HP 3458A , resistance range out of order

[HUGOMAN]

HI ALL,

I have a HP3458A with error ERRSTR205hardware failure test value OUT OF RANGE:95"; at start up it gives also this message 110, CALIBRATION REQUIRED -- ACAL.

So following some testing results i notice that there is a problem of level of nano ampere range because i applyed 1µA it gives as follows:

Range of µA
range 1 -->  001,14432 µA
range 2 --> +01,058890 µA
range 3 --> +0,659799 µA

range of nano ampers
for 1µA injected ---->  it reads 000,003nA 

the 3 ranges of micron has been tested overlimit, they are ok, but the one of nano, it doesn't give overlimit even the value is greater that it's capacity.

the 2nd probelm is the resistor couldn't be measured:
i apply resistor of 2kohms ---> it gives for different ranges those values
-00,29841 OHMS
-000,30204 OHMS
-0,0003013 KOHMS
-00,000331 KOHMS / -000,00036 KOHMS / -0,0000009MOHMS / -00,000006MOHMS / -000,00001MOHMS / -0,0000000GOHMS.
The same for 9,99 KOHMS resistor it gives almost the same values.

then i found there is in front of dmm a short circuit between LO and the input of 1A/250V.
also when i lauch cal it goes out at ACAL DCI 1/10u.

I hope receiving more feedbacks about those issues.

Kind Regards
HUGOMAN

[Kleinstein]

For the resistance ranges, one should use a second meter to check the test currents.

The ohms current source is likely also used in auto-calibration of the low current ranges. So this could be a common source of error.

Another point to test would be the input impedance / bias in the 1 V or 10 V range. So connect a 1 nF to10 nF low leakage cap and watch it charge / discharge.

[HUGOMAN]

Hi kleinstein

i don't understand well the second point, i connect a ceramic cap of 10nf
in the input voltage, i monitor then by scope the charge discharge, is this ?
if i understand without applying any voltage in the input hi lo!

can you see the schematic for the concern range and we analyse it?

Thanks in advance

[Kleinstein]

For measuring input currents there are three ways:
1) an external meter for low currents (pA range) and voltage source for the offset.
2) use a high resistor at the input and use the meter under test. This get hard with offset.
3) use a capacitor and observe the charging with the meter under test. So no external scope used.

A typical 10 pA current into a 10 nF cap would give you a 1 mV/second voltage drift. By charging the cap one can start at different voltages.

[Satbeginner]

Hi, we will continue in this thread, I will have a look at the schematics for some things to check.

First guess is:

Find and check the voltages around the Ohm current source, and as Kleinman said: see if there is current coming from the meter when in the Ohms ranges.
Basically an Ohmmeter is: sending a known current through an unknown resistor measure the voltage across the unknown resistor.
After that the meter will do the math and show the value of the unknown resistor.

So, if measuring voltages is working OK, the next step would be the Ohm ranges current source.

[HUGOMAN]

Hi Satbeginner,

Yes good, for voltages no problem, i am gonna to check the current delivred by dmm to measure the unknown resistor as you said and our friend Kleinstein said too, and i come back to you.

thanks

Regards

[HUGOMAN]

Hi Satbeginner

Yes i do the test of current and it's quite 0 mA (0,5µA) on the different ohms ranges.

Kind Regards

[Kleinstein]

So it looks like the current source or some of the relays could be the problem.

Checking the relays could be relatively easy, once you find the suitable test point on the board. Running on ohms mode one should be able to follow the current path and measure for low resistance.
One can also measure the current directly at the source (from before the protection diode).

[HUGOMAN]

Hi Kleinstein Satbeginner

I start with page of sentry ohms, i found the fuse open, after bypassing this, it still gives the same state!!

Please could You see with me the schematic analysis?

[Kleinstein]

The broken fuse suggests that the damage come from something like massive over-voltage, thus triggering the spark gap. So the inductor close to the fuse might be blown too.

