Agilent 3458A repair

[Kleinstein]

The interesting parameter is CAL78 , and one needs to ran ACAL (at least part of it) to get an updated value.
After no being used for longer time (e.g. a few weeks), it may need a week or two to stable again  (likely removing humidity from some parts).

For the reference oven the absolute reference voltage has only very little correlation with the actual temperature. I know of no easy way to directly read the oven temperature. The heater voltage may be a crude starting point, but the heater resistance also varies.
The variation in heater voltage with an external temperature change could in theory give a hint.   

Reducing the oven temperature is a 2 sided thing and also includes disadvantages:
The voltage calibration is lost.
The TC may change and get larger (the refs are likely screened for low TC at the given temperature).
For the first some 12 months there can be more drift, as the reference is settling to a new target.
 The old ref is likely well aged to the set temperature and thus relatively stable by now.

[franklin]

About oven temp; I’m not going to change that but was just interested in a possible correlation between temp and reference voltage. I understand there is a correlation but I guess nothing like that I have a lower oven temp due to a slightly higher reference voltage.

I got interested in that CAL? 78; what does it represent? All this, like CAL? 78, CAL? 59, CAL? 2,1 etc., where can I find this kind of information? Can’t say I’ve seen that in any of the manuals or maybe I’ve overlooked it. Is there somewhere to find info that’s not covered by the manuals?

[Dr. Frank]

Just started testing drift (inspired by TiN and Keysight Service Note 18A). I guess “CAL 2,1” displays the A9 reference calibration voltage and not the present value since the reading is fixed. I’ve been running it for only four days so far and it’s still running. This will be interesting and it’s something new to me just as the 3458A is.

I understand that the display should be kept OFF while the meter is ON for prolonged period to prevent wear and tear of the display. For this I’ve programmed one of the keys with “DISP MSG” and that does the trick for me but I guess this is old news.

TiN has some interesting views on oven temperature; lower it and get lower drift. When doing that I understand the reference voltage seem to raise when die temperature is lowered. Does this say something about oven temp when voltage at my reference A9 is 7,20741? Could it mean that there IS a lower oven temperature in my meter?

CAL? 2,1 (A9 voltage) = 7,20748898; since this doesn’t change with time, I suppose it is the reference voltage when it was calibrated. It matches closely to what I measured on A9 while instrument was open.

Hello Franklin,

you are checking the short term drift of your 3458A, following SN18.
The relevant parameter for you is CAL? 72, that is the DCV gain 10V auto calibration constant, see table of cal constants in the calibration manual, pages 65 - 71.
Do an ACAL DCV once a day and afterwards read this constant CAL? 72, and the internal temperature TEMP?.
This constant should not change too much over one week if the A/D chip U180 is ok, AND if the inner temperature is always the same.
Latter aspect is missing in the SN18A.
Please provide a table of these values after one week, or so.

The long-term drift described in SN18A, of the LTZ1000A, you can not measure, as you are probably lacking a stable and precise external 10V reference.

The external CAL constants 1 and 2, that are the 40kOhm and 7V reference values, are fixed and are set only during an external calibration, therefore can not change by an ACAL, see calibration manual. So your idea about that constant is correct.

The LTZ1000A reference is between 7.0 to 7.5V, typically 7.2V, see datasheet.
Therefore you can not judge from the value itself, on which temperature the oven is running.
The circuit w/o heating has a T.C. of about +55ppm/°C, a reduction of the oven temperature would then decrease this reference voltage by about 12mV, i.e. to about 7.19586 V.
You could as well not determine a modification of the oven temperature from that.


HP, agilent and Keysight always delivered references with 15k over 1k, i.e. 95°C oven temperature, so a modification would have to be done by the former owner of your instrument. Simply look at the reference board, if an additional resistor in parallel to the 15k is assembled, like in my unit, that big 100k. In the picture you made of your instrument, there seems not to be any modification.

If you would do this modification, your instrument would loose its DCV calibration.

PS: unfortunately, our colleague Kleinstein yesterday posted mostly incorrect statements, please ignore all about this.

Frank

[franklin]

Hi Dr Frank
There is no additional resistor in parallel to the 15k on the reference board on my A9 and I didn’t expect that. About the modification; I’m NOT going to modify A9, I’m happy with the present condition. Besides, I under¬stand that having it done results in an expensive calibration which I want to avoid.

Here is the result of the measurement you suggested; seems OK to me. They’re not 35,7 ⁰C all of them but 35,8 is close enough for me. It’s hard to keep a stable temperature in an ordinary house.

