USA Cal Club Round 3

[Conrad Hoffman]

FWIW, I've got about 6-8 of the L&N resistors on my bench, from various eras and sources. Not everybody used them with mercury and it's not hard to tell. I do wonder if the cal labs usually did use mercury on any that came in for certs. Some show corrosion of the ends of the terminal bars due to being used with mercury, but it was decades ago and they've been cleaned so many times I figure any mercury is long combined or removed. In any case, I don't eat off them and it's a location one doesn't routinely touch in use. If you're really worried it might be useful to rub the area with powdered sulphur, which binds with any mercury, making it easier to remove. My guess is there's no metallic mercury present by now. Some of these resistors are quite old and were standardized before the major change in the value of the ohm. I think those will be high by about 0.05 ohms per 100 ohms. I had a lab tell me my resistor was defective because they couldn't grasp that the definition had changed, not the resistor!

[bdunham7]

My guess is there's no metallic mercury present by now.

So what exactly is the compostion of those horns then?  Assuming it is copper on the inside, what are the outer layer(s) that give it a silver metallic appearance? 

[Conrad Hoffman]

The horns are almost certainly copper but I've no idea what they're plated with. Nickel or chrome seem the likely candidates, with nickel being more likely. Actually, chrome is almost always done over nickel. Not sure they took thermocouple effects into account back then. Reading the instruction sheet, it sounds like the mercury was applied to the end surface of the horns, then they were placed in mercury cups for connection. That would make a certain amount of sense and it would let you sand the surface easily to clean it. The wetting process did sound a bit nasty.

[Rax]

Well, I think it's good we've shuffled and discussed a range of thoughts and positions, from the most paranoid (yours truly - I knew I had this!! ) to the relaxed end. I only brought it here because Randall asked me, but again, I think it's best to have the information at the surface so everyone has more to work with in order to make their mind about how they want to proceed.

In any case, I am currently looking forward to receiving the package (literally in transit). My meters are turned on (well... never shut down ), elbow grease applied, etc etc. Excited!

[rhb]

When I was in grade school,  a filling was always accompanied by a pea size drop of mercury if you asked.  I've never been concerned.  I *always* wash my hands before I handle food or eat.  I'll be 70 in May.

Mercury poisoning is very serious, but  unless you're making hats the old fashioned way it will require considerably more effort tan fooling around with resistance standards.

Reg

BTW  Please add me to the tail end of the list.  It will take a while to get ready, but  I've got a set of standards.  Mostly L&N, but also another maker whom I can't recall without digging them out.

[Rax]

The bird has landed!
Not much time for much else tonight - XL day at work - and no real guarantees the goods have yet acclimated to my bench, but I thought I'd have a preliminary go at Randall's "standard resistors switcheroo" 

This is at 67F (the goods have been in the mancave/garage for about 4h before this, so WTW). The instrument doing the measurements is an 8502A, recent Z540-1. I have not tried to mitigate the readings to account for what my deviations from nominal would be per the Z540-1 "as found" readings.

• 10ohms - 10.00334ohms
• 20ohms - 20.00072ohms
• 100ohms - 100.0000ohms
• 200ohms - 200.0066ohms
• 1k - .999983kohms
• 2k - 1.999895kohms
• 10k - 10.00024kohms
• 20k - 19.99908kohms
• 100k - 100.0002kohms
• 190k - 190.0042kohms
• 1M - 1.000131Mohms

I have not tried to curate these in any way, they're fresh off the press. I did a quick zero of the 8502A, with the rig's switch on short.

[Rax]

What have you folks been using for the GPIB adapter? In terms of software on the computer?

I would use this with a Fluke 8502A, but plenty other instruments at my bench with GPIB.

[guenthert]

What have you folks been using for the GPIB adapter? In terms of software on the computer?

I would use this with a Fluke 8502A, but plenty other instruments at my bench with GPIB.

Least trouble for me was a PCI adaptor from CEC (I also have one from NI, but that one optionally offers high speed GPIB which seems to cause some confusion with older gear even though it shouldn't).  Alas, I don't have currently a PC with PCI slots close to my measurement gear, so I use an Agilent 82357B (or clone thereof, I can't tell).  Occasionally I have to do the rmmod/modprobe USB unplug/plug-in dance, but once working, it continues to work until reboot/power-loss.

