HP 34401A DMM settling time comparison / questions - Fixed

[radioFlash]

I have two 34401A units I got at separate times, and both pass all self tests without any issues. However, I noticed that one of the units doesn't seem to settle in a reasonable amount of time from power on, while the other is pretty stable from within a few minutes from turn on.

I used the DMMCheck 5V as reference. Meter A is physically stacked on top of Meter B. Both meters are connected to the same reference.

Below is a chart of the data from the two meters. The data actually starts several minutes after both were powered on after being turned off the night before.

Is there a fault of some sort in meter A?

EDIT:

March 3, 2015:
* I checked the power supplies and outputs were in spec and I found no oscillations. It turned out that U400, an Analog Devices AD706 op-amp was defective. Based on Dr. Frank's suggestion, I replaced the op-amp and it's no longer drifting like it was. Thanks all!

[Andreas]

Is there a fault of some sort in meter A?

Yes, as you can see in the 100uVpp noise.

with best regards

Andreas

[HighVoltage]

Yes, something is wrong with that meter
I have an older 34401A from 1994 and another one from 2014 and both of them settle really fast within a few minutes.

[radioFlash]

I haven't heard mention of a fault like this before--is it likely to be a failure of the heater on the LM399 voltage refrence, or something else?

[dom0]

Always check the supply first. Excess noise, drift and so on can be caused by a faulty (ripple to high, bad regulation, improper voltages) supply.

[DJ]

That's a long time constant. Doesn't look thermal.

[radioFlash]

I measured the voltages from the voltage regulators. Using the multimeter (a BM869) in DC+AC mode, I didn't see any significant ripple and the voltages seem to be in spec. No visible damage to any components that I can see.

[Andreas]

and the voltages seem to be in spec.

sorry but you have to think twice:

6.94 V at reference output and 7.0695V at buffered output of U400 do not fit together.
Either U400 has 130mV offset on one of the Op-Amps (unlikely) or there is something oscillating in the cirquit.
If U400 is working perfect all + and - inputs should have the same voltage.

Edit: and also the -15V seem to have a problem:
It is generated by a -18.2V regulated voltage and a 3.3V Zener.
So I would expect more -14.9 V than -14.5V except when the current through the zener is very high.

But on the other side the voltage drop at the +15V seems to be similar.

With best regards

Andreas

[HighVoltage]

Have you looked at the noise ripple at those DC rails.
I had once a failing LM317 in a 34401A and it injected noise to the +18V rail.
Replacing the LM317 and the attached tantalum capacitor solved that problem.

[radioFlash]

What's the best way to hookup a scope to check the rails for ripple which aren't earth ground referenced? If I connect the scope ground to the rail's ground, I would expect a possible problem.

[dom0]

The analog part of the DMM is floating.

[radioFlash]

Since the analog power supply is floating, I'll need a differential probe to check for ripple, right?

[joeqsmith]

I have two 34401A units I got at separate times, and both pass all self tests without any issues. However, I noticed that one of the units doesn't seem to settle in a reasonable amount of time from power on, while the other is pretty stable from within a few minutes from turn on.

I used the DMMCheck 5V as reference. Meter A is physically stacked on top of Meter B. Both meters are connected to the same reference.

Below is a chart of the data from the two meters. The data actually starts several minutes after both were powered on after being turned off the night before.

Is there a fault of some sort in meter A?

Nice plot.   How are you collecting the data?   

What happened at the center of the data?  Looks like they both hung for a while. 

Are you sure you have both meters setup the same way?   

[dom0]

Since the analog power supply is floating, I'll need a differential probe to check for ripple, right?

Nope, it's all floating, just connect your probe. But don't connect anything to the inputs of the DMM, ya?

[nctnico]

That's a long time constant. Doesn't look thermal.

Unless the reference heater isn't working.

[radioFlash]

I used Benchvue to capture the data. The glitch in the middle is because Benchvue has a 1 hour capture limit and I had to manually restart the capture. I've also been having flakiness problems with Benchvue and my serial connections. Benchvue often has problems starting captures and times out, but once the capture is going, it doesn't have problems. I've tried 3 different kinds of usb-serial adapters (with FTDI, PL2303, and some other kind of chipset I can't remember) and two different USB hubs, but still the same problem.

[LaurentR]

I used Benchvue to capture the data. The glitch in the middle is because Benchvue has a 1 hour capture limit and I had to manually restart the capture. I've also been having flakiness problems with Benchvue and my serial connections. Benchvue often has problems starting captures and times out, but once the capture is going, it doesn't have problems. I've tried 3 different kinds of usb-serial adapters (with FTDI, PL2303, and some other kind of chipset I can't remember) and two different USB hubs, but still the same problem.

