Solartron 7150 plus unstable volts readings

[Dunckx]

Hi Folks,

My 7150+ has been left on for 5-6 weeks and should be fully warmed up by now.  Mostly it has been used to measure resistance and in that configuration gives stable ohms readings which only vary by about five counts in the last of six digits, which is what I would expect, see ohms 1 and ohms 2. 

It is a very different story with dc volts.  With the inputs shorted and the instrument nulled, within minutes the readings vary from zero to 600uV (see volts 1 and volts 2) and this variation is the same for the output of a 2V lead acid Cyclon cell and for the output of a LM399.

I believe the ohms readings give the onboard 1N829A a clean bill of health and after forty years the reference board ought to be rock solid, so I believe the problem with the volts range lies elsewhere.  I hardly ever measure ac volts or current so can't really comment on these.   The only ac voltage source I have aside from the mains is a frequency source which I could not guarantee the output stability of.

Any ideas regarding what the problem might be?

[Kleinstein]

Neither the voltage reading with a short nor the resistance reading is very sensitive to variations in the reference.  As the Ohm part seems to work OK the ADC itself should be OK, at least low noise.
For the voltage reading the first simple question is if the meter has the correct settings for low drift. A don't know if this meter as some auto zero setting to be chosen - the meter may or may not need this.
600 µV are even poor for the non auto zero mode - so there is likely more to the problem. Quite some drift could be normal in the 200 V or 1000 V range

It could help to connect to a computer to get a bit more quantitative values to the drift / noise. A next point would than be comparing different ranges with a short at the input. This could tell if the problem is before or after the gain.
Zero current should be somewhat similar to zero volts in the 0.2 V range. So it could be worth looking at this too.

Getting so much drift and still a working ohm mode is somewhat odd, especially with 2 wire ohm mode. 4 wire ohm may have implicite auto zero and this way could be better than non AZ votlage.
I would suspect something rather close to the input, like some switching (relay or FET) not fully connecting the input.
Front / rear switching may have dirty contacts - as it is not very often used the contacts may get dirty / oxidized.

[Dunckx]

Much appreciated Kleinstein!

Yes, I thought the stable ohms implied the ref was OK.  I will check the volts ranges with the input shorted and see if there's similar variation on all scales.  I also suspected it might be a relay issue and I think I'll have to ask Coto if there's suitable replacements for the front end relays.

Somewhere I'm sure I have a usb-to-GPIB adapter, but what I don't know is which program I might use to do the data logging.  Suggestions gratefully received - it isn't something I've done before.  I grabbed this thing because it was available and I've never had cause to use it until now.

[dietert1]

For precision it may help to exercise the relays.
E.g. our HP 3457A multimeters have two important relays in the voltage input: Front/back and high/low voltage. As they run logging all the time the relays tend to remain in the same position for a very long time. In order to avoid contact problems, the logging software exercises the relays once every day. Each one opens and closes five times. That seems to work.

Regards, Dieter

[George Edmonds]

Hi Danckx

I have a Solartron 7150+ that has been sitting unused for well over three years.  I have powered it up with a short (2.5mm cable) direct link between the Hi and Lo meter terminals.  With the 7150 set to auto and then pressed Null, after the Null routine had completed the right hand digit oscillates between 0 and 1 and has done so for at least one hour.

I wonder if you are seeing the Seebeck effect from your shorting cable as it looks from your photographs as if you have an external junction for them.

May I suggest that you carryout the same test that I did using only a reasonably heavy cable directly between the Hi and Lo terminals only and then let us know what you find.

G Edmonds

[Dunckx]

Thank you all for your input on this.  I've left it overnight on the current scale and there seems to be little drift.  I've not had the chance to make up a short cable yet and further experiments may have to wait a day or two, but I am increasingly persuaded it is a relay problem.  When I cycle the operating mode and sensitivity there does seem to be some recovery, however it is relatively short lived, up to a few tens of minutes.  Perhaps cycling the settings several times would improve things, but I think I am heading towards relay replacement.  Perhaps NATO gave this one a good hammering before I got it second hand three years ago.

I'd expect at the most tens of microvolts from thermocouple effects, 600uV suggests an alarming temperature gradient somewhere, but I will make up a very short cable and try this one out asap.

[George Edmonds]

Hi

It is quite possible for the 600uV to be the Seebeck effect.  As it can be 50/60uV per degree Celsius depending on the materials in contact so you would only need some 10 degrees Celsius temperature difference between ambient temperature and the temperature inside the 7150+.  Not difficult for an instrument that has been on power for days.

Have you tried using the rear terminal for your tests?

G Edmonds

[Kleinstein]

Hi

It is quite possible for the 600uV to be the Seebeck effect.  As it can be 50/60uV per degree Celsius depending on the materials in contact so you would only need some 10 degrees Celsius temperature difference between ambient temperature and the temperature inside the 7150+.  Not difficult for an instrument that has been on power for days.

Have you tried using the rear terminal for your tests?

G Edmonds

The part from the instrument inside to the terminals is made of copper wire with very low thermal EMF. The observed voltage would have to come from the ouside part, starting from the terminals.
50 µV/K is for a materials pair like iron-Konstantan choosen to give an exceptional high voltage. The material with the plugs is more like brass or similar and thus more like in the 10 µV/K range, or even better. The expected error from Seebecke voltage is more like in the 10 µV range.

A stable zero on the current ranges suggests a working ADC and amplifier.

The main candidate are dirty relay contacts - these sometimes in combination with EMI produce some unexpected voltage. With old meters not using the relays for a long time may be a problem, more often than actual contact wear. Cycling the relays a few times can be quite effective (e.g. for a few years) and regular use can keep the contacts clean. It is not uncommon that meters were sitting unused in a shelf for years before beeing sold.
There is usually no much different in the input path betwee the voltage and resistance mode. So its only a few relays or protection elements to check.

[George Edmonds]

Hi

Although it is quite possible that it is relay problems, the point that I am making is being ignored.  Look carefully at the photographs there is an unknown type of EXTERNAL junction of the three wires from the 4mm plug wires.  No way should such a setup be used if you are looking for small voltage errors.

I used to use Solartron DVM’s in my working life, we called them Solarrotten as a result of all the repairs required

G Edmonds

[Dunckx]

Hi George, I do intend to investigate this asap, at the moment I've not had the time to play.  Expect that to change this weekend

[Dunckx]

OK, I have now had a chance to play, and with either my three way short, or George's suggestion (I ran this last night) the dc volts readings are now stable.  I have cycled the modes between dc volts, ac volts, ohms and current and all across the various ranges in each mode and it is this which has now resulted in stable readings.

Those of you who suspected relay problems appear to be vindicated!  ATM I have a working multimeter but this episode has alerted me to the possibility that in the not-so-distant future I may have to replace a relay or two.  In the meantime I will be following Dieter's advice and cycle the relays regularly.

Thank you all for your input.  This case is now happily closed