USA calibration club

[cellularmitosis]

Sorry I fell off the face of the earth!  Life got in the way of my hobbies.  I’ll make time this week to get the reference into the hands of the last few people who jumped on the band wagon.

[bitseeker]

Welcome back to planet Earth. We were afraid you may have become an ex-pat to some wonderful new planet and left us all behind.

Yeah, real life does have a way of interfering with important stuff like tinkering with electronics and T&M equipment. No worries. Glad to have you back.

[technogeeky]

I'd like to throw my name back into the ring, too. I was one of the first people to get my hands on it, but I couldn't make use of it so I just forwarded it quickly. I have finally fixed at least one of my precision multimeters so I can make a useful measurement of it now. Well, useful for me. I doubt it's good enough to help any of you guys.

Let me know if it's coming back to this quarter of the country, and if someone can intercept.

[bitseeker]

I believe everyone who was in Round 1 will be participating in each subsequent round. There were a few late additions to Round 1, which is what cellularmitosis is taking care of now.

Last I recall, for Round 2, the route will be changing from a loop to a star/hub, returning to cellularmitosis after visiting each point/spoke to check for post-shipping variation.

[dr.diesel]

If we decide to do a round 2, a testing standard definitely needs to be developed.

[CatalinaWOW]

bitseeker very generously paid all of the postage for the first round.  I know it wasn't a huge sum, but he shouldn't have to carry the whole load.  Early in the thread there were some suggestions about how to cover postage costs.  At the very least I would suggest that in the next round each participant pay postage to the next location.

The easiest implementation for this might cause privacy issues for some, as it would require a complete name and address for each participant.  Would this be a problem for anyone in the club?

[bitseeker]

bitseeker very generously paid all of the postage for the first round.  I know it wasn't a huge sum, but he shouldn't have to carry the whole load.

I believe cellularmitosis covered the postage. In any case, I agree that we should distribute that burden.

Early in the thread there were some suggestions about how to cover postage costs.  At the very least I would suggest that in the next round each participant pay postage to the next location.

The easiest implementation for this might cause privacy issues for some, as it would require a complete name and address for each participant.  Would this be a problem for anyone in the club?

It depends if the route for Round 2 is chain or star. A chain does leak address information, but that can be minimized by everyone creating their own label to the person after them.

For Round 1, since all the postage was pre-paid, each person along the path could see the mailing labels for all the destinations after them. Forum names were used on the labels (which caused a USPS employee to write a "?" on the label that was addressed to me ), but real addresses, of course, had to appear there.

In my case, the address information leakage is OK as inbound mail is held at the post office. If anyone wants to send me free goodies, let me know.

[vindoline]

At the very least I would suggest that in the next round each participant pay postage to the next location.

The easiest implementation for this might cause privacy issues for some, as it would require a complete name and address for each participant.  Would this be a problem for anyone in the club?

I agree that we should share in the costs. My address is no secret. Anyone is welcome to visit, as long as they like large dogs...

[vindoline]

I REALLY need to get my GPIB datalogging worked out before it comes around though!

[kj7e]

At the very least I would suggest that in the next round each participant pay postage to the next location.

The easiest implementation for this might cause privacy issues for some, as it would require a complete name and address for each participant.  Would this be a problem for anyone in the club?

I agree that we should share in the costs. My address is no secret. Anyone is welcome to visit, as long as they like large dogs...

Never met a dog I didn't like.

[bitseeker]

If we decide to do a round 2, a testing standard definitely needs to be developed.

I'm all for learning about such things. Let the defining begin.

[dr.diesel]

If we decide to do a round 2, a testing standard definitely needs to be developed.

I'm all for learning about such things. Let the defining begin.

It's up for discussion, what is the goal for round 2?  A few first thoughts:

 - Is anyone actually using this as a cal standard?  What can we reasonably expect from this as a mobile transfer standard?
 - The reference has two separate outputs posts, are we all utilizing the same one?
 - Cabling consistancy
 - Meter setup, NPLC, Math, temp, polling frequency
 - Documentation file format
 - Other notable potential environmental factors?

If this goes around again, and we end up with say 20 datasets, how do we make this data useful to the group, or the reference?

[TiN]

Lack of transparency still an issue, from non-participant POV  .

Cabling consistancy

Resolve by shipping with cable set.

Meter setup, NPLC, Math, temp, polling frequency

I'd expect settings best for accuracy for specific meter?
Temperature should be fixed to metrology level range : +20-+24C.