To go around the 1hr limit of BenchVue, use BenchVue's predecessor DMM Connectivity Utility. It shows one fewer digit on the 34401A  but the graphs still have all the precision. The utility also has histogramming, which is not in BenchVue's basic version.

As for the RS232 issues, unfortunately, it's a general problem due to the 34401A's old school flow control. There are some discussions about it here:
https://www.eevblog.com/forum/testgear/34401a-through-rs232-with-matlab/

I built a cable with the flow control rewired. If I use Matlab first to set the serial port in the right mode, then the Agilent software (either version), it's 100% success. If I launch the Agilent sw first, it's hit or miss.

[radioFlash]

Thanks for the info on the serial port, I'll have to look into it.

I measured the output of the voltage reference for 15 minutes after turn out, see attached. I don't notice the same trend as when I was measuring the 5V input. I didn't have the best setup in the world for measuring it and there may be noise because of vibration (my measurement platform wasn't particularly steady and of course the cover and shield were off the unit).

[Andreas]

Hello,

is the zener warm if you touch it (carefully its hot).
Is the same noise measured directly at the zener too or only at the buffered output.

With best regards

Andreas

[Zucca]

sed the DMMCheck 5V as reference. Meter A is physically stacked on top of Meter B. Both meters are connected to the same reference.

Before to dig into the schematics I would have done different tests:

0) VERY IMPORTANT! As Dr. Frank  suggested me many times, try to measure R  with inputs in short circuit; both in 2W and 4W configuration. Is the measure stable? Check/compare the results also rear vs front inputs. (The input switch is famous to fail for oxidation.. check the other 34401A posts in this forum)

1) change the voltage range and test another stable voltage source, your DMM supports those DC V ranges:

100.0000 mV
1.000000 V
10.00000 V
100.0000 V
1000.000 V

depending on what V source you have available check if the transient is still there also in different ranges.

2) Try to check the measure the VAC voltage mains of your home with the 2 multimeter. Is the transient still there?

3) does the 4 wire configuration solve the problem or not?

4) What about a known R measure? Is the  transient still there?

5) and what about a current measure? Is everything OK there?

I suggest to understand better what is working and what is not in the DMM before to look for a HW failure inside. The smaller the circle, the less you need to check/measure on the DMM board.

PS: Create a new post in the repair section will maybe help you better!

Good Luck!

[radioFlash]

Thanks all for the suggestions and input.
* The reference is warm to the touch, so the heater is active
* I have checked both front and rear input terminals and the behavior is the same
* I have replaced the electrolytic capacitors and input switch and still see the same behavior

I've run some other comparison tests between my two 34401A's, and below is what I've done so far. All measurements were done after turning on the meters for a few minutes (it takes me a while to get Benchvue loaded and setup before capture). All readings were normalized against the average reading for the hour to make it easier to compare between the two meters. The suspect meter is in blue and my reference, presumably good meter, is in orange.

In summary, the issue seems to be present on DC Voltage  in toe 10 V and 100 V range, and to a lesser extent in the 1 V range. It doesn't seem to be an issue in the AC Volt .1 V, 1V, and 10V range. It also doesn't seem to show up in the 2W resistance mode testing a 1K resistor.

DC Voltage
.1 VDC Input / .1V Range: Meters agree almost exactly
1 VDC Input / 1 V Range: suspect meter's readings change more over the hour, but overall is a very similar shape to to the other meter
5 VDC Input / 10 V Range: suspect meter's change significantly  over the hour, while other meter is nearly flat
10 VDC Input / 10 V Range: suspect meter's change significantly  over the hour, while other meter changes very little
30 VDC Input / 100 V Range: suspect meter's change significantly  over the hour, while other meter changes very little

AC Voltage
.1 VAC Input / .1 VAC Range: Meters agree very closely
3.2 VAC Input / 10 VAC Range: Meters agree very closely, but there is some odd differences in the first 5 or so minutes. (periodic dips in the readings are likely due to my cruddy signal generator)

DC Resistance
2W / 1 KOhm Resistor: Suspect meter takes about 20 minutes to settle, then it gets to within the range of the other meter
4W Short: Suspect meter is pretty flat, but my other meter takes about 40 minutes to settle down and shows periodic oscillations in readings (but that odd behavior will be for another thread...)

[Dr. Frank]

Thanks for the info on the serial port, I'll have to look into it.

