USA Cal Club: Round 2

[vindoline]

TiN, That's great news! I can see how it may be a bit confusing to correctly power the FX, but this is the first time it has happened. Great news about the stability. Regarding the hard case, the USA cal club kit is always shipped in a custom padded hard case to protect all the gear. No problem if you want to send a "special" one just for the FX though!

[TiN]

Reference as received, connected to output spade lug cable. Hard to see from photo, but crappy 3770 posts don't hold connectors well.



Connection at PSU shown and DMMs.



Right now sitting in that poor looking styrofoam thermal chamber with watercooled TEC module to sweep thru 18 - 45 °C, being logged by triple 3458A.

[uski]

logged by triple 3458A

Are you sure you have enough 3458A ? What if all of them fail ? Are you sure you don't need a few more to be sure ? 

[maxwell3e10]

logged by triple 3458A

Are you sure you have enough 3458A ? What if all of them fail ? Are you sure you don't need a few more to be sure ? 

A man with one watch knows what time it is, a man with two watches is never sure. A man with three watches can pick the best two and calculate a confidence interval.

[CatalinaWOW]

If you assume independence, which is probably a fairly good assumption, particularly if tempco of the instrument is compensated you can average the readings for an improved estimate of the value.  And the confidence interval reduces by the square root of the number of 3458s.  All you need is 100 to add a whole digit of precision.

[SilverSolder]

If you assume independence, which is probably a fairly good assumption, particularly if tempco of the instrument is compensated you can average the readings for an improved estimate of the value.  And the confidence interval reduces by the square root of the number of 3458s. All you need is 100 to add a whole digit of precision.

On the bright side, you only need 4 to double the performance!

[uski]

Okay. I saw 3 3458A for sale. Should I buy them all ? 

PS: Just kidding. This stuff is way above my metrology skills. But I enjoy seeing it. And yes there are 3 3458A for sale in China if anyone is interested

[bitseeker]

Got the box. Autopsy revealed that patient died due to autopsy.

Jokes aside, FX reference still works just fine. I've connected it to powersupply for +12/-12V , hooked meter to it and on power on its drawing expected 42mA and 5mA on rails. Output is about +1.5ppm so far wrt assigned value in May 17, 2019. Live data: https://xdevs.com/usac_fx_as_received_k2002x_nov2020/

Why it didn't work before? It was an operator error, now that I looked closer on the posted photo. I think CatalinaWOW will feel relief that it's not borked on his watch I see connected +12V to PSU ch2 output HI and -12V to ch2 LO. Common return wires from either power rail are not connected to anything. So there is no ground provided to reference and no workie as result.

Perhaps instructions on connecting were not clear enough to stress that FX require bipolar power, it cannot work with single +12 supply and was not designed for single supply.

Since I have it here anyway, i will replace bad cracked pomona 3770 posts (very common issue with these low cost posts, sadly) with something better and fully recalibrate the unit. It may take few weeks. Will also donate hard-case for it and make users manual with proper connections shown with photos.

That's a relief, TiN. Glad the patient is fine after all.

If you have hookup photos, I'll add them to the guide along with more explicit info about power requirements.

[TiN]

USA Club revive. 

Reference is happy and ready to go. I have recently revalidated the xDevs Volt, so uncertainty is real nice at the moment

According to my measurements this FX reference survived just fine and after all the shipping and labs it drifted up +0.800 ppm since my last calibration 572 days ago. Measurements done against my primary +10 VDC using my trusty golden Keysight 3458A units. Alternatively measurements also verified with Keysight 34420A nanovolt meter and low-thermal DP 160 scanner with support from ManateeMafia. 

Today's data, temperature sweep 20-30 °C:



Original data, wider temperature sweep 20-50 °C:



If we make huge assumption that LTZ1000A module was drifting linearly during this time frame, we get 0.800 ppm / 572 days = 1.399 ppb/day = 0.51 ppm/year stability. I'd say that is amazeballs for a reference shipped all around hobby labs.

USAC FX PDF calibration report

Also cooked powering note for better clarity, hopefully it will make stuff easier for new members.

Power up procedure guide

I have replaced output crappy 3770 binding posts that all go bad fast with nanovolt-grade pure-copper spade lug only terminals. Banana people will surely hate me now, but I don't mind  Maybe somebody in USA club can make adapter spade lug -> banana if anyone still wants that. 