The Z300 transistor block is overvoltage protection for one direction. One might measure the current before that, from TP300 to GND, maybe with something like 1 K in series for protection (in case the source gives to much current and blew the fuse this way). Another point would be checking the 10 and 13 V ref voltages used for the current source. So 13 V at TP301 and 10 V and TP303.

[HUGOMAN]

Hi Kleinstein

i am very happy about your help, thank you very much.

So i perform the measurement it is as follows:

TP301 at 12,87Vdc
TP303 at  9,60Vdc
TP300 at 16,63Vdc

The inductor one leg removed from board, then it measures 0,04mH !! i don't trust fully this tester.  with my fluke gives 2,3ohms!

to read.

Regards

[Kleinstein]

The Inductor values 2,3 Ohms and 40 µH look plausible - at least not open. It should have been possible to measure that in circuit.

TP301 and TP303 look about right, but TP300 should not be at 16 V. If current is flowing, it should be more like 10 V + the gate threshold for Q305, so more like 12-14 V. 16 V is more like the maximum the OP can deliver. Is the reading with open terminals or with a low resistor at the terminals ? With open terminals one would expect an even lower voltage.

One suspect would be U300 - but a little difficult to test in circuit and it could be delicate part / area. Alternatively it could happen that non of the current setting resistors is activated. This second possibility could relatively easy to check and a non critical area : The corresponding LM339 (comparator to control the JFETs) outputs should be high, at something like 14 V.  I think TiN had a failure of a 339 before. Pin 15 of U302 could be an interesting test point too - should be about 14 V. If it is more than about 15 V, it points to non of the currents is selected (e.g. a problem with an LM339).

[TiN]

I've also had issue with Z300 hybrid in my original meter, causing random resistance readings. If I wiggle network a bit - readings normalize.
After replacement Z300 from dead A1 everything become normal. So that worth checking too.

[HUGOMAN]

HI TIN

Yes i am totaly agreed with you about the Z300 but i hope that won't happen with me because this hybrid is unfoundable unless if we can manually design ours in case of if it is only transistors and resistors!!!

Any way i am concentrating with analysis of our friend Kleinstein because he has very profesional approch to the problem, in my part i've changed from before the all of the lp365 anld lm339 and 4096 but nothing change.  so i am now on research of the spare.

coming back soon

All my thanks to you

Regards

[Kleinstein]

The Z300 part could be replaced with the internal circuit and more or less generic parts. The transistors are not even critical (e.g. leakage, gain and speed do not matter), just higher voltage (e.g. > 200 V) PNPs.

[HUGOMAN]

Hi Kleinstein

I've checked this opa111 with my tester, it was found ok recognized as LM741; then i look around this area near Z300, i found a leaky Q304 i replaced it, but in this time the error that's was given by DMM is Erestr210 hardware failure, it means A5 board is defective too with faulty outguard controller!! so the checksum rom 0, rom1 and rom2 low!!

but i notice during the autocalibration it goes out at 10µA calibration step.

Please help on this new status.

Regards

[TiN]

Replace Dallas NVRAMs.

[HUGOMAN]

Hi to All,

but i don't know if i could really fix this new Error 110 CALIBRATION REQUIRED --ACAL who is appeared finaly with only replacement of Dallas Nvrams?

and shall i replace all of them the 3 ics or less ? the dmm could be working with the empty virgin rams ?  there will be a possibility to restore data systematicaly or what?

But also after performing AUTO Cal it initialse at 10µA ---> Error 205 out of range 95 still appearing.

i attached my board photo for nvrams existing there!

Sorry for correcting several times the text, ok now , the problem is like that.

Regards

[Kleinstein]

The RAMs look rather old (1997 and 1995). So even with no error message it would be good idea to have them replaced. AFAIK one of the 3 holds the important calibration data - here it would be really good to keep the contend if it is still there. The other two should be less critical and the DMM should start even when they are empty (might need a ACAL, but this could be the least problem).