I measured the output of the voltage reference for 15 minutes after turn out, see attached. I don't notice the same trend as when I was measuring the 5V input. I didn't have the best setup in the world for measuring it and there may be noise because of vibration (my measurement platform wasn't particularly steady and of course the cover and shield were off the unit).

The stability figure of the DUTs LM399H is not stable enough.
Should be < 1ppm stable.
I assume, you  use the Brymen 869 for that measurement. So you see the instabilities of the Brymen only!

Please use your 2nd, known good 34401A for such measurements only.

Cover off and any "vibrations" (??) should not make any deal, as far as you have stable contacts to the test points.

Frank

[Dr. Frank]

I measured the voltages from the voltage regulators. Using the multimeter (a BM869) in DC+AC mode, I didn't see any significant ripple and the voltages seem to be in spec. No visible damage to any components that I can see.

As others already said, something is fishy here!

Again, only use your 2nd 34401A for measurements!

Use 6 digits, high impedance (> 10GOhm)  DCV mode. Let the known good instrument warm up for 1h.

Please check all internal reference voltages again, see the six red marked locations.

That is LM399H direct out, divided 5V reference, +/- 10V A/D reference, and buffered / unbuffered 7V reference .



Please measure all six points several times, i.e. directly after turn on, and then maybe 5min and 10 min after turn on.

The LM399H should be stable to ppm level, 1minute after turn on.



I suspect U400 to be defect, as these 7.0695V should be exactly 6.94..V , like the direct LM399H output..

Frank

[Dr. Frank]

Thanks all for the suggestions and input.
* The reference is warm to the touch, so the heater is active
* I have checked both front and rear input terminals and the behavior is the same
* I have replaced the electrolytic capacitors and input switch and still see the same behavior

I've run some other comparison tests between my two 34401A's, and below is what I've done so far. All measurements were done after turning on the meters for a few minutes (it takes me a while to get Benchvue loaded and setup before capture). All readings were normalized against the average reading for the hour to make it easier to compare between the two meters. The suspect meter is in blue and my reference, presumably good meter, is in orange.

In summary, the issue seems to be present on DC Voltage  in toe 10 V and 100 V range, and to a lesser extent in the 1 V range. It doesn't seem to be an issue in the AC Volt .1 V, 1V, and 10V range. It also doesn't seem to show up in the 2W resistance mode testing a 1K resistor.

DC Voltage
.1 VDC Input / .1V Range: Meters agree almost exactly
1 VDC Input / 1 V Range: suspect meter's readings change more over the hour, but overall is a very similar shape to to the other meter
5 VDC Input / 10 V Range: suspect meter's change significantly  over the hour, while other meter is nearly flat
10 VDC Input / 10 V Range: suspect meter's change significantly  over the hour, while other meter changes very little
30 VDC Input / 100 V Range: suspect meter's change significantly  over the hour, while other meter changes very little

AC Voltage
.1 VAC Input / .1 VAC Range: Meters agree very closely
3.2 VAC Input / 10 VAC Range: Meters agree very closely, but there is some odd differences in the first 5 or so minutes. (periodic dips in the readings are likely due to my cruddy signal generator)

DC Resistance
2W / 1 KOhm Resistor: Suspect meter takes about 20 minutes to settle, then it gets to within the range of the other meter
4W Short: Suspect meter is pretty flat, but my other meter takes about 40 minutes to settle down and shows periodic oscillations in readings (but that odd behavior will be for another thread...)

Sorry, that comparison stuff is much too complicated and (for me) not understandable.

The error shows up in the 10V DCV mode /range already. That's the most simple one, and therefore you should only track the error in that mode. That's much more systematic, then.


Supplementary to the reference voltages, now check the input path, also.

For that, please put the DUT 34401A in DCV, 10V range, 6 digit SLOW mode, .
Apply 10V externally, from your 731B. 5V externally, from your DMMchecker.

Check briefly the stability and value of the 731B DMMchecker by the other, known good 34401A.

Then check if the input amplifier / multiplexer is ok, you should see exactly these 10V   5V also on pin 4 of U101.

It may be, that this buffered output is also switched on /off, due to AZ measurement, then put your other 34401A in fast DCV mode, so you will see 5.00V briefly flashing, maybe.


You may also measure the 5.000V input voltage before the buffer (maybe relays K103 is defect), that is marked Buffer input, pin 5 of U101.

For that measurement, you have to put your known good 34401A in high Z mode again.

Frank

PS: The last step, if that up to now would all be ok, would be checking the MC output of the multiplexer, and the DCV amplifier.

[radioFlash]

Thanks, Dr. Frank, for the suggestions. I'll measure at the points you suggest.