Low ohm precision resistor standard and testing

[HighVoltage]

For my work, I need to measure lots of low ohm resistors in the range of 0.01 to 0.1 ohm
And for reference I found a couple really nice calibration resistors on ebay for a low price.

This one is a Burster, Made in Germany
Type 1240-0,01
- 10 mOhm
- Tol. : 0.03 %
- Imax: 14 A

I hooked it up to my Keithley 2460 SMU and send 1.00000 A source current through the resistor
The resistance shows at 0.01000 Ohm without any movements of the last digit

So, either I was lucky to have a "perfectly" trimmed resistor or the calibration between the resistor and the instrument are just on spot.

Anyone else here on the forum that has to use low resistance standards?
What are you using and how stable are those resistors?

[plesa]

I'm not using nothing so fancy. Only Vishay VCS332. Values from 1R-R1 mounted on piece of aluminium.With measuring with continuous current I see the temperature drift, one of good features of SMU's - easily you can setup measurement with pulses to eliminate self-heating.
What did you paid for such nice resistance standards?
Hookup the 3458A and measure voltage with it to see how stable it is.

[lukier]

I have Leeds Northrup 4385 box, 0.02%, multiple ranges for various currents.

Like this one: http://www.ebay.de/itm/171945773961

When I get home I can measure the resistances, but none of my meters (K2001,K2015,34401A,3457A) are calibrated and I don't have an SMU to source the current so precisely :/ Maybe a mini-teardown instead?

I bought it, as it was cheap on eBay to measure higher currents than my DMMs allow (2-3A max) and more precisely. For example to calibrate my 6632B in the manual they require Guildline 9230/15 shunt and these Guildline are difficult to get and expensive and I think my L&N is almost as good.

[kxenos]

I think you need one more digit of resolution to judge if it is on spec, no? Is there a typo?

edit: awesome meter and cables there! 

[gamalot]

You may need a low resistance meter. 

[VintageNut]

Instead of meauring ohms, measure voltage on the lowest voltage measure range. You will see 10.0000mV; 5 1/2 digits for 10mV. The ohms function lacks the resolution that you can achieve with voltage measure for low ohms.

The 2450 has the same behavior.

[HighVoltage]

I'm not using nothing so fancy. Only Vishay VCS332. Values from 1R-R1 mounted on piece of aluminium.With measuring with continuous current I see the temperature drift, one of good features of SMU's - easily you can setup measurement with pulses to eliminate self-heating.
What did you paid for such nice resistance standards?
Hookup the 3458A and measure voltage with it to see how stable it is.

It was Euro 125 for this one, a while back on ebay Germany
This was also a test for the 2460 and the resistor of how they would behave under constant 1A current
I am surprised how stable this combination is. 1A seems to be idle for the resistor and SMU

Yes, I will hook up the 3458A soon to this setup. 

[HighVoltage]

I think you need one more digit of resolution to judge if it is on spec, no? Is there a typo?

edit: awesome meter and cables there! 

Actually the meter was set to 5.5 digits
Here are some pictures for 6.5 digits and 10 NPLC setting

[HighVoltage]

You may need a low resistance meter. 

That Hioki meter looks really nice too.
What current is it pushing through the DUT at 1 mOhm?

[HighVoltage]

Maybe a mini-teardown instead?

Yes, please.
Show us some pictures from the inside.

[HighVoltage]

Instead of meauring ohms, measure voltage on the lowest voltage measure range. You will see 10.0000mV; 5 1/2 digits for 10mV. The ohms function lacks the resolution that you can achieve with voltage measure for low ohms.

The 2450 has the same behavior.

Yes, I am getting 1 more digit in the 200mV rage, reading 10.0001 mV for this resistor
At 10 NPLC, only the last digit is changing.

[kxenos]

Cool! Right down in the noise! 

[VintageNut]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

[gamalot]

You may need a low resistance meter. 

That Hioki meter looks really nice too.
What current is it pushing through the DUT at 1 mOhm?

It is 1A at 10 mOhm range.

