DIY 34401A calibration

[baoshi]

Does anyone know, how is the internal wiring of Agilent's 34172A/B short block ?

I have made this but many tell me it is not the correct way for 4 wire short.

Thanks


Sent from my iPhone using Tapatalk

[Zucca]

I have made this but many tell me it is not the correct way for 4 wire short.

Why? It should be fine, I would have done it differently anyway. Sense wires carry low current, no need of big copper there. 

[jlmoon]

That would be an industrial strength shorting block. indeed.

[jpb]

For my Keithley I use two separate standard two pin shorts (see attached photo) which I connect by thick copper wire, it seems to work ok. Doing it with 4 pins as you've done should be fine - the only thing is that you need to be more accurate on placing to avoid perhaps mechanically putting strain on the sockets and I wouldn't trust my own metal working skills to do this - but you look to have done a good job.

NB I put the shorts across force and across sense so the copper wire shouldn't carry any significant current so any contact resistance effects should be negligible.

[macboy]

For my Keithley I use two separate standard two pin shorts (see attached photo) which I connect by thick copper wire, it seems to work ok. Doing it with 4 pins as you've done should be fine - the only thing is that you need to be more accurate on placing to avoid perhaps mechanically putting strain on the sockets and I wouldn't trust my own metal working skills to do this - but you look to have done a good job.

NB I put the shorts across force and across sense so the copper wire shouldn't carry any significant current so any contact resistance effects should be negligible.

It isn't contact resistance that you need to avoid, it is thermoelectric potentials. With all those plated metal surfaces, that is a lot of possible thermocouples. With the right (wrong?) combination of metals, even a fraction of a degree difference can result in microvolts of offset. This is an issue when calibrating null on a 6.5+ digit meter.

Keithley's shorting block is built on a PCB like attached.

I think the solid copper block is a good idea. I would prefer a plastic standoff instead of a heat-conducting copper screw.

[jpb]

It isn't contact resistance that you need to avoid, it is thermoelectric potentials. With all those plated metal surfaces, that is a lot of possible thermocouples. With the right (wrong?) combination of metals, even a fraction of a degree difference can result in microvolts of offset. This is an issue when calibrating null on a 6.5+ digit meter.

I'm no expert, but I'd be surprised if it was a problem for 6.5 digits. I used the approach for the self test on the Keithley which it passed happily so presumably any effect was within the allowed spec.

Out of interest I just experimented a bit with different combinations of connecting the short both directly and via a copper wire between two shorts and with every combination my meter was reading somewhere between 3.3 and 3.8 microVolts (Edit : this drops to around 0.4 microVolts on warming up) which I think is just noise rather than thermo-couple effects.

I accept that you have to be rather more careful with 7.5 and 8.5 digit meters.

[quarks]

probably this is interesting for you to read
https://www.eevblog.com/forum/testgear/lking-for-gold-plated-tellurium-copper-banana-plugs/msg246371/#msg246371

[Dr. Frank]

baoshi,
your short is just perfect!

The big copper plates takes care for low resistance < 1mOhm between hi and lo, and it also takes care for minimum temperature differences between all four plugs, due to its low thermal resistance. This mitigates possible thermal voltages. Also, each plug only has one relevant contact point to the base plate - underneath.
If you wait several minutes after application, you will measure no thermal offset at all.
 
Look there, that simply bent copper wire is how the 3458A should be shorted, as it's officially drawn in the calibration manual.
That is good for <10nV offset calibration for 8 1/2 digits DCV, and for 7 1/2 digits low ohm (<1mOhm) so that mainly the internal front/rear switch and internal PCB connections can be zeroed during calibration (the 34401A is very similar concerning the front end).

For the 34401A, I soldered that clumsy short on the right, accordingly.

But your design is much better, because it is more simple, and elegantly designed.

Frank

[LaurentR]

http://us.flukecal.com/products/accessories/test-leads-probes-and-clips/884x-short
$30 - a tad bit cheaper than the $200 Agilent shorting block   

Or stack three Pomona 2035 (or two 2035 and one or two 3030) - ~$7 each
http://www.pomonaelectronics.com/pdf/d2035_100.pdf
http://www.pomonaelectronics.com/pdf/d3030_1_01.pdf

[Vgkid]

Dr. Frank, who made those bananna plugs?

