Precision Resistors for Checking Calibration of DMMs

[Humber888]

This may be a silly idea but.....

Some time ago I bought a set of 26 precision resistors by Kelvin Industries, with a view to using them to check the resistance ranges of my DMMs, etc. All were unused and long leaded. They weren't cheap, but much better value than buying a few new precision Vishays. They range in value from 1 Ohm to 1M Ohm, generally 4 per decade, where the central values are all 0.002% tolerance. For example, the lowest decades have values of:

1 Ohm, 1%,
2 Ohm, 1%,
4 Ohm, 1%,
8 Ohm, 0.5%,
10 Ohm, 0.5%,
20 Ohm, 0.2%
40 Ohm, 0.1%
80 Ohm, 0.05%,
100 Ohm, 0.05%
200 Ohm, 0.02%,
400 Ohm, 0.01%,
800 Ohm, 0.005%,
1.0 kOhm, 0.002%,
etc.

I cannot find a datasheet on these resistors so don't know the tempco or ageing characteristics but, given the tolerances, these should be commensurate. Needless to say I cannot measure the actual core values with my instruments to anything like their stated tolerances.

So here is my half-baked idea. If a forum member has access to something like a recently calibrated 6-digit DMM or anything with a better accuracy figure than the resistors themselves and is willing to measure them and tabulate their current values, I would then be prepared to let others use them as secondary 'standards' to check the calibration of their instruments.

This would involve a level of trust on everyone's part and a few obvious provisos such as no soldering or shortening of leads, no stressing of the components with high voltages or current i.e. calibration purposes only and not Meggers! The primary costs involved would be postage and probably insurance. Obviously I would cover P&P costs if someone did volunteer to accurately measure the resistors initially.

So what does the team think?

Mike

[The Philpott]

A generous idea. Anything with a tolerance of 0.2% or less may be worth assembling into a resistance box, which you can then take to *get-togethers and swapmeets, and charge a bit per test for charity? This would protect them all from rattling around and getting leads bent.

Anyone with 'mains only' test equipment would be out of luck, but then again checking things to four or five decimal plates could be said to be a bit of a black hole which tends to rob sufferers of their spare time.. ?

(*When normality returns)

Dave

[wireman]

Originally Posted by The Philpott

A generous idea. Anything with a tolerance of 0.2% or less may be worth assembling into a resistance box, ...

I did exactly that last year, I purchased one each of 1k/10k/1M 0.1% 5ppm/C resistors and put them in a box with terminals.

On my two meters I measure...

---MS8221--87V---
1k 1.000k 0.999k
10k 10.02k 10.00k
1M 1.001M 1.000M

I don't yet have a bench DMM.

I had arranged to borrow an Agilent 34401A from a work colleague to obtain a more accurate measure of the resistance for them but the first Lockdown put paid to that.

I also purchased a (cheap) voltage reference around the same time.

I did this as I wanted a way to check if any of my meters were going out of calibration.

[pmmunro]

How well your scheme will work depends on the accuracy of the meters you want to check.

If the values above 800 ohms are to the same tolerance, even a bench multimeter will probably only be one order of magnitude more accurate, depending on the model, conditions of use and how recently it has been calibrated.

The resistance range of a bench multimeter is the least accurate, but this will also be true of hand-held meters. I have attached the specification for my Fluke 8088A 5 1/2 digit bench multimeter which might help with assessing what might be done.

Without seeing a data sheet, its not easy to know how stable your reference resistors are, but if they are good quality components, I think you might find they are verifying the bench multimeter rather than the other way round.

I would guess that you paid quite a lot to Kelvin Industries for the set of resistors and so that price should entitle you to a data sheet.

PMM

[Humber888]

To answer PMM's question, I did not buy these resistors direct from the manufacturer but as unused, sealed old stock - hence them being under £100 in total but also the lack of data. I can find no real references to Kelvin Industries on the web. Some of the resistor values are plastic encapsulated but many are of a sealed metal and glass construction (all are axial). All values above 800R are 0.002% tolerance except as you get to around 1M when this slightly worsens.

Looking at your Fluke's specs, if it can be shown that the resistors are still pretty much within tolerance, they would be ideal for checking the resistance ranges - being approximately 10 times better toleranced than the basic specs for the DMM.

I am in two minds about boxing them up vs. leaving them to be used individually with test leads. You would need a lot of terminals for the whole 26 of them, plus the low values should really be able to handle 4-terminal measurement. Then there is the question of how much any soldering would impact them and how good the terminals should be. This is less of a problem with 0.1% resistors.

Finally, I have a few National Semiconductor LH0070-2H 10V dc references dating for the 1970s. I can't remember their exact specs off hand, but they are pretty good and at least the annual drift should be very low by now!