Date / Time          DMM   Amb.   RH     CAL? 72        Deviation
(dd-mm-yyyy/time)   (⁰C)   (⁰C)   (%)   (CAL const.)   (PPM)
29-03-2020/2045   35,7         1,00125178   Ref
31-03-2020/2045   35,8   23,4      1,00125174   -0,04
01-04-2020/1010   35,7   23,3      1,00125179   0,01
02-04-2020/1335   35,7   22,6   38   1,00125187   0,09
04-04-2020/0845   35,8   23,2   33   1,00125174   -0,04
05-04-2020/1545   35,7   23,3   34   1,00125176   -0,02
07-04-2020/1215   35,7   24,0   36   1,00125177   -0,01
08-04-2020/1555   35,7   22,8   38   1,00125184   0,06
09-04-2020/1835   35,7   23,6   36   1,00125173   -0,05
10-04-2020/1920   35,7   23,1   38   1,00125173   -0,05
11-04-2020/0830   35,7   23,8   37   1,00125188   0,10
11-04-2020/0905   35,8   23,5   37   1,00125176   0,02
11-04-2020/2135   35,8   23,2   37   1,00125181   0,03
12-04-2020/1130   35,8   23,0   39   1,00125194   0,16
12-04-2020/1730   35,4   23,1   37   1,00125182   0,04
12-04-2020/1845   35,6   23,1   36   1,00125191   0,13
13-04-2020/1720   35,7   23,5   30   1,00125182   0,04
14-04-2020/1630   35,8   23,7   34   1,00125179   0,01
15-04-2020/1045   35,8   23,2   35   1,00125186   0,08

When doing this test, I was running ACAL DCV a lot of times and more than what the list indicates. Something interesting happened and that was the error message which turned up every 5th to every 25th run. Made me wonder if it’s coincidental or maybe a latent error somewhere. The error message is always:

-   ERRSTR 204, “HARDWARE FAILURE - - DC VOS DAC CONVERGENCE:61
-   ERRSTR 210, “HARDWARE FAILURE - - CALIBRATION REQUIRED
-   ERRSTR 101, “CALIBRATION ERROR

Next run and there’s no error. What’s your view on this.

[Kleinstein]

The CAL?72 values look very good. So there is likely no problem with U180.

The message  “HARDWARE FAILURE - - DC VOS DAC CONVERGENCE:61" suggest that there is some iterative method adjusting an offset. The parts may be boarder line here so that sometimes it does not find a good setting or get a residual error close to the edge. So there may be something like an offset that is a little high or DAC problem.

The service manual may give a few more details which DAC / offset is in question.

[TiN]

Most likely that error related to elantec parts on A2 AC board, but my guesstimate may be wrong.

[Dr. Frank]

Hello Franklin,
happy to hear again from you.. good messages.. U180 also to my opinion is fine.

This sporadic failure, I agree to Illya, maybe caused by the ELANTEC comparator chip.
As long as this does not occur too often, simply don't bother.
Try keeping the instrument as cool as possible (it's probably a heat problem).
Please keep the fan clean cool, so that the inner temperature does not rise further (I got it at the same level)


Maybe you would like to become a volt-nuts, also?

Stay healthy!

Frank

[Kleinstein]

For the cooling fan, one may want to check / clean the air filter, if not already done.

[franklin]

Hi Kleinstein
Checked the cooling fan and filter when I got the DMM; no dust at all just as if it was never used.

[franklin]

Hi.
Still running the stability test here.

Now, when reading the Operating Instructions for Agilent 34137A Test Lead Set (which accompanied the DMM I got), there is a statement which is:

This Test Lead Set is designed for use only with the Agilent 3458A Multimeter.
Do not use with other Agilent meters.

Does anybody know the reason for not using these test leads with a different DMM?

The ELANTEC comparator chip, I can’t seem to locate that on the A2 schematic. Am I overlooking something here?

[Kleinstein]

There are a few of the Elantec comparator on the A3 board and there is likely also one on the AC board.

For the test leads the limitation to the 3458 is odd. It may be because they have a simple banana plug without a shroud. The modern DMM usually support the newer safety plugs and should use them if possible.  However there are also still some older meters around with old style terminals.

[Dr. Frank]

Hi.
Still running the stability test here.

Now, when reading the Operating Instructions for Agilent 34137A Test Lead Set (which accompanied the DMM I got), there is a statement which is:

This Test Lead Set is designed for use only with the Agilent 3458A Multimeter.
Do not use with other Agilent meters.