[joeqsmith]

For my work, I need to measure lots of low ohm resistors in the range of 0.01 to 0.1 ohm
And for reference I found a couple really nice calibration resistors on ebay for a low price.
...
Anyone else here on the forum that has to use low resistance standards?
What are you using and how stable are those resistors?

Equipment is cal'ed with a report.   I use these as a sanity check:
http://www.ietlabs.com/decaderes/resistance-standard.html

[Vgkid]

Which one Joe?

[lukier]

Yes, please.
Show us some pictures from the inside.

OK, quick teardown then.



It's a very simple device. 4 binding posts for force and sense, L&N made rotary switch and a bunch of wires/bars inside.



I got my unit from eBay very cheap ($45 total), so I'm not surprised it has a long history. High ohm resistors look a bit burned, probably someone tried to push too much current. But nothing is broken and the "event" happened probably long time ago so these resistors aged again.





I've measured them with my HP 3457A (NPLC=100, NRDG=10, OCOMP, AZERO) and Keithley 2001 (NPLC=10, OCOMP, AZERO, LINESYNC etc), nothing calibrated at a lab since I got these meters from eBay:

Range [Ohm]	HP 3457A Mean [mOhm]	HP 3457A StdDev [uOhm]	K2001 [Ohm]
0.01	10.128	43.57	0.010000
0.02	20.103	45.47	0.020008
0.05	50.138	27.66	0.050006
0.10	100.127	44.48	0.100007
0.20	200.148	55.73	0.200056
0.50	500.116	50.82	0.50067
1.00	1000.901	42.52	1.000823
2.00	2000.852	53.89	2.000906

I've also tested this 0.001 Ohm, 1%, 25ppm/C shunt:
http://uk.farnell.com/powertron/fhr-4-2321-0r001-1-q/resistor-metal-foil-0-001ohm-3w/dp/2419239

Range [Ohm]	HP 3457A Mean [mOhm]	HP 3457A StdDev [uOhm]	K2001 [Ohm]
0.001	1.138	50.72	0.001008

I don't have an SMU, but I got curious how the L&N 4385 will behave under load (as multimeters only source ~1mA (3457A) - ~10mA (K2001) for 4W Ohms on this range), so I've selected 1.5A range (0.1 Ohm) and sourced 1A from Agilent 6632B, measured with K2001 as: 1.0007006A, connected that to the shunt box (assumed cold, as 1-10mA from previous measurements probably didn't heat the resistors a lot), sense outputs to K2001 in DC mode, with 1 NPLC and the voltage started with 100.120 mV and after 20 seconds stabilized at 100.105 mV. Therefore, assuming the current from 6632B stayed the same the resistance went from 99.423 mOhm to 99.408 mOhm, a delta of 14.89 uOhm.  Now it is impossible to say anything, K2001 current ranges are not calibrated and not very accurate, 6632B was stated cold and might be unstable and the 14.89 uOhm delta is difficult to measure with my equipment anyway (see the 3457A StdDev).

[joeqsmith]

Which one Joe?

I think they are the SRL series.  Under $2K/ea.  I just use them for a quick sanity check.

[HighVoltage]

OK, quick teardown then.

That seems to be a very nice shunt resistor made by Leeds & Northrup
Thanks for the teardown.

Looks like your K2001 is well calibrated and the HP3457 is a little off.

[HighVoltage]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

That is interesting.
This is probably an expensive zero ohm standard.
But how complicated can it be to make a zero ohm standard?
Most likely, it has a copper bar connected between the 4 binding post threads?
Would be interesting to see the inside.
Are you allowed to open that one up in your office? Mini teardown may be?

From their website:
http://www.ohm-labs.com/resistance-standards/100-0-standard.html
"Zero ohms in any connection configuration is achieved by using a tetrajunction. A tetrajunction is an ideal connection of four wires, such that current passed through any two creates no potential drop across the other two.

Based on work performed by B. V. Hamon and others at NIST (formerly NBS), a 'copper disc and post' construction was evaluated, and subsequently commercialized in Leeds & Northrup's Hamon Transfer Standards.

Verification of the 100-0 standard can be performed by passing current through various pairs of terminals and noting the potential at the remaining pair.