Mike

[David Simpson]

Hello Mike,
I'm sure there must be a number of folk down your way with access to high grade Ex Cal Lab T/Eq. to suit your purpose.
I've an ex C/lab Cropico PW3/E Resistance Bridge which I use for accurate resistance measurement. A Marconi TF2700 Universal Bridge which is pretty spot-on. And, using Ohm's law - an ex C/lab Time Electronics 2003N (0.02 Grade) Voltage Potentiometer (1uV to 9.9999V) coupled with a Solartron 7040 Bench DMM on it's uA DC range. This 2003N/7040 set-up I also use for testing meter's fsd's. I haven't phooked a meter or resistor yet. For my own week by week DMM or AVO8 resistor cal. - I use ex AVO high spec. resistors out of scrapped AVO 8's, etc.
If I can be of any help, let me know. Just P&P involved.

Regards, David

[Sinewave]

It's not a silly idea at all, in fact it's how calibration labs do just this, with resistance decade boxes. I use one from Time Electronics to do this.

[Pamphonica]

I suggest you box up your precious resistors and enjoy them, knowing they are safely protected and not subject to passing draughts.
I also acquired (NOS) a set of resistors which were specified as 0.01% or 0.005%. For years I had them loose but finally bit the bullet during lockdown and put them in a diecast box. The simple layout is the same one used by Vishay for boxes based on their precision resistors.
I now have a very compact 4-terminal resistor source that I trust reasonably well.
I have meters up to 8.5 digits, all out of cal. But at 2 years+++ out of cal even the best would be at about +/- 50-100ppm on resistance. So that's about the same spec as the resistors I have.
So I sit back, relax and stop worrying about precision, and know that it will be close enough!

[The Philpott]

A forum member distributed a broad selection of precision resistors to several members about a year ago (i think it may have been Norman?) They have been useful for peace of mind. Quite large with hand painted values and tolerances, it's about time i put them into a box, possibly using setscrews as terminals and just jab onto them with probes.

The 30kohm shunt in one of my old analogue meters was found to be o/c, and got replaced with 3 good quality resistors of appropriate value in series- each is marked 0.01% tol. so in combination they should be very accurate- although installed in a meter they still get used as my 30kohm reference whenever a DMM is suspected of wandering.

Dave

[David Simpson]

I forgot to mention that the T/Eq. I mentioned, apart from the S.7040, are all battery powered, hence jolly stable.
Unlike lazy me, Jeremy's & the other guy's recommendations to fit your resistors into a decade type box is spot-on, & Jeremy's looks the bee's knees.
I've also got three Decade resistance boxes, two cheapo ones & a lovely old Muirhead Lab - Grade one. So I've plenty of belts & braces, resistance accuracy-wise. Quite frankly, I only ever really need such accuracy, unless its determining shunts & multipliers for low uA fsd meters in homebrew projects.

Regards, David

[Humber888]

David, thanks for your offer. Nobody has expressed much interest in the idea of 'post-a-res', so I will just keep the resistors for my own use. I may box them up if I can raise the enthusiasm - the current lockdown state does seem to have made me even lazier than usual (if that's possible).

[David Simpson]

Don't be put off Humber888, its a jolly benevolent idea. Perhaps if you wait till BVWS Swapmeets etc. re-commence &/or the R&TV Museum opens, then arrange to meet up with a Forum chap who has an NPL Standard Resistance Bridge or some high-spec Agilent gear with current certification.
I ken SFA about Kelvin Industry's products, but I suspect that your NOS resistors need not be doubted as to their stated values.
The only reason I delved into high spec resistors, meter fsd accuracy, etc., was because several years ago I was given over 20 odd AVO8's, when a college lab undertook refurbishment. Being over 40 years since I'd been involved with servicing & repairing RAF ones, thought I'd better get my arris in gear. However, my AVO experiences pale into insignificance compared to PMM & a handful of other Forum metering enthusiasts.

Regards, David

[trobbins]

There can be many accuracy issues starting to risk a 0.002% tolerance.

A setup like in post #8 manages a few accuracy issues like where on the resistor lead you are probing; keeping the resistor bodies clean and free from touching; keeping the temperature stable; managing humidity (eg. with a sachet of moisture absorber); and using a known connection terminal to manage junction voltages.

Another source of stable resistances for the home calibrator can be to pick up a vintage Wheatstone bridge. Apart from cleaning up the beefy rotary contacts, and possibly repairing a step resistor or two (due to past accidents), the manganin coils have likely drifted to a stable spot in their life, and they can offer hours of fun checking the drift of each coil and noticing the odd coil that has gone out of original tolerance, and how higher resistance coils typically drift more due to a higher voltage/power test setup.

There is a strong metrology group on eevblog.com - it can be gobsmacking to read some of the posts there.