DMM Volts/Ohm Calibration Club - anyone interested?

[enut11]

SEE CHANGE OF SCOPE REPLY #28

I have a HP3456A which is un-calibrated but still nice to use for many hobby metrology projects or for just learning about small signal measurements. It would be nice if it were calibrated at minimum and adjusted at best. However, I am not about to spend $2000 (or whatever) to have it tested and adjusted but I might consider spending ~$200 if I had access to a known stable voltage and resistance standard.

Proposal:
Who would be interested in being part of a Calibration Club? The idea is that 10 people pool an amount of money sufficient to calibrate and adjust one of the members' high end DMM (say a HP 3458A). In exchange, that member would make available by post a voltage standard and a resistor standard checked against his 'reference meter', to the other Club members.

Once a Club member has checked his own DMM he then returns the volt/ohm standards to the reference member who would then re-check the standards ready to ship to other Club members.

Could something like this really work?
What voltage/resistance standards would be required to satisfy potential members?
What are the pitfalls?
Would the outcome be better than no calibration (current situation for many)?
What would YOU pay to have more confidence in your 5-8 digit DMM?
enut11

[JS]

  I like the idea, while is no use for me right now and I don't think I could join from here. That could start an extraoficial network of traceability, without the liabilities or the cost but without the paper work either. Maybe it gets a day which becomes a GNU version of the standards... 

JS

[TiN]

What voltage/resistance standards would be required to satisfy potential members?

1. Characterized 10V standard based of LM399 for DCV (over 20-40C ambient temp, over 12-20VDC input)
2. Characterized 10K standard based on hermetic foil or wirewound resistor for OHM (over 20-40C ambient temp, over 1-10mA current)
3. Perhaps 1mA current source based of same LM399 as self-check point
However this would not help to calibrate most of the meters other than 3458A, so it's rather only for base range check.

What are the pitfalls?

Few : Shipping cost. Module can be made small, but still would be 40-80$ to ship internationally.
Risk of loosing the package. Also frequent shipping back is required to make sure reference still agree to original values.

Would the outcome be better than no calibration (current situation for many)?

If one after absolute accuracy - yes, of course it would be better.
If just need ratio measurements - not so much.
But perhaps it's less useful for more popular 6.5-digit DMM owners, because even for simple check you need 1V,10V,100V,1kV/10ohm-1Meg in decades, not just 10V/10KOhm.

The idea is that 10 people pool an amount of money sufficient to calibrate and adjust one of the members' high end DMM (say a HP 3458A)

I'd skip on this idea, and modify it into "few members that already own calibrated reference meter help to check travelling reference box". Not much would like to pay for calibration of someone's meter with unknown outcome, as this "club program" would be a long-term thing, not a "all done in one month".

I can jump in and help, but will need exact and defines conditions of such activity to be ran, so it would be helpful.
Ideally I can see this as next path:

A. Design and build few pieces of "Travel-box" with DCV,OHM reference in it and perhaps Raspberry Pi Zero with EEPROM to store measurement data/temperature/environment condition and test member name.
B. Send it to well-known volt-nut "A" to calibrate for DCV,OHM reference value over temperature range
C. Assign measured data to box
D. Send it to club members "A" -> "B" -> "C" -> "D" -> "E" -> "F". Set hold time e.g. 2 weeks for each member, so it would take 5 months for 5 members to loop a reference (2 wks shipping, 2 wks test)
E. Last member "F" sends unit back to "A" to recheck and recalibrate if required. Deviation and delta of each member data would be recorded
F. Repeat cycle again.

[enut11]

Thanks for the comprehensive feedback TiN. I like the all-in-one roving standard. Several voltages and resistances would be ideal as well as a current standard. Perhaps a pool of money to fund this box is in order. I do not have the skills (yet) to build a 'metrology' grade reference unit.
To reduce travel time and cost, each country could have its own calibration club? I live in Australia. enut11

[2N3055]

Sounds interesting.. Subscribed...

[Dr. Frank]

In metrology community, such Key Comparisons with Travelling Standards like 10V zeners, 10kOhm, ratio divider and AC/DC convertrers, are very common.

This is useful for participants, who would afterwards be able to transfer at least 10V and 10kOhm to all other ranges and modes of their DMMs, i.e. who have appropriate transfer standards available in their own labs, requiring a 3458A, or calibrators like the Fluke 57x0A.

Otherwise, the bare exchange of 10V or 10kOhm will not really calibrate DMMs, as these usually need about 35 different artefact values.
The "check" of the say 10V and 10kOhm range of your DMM also has absolutely nothing to do with "calibration".
That's why I regard these DMMCheckers and alike as being quite worthless.