Does anybody know the reason for not using these test leads with a different DMM?

The ELANTEC comparator chip, I can’t seem to locate that on the A2 schematic. Am I overlooking something here?

Hi Franklin,
Meanwhile, you should have had collected enough data to calculate drift and T.C., and to decide how well your 3458A is doing. Any hints or data that you might want to publish here?

Your Test Lead Set looks very similar to the 34138 set which is delivered together with recent instruments like the 34465A, which has recessed jacks, see pictures, like probably any other handheld DMM also.
As the lugs of your set are also isolated but have to fit over 'normal' banana test jacks (similar to the Pomona CuTe ones), either set does not fit mechanically these other instruments.

A photo showing the mechanical setup of your lugs would be highly appreciated.

There are even two ELANTEC chips on the AC/DC board. I marked the comparator EL2018 with that red arrow, it's in the right upper part of the schematic.
You've overlooked this chip, because the designator is written over other items .

Frank

[franklin]

Hi Frank
Spring has come so there has been some assistance (and delay) in gardening on my part. Yes, I will publish data but I do have a couple of questions concerning these measurements. In the start data seemed to be fluctuating a little around the starting point (centre) but this centre seems to be drifting very slowly away but still within limit. I feel I can assume that this is to be expected or am I wrong?

About the test leads; as far as I can see they are suited for any kind of multimeter since the connectors have retractable shrouds. Nice set but I like the probes from Probemaster in USA better. I have enclosed pictures of the probe set.

About the ELANTEC chip; is replacing it something to think about if it proves to be a problem or will that ruin the calibration?

[Dr. Frank]

Hello Franklin,
to understand and judge your drifting data, I need to see them. No chance for me to imagine anything this way. Your data from 15th of April already were fine.
These retractable shrouds may get stuck in the jacks of other DMMs, or may not retract far enough, so that the contacts can not stick deep enough into the jacks.
The ELANTEC chips are not available any more, the ones offered on the common marketplaces are mostly or very probably fake ones.
Their exchange would not affect the AC calibration, as this chip is not inside the core AC circuitry.
The Volt and Ohm calibration of course would not be affected at all.

Frank

[franklin]

Here are data for a 28 days period. I’m guessing short time drift of about 0,2 ppm/day is not unusual although it seems excessive to me (with little experience).

Date / Time   DMM   Amb.   RH   CAL? 72   Deviation   
(dd-mm-yyyy/time)   (⁰C)   (⁰C)   (%)   (CAL const.)   (PPM)   
29-03-2020/2045     A   35,7         1,00125178   Ref   
31-03-2020/2045   35,8   23,4      1,00125174   -0,04   
01-04-2020/1010   35,7   23,3      1,00125179   0,01   
02-04-2020/1335   35,7   22,6   38   1,00125187   0,09   
04-04-2020/0845   35,8   23,2   33   1,00125174   -0,04   
05-04-2020/1545   35,7   23,3   34   1,00125176   -0,02   
07-04-2020/1215   35,7   24,0   36   1,00125177   -0,01   
08-04-2020/1555   35,7   22,8   38   1,00125184   0,06   
09-04-2020/1835   35,7   23,6   36   1,00125173   -0,05   
10-04-2020/1920   35,7   23,1   38   1,00125173   -0,05   
11-04-2020/0830   35,7   23,8   37   1,00125188   0,10   
11-04-2020/0905   35,8   23,5   37   1,00125176   0,02   
11-04-2020/2135   35,8   23,2   37   1,00125181   0,03   
12-04-2020/1130   35,8   23,0   39   1,00125194   0,16   
12-04-2020/1730   35,4   23,1   37   1,00125182   0,04   
12-04-2020/1845   35,6   23,1   36   1,00125191   0,13   
13-04-2020/1720   35,7   23,5   30   1,00125182   0,04   
14-04-2020/1630   35,8   23,7   34   1,00125179   0,01   
15-04-2020/1625   35,7   23,4   35   1,00125181   0,03   
16-04-2020/0935   35,7   23,5   33   1,00125180   0,02   
16-04-2020/1330   35,6   23,4   31   1,00125171   -0,07   
17-04-2020/1710   35,7   23,0   31   1,00125162   -0,16   
18-04-2020/0935   35,5   23,4   33   1,00125171   -0,07   
19-04-2020/0655   34,6   22,3   35   1,00125168   -0,10   
19-04-2020/2120   35,0   23,1   34   1,00125161   -0,17   
20-04-2020/1245   37,6   26,0   33   1,00125168   -0,10   
21-04-2020/2035   37,4   25,5   33   1,00125157   -0,21   
22-04-2020/1805   35,0   23,7   34   1,00125178   0,00   
23-04-2020/1310   35,7   23,2   35   1,00125176   -0,02   
23-04-2020/2005   35,7   23,1   37   1,00125169   -0,09   
24-04-2020/0945   35,7   23,6   37   1,00125166   -0,12   
25-04-2020/1440   35,7   23,4   38   1,00125168   -0,10   
26-04-2020/1115   B   35,5   22,5   36   1,00125163   -0,15   
                  