Is there a photo of whole current setup?
Round two should be testing current setup by somebody who has reference standard with lower uncertainty, so travel box can be verified that it's still in spec (whatever that is?) and comparisons made by previous members from run 1 are in line.

So essentially it should be few "golden members" with calibrated equipment to verify travel box, and "green members" who compare their gear to travel box.
I have L&N 4030B 10KOhm resistor to donate for 2nd round, which was measured by 3458A/002 over a month.

[CatalinaWOW]

First, my apologies to cellularmitosis for not correctly thanking him for the first round postage. 

Second, I have no concerns about sharing real world name and address information with other members of the club.  The only risk is to those who use the information as my lab and shops DO have materials known to cause cancer and other hazards to health.

Finally, I will clearly be a green user, but would like to contribute something to this.  While I can't provide traceability, my measurements can possibly make assertions about noise and stability.  One suggestion would be that in addition to the raw data each user provide brief summary information of the results.  Things like mean, standard deviation, shift from prior observation for each instrument used.

[bitseeker]

Lack of transparency still an issue, from non-participant POV  .

What would help with transparency? What can we provide?

Cabling consistancy

Resolve by shipping with cable set.

Some cables were included in the package, though I suspect that not only do we want to use them specifically, we also want to use the same ones for the same purpose each time (i.e., pair 1 between power supply and standard, pair 2 between standard and DMM1, pair 3 between standard and DMM2, etc.).

Meter setup, NPLC, Math, temp, polling frequency

I'd expect settings best for accuracy for specific meter?

I posted some of my settings with summary data in the thread. Additional info as well as raw data is in the log files uploaded to xDevs. Sampling was at the maximum rate that BenchVue would capture, which I assume is based on the instrument's NPLC.

Temperature should be fixed to metrology level range : +20-+24C.

That sounds prudent. Unfortunately, for me, there was a heat wave when the standard arrived, so my time with it was well above this temp range.

Is there a photo of whole current setup?

Taking photos of all the items in the roving package crossed my mind, but alas, I did not take them before sending everything to the next participant. Perhaps cellularmitosis can take a photo inventory before sending it out again.

Finally, I will clearly be a green user, but would like to contribute something to this.  While I can't provide traceability, my measurements can possibly make assertions about noise and stability.  One suggestion would be that in addition to the raw data each user provide brief summary information of the results.  Things like mean, standard deviation, shift from prior observation for each instrument used.

I, too, am a green user. Summary info of setup and results from Round 1 are in the following posts.

• Initial setup and checking the power supply for powering the standard
• First run summary and graphs
• Second run summary and graphs (post-heatwave, but still a few degrees higher temp than desired)

Raw data is in the xDevs share that was set up for our cal club (/USA_Cal_Club/Round1). CatalinaWOW and I created individual subdirectories in there for our results.

Recommendations on improvements to methodology, recording, reporting, etc. are most welcome.

[Conrad Hoffman]

I'm green and am using it for calibration, not being able to justify using a local lab right now. I make no effort to hide myself,  hoping it will attract people that need machine or design work, though the result of that has been zero for years, save for a huge amount of junk mail. 

In any case I'd be happy to pay for postage. The last package was already at the limit for weight, so I couldn't add much of anything. It would be nice if we had a bit of freedom there.

[nikonoid]

Ideally here is what I would want to have in a kit, in no particular order.

1) 10V Voltage Standard  (Geller Labs)

2) Metal shield for Voltage Standard

3) Power supply for Voltage Standard

4) GPIB to USB controller. Smallest and lightest available (initially this would only be driving one instrument at a time). 

5) Temperature/Humidity sensor (BME 280) and cabling. We may also consider two sensors (one inside voltage standard shield, another next to DMM being evaluated).

6) Raspberry Pi with software that already supports: GPIB, Temperature Sensor, and wide and hopefully growing number of DMMs. It will provide HTTP interface, common logging format and whatever else we need, including posting files to public server.

7) Low thermal leads leads for voltage standard (possibly just pure copper wire with Teflon insulation). It must be easy to connect to both voltage standard and banana plugs, since many participants do not have binding posts). Maybe use Pomona 5405 and pure copper crocodile clips. Some input from more experienced members is needed here.

8 ) 10k Ohms resistor. 10K is probably the best value. Anything much higher than that is harder to measure precisely. Ideally this resistor should be small and light (easier and safer to ship), have 4 wire low thermal connection and wiring, have fairly low thermal coefficient because many members cannot control temperature and be time stable. Maybe this could be a custom wire wound or z-foil resistor in small metal case with copper wires permanently attached. Help of experienced members is needed here too.   
 