Updated module appearance  front side :



Updated module appearance back side :



 

[hwj-d]

@TiN
Hi TiN,
is the PX with you as well? Because i've had the PX only at the beginning, can you please measure the PX with one of your precise 3458A and take a foto from it? I still don't know how to estimate my dmm's.
Thanks a lot.

[TiN]

No sir, I got only FX without the packaging/case or anything but cable.

[Grandchuck]

@TiN
Hi TiN,
is the PX with you as well? Because i've had the PX only at the beginning, can you please measure the PX with one of your precise 3458A and take a foto from it? I still don't know how to estimate my dmm's.
Thanks a lot.

I will compare the two references when the FX arrives and report my 3458A readings here.

[hwj-d]

@TiN
Hi TiN,
is the PX with you as well? Because i've had the PX only at the beginning, can you please measure the PX with one of your precise 3458A and take a foto from it? I still don't know how to estimate my dmm's.
Thanks a lot.

I will compare the two references when the FX arrives and report my 3458A readings here.

Thanks Grandchuck, that's nice.

[hwj-d]

Something happening here?

[Grandchuck]

It is scheduled for pickup from TiN.  I will post here when it arrives.

[TiN]

It was also recalibrated and relabelled few days ago for fresh volt. I've included shorting dual banana plug back with reference, but obviously now it has no any use. New copper low-thermal posts are not banana compatible 

Callout for everyone participated in round 2 already. Please, upload all data to server, so we can plot a measurement agreement between everyone and guesstimate deviation of each person, as now trend of FX reference over 600+ days is guesstimated 

[Grandchuck]

The FX arrived safely and works as expected.  TiN included a 4 page calibration document.  The first page of this document is attached with this post.  Pages 2 and 3 show uncertainty error in ppm (not attached).  Page 4 has probability histograms: one for UUT- and one for UUT+ (not attached).  Of course, the entire 4 page calibration document will be included with the cal club kit when I ship it to the next member in line.

Hats off to TiN and many thanks to him for taking the time for providing this valuable reference to the US CAL CLUB!  Not only did he design and build this particular reference, he put a lot of time and effort into verifying its performance … again!

Note to hwj-d:  The PX reference is showing 7.045639 volts on my 3458A (recently calibrated to the International TiN Standard [ITS]).

[Grandchuck]

Here is some data for the FX reference.  2,692 readings (NPLC = 100) with a 3458A that was calibrated using the FX reference itself (I guess it is what one calls a bootstrap operation).  The XL file is also attached.  Later, wondered if the data were normal so I prepared a histogram plot.

[SilverSolder]

Here is some data for the FX reference.  2,692 readings (NPLC = 100) with a 3458A that was calibrated using the FX reference itself (I guess it is what one calls a bootstrap operation).  The XL file is also attached.

Very neat.   Is the X axis in Seconds?

[Grandchuck]

The X axis is the number of the reading ... from one (the first) to 2,692 (the last).

[SilverSolder]

The X axis is the number of the reading ... from one (the first) to 2,692 (the last).

Ok, so one reading every 1.667 seconds (60Hz) with 100nplc?   1 hour and 15 minutes total test time?

I'm trying to understand how stable it is over time...

[Grandchuck]

The data collection runtime was about 4 hours.  You can see it in the data file (there is a column for time).

[SirAlucard]

It's hard to believe any switch could have a resistance that high, if it's clean and if it uses metal on metal contacts. Have you ever disassembled one to see how it's made?

Yea, it's how I was able to throw some fresh solder down to kinda beef up the pad on a test unit, I think my plan is to copy and redesign the PCBs a little bit for these and get them made and just replace all the PCBs in the switches, would be cheaper than getting the nice push button switches.

[Grandchuck]

Not able to upload to the club repository?

FTP server: ftp.xdevs.com
Login: usac_run2
Password: same as login, case-sensitive

Anybody?

[CatalinaWOW]

Not able to upload to the club repository?

FTP server: ftp.xdevs.com
Login: usac_run2
Password: same as login, case-sensitive

Anybody?

I had a similar problem.  Went and got an FTP server program and the problem was solved.  Never took the time to debug what was wrong.