The 100-0 is factory adjusted to less than 50 nano-ohms in any connection configuration. The tetrajunction will not measurably drift over time.

The binding posts are of gold plated, low thermal emf tellurium copper. A case ground is provided for reduction of electrical interference."

Very interesting, thanks for sharing this zero ohm standard

[HighVoltage]

Here is my second Burster Resistor, it is a 100 mOhm

Type 125.100
- 100 mOhm
- Tol. : 0.02 %
- Imax: 5 A


The graph shows the warmup time of resistor and SMU within 1/2 hour, both started cold.
Starting at: 100.019 mOhm
Ending at: 100.024 mOhm

It is also kind of amazing how fast the SMU is up to specs in only a few minutes.

My cables are DIY with MultiContact gold connectors and LiFY cable.

[lukier]

Looks like your K2001 is well calibrated and the HP3457 is a little off.

Pure luck. I've actually calibrated this K2001 against 3457A after I repaired it, but my transfer accuracy wasn't that great (two 10V references in series on a protoboard, LT5400 1:10 divider, some Vishay 20K and 1M Ohm resistors) so by pure chance it seems OK. Also, L&N shunts are specified to 0.02% and god knows how the resistors drifted over the years.

That's the reason I haven't sent any of my gear to cal lab - everything seems "roughly" OK and agreeing. But at some point I'll probably send at least K2001 for calibration, because it is silly to have 6.5/7.5 digit meters and only have 0.X% accuracy. Another issue is that I often got things for parts/repair for very low prices, therefore calibration cost is significant in proportion

[HighVoltage]

Hookup the 3458A and measure voltage with it to see how stable it is.

OK, here is the first setup with the 3458A measuring the voltage at the sense terminals of the 10 mOhm resistor.
I will hook up the PC and take some measurements over time.
Although my 3458A is not calibrated, it is in very good agreement with the SMU

[VintageNut]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

That is interesting.
This is probably an expensive zero ohm standard.
But how complicated can it be to make a zero ohm standard?
Most likely, it has a copper bar connected between the 4 binding post threads?
Would be interesting to see the inside.
Are you allowed to open that one up in your office? Mini teardown may be?

From their website:
http://www.ohm-labs.com/resistance-standards/100-0-standard.html
"Zero ohms in any connection configuration is achieved by using a tetrajunction. A tetrajunction is an ideal connection of four wires, such that current passed through any two creates no potential drop across the other two.

Based on work performed by B. V. Hamon and others at NIST (formerly NBS), a 'copper disc and post' construction was evaluated, and subsequently commercialized in Leeds & Northrup's Hamon Transfer Standards.

Verification of the 100-0 standard can be performed by passing current through various pairs of terminals and noting the potential at the remaining pair.

The 100-0 is factory adjusted to less than 50 nano-ohms in any connection configuration. The tetrajunction will not measurably drift over time.

The binding posts are of gold plated, low thermal emf tellurium copper. A case ground is provided for reduction of electrical interference."

Very interesting, thanks for sharing this zero ohm standard

Definitely not a copper bar. I have constructed a copper bar low ohms device and it is difficult to get good results below 1 micro-ohm. You can see from my DIY thread that I achieved almost two orders of magnitude below 1 micro-ohm.

The Ohm Labs device is stated on the front plate to be a balanced tetrajunction. I posted the drawing on my thread to illustrate the geometry of a balanced tetrajunction. It is not a simple linear bar or ingot or rod. It is a balanced geometric shape that is the same very low resistance no matter how you  connect it. A bar or rod could never meet this requirement.

What is done after initial manufacturing is that if there is an imbalance in one orientation, a small amount of material is filed off of one particular face of the device and that orientation is re-measured to see how it behaves. This process is repeated until the device measures the same low number in all three unique orientations. That is why I posted all three unique orientations for my DIY tetrajunction thread.

I paid something like USD$400 for my model 100-0 device. I do not consider that expensive.

Sadly, I cannot break the seal. It is under warranty and is used for my day job.

