what is the deal with DC shorts and offset

[coppercone2]

damnit I made another one with new banana plugs that have more leafs that I bought a few months ago from DK with a thinner piece of bus wire that fits in there and it works OK (but its not stiff enough to insert good, its not a good 'standard'). Maybe I will drill it to accommodate that big ass wire that is in the bad banana and try it again. The new plugs go in deeper because their not threaded for the insulating collar

I will make a big long copper loop that sticks out like an inch to compensate for the shorter plug. If the same wire (longer and extending more for thermal mass) on different banana plugs gives me a better comparison for thermal mass and heatsink. Or maybe I should just solder it to a big 1/4 inch copper U shape to make a really exadurated heat sink Or I can solder it to a freaking bus bar


Hmm I noticed this one has a temperature meter on it. I wonder if it measures thermocouples.


Can the thermocouple meter on the K2001 tell me something about my jumpers? It has alot of settings.

I put it on T mode and plugged in various jumpers. The bad banana reads 0000.2 c immediately . The other ones are 0 or 0.0001


Ehh its not really conclusive.


What can I use to study the different jumper wires behaviors, etc?

Something like a special thermocouple meter? I mean to get more resolution/clarity.

Like some rig with heater resistors and temperature sensors attached to some faux banana connector ior something

[Kosmic]

For reference, This is how the Agilent 34172A is constructed (I was expecting something better).



I'm also including some documentation from Fluke about thermal EMF and the characteristics of a good short (low thermal mass).

Watch Out for Those Thermoelectric Voltages!

[coppercone2]

My feelings on the matter so far is that you might be paying for the insulation ,the alloy choice/plating *thick gold, hopefully?), but not their handywork . And that they probobly tested in in 10,000 meters before putting it on the market. And the injection mold lol... with ENGRAVED lettering

I have been chronically disappointed taking apart various adapters and such lol. I need to remind myself its the ergonomic clam shell I paid for, not what is inside. And that I can write a official sounding PN in a document.

I had my own banana clam shell thing made up for some specific thing we do. 70% of the work was on the plastic enclosure , and this one was 3d printed. Two of them kept not fitting right into the front panels and it got reprinted 50 times. Good puzzle for bored co worker (they just want a extra measure of 'touch safety' for low voltage things, personally I just used it 'bare', though it looks like kind of shady to have it sticking out of a expensive device without a housing ). Now he knows the specifics on why 3d printer is a wild animal

I am wondering if that Agilent one on purpose does not have a PCB inside. Because mine used a BAKELITE board (screw mountings for the parts) inside of it with a plastic covering. But theoretically theirs has higher thermal isolation between the connectors if its all polystyrene plastic. that might be the magic of it.... IIRC its great insulator . But mine is a high current device and I don't care about microvolts there, its just essentially a weird adapter for some test probes, and its not like bakelite is a heat sink anyway, and @ 3/16 inch thickness its really tough and also easy for machine shop to make, unlike PCB (no glass)

[Ismsanmar]

My feelings on the matter so far is that you might be paying for the insulation ,the alloy choice/plating *thick gold, hopefully?), but not their handywork . And that they probobly tested in in 10,000 meters before putting it on the market. And the injection mold lol... with ENGRAVED lettering

That is just a 34171A repurposed. Are you sure that you are not paying for their handywork? One of the current model connectors, 34171B, costs $170.5 directly from their website, and the 34172B costs $341. That's exactly $123.5 for their handywork (plus the wire, solder and the sticker with the model number), and they also take the gold-plated copper nuts as part of the payment.

[coppercone2]

What I mean to say is that you won't find like anything expensive or difficult inside. all of those you take em apart and its like what? especially the LCR things, like $1000 component testers/adapters that are like a spring and a capacitor with some bus wire or whatever

You might get the impression that it will be something legendary like a resistance box switch or something in there that makes your jaw drop for the price


I feel the same way about pump things from the chemistry side of it. You get the worksmanship of a childrens china 1990 toy with some dead bug parts and a cheap motor for like $1000+ . For the chemistry side of 'lab tools' . From like VWR etc. I love the ones that don't bother with wire (they use component leads) or insulation (try to wedge it between plastic pegs and stuff. or PCB's, brackets (they like foam), connectors, cheap battery contacts, etc

Sometimes those companies are kinda selling junk on the side (even if they have some really nice equipment)

I took apart both a scandously priced electronic motorized pipette and a 'micro pump' that were like that from the big expenisve brands


I have a feeling that its not the same designers either. Just the badge



Have you never encountered the brand image part of product design where you get some enclosure that is basically selling because of looks and then there is no money to work on the inside and stuff, and its super irritating because of the enclosure features ? The way I always saw it work is sales says "we need something that LOOKS like this" and then engineering goes "well....." and then sales go "it has to look LIKE THIS or it WONT SELL" and then when it comes down to the remaining money for the parts  there is none? Or you can't fit anything in there because of decorative features?

