Solartron 7150 Plus Out of Cal

[Sam__]

Hey everyone. I have recently purchased a Solartron 7150 Plus in the hope of using it as a transfer standard for my hand held multimeters, and also generally as another multimeter. I was aware of the problems with the mains filters, and mine blew up on me. It's finally starting to smell a little less after a few weeks now. 

Anyway. I decided to look into getting it calibrated, but before that I thought I'd test what I could to see if it was wildly out. I has some 0.01-0.1% Vishay resistors to try in 100ohm,1k,10k and 100k values so I hooked them up. The 100ohm read as ~93!!! the rest were >0.75% from the absolute value they should have been. I was shocked. Could this have something to do with the filter going? Sadly I have no voltage or current references at the moment to try with it to see how far they are out.

I'm in the UK and have contacted Calibrate.co.uk to ask if they can take a look and adjust it but they told me they can only test and not adjust (it uses a GPIB to make adjustments), so not overly useful. I don't really want to spend £70 to find out how far out of cal it is. Does anyone here know of somewhere in the UK that might be able to adjust the meter? Is 7% too far to adjust? I do have the option of sending it back from whence it came and getting a refund if it comes to that.

Thanks in advance.

[Paul Price]

It is way out of cal, send it back, it's now nothing you need or lack. Get a used Fluke or HP that can be cal'd. Burnt IC's can make you puke.

[Sam__]

So what am I looking for which can be easily adjusted? Something with pots to twiddle with my tongue at the right angle??

[Paul Price]

yeah

[krivx]

Just to be sure, you have checked these resistors on another meter, right?

[ecat]

Firstly, the 7150 Plus can be calibrated from the keypad, iirc this is the entire reason for the Plus as the base 7150 is IEEE calibration only. If you do go for calibration, as another forum member discovered it is wise to check your chosen calibration service has the equipment needed to deal with 6 1/2 digit meters.
 
Secondly, your problem may not be due to calibration. If my understanding is correct every range has its own pair of calibration factors, the offset at 0 and the slope required to match the calibration value, y = mx + c. If the calibration is out on all ranges then all the factors are wrong, it's certainly possible that the unit was incorrectly calibrated at some point, its also possible that some common component has become marginal.

Thinking about your resistance tests. The meter has the ranges 2k, 20k, 200k, 2M, 20M, so your 100r resistor and your 1k resistor were both measured on the 2k range? The 100r was out by 7% and the 1k by > 0.75%, unfortunately '>0.75%' does not tell me a great deal but if 100r is out by 7% and 10 x 100r (1k) is out by approximately 0.7% there is certainly a pattern, possibly a constant offset? What does the meter read when the inputs are shorted?

[Sam__]

Just to be sure, you have checked these resistors on another meter, right?

I have multiple resistors and have tested all on multiple meters so it has to be the Solartron.

[Paul Price]

Didn't mean to be rude, just sensible.  What good is a 6-1/2 digit lab standard if it is only accurate to 1 or 2 digits?  If the calibration lab can't touch it, how are you going to cal it?  And if there exists a circuit malfunction creating this way out of expected behavior, how are you going to fix it, and if you have a chance to send it back then, (considering these problems of repair and even calibration) and get your money back to buy something that works, why don't you???

[Sam__]

Firstly, the 7150 Plus can be calibrated from the keypad, iirc this is the entire reason for the Plus as the base 7150 is IEEE calibration only. If you do go for calibration, as another forum member discovered it is wise to check your chosen calibration service has the equipment needed to deal with 6 1/2 digit meters.
 
Secondly, your problem may not be due to calibration. If my understanding is correct every range has its own pair of calibration factors, the offset at 0 and the slope required to match the calibration value, y = mx + c. If the calibration is out on all ranges then all the factors are wrong, it's certainly possible that the unit was incorrectly calibrated at some point, its also possible that some common component has become marginal.

Thinking about your resistance tests. The meter has the ranges 2k, 20k, 200k, 2M, 20M, so your 100r resistor and your 1k resistor were both measured on the 2k range? The 100r was out by 7% and the 1k by > 0.75%, unfortunately '>0.75%' does not tell me a great deal but if 100r is out by 7% and 10 x 100r (1k) is out by approximately 0.7% there is certainly a pattern, possibly a constant offset? What does the meter read when the inputs are shorted?