The error in the 10 µA step could still be a problem with the current source, but might also be due to a RAM problem.

[HUGOMAN]

HI KLEINSTEIN,

First thanks for your help.

So as you said the nvrams are old then it required to be replaced sure (because of battery life) anyway, i am not yet removed the chips outside from the board, and i am aware about those memory keeps the calibration data ok.

1/how can i copy the content safely of the critical one? as you said; by opportunity i have a programmer Minipro.
2/ if i loose all the data the dmm won't display anything? the companies who are doing the calibration in this case they will put the data on the nvrams then during calibration they adjust the correct values, is it right?
3/ when we finish to replace those nvrams with new ones and may be the error of current source still there, can we make a manual assembly for this Z300 if yes could you fix me which Transistors reference and the power of resistors, does the schematic of z300 is it real or generic?

Best Regards
Hugoman

[Kleinstein]

It should be possible to replace the Z300 element. For a first careful (no more protection against voltages below about -30 V) test, one could just replace or bypass the Z300 with a wire.
This part is not at all critical with resistor values and transistor types. So more or less generic circuit would work too. It is just there to allow a large negative voltage at the input - leakage is set by the external FET used as a diode. As this passes a current, there is no problem with thermal EMF or similar. The transistors should be high voltage PNPs (good for 1 mA (10 mA short time) at about 250 V) in a small case.

There are a few instructions on how to read the RAMs. AFAIK there is a way to read the calibration data over GPIB - but I am not sure about how to write those data back. So maybe nobody was desperate to do so. In theory the programmer might do that. AFAIK the usual way is to carefully unsolder the RAMs and read with the programmer and transfer to a new chip. I would expect instructions on TiNs website. There is also a thread here on how to replace the RAMs with FRAM - so no more problem with batteries in future.

Not sure on how the DMM would behave with lost data: likely it would display an error on startup.  So a calibration lab should be able to do calibration and this way restore the values. With a suitable reference (10 V + 10K Ohms) this could even be done yourself - though limited by the accuracy of the reference. I  am not sure if AC might need a little more calibration, but I don't think so.

[TiN]

MiniPro can read and write DS1220 (that's where CAL data is stored).
Anyway since your meter have hardware issues , it will not be able to run ACAL/CAL procedures. So I wouldn't worry about calibration at all at this point. After you get meter to a point when it pass SELFTEST and working, then you can just run 10VDC, 10kOhm calibration using source, such as Fluke 54xx/55xx/57xx etc to get out-of-spec initial calibration.

I have mentioned and described test procedures/conditions and NVRAM replacements many times, you can just reference worklog here, here or here or even here.

[Kleinstein]

Even if there is a problem with the ohms ranges, it might be worth keeping the old cal data (those for the 10 V and 10 K value). It won't be so important if there will be a calibration after the repair anyway. Not so old calibration data can still have a value even if the calibration is expired.

[TiN]

Parts replacements on A1 already invalidate old calibration , be it expired or not. It's very easy to bring 3458 out of spec, I learned that many times poking with iron inside the meter. Just to be clear on this for newcomers.

[HUGOMAN]

HI Mr KLEINSTEIN,

COME BACK TO YOU WITH SOME NEWS!!!

The dmm could finally measure the resistance, see the atached photo, awesome, the problem is effectively the z300. i could by-passed we correct it after by a new home module.

so after this i perform the calibration it passe successfully the 10µA it pass over 100KOHMS, 1MOHMS 10Mohms, but unfortunately it stucks on 100n?

I will work on 0,5µA branch precisely on rp302v/u302/ cr308/rp303/cr301/Q308/rp300! Please Kleinstein advice on this step.

for the 210 HDW FAILURE CALIBRATION REQUIRED, i will changes the nvsram mentioned by TIN, i understand i can pull it up and read the content but my question is it the content is good then if it is read with success it means that the ram is in good state!
in this case perhaps the others should be obligatory changed to elimination this error?

Thanks in advance
Best Regards
HUGOMAN