Their always expensive lab tools that would look good on a star trek set. It's what apple does but for test equipment


I think its especially valued in chemistry labs because it looks cleaner and visitors are less anxious about being in there if its some total desk jockey that thinks a belt pump is going to explode and spray them with plutonium or something. They love doing lab tours and facilities tours and as a result of that there is pressure to make everything look as chic as the board room. We are NOT just another company boiling metal cuttings in acid because our hot plates and microscopes have sloped beveled miters ! (process was unchanged since 1942)

[coppercone2]

I am going to try to make spider-short

I lashed 4 spring copper wires together using some kind of soft brass jewlery 'lacing' wire and I will dip it in a silver pot. Then I will shape it and trim the excess. It's not the same idea that they have now with using thick traces for everything but the short, and instead its a bunch of thin wires that converge at one point in hopefully a low temperature offset braze.


I am wondering if there is data about brazes and thermal EMF. I have 56% silver braze and also pure silver. Its hard to do pure silver. I suspect its going to have a better result though?

Does anyone have a table for this weirdness? For all I know it can be horrible performance. I just wanna try some experiments. I have some Cd solder too IIRC (low temp)

https://www.ni.com/docs/en-US/bundle/ni-switch/page/thermal-emf-and-offset-voltage.html

Though I wonder if I can just weld it together.......I suspect that it will be difficult to control. The braze version has zinc and nickel in there, which look like bad performance. It seems hardly worth doing over just soldering it with electronics solder.



Damn I did it in pure silver but then I thought Oh I will just give it one more quick dip to make sure and it totally melted and contaminated some valuable silver. And zirconia for ghetto work is no good (it cracked), graphite however is excellent. The melting point are so close

Anyway I did it with the braze alloy. I wonder how this will read

[coppercone2]

spider 4w short using 4x becu spring wires and a 56% silver dip braze in the middle reads 500 micro ohms on 34401A slow 6 digit. At least it was fun to make 

Though I don't have a commercial 4W short to test it with and htis meter was calibrated long ago

200 now. sometimes it gets to all 0

0 to 100 micro ohms now so it looks like it went to the last digit. I wonder if it will go negative


Ok so even without a box it looks like it can get a 34401A to all zeros. Nice. IDK if the 34401A reads negative like some SMU do

lets see what k2001 says

It says its 300 micro ohms. I wrapped it in a tissue or possibly toilet paper and will let it settle

If its inserted and I put it on DCV it reads 500nV


I don't remember if i ever tried the keithley 4W short with the 34470A when it was available. But I do recall someone having problems with this at work, they ended up getting one that is an acrylic block to I guess block air currents. Not sure what the 'passing grade' is for a calibration


For  3458A the test limit is 0.00007 ohms for the 4 wire short

damn that means my short sucks if their fail is 70 micro ohms and I can't get it under 233




The conductivity of spring copper is bad however, I just thought it was cool that it has good contacts for the banana holes and it should like stay.. dimensionally correct for longer. I can try to make the same thing with just copper wire. I think the BeCu could have conductivity as low as 20% of copper. But I don't even know what I am measuring here, because their so tight together. The silver solder? Parts of BeCu and silver solder intermetallics etc?
Do you actually see all 0 on a 7.5 or 8.5 digit meter from putting the factory shorting 'block' in?

Or are they just used to set zero? What is their actual resistance. I am thinking about buying one for $100

They just say its low "However, the absolute resistances of 2- or 4-wire configurations as well as
thermal voltage offsets are not specified, and metrology traceable applications are not recommended."

What is a specified one at?

Mine is down to 182 micro ohms. I guess the trick is to not measure a solder joint. Unless I can get the pure silver to work with very careful dunking it should be made out of one copper piece in that location. Perhaps I can try welding this

[coppercone2]

this measures 180 on keithley, but it does appear to totally zero a 34401A out. 


How do you get something to less then 5 micro ohms?

[laichh]

For up to 7½ digit DMM, I use either Keysight 34172B or Stäubli KS4-19 (Item No: 24.0002-21).


Keysight 34172B Calibration Short for Digital Multimeters

https://www.keysight.com/us/en/product/34172B/calibration-short-digital-multimeters.html


Stäubli KS4-19 (Item No: 24.0002-21)

https://www.buerklin.com/en/p/staeubli-electrical-connectors/short-circuit-connectors/24-0002-21/40F436/



For 8½ digit DMM, nothing beats a piece of AWG14 pure Copper wire.

[Vgkid]

Since you are now wanting Zero ohm resistance VS zero thermal emf see this thread:
https://www.eevblog.com/forum/metrology/zero-ohm-diy-4-wire-standard/

[coppercone2]

what do you actually get when you plug it in though. Is it actually all zeros?

I like sawed at the silver with a sharp knife because it wet two wires, and now its at 160 micro ohms.

I am asking how much better the number will be If I buy the commercial thing


Like if you bought it and plugged it into some NIST 8.5 digit super meter, what would the front panel say ?
The only number I found is that the 3458A has a value that says "less then 0.00007 ohms" for "pass" on their test sheet.

