

P.S. the battery is at 3.4V

[plesa]

Hi Matthew,

I'm glad that you found the root cause and meter can be fully working soon.
It is quite good meter for decent price now.

Relay should be
http://uk.farnell.com/panasonic-electric-works/s4eb-l2-5v/relay-4pst-no-250vac-30vdc-4a/dp/1191778
Battery voltage is  quite high, it should be more close to 3V.
Maybe consider calibration data backup anyway.

[Kugelsicher]

I removed that relay and tested it. The switching was working. I couldn't find a 90MOhm short across the channels. So, I replaced that relay with a couple pieces of wire and two 150 Ohm resistors (to replace the coils). The symptoms remained, unchanged.
Selecting the rear terminals or open terminals still resolves the resting 90MOhm reading.
Also, The 30MOhm range works; it is just horribly miscalibrated. If I null the -1.00061 MOhm shorted-state reading, all of the relative readings are accurate.
Could it all just be due to miscalibration?

[Kugelsicher]

To answer my own question: Yes, it was all due to miscalibration. I shorted the inputs and calibrated the 30MOhm range to 0 Ohms (Calibration code was still 3457). Now, every range works.
Crazy that a bad calibration caused an isolation error; but, I couldn't be happier.