So yes, that idea is nice, but this will have a different scope than the title implies.

Here in Germany, we already have quite some exchanges of artefacts, instruments, or better, we sometimes meet at some volt-nuts homes to do elaborate comparisons or calibrations.

Frank

[enut11]

Thank you Dr Frank. How this evolves, and I hope it does, will depend on what the members want. For myself, I'm after more confidence in what my DMM is telling me without spending a large amount of money. I am open to suggestions on how to go about it. enut11

[VintageNut]

I am interested as well.

I can some checks now and am in process of building a divider to transfer voltages.

I  have a calibrated KE2000 and I borrow calibrated gear on a regular basis. Recently I have been working on learning how to construct and measure a divider. I use a DMM7510 which for resistance between 10 ohms and 1 M is 27ppm to 35 ppm. 10VDC is 15.2 ppm.

As good as the DMM7510 is, its obviously better to have a 10V standard that is known to better than the DMM uncertainty.

To be serious about this, temperature has to be known. Should there be a uniform device to measure temperature for all of the participants?

There should also be standard practices for cables and measurement techniques. My cables are not great now. I can see the offsets from moving connectors.

[Theboel]

I do not know what the story in other parts of the world.
but here in Indonesia I think I am lucky about calibration (depend with Your point of view) for calibrated voltage standard (10v), Resistor standar (cardinal value), Frequency standard (10Mhz).
I can calibrate it in our NMI (KIM-LIPI) against their standard traceable to BPIM with minimum of money (150USD to 200USD) and because I am the only person in my country have ask to calibrate my standard to national standard they give me extra effort for me.
1. I do not need to wait up to 6 month like their company customer.
2. They help me to do regular calibrated (in my case every 90 days) so I can made a history about my standard.
all this thing may be a very special case in my country but I think no need to shame to ask Your NMI about what service they can do for You. I already talked to 2 foreign NMI basically they surprised when I tell them who I am and why I need their services and their very helpful about my question but they suggest me to talked to my NMI first before use their service.

[Alex Nikitin]

For a while now I am thinking about a portable DC only "calibrator" with the following values available : 1V, 10V, 100 V (up to 1mA max output so it is safe) , 10mA, 100mA, 1 A (up to 1V max output), 1K, 10K, 100K, 1M . It should work from an easily available supply (say, a 5V USB adapter or an external battery pack with a USB connector) so you don't need to worry about power arrangements, <10ppm accuracy on voltage, <20ppm on resistance and <50 to 100ppm on current ranges. A fairly small size (1x2x4'')Captive 5-way leads with low thermal emf banana plugs, so it could be just plugged in. This kind of a unit can be easily posted by a recorded small packet post. Does it look like a good idea?

Cheers

Alex

[2N3055]

For a while now I am thinking about a portable DC only "calibrator" with the following values available : 1V, 10V, 100 V (up to 1mA max output so it is safe) , 10mA, 100mA, 1 A (up to 1V max output), 1K, 10K, 100K, 1M . It should work from an easily available supply (say, a 5V USB adapter or an external battery pack with a USB connector) so you don't need to worry about power arrangements, <10ppm accuracy on voltage, <20ppm on resistance and <50 to 100ppm on current ranges. A fairly small size (1x2x4'')Captive 5-way leads with low thermal emf banana plugs, so it could be just plugged in. This kind of a unit can be easily posted by a recorded small packet post. Does it look like a good idea?

Cheers

Alex

That sounds like a fantastic idea..  I think ideal would be to have one that travels and also to have one of my own in the lab permanently..... I would be interested..

[2N3055]

The idea is to ship voltage reference.. LM399 voltage refs have very low thermal hysteresis, and do not drift when off.. The rest of the circuit is not so stable though..
Many people send their equipment via packet to get calibrated...
For me it wouldn't be so much a source of high end Voltnut quality reference voltage, but something I would use as a double check to check my voltmeters... To characterise at least partially meter's reference drift and to know when to send meter for cal+adjustment...

[HighVoltage]

I have several Geller Labs 10V references and could make one available for this purpose.
It would actually be interesting to see a change in the reference a few weeks later, after
it has traveled through some very cold places and airplanes.

My reference equipment would be an Agilent 3458A, a Keysight 34470A and a Keithley DMM7510
I could set the Geller Labs reference as close as possible to 10V and keep it running for 24h before sending it out.
I would also enclose a printout of the values measured buy the various instruments.

[enut11]

Thanks. Some great suggestions and ideas are being posted. Some of you have already built your own calibrators. Would you mind sharing a link to the description and construction in this thread?

I definately want more than a simple DMM checker - looking at those LSDs!

