3458A resistance measurement off by quite a lot + calibration questions?

[daqq]

Hi guys,

I bought an HP 3458A, backed up the calibration memory and put it through the standard checks. As far as I can tell the voltage/current range is accurate, but the resistance is off by quite a bit.

As a known resistance I have used a 0.005% 100 Ohm resistor (Vishay PG VFR). The measured value by a 6.5 digit DMM was 100.0005 Ohm and can be assumed to be accurate.

The 3458A measures:

100 Ohm range: 099.93925 Ohm
1 kOhm range 0.0999389 kOhm

So, something's off. After looking at the calibration table I got by the awesome GPIB utility ( https://xdevs.com/article/hp3458a_gpib/#calgpib ) I saw the following line:

19:41:24 60000 - 39,972.9570342778 (39,997.51395) - 40K reference
19:41:24 60008 - 7.11940040146345 (7.170674313) - 7V reference

Now, the 40k Reference resistor (I assume) is quite different from the typical value (in brackets). They have a ratio of 1.0006143382 . Now, here's the funny thing:

Multiplying the measured 99.93925 Ohms by the ratio gives me 100.0006465021 Ohms - which is much closer to the (assumed) real value of the measured resistance. 

So, somehow, the calibration got messed up or switched, or done wrong? It seems that the typical value should be the one in the calibration. What could have happened? Wrong calibration procedure?


Also, I've been looking into the other lines of the calibration. These look a little worrying:

19:41:24 60626 - 73 (27) - Destructive Overloads
19:41:24 6062A - 2,920,366,096 (0) - Defeats

First of all, shouldn't 'destructive overloads' me 0 or borked? Also, what does the Defeats line mean? Mine seems to have had a lot of them. Whatever they are...

Thanks guys,

David


Note: All resistance measurement was done 4W.

[TheSteve]

Is the meter supposed to have a known good/current calibration? It is quite easy to use any ~10K resistor someone wants to "calibrate" it.

[dacman]

Have you tried a CAL 10000 (Ohm)?

Destructive Overloads will increment anytime the 3458A auto-ranges from the 10 V range or below to the 1000 V range (ac or dc).  It can increment several times if ac is set to a slow response.

Defeats should only increment if DEFEATS is turned ON and the 10 V range or lower is exposed to a peak voltage of +/- 100 V or more.  It is off by default, so I would guess that someone either did a lot of programming (DEFEAT is listed several times in program examples) or just wanted it to respond faster.  It is supposed to be faster if DEFEAT is turned on, but the input protection is turned off if this is done.  Defeat is not supposed to be turned on unless it is known that all the syntax is correct and that the 10 V range and bellow will not be exposed to more than +/- 100 Vp.  It has a Defeat counter for warranty purposes.  I'm not familiar with the syntax of your post, and if Defeats is listed in parenthesis (0) then it doesn't have any.

[ap]

You should do a new calibration. It is not worth analyzing the meter, since you just bought it and have no CAL history (I assume) from which you could derive something. A precision meter with no proper calibration is not of much value. Keep in mind, also the 100ohms resistor only has 50ppm accuracy, so if you do it yourself, try to use a better 10k one, as the 50ppm will add to ohms and currents uncertainty.
I have never seen a used 3458A that did have destructive overloads with less than a 2-digit number, I would just ignore this initially and do the cal and a full verification.

[dacman]

daqq,
Could you enter CAL? 2941 and scroll the display and post the result?
Also, could you enter CAL? 2437 and do the same?

Thanks.

[daqq]

Thanks for the tips guys.

Could you enter CAL? 2941 and scroll the display and post the result?
Also, could you enter CAL? 2437 and do the same?

When I get home I'll try.

Is the meter supposed to have a known good/current calibration? It is quite easy to use any ~10K resistor someone wants to "calibrate" it.

Well, the meter was calibrated/checked in 2013 and there everything passed.

I will probably try the 10k calibration, maybe use this fellow: http://sk.farnell.com/vishay-foil-resistors/y145310k0000v9l/resistor-precision-10k-0-005/dp/1867937 as a reference.

[mimmus78]

If 10V is right I suppose they swapped the A3 (?) board or the 40K resistor has drifted. Current ranges uses this 40K ohm resistors during autocal ... are you sure they are right?

Inviato dal mio Nexus 6P utilizzando Tapatalk