What do they actually measure at??

I want to understand the value I am getting. The best I can try to make vs some thing that costs money online. In micro ohms or nano ohms per dollar.

1 inch of 14AWG copper wire is 200 micro ohms. Its around what my thing is getting. How much better do they  do?

Like I can buy the keithley one, I can buy a chinese one, I can try to rig some shit up with PCB, etc. I will buy the real one if it actually gives a low number instead of messing around

[TUMEMBER]

this measures 180 on keithley, but it does appear to totally zero a 34401A out. 

How do you get something to less then 5 micro ohms?

You built a nice thermocouple. I showed you how to make "zeroVolt" but with the stubbornness of a maniac you do everything "your way" and against the grain. There should be only one metal - pure copper. There should be no soldered joints - the second metal creates the thermocouple. A 2 mm wire bent and pulled together creates a closed loop without the need for soldering.

1. There are no different metals in contact with each other, there is no difference in potential.

2. There is no temperature difference "at the beginning of the wire and its end", there is no difference in potential. Work on the Seebeck effect.

3. Work on the "Law of the Third Metal".

In terms of whether you can have "all zeros on 7 1/2 digits on the lowest range" - see photo. Similarly, when it comes to the "zeroOhm" range in multimeters, you can when measuring using the zeroing function + 4W measurement. Multimeters are not resistance bridges, they have "limited resolution and accuracy of resistance measurement" they are "blind" to certain minimum measurable values.

[coppercone2]

I just bought the keithley one its easier then trying to get a ohms number from aggressive wacko with a vendetta, I will just plug it in and find out

[Kleinstein]

A proper 4 wire short for the ohms zero is mainly connecting the 2 sense wires to each other. The 2 drive wires are connected separate. Between the 2 parts there is a weaker link connected at 1 point for each end.  The main point is avoiding voltage from the current to enter the sense link.
A simple version is a wire from drive H - drive L - sense H - sense L.

[miro123]

Keithley DMM6500 has default option to compensate the thermocouple effect.  Many discrete sd-adc from TI has those option too. just do the measurement with and without excitation current

[J-R]

I picked up one of these a while back, seems to be decent: https://www.ebay.com/itm/326409377777

[zlymex]

A simple 4-w short for Hioki RM3545 (buttom left PCB) made by bg2iax.

[coppercone2]

I got the keithley 4 wire short for a not so low price and its reading like 190 counts on 4 w high accuracy mode. not too different from my spring spider short (233 counts) but its built very differently. Whatever parasitic effects it has don't contribute more then like 100 micro ohm of error 

so its like a 50 micro ohm difference ON drifty startup. I just need to get the voltage shunt (the $10 pipe banana connector) and this issue is settled

After about an hour the acrylic block keithley is reading 210 and IIRC the DIY one was like 233, so there is a 20 micro ohm difference between them. I don't have what it needs for the calibration yet, but anyway I think that means my spider spring design idea is not that terrible. So this instrument has offset that can be calibrated out, but this still IMO says that the difference between the shorts is small, considering one is about a 20 cents of spring wire, a drop of silver braze+1 tissue (with no design work other then being handy) and the other costs $150


If I had 4x banana to binding post adapters (expensive) I would try the U short method using  bare copper wire. Sadly this instrument does not have binding posts on it, understandably so because it might be connected up to 1kV

https://www.digikey.com/en/products/detail/pomona-electronics/2894-0/736408?gQT=1

or POMONA 5165-0 or 1286

after many hours its at 142

[TERRA Operative]

For up to 7½ digit DMM, I use either Keysight 34172B or Stäubli KS4-19 (Item No: 24.0002-21).


Keysight 34172B Calibration Short for Digital Multimeters

https://www.keysight.com/us/en/product/34172B/calibration-short-digital-multimeters.html

Do you mind showing us inside the 34172B?
I'm interested to see how it compares to the 34172A.

[laichh]

For reference, This is how the Agilent 34172A is constructed (I was expecting something better).



I'm also including some documentation from Fluke about thermal EMF and the characteristics of a good short (low thermal mass).

Watch Out for Those Thermoelectric Voltages!

Same as 34171A as shown above, just that the company name changed from [Agilent] to [Keysight].

[TERRA Operative]

Great, thanks!

Are you able to measure the dimensions? I wouldn't mind making a 3D printable version.

[laichh]

My solution:

https://www.eevblog.com/forum/metrology/diecast-enclosed-banana-plug-terminal-block/

More picture below:

[bastl_r]

That looks really high-quality!
But how do you implement the zero-crossing points?
With solder pads?

[laichh]

Yes, just simply bridge solder the pads between the two adjacent banana plugs. For low-thermal application, a gold-plated plug is recommended.

[KungFuJosh]

What's the point of the 5th plug (amps) on the 34172A / 34172B?

It's not connected to anything, so what purpose does it have?

Thanks,
Josh