How can it be done via the keypad?? Maybe I didn't read the maintenance manual hard enough but all I remember seeing was how to do it via GPIB.

Sorry I wasn't very precise with my numbers. 100ohm (0.01%) was out by 7.31%, 1K (0.05%) was out by 0.74%, 10K (0.01%) was out by 0.15% and 100K (0.1%) was out by 0.12%. This was averaged over 3 resistors.

So the higher ranges aren't too bad. It's just the lower 2k by the looks of it. Which as you said could be a failing component.

The meter read pretty much zero, maybe 6-7 counts off zero, when shorted.

[Sam__]

Didn't mean to be rude, just sensible.  What good is a 6-1/2 digit lab standard if it is only accurate to 1 or 2 digits?  If the calibration lab can't touch it, how are you going to cal it?  And if there exists a circuit malfunction creating this way out of expected behavior, how are you going to fix it, and if you have a chance to send it back then, (considering these problems of repair and even calibration) and get your money back to buy something that works, why don't you???

This is what I'm leaning toward. But if if can be adjusted via the keypad then there is hope. Unless something is failing inside that is, which could be fun/impossible to track down.

I do think however it's worth trying to learn a bit more about it, even if I do end up send it back.

[ecat]

How can it be done via the keypad?? Maybe I didn't read the maintenance manual hard enough but all I remember seeing was how to do it via GPIB.

Section 5.2 and 5.3 of the "7150 Plus User Manual"
http://www.ko4bb.com/manuals/index.php?dir=09%29_Misc_Test_Equipment/Solartron

Sorry I wasn't very precise with my numbers. 100ohm (0.01%) was out by 7.31%, 1K (0.05%) was out by 0.74%, 10K (0.01%) was out by 0.15% and 100K (0.1%) was out by 0.12%. This was averaged over 3 resistors.

So the higher ranges aren't too bad. It's just the lower 2k by the looks of it. Which as you said could be a failing component.

The meter read pretty much zero, maybe 6-7 counts off zero, when shorted.

7.31% and 0.74%, both on the 2k range? That certainly looks like an offset, eg, 7r equivalent offset is 7% of 100r and 0.7% of 1k. The why, where and how of it appearing is a mystery for now.

You could try measuring the 1k and 10k resistors on the 20k range and the 10k and 100k resistors on the 200k range. It is important that you manually select the ranges for these tests, do not use auto.

Edit:
If you get around to trying a manual calibration you should use values near the top of each range, so near 2k, 20k etc. It's in the manual iirc.

 

[Sam__]

7.31% and 0.74%, both on the 2k range? That certainly looks like an offset, eg, 7r equivalent offset is 7% of 100r and 0.7% of 1k. The why, where and how of it appearing is a mystery for now.

You could try measuring the 1k and 10k resistors on the 20k range and the 10k and 100k resistors on the 200k range. It is important that you manually select the ranges for these tests, do not use auto.

Edit:
If you get around to trying a manual calibration you should use values near the top of each range, so near 2k, 20k etc. It's in the manual iirc.

Well silly me for looking in the maintenance manual haha. I will get on with reading those sections. Sadly I don't have the meter at the moment so I will have to do those tests at a later date.

Could it be that the EAROM has lost the data for that range? The manual says only trust them for 10 years of retention and I'm not sure when mine was last calibrated.

[ecat]

The EAROM can be re-written without going through a calibration but, yes, it is safer to assume the last EAROM refresh was during the last calibration.

I would hope the data in the EAROM includes a checksum and if any single byte were to be corrupted the meter would display an error message. I cannot find that error message 

[Paul Price]

EEPROMs do not lose there data in one small section, extremely unlikely.  Try to find another reason for your off-readings.

[Sam__]

Good to know that it's probably not the EAROMS then. My thought is to set it back to 100ohms from the 93ohm reading using the front panel cal and a resistor. Then monitor any drift over the next few weeks. If it's really bad then I assume it could drift noticeably over a couple of weeks? If there's no drift I can try and cal it more accurately but for the moment I'm just looking for a good starting point.