9) Low thermal short, to measure zero offset for volts and resistance, and also noise. Fluke 884X, plus possibly draft guard for it.

10) Padded case, like cheap Pelican copy from harbor freight: https://www.harborfreight.com/watertight-protective-case-small-63518.html. Alternatively we can take a look at custom padding for Small or Medium flat rate box from USPS.

11) If the Voltage Standard is capable of driving 1mA without loss of precision and resistance standard is not compromised by continuous 1mA current, we can also have 1mA current standard, making sure required cabling is included.

Essentially everything but your meter will be included in the package, guaranteeing the highest consistency possible.

Items that we do not yet have are fairly inexpensive and I will be happy to contribute to the cause.

Since everything is already prepared the time per hop will also be minimal. What do you guys think?

[CatalinaWOW]

Ideally here is what I would want to have in a kit, in no particular order.

1) 10V Voltage Standard  (Geller Labs)

2) Metal shield for Voltage Standard

3) Power supply for Voltage Standard

4) GPIB to USB controller. Smallest and lightest available (initially this would only be driving one instrument at a time). 

5) Temperature/Humidity sensor (BME 280) and cabling. We may also consider two sensors (one inside voltage standard shield, another next to DMM being evaluated).

6) Raspberry Pi with software that already supports: GPIB, Temperature Sensor, and wide and hopefully growing number of DMMs. It will provide HTTP interface, common logging format and whatever else we need, including posting files to public server.

7) Low thermal leads leads for voltage standard (possibly just pure copper wire with Teflon insulation). It must be easy to connect to both voltage standard and banana plugs, since many participants do not have binding posts). Maybe use Pomona 5405 and pure copper crocodile clips. Some input from more experienced members is needed here.

8 ) 10k Ohms resistor. 10K is probably the best value. Anything much higher than that is harder to measure precisely. Ideally this resistor should be small and light (easier and safer to ship), have 4 wire low thermal connection and wiring, have fairly low thermal coefficient because many members cannot control temperature and be time stable. Maybe this could be a custom wire wound or z-foil resistor in small metal case with copper wires permanently attached. Help of experienced members is needed here too.   
 
9) Low thermal short, to measure zero offset for volts and resistance, and also noise. Fluke 884X, plus possibly draft guard for it.

10) Padded case, like cheap Pelican copy from harbor freight: https://www.harborfreight.com/watertight-protective-case-small-63518.html. Alternatively we can take a look at custom padding for Small or Medium flat rate box from USPS.

11) If the Voltage Standard is capable of driving 1mA without loss of precision and resistance standard is not compromised by continuous 1mA current, we can also have 1mA current standard, making sure required cabling is included.

Essentially everything but your meter will be included in the package, guaranteeing the highest consistency possible.

Items that we do not yet have are fairly inexpensive and I will be happy to contribute to the cause.

Since everything is already prepared the time per hop will also be minimal. What do you guys think?

If this route is followed then items 1-3 should be permanently wired together to avoid any cabling variations.  Thinking about that leads to concerns about how to investigate possible future contributions of error from each of the three elements.

Supplying GPIB controller has both benefits and problems.  The major benefit I see is assuring everyone has a chance at data logging.  The problem is that compatibility with various instruments varies, and will force those who have already resolved those issues to do further debug.  If the controller is only single instrument capable it also precludes simultaneous logging of multiple instruments.  Not good.  But these issues are somewhat resolved if the Rasberry Pi is included.

Item 7 is resolved if items 1-3 are permanently wired.  The travelling kit already includes a pair of binding posts with a banana pin out.

A pelican case ups the shipping bill automatically through both size and weight.  I do think it is worth biting the bullet to go to a standard sized USPS package.  We should be able to fit a medium box easily (~$15) but might be able to get into the small (~$8) one.

I do like the idea of a standard data format (must be extensible for multiple instruments and future adds like current and resistance).

[Conrad Hoffman]

What is our purpose? I'd think if the device is well characterized, its performance should be predictable. Are we mostly measuring the test environment to confirm that the measurements change as expected? Or, are we doing a long term evaluation of the device, with the added variable of doing it in multiple locations? I'm not trying to promote any particular thing, but I think it's important to be able to clearly state the goals.

[nikonoid]

Conrad, I think we are doing multiple things.
First of foremost we are collecting a rich and consistent dataset. By writing a little bit of a software we can determine a number of things, like:

1) When device reaches our "calibrated" members, then other members can see where they stand in terms of calibration.