[baoshi]

Dr. Frank, who made those bananna plugs?

If you have access to taobao.com here is a link
http://item.taobao.com/item.htm?spm=a1z09.2.9.63.ewdDg2&id=22233300725&_u=s15rvgic7df

[baoshi]

Thanks everyone.
My idea is the link between the "force" terminals (left) and "sense" terminals (right) need not see current flow. So the connection should be thin. Something like below. Apparently the 3458A calibration manual suggests the same layout.

[baoshi]

Are you writing software to perform the calibration?  If so, any plans to open source it?

Afraid not. See no point to automate it. Just picked up a 34401a and some 9-output reference categorized by a 8.5. Trying to calibrate it.


Sent from my iPhone using Tapatalk

[Dr. Frank]

Dr. Frank, who made those bananna plugs?

That's stuff from mc, multi-contact.
afaik, it's a copper beryllium bronze, gold, hard plated. LS4 + KT4-B

Frank

[HighVoltage]

Your copper plate seems perfect.
I wanted to do the same, but never got to it.

The 4mm studs are made by MC (Multi Contact) and I got them already.
So far I have used MC gold plated banana plugs and soldered heavy copper wires on to them, similar to the version by Dr.Frank
It works perfectly.

[TiN]

Not 100% relevant, but since we talking about shorts, I have some PCBs with various shorting configurations.





Fluke have note advising against big thermal mass shorts

[HighVoltage]

Fluke have note advising against big thermal mass shorts

Hello TiN
Your link to the Fluke document is not working
Could you please fix that.

[baoshi]

Is this the article TiN has mentioned? http://us.flukecal.com/literature/electrical-calibration/watch-out-those-thermoelectric-voltages-cal-lab-journal-reprint
I start to buy in the theory.

[radioFlash]

I got the Fluke 884x-short. Attached are photos which show the layout. There are thick traces between the Hi-Lo sense terminals and the Hi-Lo force terminals (about 60 mils) and a thin trace (about 10 mils) connecting the force and sense sides.

* PCB is .1 inch (2.54 mm) thick
* Weight: 14.7 g

[acbern]

I am using these (fluke 884x, keithley has the same under its own name) as well and moved away from thick copper plate shorts as shown in this thread. mostly driven by the fluke app note.
while it is ok (convenient) to use a bare copper wire to calibrate a 3458a, for day to day work with it, the fluke is just better.

for a 34401 the issue is neglectable though, because of its resolution/accuracy.

[Tj138waterboy]

Not 100% relevant, but since we talking about shorts, I have some PCBs with various shorting configurations.





Fluke have note advising against big thermal mass shorts

Did you happen to make modules for different manufacturer spacings and maybe still have the gerbers?

[Kleinstein]

The pin configuration is pretty much standard on modern meters. So the Fluke short should also fit Rigol and Keysight.

there is nothing really magic about the simple 4 wire short to adjust 4 wire ohms. It can be carved from a piece of copper clad board (not just leave all the copper), or made with a piece of isolator and than using 3 pieces of wire in an H shape.

The tricky version it the symmetrical 4 contact short that could be turned in any direction. These do exist, but are more tricky and less accurate.

[Tj138waterboy]

 :-//Just had a bit of an idea for cal club rd3 with making a standard resistor plug n play set with this design. Would definately need a shield or cover though. Seen this and they aren't same dimensions

[The Soulman]

Some more info is here:

https://www.eevblog.com/forum/oshw/4-wire-short/

The spacing between force and sense terminals isn't standardized.

I'm liking your idea tho.

[iMo]

How the "adjusting" of the 34401A actually works?
Provided I've got 10V source and 10k resistor, and I want to "DIY adjust" my 34401A's 10V and 10k ranges.

I cannot find the details on the procedure, though. I do not want to start to play with the "Calibration" Menu items before having a good understanding what the meter would request me.. 

My current understanding is it will ask me to connect 10V (or 10k) and then to enter (via the front panel keypad) the "exact values", like "10.001.345V" or 9.999.887ohm".

Is my above expectation somehow correct?