Drift   (A-B)*1E+6 / A*D         
      A   B      A          D=28 days
   (1,00125178-1,00125163)*1E+6 / 1,00125178*28 = 0,00535

[Kleinstein]

The short time variations can be just noise, temperature, magnetic field or humidity effects and maybe some hysteresis. For this reason it usually takes longer to establish a long time drift value.  The 0.005 ppm/day look good. So I would say no need to worry so far.

[TiN]

Yup, good healthy ADC. 

With precision instrument like 3458A I would consider anything changed under the hook as automatic need of recalibration. But if your precision/traceability requirements are not high, ACAL ALL after replacement would factor in changed comparators, since AC board is part of internal calibration procedure anyway.

[franklin]

Yup, good healthy ADC. 

With precision instrument like 3458A I would consider anything changed under the hook as automatic need of recalibration. But if your precision/traceability requirements are not high, ACAL ALL after replacement would factor in changed comparators, since AC board is part of internal calibration procedure anyway.

I’d like to keep the 3458A up to peak both in precision and traceability but I’m not quite aware of what it takes in terms of effort, professional calibration or what. Professional calibration I would say is out of the question (there goes the traceability) for an amateur like me; the reason is the cost which is about 1600 USD. So, aside from that, what would you suggest I do?

[Dr. Frank]

I’d like to keep the 3458A up to peak both in precision and traceability but I’m not quite aware of what it takes in terms of effort, professional calibration or what. Professional calibration I would say is out of the question (there goes the traceability) for an amateur like me; the reason is the cost which is about 1600 USD. So, aside from that, what would you suggest I do?

Hello franklin,

please excuse me, that I answer this question, on behalf of Illya (TiN).
Me, like many other so called 'volt-nuts' (metrology amateurs) around the Globe are in the very same situation, i.e. being owner of a 3458A, needing import of Volt and Ohm in their labs, but not needing and not able to afford official calibration or traceability.
Therefore, there are several Cal-Clubs, which exchange 10V and 10kOhm references, which allow to check the 3458A, and even might allow to adjust it to a good enough estimate of uncertainty.

I recommend that you 'simply' try to contact other volt-nuts in your country, or in Europe, whether an exchange of standards is possible, and which (estimated) uncertainty is available.
The 3458A has the big advantage, that all analogue functions are auto-calibrated from only two references, apart from the RF adjustments, which you probably might not need.

If you would already own a DIY LTZ1000 and 10kOhm reference, the exchange would be easy, for the cost of these references, and for the postage costs.

Frank

   

[BU508A]

Hello franklin,

if it's doable for you, there is another option what you can do:

The Maker Faire in Hannover. The PTB is one of the exhibitors there and they are coming with calibrators. Sadly, for 2020 the Maker Faire in Hannover is cancelled because of Covid-19. But 2021 can be an opportunity for you to compare your 3458A with some gear of the PTB. And the best thing is: it is free of charge as far as I know. And even if you don't want to travel with your 3458A perhaps you can assemble a 10kOhm reference resistor and a 10V reference as Frank has already suggested.

This is, of course, no official calibration, put perhaps a good starting point.

PTB @ Maker Faire 2019: https://maker-faire.de/maker/ptb-natuerlich-genau/

Some impressions:





Here are some further explanations (in German) from the German mikrokontroller.net forum:

https://www.mikrocontroller.net/topic/398160

[franklin]

Hello Dr Frank and BU508A.
Thank you very much for the info and advice. I will try to locate somebody who share this interest. And the 10 V reference and 10 k resistor, I have neither so there is a challenge. About the Hannover fair; I visited back in 1970 but today this would be a little difficult.

[KingSolomon]

I remember that Marco Reps did a video several years ago where he brought his LTZ1000A reference to check it against their 3458a.