A hard blow moulded plastic foam lined case should survive the rigors of the postal and shipping services and protect the delicate calibration gear.

Nice as it would be to participate in an International Club, it may only work within a country if what DiligentMinds says is true of customs problems.

[retrolefty]

I've never come to terms of the 'calibration' Vs 'adjustment' definition.

 I would think what is being proposed here is just the sharing of a specific value voltage and resistance standard, they are calibration sources only if used to make adjustments such that when 'adjusted' the meter DUT readings match the published measurements of the shipped references? Otherwise it's just some kind of forum scoreboard if users wish to share their DUT readings for comparison?
 
 Are these two simple references all one needs to 'adjust' one's specific DMM, probably not?
I have no problem considering this proposal as a 'calibration validation program', but to consider and call it some kind of 'calibration program' still seems wrong to me.

 

[2N3055]

Well, as someone more knowledgeable than me explained, in metrology circles term "calibration" is actually measurement of your instrument in comparison to standard and documenting data..
A "check of calibration" would be more appropriate term..
In layman terms they give you table of how much your instrument is in error... To actually make your instrument accurate they call it adjustment or even repair sometimes, because if a component drifted out of adjustment range it needs to be replaced before it can be adjusted... Some labs do check and automatically adjust instrument if it is out of it's specified range, if you have such an agreement with them..
Many companies have such contracts with calibration places, so most of us equate "calibration" with "check of calibration and adjust or repair if necessary to make it accurate" because we wouldn't  worry about it.. You would give it to the instrumentation guys to take care of it..

So that is how it was explained to me..

[2N3055]

To explain my rationale: I have one good 6 1/2 meter still in cal...
I don't have 2 or 3 good meters like some other people here and frankly I don't think I can justify getting them..
But I could have one (or three) decent LM399 based references, and I think that would give me some certainty that I have something to check upon..
When my meter drifts enough as evidenced by those references ( and they agree) I will send meter for cal...

When it comes back I would start monitoring again...

I would be happy with just 10V for beginning and I would mainly like to check upon meter's reference drift. 

I'm not expert, can someone explain me how much of instrument drift comes from reference itself (for instance LM399 in RIGOL DM3068) and how much from other sources (dividers etc)

Best regards,
Sinisa

[enut11]

OK, if all you want is just a "DMM Check Standard" that provides some known artifacts for the meter to measure, then that is possible.  The price for this depends on how accurate you want these check standards to be.  You might be able to get away with some PWW or epoxy-packaged foil resistors (or maybe even good quality metal-film), an LM399A based voltage reference, and DCI, ACV, and ACI derived from that.  Maybe we can price the resistors (up to 1M\$\Omega\$) from Edwin, and then use whatever we can find for the rest.

What voltage(s) and/or current(s) and/or resistance(s) would everyone want?

1) The very minimum that I would want is a calibrated 10v DC and 10k that is within the performance tolerance of my HP 3456A, taking into account any temperature difference when these standards were calibrated.

2) Beyond that, it would be better if 1v were also available along with 1k and 100k

3) Wish list, but not necessary, would be also 100mV and 100v along with 100r and 1mohm and a current standard.
enut11

[TiN]

I'd start with very detailed and realistic (not like 5V USB-powered 100V 10ppm source specification for such travel-box.
To mitigate customs/shipping damage issues it should be priced relatively low (300-400$USD tops?) so loosing a box would not be end of the world.
This rules out expensive options, like using LTZ1000, hermetic 0.005% foil resistors and such, as just parts alone for these would overwhelm project cost by a lot.

Few months ago I actually did something which this thread imply (sorta), sent my one of my 2002s via post, and took second 2002, all LTZs and resistors std's with me on airplane to perform a calibration against fresh calibrated standards. It took a week of ~14 hour/day work to get just two 2002's and LTZ modules/chips (7pcs) calibrated, and that's not including ACV/ACI at all.
Then I took both 2002's and references back to my homelab, crosschecked them and confirmed that DCV still agree within ~2ppm on both meters and LTZ modules (unless all of them drifted same way from transport shocks, which is unlikely) and within ~8ppm for resistance. This allowed me to adjust 3458A to these absolute accuracy levels, and I could do third round of checks on references using 3458A (which stayed home and ran 24/7).

Of course there was some amount of luck that 2002's did not get smashed and destroyed or lost, and similar stuff. So I'd rather call this extreme way to calibrate homelab gear.
Again, if I did not have 3458A (which can calibrate from 10V/10KR), usefulness of such experiment would be much much less, as most 6.5-digit DMMs require proper MFC with many points to calibrate, so whole thing would boil down to check few voltages only.

