USA calibration club

[IconicPCB]

Gnumeric is much quicker than Ope Office if You are using Linux

[Andreas]

Perhaps someone can point me to a good (non-excel, prob. not python) windows or cloud way to show  csv data graphically.

Hello,

I often use plotter from Ulrich Bangert.
(does also some statistics and analysis).

http://www.ulrich-bangert.de/html/downloads.html

with best regards

Andreas

[bitseeker]

Thanks, Andreas. I've been intending to try Plotter (saw when programming EZGPIB), but it slipped my mind.

[RandallMcRee]

Trying Plotter, now. Should have thought of it since I'm using EZGPIB.

Ulrich Bangert seems like a stellar individual. I was really sorry to read that he had passed away. There is no emoticon to express my feeling of loss.

The run that I was doing crashed out after 9,656 samples with some software exception. Is the EZGPIB code available somewhere to debug? (It has happened consistently).

Thanks again,
Randy

[CatalinaWOW]

The EZGPIB source code doesn't seem to be available currently.  Ulrich's brother has control of the source code, and may be amenable to releasing it.  I haven't contacted him for fear that I can't do justice to it.

I have run logs longer than 9,656 samples, so the problem may not be purely with EZGPIB.

[RandallMcRee]

Plotter output from run1. Strong temperature dependence as expected. Obviously, I need some help with Plotter....

First plot is voltage which only varied from 10.00004 to 10.00007 volts. Second plot is temperature. Around 5pm I covered the reference with poly batting to dcrease the influence of air currents, but the temperature sensor remained uncovered. I think the big temp event is me closing an open window (not sure when it *was* opened--not by me anyway).

CatalinaWow--yes, my script could be at fault, memory leak perhaps? Hard to know, without better debugging.  I would be interested in helping maintain it, though I'm a Java guy.

[bitseeker]

Randy, what error did you receive when the logging crashed? Which script file were you running? If you're using TiSensorLogger.488, then there could be a bug that I introduced as well.

I logged for a little over two hours at a time, so my logs were between 8k and 9k. But if CatalinaWOW was able to run longer, then it could be something in the script or perhaps a driver issue. Details on the error may help reveal more.

[RandallMcRee]

The EZGPIB error was in german and I id not capture it. Next time!

Here is the script--

[bitseeker]

OK, I only noticed two things in the script:

• Lines 63-64: A '1' is being sent to the HDC1010 and the returned value stored as temphex. However, '1' is humidity. The value doesn't appear to be used before it's overwritten on line 86. I assume this portion of code is just to fetch and discard the first value, which I found to be garbage data when I was working with the HDC1010. So, this issue is more an FYI for the sake of consistency.
• Line 71: The header is being output with a comma as the delimiter. However, when the captured values are sent to the file buffer (lines 81, 102, 103, 110), EZGPIB_FileAddToBuffer() uses tab as the delimiter. So the header and data use different delimiter characters. Do you see that being the case in your log file?

Other than that, the script code looks fine.

[bitseeker]

Oh, one other thing to note about EZGPIB is that each time EZGPIB_FileWrite() is called, it opens and closes the file. Since the script does this on every loop, if something goes awry with the file or the file subsystem, an error could be generated.

The EZGPIB docs have a note in the revision log about how you can restructure the code to buffer all the output during logging and write it at the end of the session. However, you may not want to do that if power is an issue, otherwise the data could be lost.

[RandallMcRee]

EZGPIB crashed again on yesterdays run, I did a screen capture this time....

I have noticed that the prologix adapter I picked up on ebay does not have the latest firmware. So will update firmware and try again, later.

Will do a separate post on the measurement itself.

Randy

[RandallMcRee]

So here is the data from my second run. It's an experiment. You can judge for yourself whether there is any value to it.
I have recently finished an LT1027-based ovenized reference. There are five LT1027D chips averaged. The averaged output is about 5.00005 volts. I constructed an LTC1043 /2 circuit (and a *2, but that was not used in experiment). The geller was the input to the 1043/2 and the output was 4.99999 volts. The geller, the LTC1043 circuit and the LT1027 ref are all hooked up to the same power supply (Analaog Devices +-15v power brick).

So I simply hooked the K2015 to LT1027 averaged vout (+) and geller out to (-). The difference was about 0.05 mV. So I logged this difference.
The idea was simply to see if there was or was not a temperature correlation (hoping not). As you can see it was a bit ad-hoc!
Mean of run was 4.91e-5 and SD was 7.47e-6.