2) We can compare noise generated by our equipment. Is K2000 generally better than 34401a? Is my K2000 noisier than some else's K2000.

3) When standard returns to calibrated members with very stable equipment in stable environment we can judge stability of the standard itself and any problems the shipping might have caused, power up/downs temperature in transit, mechanical stress.

4) We have a lot of old aged equipment. In my experience old equipment runs better than manufacturers specs that had to be conservative and include "new device" drift. If we run this club for several years and have large enough membership, we might be able to ascertain if our equipment is better then spec and how much. For example if 34401a 10V spec is 40ppm per year, we might discover that typical drift for old instrument is 2ppm. In my mind that would be a useful information.

These are all examples how we can crowdsource the data, so that every contributor can benefit from data that others collected.

As logging from multiple instruments is concerned, I believe seeing something about instruments sometimes affecting each other, with noise, extra cabling or mismatched impedance. Logging them individually may take more time, but would be more consistent.

For GPIB, I started with a $30 logger from Amazon that "sometimes" worked. Then I bought old HP 82357a for $80 (original, not fake) and it supports every device I have, no fiddling required. It supports multiple devices too if needed.

Between GPIB, low thermal short, raspberry pi, 10k resistor, etc. we are probably looking at around $200. Shared metered members, it will be compatible to postage cost. I am certainly willing to contribute. We can PayPal contributions to CM.

With proper prep, round 2 will be great.




Sent from my iPhone using Tapatalk

[TiN]

I'd say verifying any spec even at 6.5 digit meter, like K2000/34401 mentioned need a standard bit better than AD587 or attached environment data for both calibrated condition and DUT setup.
And since there is multiple mentions about shipping cost, seems like 20$ of whatever it is now is bit much for participants (?), that already rules better gear out?

Leave alone noise, which is tricky to get correctly even at one location by single member. How would you know that member's B K2000 noise larger because of his meter , not the Wifi router nearby under the desk? Sure, member B can evaluate possible noise sources and try to obtain lowest noise possible on his setup, but for that he need no cal club reference..

Most valuable outcome in my opinion is going after 3) to estimate how much stress/condition cycling affect the DUT. And this means that all non-calibrated members would just need to be a transit points, with their time on the gear (doing whatever they see fit for their own benefits), and all major data collected by high-end gear (more stable than a DUT box) at calibrated members.

Anyhow, I have few ideas for next year calibration club, on which i'll share the details as right time comes (other that it wouldn't be low-cost, that's I already know  ).

My offer of L&N 4030B stands (free of charge  ), I bought it specifically for community and it's just collecting dust somewhere in USA right now.

[ManateeMafia]

Yeah, that 4030B is collecting dust. So are a few of his other things. Sorry! 

[cellularmitosis]

Howdy gents!

I hooked up the SVR-T this weekend and took about 20 hours of data with my 34401A (and a Si7021 temp/humidity sensor), and now I've got the reference packaged up and ready to ship after work tomorrow -- I believe nikonoid was next in line.

I plan on doing a "star" configuration this time around, where the reference will hop back to me in between each participant.  I'll take at least 8 hours of data each time it comes back to me, so that we can start to get a better handle on the repeatability / consistency of these measurements.

As far as shipping costs go, that's great if anyone would like to pitch in.  How about this: I'll pay to ship it to you, you pay to ship it back?  So far, I've spent about $50 in shipping over 6 months, so this is a pretty cheap endeavor

Lots of great discussion in this thread, and I totally agree that it is time to start tightening up the process, in terms of trying to get each participant's setup as similar as possible, etc.  I've got some ideas I'll be trying out in that vein.

More details to follow soon!  Excited to get back into the swing of volt nutting 

(attached: ~20 hours of 34401A data, slow 6-digit mode, 10 GOhm input impeadance, plotted with a 100-sample rolling average filter, units are microvolts)

[cellularmitosis]

Also, I decided to take the plunge and start working on a resistance transfer standard, so I ordered a few VHP202Z's.

I decided to go with a 10k and 20k resistor in series, so it should be useful to calibrate the 10k, 20k, or 30k range of a meter, which I think will cover all of the meters we are likely to use.  For example, a 34401A would use the 10k value, the Keithley 196 would use the 30k value, and I believe the Fluke meters use a 20k scale.

I got three pieces of each, so that I can mail out one unit while keeping the other two in-house, which should allow me to figure out which one drifted.

[bitseeker]

Very cool, CM. I think TiN has a 10K resistor to add to the mix as well.