I can help on hardware design for such a project and testing if community desire, but I'd prefer not to do it all myself at this point, as I have few other homelab projects to finish up first.

[Nuno_pt]

I'm with Diligent Minds on the problem that can be shipping International because of customs, in EEC there is no problem but if it's going/arrive from US or Asia that could be a problem, because it could be retain in customs almost 2 months, and one would have to pay custom tax plus 23% VAT, but if this project would go forward I'll ask the customs if there can be an exception for the product stay here for 2 weeks period, and share the replay here.

In my understanding we've to agree on some DCV and Ohm values, then agree on components, someone designing the PCB's for the DCV, order the PCB, assemble everything and then initiate the traveling for the box, preferable all in one box kind of thing.

The person(s) that is going to do the measurements should have the DMMs calibrate, if not the readings will/can be wrong, so would it be better if we could arrange first a network of unofficially adjust the DMM, some of the guys here have the proper equipment to adjust 6.5 digits DMM, if they will be willing to do that it's another story, but as an example, shipping one of my DMM (K2000 or 3457a) to anywhere in EEC would be between 30€ to 50€, that way I could have the DMM calibrate, and could do the checking properly.

For logging the parameters we could use an Rpi and BME280, the BME280 will log Temp, Pressure and Humidity.

[Alex Nikitin]

I'd start with very detailed and realistic (not like 5V USB-powered 100V 10ppm source specification for such travel-box.

That is not a very difficult bit. 10ppm accuracy and 90 days stability for 10V is do-able even without a heated reference, using, say, LT1021-10 in a metal can + LM35 to measure the reference temperature, and to scale it up to 100V with a self-calibrating Hamon divider can be accurate to <1ppm short-term. For me a difficult bit would be the current up to 1A with a reasonable stability from 500mA 5V available from USB supply  . I took the requirements for a full DC calibration of K2000/K2015 as that is what I have here and it would be nice to be able to check these easily. For the voltage only I can do a travelling reference fairly soon and test it for tempco and hysteresis.

Cheers

Alex

[TiN]

I'd avoid USB-powered anything for precision analog stuff to begin with. Same goes to battery, as postal services usually don't allow battery stuff for shipping. After Note7 shenanigans I'd expect rules to harden even more.

[Alex Nikitin]

I'd avoid USB-powered anything for precision analog stuff to begin with. Same goes to battery, as postal services usually don't allow battery stuff for shipping. After Note7 shenanigans I'd expect rules to harden even more.

No battery or power supply to send - that is the point. 5V from a USB charger/supply or a battery pack (sourced locally) is easily and cheaply available. An isolated DC/DC converter to provide regulated +15 and +110V for the reference and outputs, with a good isolation so the polarity can be swapped easily. As the power consumption is low, a simple (and cheap) low noise resonant converter will do the job (even if LM399 is used as a reference). 

Cheers

Alex

P.S. - with the LM399 it would be also possible to provide a direct reference output and anybody with a good linearity 6.5 + digits meter will be able to check 7v to 10V conversion ratio, to verify the 10V value, so the uncertainty will be reduced down to the reference drift and the linearity error for 10V, plus Hamon divider error for 100V. All resistive dividers errors could be controlled that way. A little bit more involved but a good cross-check for a travelling reference.

[2N3055]

I personally have no problem providing 24 V for the box.. for 100V would  still would need upconverter...

[enut11]

I searched www.eBay.com for "0.005% resistors" and got the following hits. Would any of these be suitable for this project?



1x Z204 100K000 0.005% Vishay Z-Foil Series Metal Foil Resistors Y1441100K000V0L
http://www.ebay.com/itm/1x-Z204-100K000-0-005-Vishay-Z-Foil-Series-Metal-Foil-Resistors-Y1441100K000V0L-/111862808308?hash=item1a0b8afef4:g:CCEAAOSwX~dWhNkB

NYT ULTRA PRECISION RESISTOR ARRAY 3PCS 10K OHM 4PCS 100K OHM 0.005% P444A
http://www.ebay.com/itm/NYT-ULTRA-PRECISION-RESISTOR-ARRAY-3PCS-10K-OHM-4PCS-100K-OHM-0-005-P444A-/401193617684?hash=item5d6900c914:g:4TwAAOSwOyJX4~h3

1X 10K/1K111111 0.005% VISHAY Precision Voltage Divider Resistor Metal Foil seal
http://www.ebay.com/itm/1X-10K-1K111111-0-005-VISHAY-Precision-Voltage-Divider-Resistor-Metal-Foil-seal-/111726922486?hash=item1a03718af6:g:oukAAOSwMmBVs8Dz