Output seems to show no, or negative correlation with temperature. But, there are concerning unexplained voltage spikes (noise of some sort I guess).  Anyway here are the pics,
(Ran into data limits--will need to break up this post)

[RandallMcRee]

Pics of setup

[RandallMcRee]

Pics of setup without polyfill.

More on the LT1027 based "ovenized" reference. There are five LTC1027 refs in LCC package (no humidity sensitivity). Three on each side. (One was disconnected because it was unstable). These are mounted on a piece of 0.25 inch silvered copper bar. This bar is ground. There are three drilled holes in which are mounted heater resistors. The heater is the simple wenzel TL431 set to 35C.  I just surrounded this circuit board with thermal blanket and polyfill -- it worked well enough that I nave not so far mounted it into the bud box.

Randy

[cellularmitosis]

Very cool!   

Can you log while you are asleep or otherwise out of the room and see if the spikes are still present?

[RandallMcRee]

No one in the room. I setup the logging and went to bed!

[Conrad Hoffman]

I've always thought the 1043 was a neat part. Just curious, have you hooked both inputs to the same reference (not ground) and logged data for a while?

[RandallMcRee]

Ummm, as I came up from my lab, I remarked to my wife--I need another 6.5 digit meter.  My remark came about exactly because it seemed the quickest way to see how that 1043 circuit was doing--I have no way of precision logging more than one source at the moment. What I have been doing is flipping from back to front on the meter....sigh.

Wife was not amused. Discussion about budget. No new meter coming.   

So, I've logged reference outputs only. Have not done the same on 1043 outputs. I do have good logs of the exepected standard deviation on my LTZ1000 references and LT1027. So I could hook up the 1043 and look at the SD. Should be correlated, right?

Also I completely trust that datasheet.

Seriously, I've compared the 1043-implemented instrumentation amplifier (first page of datasheet) with an AD8429 IIRC. The 1043 showed better CMRR, which was nice.

[Andreas]

I've always thought the 1043 was a neat part.

Hello,

The 1043 is a neat part.
But you must not load the output. (and a 1nF capacitor at the input or a buffer keeps EMI away).
A output buffer amplifier is mandatory. (I see none on the perf board).
Unfortunately the buffer amplifier limits the accuracy of the LTC1043. (offset temperature drift)

https://www.eevblog.com/forum/metrology/t-c-measurements-on-precision-resistors/msg528192/#msg528192


with best regards

Andreas

[cellularmitosis]

The Geller Labs reference has finally made its way back to me

We have a few people whom have jumped on the train, and for them I'd like to try something slightly different.  Until now, we've used a ring or round-robin pattern of mailing the reference -- it goes straight from one person to the next.

For the next few hops, I'd like to try a star pattern, where the reference is sent back to me after each hop.  I'll use this to try an get an idea of the repeatability of the reference after each shipping hop -- i.e. does shipping induce any hysteresis?

I'll take some measurements over the next few days and then send it on its way again!

[nikonoid]

How about keeping a powered logger with it to track temperature, humidity and pressure as well as 3D acceleration (bumps) in transit? Are there any regulations against shipping powered devices?


Sent from my iPhone using Tapatalk

[bitseeker]

Backing up a bit, to log while in transit requires power. For domestic shipments via USPS, household dry-cell batteries are OK. See 348.22.b:

https://pe.usps.com/text/pub52/pub52c3_026.htm#ep900090

Lithium-based cells (rechargeable or non) have restrictions. See 349.221 and 349.222:

https://pe.usps.com/text/pub52/pub52c3_027.htm#ep900090

It's also interesting to note that 349.221.b.2 and 349.222.b.2 state that, "the equipment must be equipped with an effective means of preventing it from being inadvertently turned on or activated." So, no running devices that use lithium batteries.

However, a similar requirement isn't listed for dry-cell batteries.

Disclaimer: IANAL

[SvanGool]

... "the equipment must be equipped with an effective means of preventing it from being inadvertently turned on or activated."
So, no running devices that use lithium batteries. ...

Sorry if I am nitpicking and I am not a native speaker, but wouldn't this mean that if you would advertently turn on the equipment, it is allowed ?

[bitseeker]

Yes, semantically they are referring specifically to inadvertent activation. However, I don't think they intended for it to mean that it is therefore OK to activate it beforehand since the end result is the same, a running device powered by a lithium battery. Even though I'm a native English speaker, I'm not a lawyer, so my interpretation isn't worth much.

[bitseeker]

Hey there, fellow cal clubbers. What's new with the roving standard, etc.?