HP 34401a - Drift in the 2W R mode / 100 Ohm range

[ManateeMafia]

I re-read your readings and must have had my blinders on.  Luckily your meter showed up better than a Fluke 510a I bought. The transformer broke free of its mounts and was rolling around inside the reference.

Good luck with your repair.

[coldframe]

Hello Z,

I’m looking forward to seeing your job.   
Good luck!.

[HighVoltage]

Statistics of the good one (front side)

Mean   0.002524814
Standard Error   4.86997E-06
Median    0.00252
Mode    0.002609
Standard Deviation    0.000196435
Sample Variance    3.85869E-08
Kurtosis    -0.288419334
Skewness   0.063261047
Range    0.001202
Minimum   0.001919
Maximum    0.003121
Sum   4.107873
Count    1627
Largest(1)    0.003121
Smallest(1)    0.001919
Confidence Level(95.0%)   9.55207E-06

Hello Zucca
May I ask how you got these statistical data from BenchVue ?
Thanks

[Zucca]

May I ask how you got these statistical data from BenchVue ?
Thanks

Export in Excel, use the data analysis toolpack

http://office.microsoft.com/en-us/excel-help/load-the-analysis-toolpak-HP001127724.aspx

and run the descriptive statistic on the data

http://www.excel-easy.com/examples/descriptive-statistics.html

[HighVoltage]

Thank you, Zucca

[Zucca]

Short update (don´t ask me the entire story please). It looks like the drift is gone (I think the reflow helped here) BUT the 100R 2W adjustment process does not work.
Basically before and after the 100R 2W adjustment the meter is reading pretty much the same value out of tolerance.

I allow myself to declare the issue is not an analog one (that´s why I was going in circle without huge steps forward) rather a digital one (bad cell memory on the 2W 100R value??, it won´t overwrite that value anymore???).

Anyway I am close to put that unit on ebay as not working (just because I like to be honest), since 99% is working perfecty and only the 100R 2W is slightly out tolerance I think somebody of you guys can make a good deal.

A friend of mine is an exorcist, but he told me he won´t deal with electronic devices as a 34401A 

[Gleb]

Hi there,

Where did you find information that the Specifications of Agilent 34401A are given at 95% confidence interval?

Because I have had a look through the whole User's Guide for 34401A and couldn't find that information anywhere.

Gleb

[Zucca]

Where did you find information that the Specifications of Agilent 34401A are given at 95% confidence interval?

That value is not provided by Agilent, it is an empiric statistical number based on the data collected.

[Kleinstein]

Doing adjustments may need special codes or a hidden switch to be operated. So I would double check the instructions on how to do the adjustment. A memory fault is less likely than a user misunderstanding the instructions.

Even with 100 Ohms in 2 wire mode is not perfectly working. It's still a good unit, if you know the limitations.

[Zucca]

Doing adjustments may need special codes or a hidden switch to be operated. So I would double check the instructions on how to do the adjustment. A memory fault is less likely than a user misunderstanding the instructions.

The Cal Lab was not able to adjust the "100 Ohms in 2 Wire", all the other parameters were successful adjusted. 

Even with 100 Ohms in 2 wire mode is not perfectly working. It's still a good unit, if you know the limitations.

Right. It is a very good unit, even if 99,9% working.

[Gleb]

So is it then safe to assume that all accuracy specifications of instruments are given at 95% confidence level unless otherwise specified??

[Zucca]

So is it then safe to assume that all accuracy specifications of instruments are given at 95% confidence level unless otherwise specified??

No, the 95% confidence level was calculated like this:

https://www.eevblog.com/forum/repair/hp-34401a-drift-in-the-2w-r-mode-in-the-100-ohm-range/msg518734/#msg518734

it can not be interpreted like a sort of specification value.

The specification numbers are based on the device design, with statistical measures you can verify them not create new ones.
This is what I believe but I´m open for corrections.

[Gleb]

The uncertainties in accuracy specifications are all given at a certain confidence level, however some manufacturers don't specify the confidence level, I don't know why.

i.e. if uncertainty is say 0.002% at 95% confidence, then you can calculate uncertainty at any confidence by dividing by coverage factor k, you can also find standard deviation by bringing uncertainty to 68% confidence.

 k=1 -> 68% confidence = standard deviation (sigma)
 k=2 -> 95% confidence
 k=2.58 -> 99% confidence