[rhb]

Check out @WaveyDipole's  AR488.  You can daisy chain up to 10 devices for under $60 and access via USB.  @artag designed an  Arduino shield which takes care of the wiring so all you need to do is space IDC connectors at the right spacing on the ribbon cable.

https://github.com/Twilight-Logic/AR488

There's a long thread on EEVblog

https://www.eevblog.com/forum/projects/ar488-arduino-based-gpib-adapter/

@artag's board can be found here.

ttps://oshpark.com/profiles/artag

I downloaded the files and got 5 for $10 from JLCPCB.

BoM for a 2 port unit is under $15.  I estimate that ribbon cable and IDC connectors are ~$5 per port.  The ribbon cable works best using one AR488 per stack of instruments

AR488 is pure OSSW/OSHW, so every time I see people buying the counterfeit HPAK devices on ebay I'm always puzzled. 

Have Fun!
Reg

[Rax]

Thank you both, but I was really hoping for more of a quick guide or pointers on this device specifically. This has been in enough hands that I'm hoping a quick start guide has been figured and possibly exchanged. Please keep in mind I only have this for two weeks tops, so a BOM and boards is not what I'm looking for.

I actually have all I need for a whole "GPIB to USB" kit with enclosure, parts, everything, I just never found the time to put it together. And given this kit comes with an assembled interface, it'd be particularly misguided to try doing that now.

[Rax]

On further reporting, the FX seems to be doing very well. I will put more info on the uploads, but this is fed by an HP 6236B, and is hooked with all copper on copper connections, including Adrian B's shielded low-emf banana cable on the hookup to the meter. On a point of wider interest on this - it seems the banana connector on the FX ref has been replaced with a "spade only" connector, which is not represented in the instructions or anywhere else I can tell. In my case, I have no spade connecting provisions, so I improvised with lo-emf bridges from my meters and a couple of quality female banana connectors I have here. Will enclose pictures later.

It started a bit high from dead cold - my meter, per its "as found" Z540-1 data, sees 10.00010V for 10V on the dot. The FX started at about 10.00011V in slight fluctuation with 10.00012V, and within an hour or a maybe bit more, has stabilized to a bit "better" than expected 10.0009V. Per my math, due to the above data, the meter should see 10.00006V for the characterized 9.99996194V (per label on FX). It's currently "rock solid stabilized" at 10.00009V, so a bit closer to 10.00000V than what's stickered on the unit, as far as I can tell.

Any flaws with my math?

[bdunham7]

I don't know if this has been discussed already in great detail (it probably has) but does anyone have a good calculation of the total uncertainty of the FX 10V reference?  Is it quasi-traceable in a manner that we all can look at? 

[CatalinaWOW]

There is some discussion on TiNs website of the traceability flow for the PX reference. 

https://xdevs.com/article/usac_2023/

 I am not skilled enough to evaluate the degree of traceability this gives.  Still trying to wrap my head around it all.
 

[CatalinaWOW]

In my first bit of analysis of my data, I notice in my first test run that there was a substantial settling in period for the PX reference.  The data is in my folder on the xdev site (USA_Cal_Club_Round3_Test 1.Txt).  This was not unexpected since the unit was still quite cold from winter shipping and had just been powered on.  I have no meaningful way to estimate the initial temperature of the reference, at best I could have measured the surface temperature.  It was cold to the touch, outdoor outdoor temperatures were around freezing and the box was delivered in a mail truck driven with doors and windows open (tough mailmen in our neck of the woods).

I went ahead and evaluated the time constant for the stabilization.  Though there is not enough data to separate temperature stabilization and power on stabilization at least it gives a figure to use in planning your tests.  I came up with a best fit time constant of about 76 minutes.   The second digit on that estimate is misleading, the error bars are at least several seconds.

[CatalinaWOW]

What have you folks been using for the GPIB adapter? In terms of software on the computer?

I would use this with a Fluke 8502A, but plenty other instruments at my bench with GPIB.

I use a NI adapter (possibly a clone) with EZGPIB (under Windows) as the software platform.  EZGPIB is free and fairly easy to find.  Unfortunately the developer passed some years ago, and there are barriers to anyone else taking it over, so the present (very usable) state is the way it will be forever.  If you look through the threads for the first two rounds of the club there are references to other software platforms and at least fragments of Python code to do the job.  If you look in the Metrology section there is reference to a program that "Can log data from many kinds of voltmeters."  that may be of interest to you. 

[Rax]

Further reporting (now on to the PX).

But before that, I probably need to clarify that my 8502A has had this Z540-1 I've been talking about done roughly a month and a half ago, so its cal is as fresh as they come.

At 22.5C (starting from dead cold) my meter says 7.04574V. About 2.5hrs later, my meter thinks 7.04572V. So the PX is just as solid as a thick brick wall. This is with an HP 3610A dialed in very close to 15V.

The Vout value on the unit is listed at 7.0456V.

[bdunham7]

What does the 8502A read if you manually select the 10V range, put a short on the input and let it sit for an hour?

[Rax]

What does the 8502A read if you manually select the 10V range, put a short on the input and let it sit for an hour?

Without performing the zeroing procedure, in all scales I'm seeing a perfect zero across the board, except for some occasional 1s in the last digit on the 1V scale, I think.

[Rax]

At Randall's recommendation, I've been keeping track of both the DP8200 and the FX at different temperature points. As expected, the 8200 exhibits more delta, but frankly, less so than I expected. Will plot this out later.

I'd welcome some other ideas as to what I could use the time with the standards for. I'm not really able to automate logging data (my GPIB success seems to be somewhere in my future), so all of this is very analog (paper and pen). But overall it seems both my calibrated meter and the standards are extremely stable (well, duh...) and so I feel I'm a bit running out of things to log, measure, submit.

I am eager to be a useful contributor so please assist with suggestions. I got about a week allocation left, so I'm hopeful to make the best of it. .

[SilverSolder]

Re GPIB, I have found the National Instruments GPIB-232CT-A Serial to GPIB box to be a very easy device to get along with for logging data and/or controlling instruments.

What I like about it:

* Often found on eBay for a good price.
* Old school build quality is excellent - steel box with solid connectors that can definitely take a beating.
* Compact, easy to hide away.
* Easy to connect to a serial port, or you can use a USB to Serial adapter if your PC doesn't have one.
* No drivers needed. Just plug it in, and you're good to go.
* The high-level commands are simple, you can write VBA commands in Excel for data logging or use a terminal program, the box manages the GPIB .
* Compatibility is top-notch. I've used it with a bunch of vintage HP and Fluke gear without any hiccups.
* Comes with old school documentation of all functions, features, and commands.

The only thing that could be seen as a downside is the speed of the serial connection, but I haven't had any issues with it. The simplicity and the freedom from dealing with bulky drivers or specialized software has made this device a game-changer for me. Well worth a look if you want something that "just works" without having to attain a degree in computer science to get it up and running!

[RandallMcRee]

Although I have not received the kit back from the last user (was out of order) I am expecting to be ready to send on to rockstedy40 in another week or so.

Please remember (admonish yourself) to update this thread with any results, etc.

Thanks,
Randall

[Rax]

Please remember (admonish yourself) to update this thread with any results, etc.

Admonished! There's a good amount of data I am still struggling to cleanup and log, but among the outstanding results I've seen is the agreement between the Prema 6048 and the FX. I've seen them further away from each other than this, but I was at the time still learning how to do a proper zeroing of the Prema (I had it for a minute before I had to send the FX away), so the below is probably as close to the truth as my bench and skill allowed during my custody.

What is this... .4ppm off? I know we're getting into gambling territory, but it's still really nice to see such agreement between references. 

[SilverSolder]

I'm delving deep into all the goodies in the USA Cal Club Kit!





Currently running a test with the FX reference.  The laptop in the foreground does two things:

1) Runs and logs the kit temperature/humidity tracking Arduino, using Tera Term for this, which supports logging with timestamps out of the box.

2) Runs a little home brewed Delphi program that sets up the four DMMs, then goes in a loop triggering them concurrently and fetching their measurements (via GPIB) and logs the results to a file.


Since my DMMs are not really capable of measuring up to the level of quality of the supplied equipment, I am doing it "the other way round", i.e. the objective of the test is to characterize the four DMMs by measuring as close as possible the same input at the same time in the same environment, over an extended time period.  For each trigger, all four DMMs take 8,192 samples in 34.1 seconds and averages them, which is the number that is logged.  The two units on the right were DIY calibrated three years ago, using a battery powered reference from VoltageStandard.com, while the two on the left have had parts swapped and were not calibrated afterwards.

I want to thank all those who have contributed to putting this kit together, it is nothing short of amazing to have access to this level of quality references in a home lab setting!

[jjoonathan]

The kit has arrived in North Carolina and I'm logging away! 

One of the banana jacks had a strain relief fatality. The connection is redundant so it isn't an immediate problem, but since I am not very far down the volt nut rabbit hole I thought I'd check in before trying to fix it. Can I just strip back the wire, trim its partner to match, and re-seat the terminals? If I need to apply special anti-corrosion schmoo or perform ppm-preserving rituals I want to make sure to do it right.

[Rax]

The kit has arrived in North Carolina and I'm logging away! 

Nice working space, BTW. Have fun with the kit!