[plesa]

Although my 3458A is not calibrated, it is in very good agreement with the SMU

After 3 years it will be still good enough to calibrate Keithley SMU
BTW 3458A is only one needed equipment to calibrate SMU. TiN created Python scrips for it.
You are using Pomona 2305 on picture?

What is stability of this standard if you reach 14A ? It can be quite interesting measurement.

[HighVoltage]

What is stability of this standard if you reach 14A ? It can be quite interesting measurement.

That might be a challenge, how do I get a stable 14A and how to measure it?

I had the 10 mOhm resistor hooked up to the SMU for 18 hours
Here are the results:

[plesa]

What is stability of this standard if you reach 14A ? It can be quite interesting measurement.

That might be a challenge, how do I get a stable 14A and how to measure it?

You have 2460, so it should deliver 7A or it is only in pulse mode?

[HighVoltage]

Here I did a 10A test with an Agilent high current PSU,
The 34470A is used to measure the voltage at the Resistopr
The 34461A is used to measure the current (Limited to 10A)

OK, this following setup failed, if looked from a metrology point of view.
The current and voltage is fluctuating too much.
Banana plugs are not good to be used for 10A current transport in to the resistor, because the binding posts get hot.

Results are:
Average Voltage: 100.17695 mV
Average Current: 10.004859 A
Average Resistance: = 10.0128 mOhm

It seems I have to look in to the metrology way of measuring high DC currents exactly with high precision.
Not easy and probably be very expensive.

OK, back to the Keithley 2460 and up to 7A

[plesa]

If you needs to make test at higher current than 7A (Keithley limit) you can still can use your Agilent/TDK PSU.
But you will needs to use one of your fancy precision resistors for monitoring current and set up feedback to voltage or current regulation terminal ( on rear DIP switch you you needs to set it to voltage controlled voltage or current source).
Input signal level is  between 0-5V or 0-10V, so some OpAmp gain needed.
This will make your setup far more stable.

[HighVoltage]

The 2460 SMU does go to 7A continuous current, although it does sound like an airplane when you do.

So, I started with 5A for about 15 min
But for this I had to build a new set of cables, starting at 4A the binding post of the resistor got warm.
Now I am using good quality lugs.

The readings are very stable at 5A
Actually standard deviation is 65 nOhm
Peak to Peak is about 500 nOhm

Interestingly, the voltage of the 34470A increased by about 1.3 uV
The resistor seems to not warm up at all, (not measured) and stays at room temperature.

[HighVoltage]

Ok, here is the last set of measurements at 7A.
At least this SMU has reached its limit

Also interesting, the fan stays at full power for a long time, after the current has been reduced from 7A to 100 mA and the instrument is very loud.

This cable I assembled seems to work very well for this test.

[zlymex]

It seems I have to look in to the metrology way of measuring high DC currents exactly with high precision.
Not easy and probably be very expensive.

One way to do this is to use a LEM IT-600S to scale down the current by 1500 times. The 1500:1 ratio is guaranteed and the linearity is ppm level. Even if used as a 15A:0.1A(10 turn primary and 10 Ohm secondary resistor), the parameter is also not degrade very much.

[Conrad Hoffman]

This thread triggered a memory of an old L&N low resistance box I keep stashed on the lower shelf of a scope cart. It needs a good cleaning, but I brushed off the dust. It doesn't seem to have a model number, but it's 300 ohms max in 0.001 ohm increments. I can't believe the cal sticker is original, as I doubt they made that style in '92.

[HighVoltage]

This thread triggered a memory of an old L&N low resistance box I keep stashed on the lower shelf of a scope cart. It needs a good cleaning, but I brushed off the dust. It doesn't seem to have a model number, but it's 300 ohms max in 0.001 ohm increments. I can't believe the cal sticker is original, as I doubt they made that style in '92.

That Leeds & Northrop resistance box looks fantastic.
Just amazing, what they built in those days.
Do you know, when it was built?
I would not be surprised, if it was right on to specification as well.

Reading up on it, Leeds & Northrup did built a Thomas 4210 with one Ohm and they built it to 1 ppm specification.
I bought one of these 4210 a few weeks ago  but it has not arrived.

[manganin]

I can't believe the cal sticker is original, as I doubt they made that style in '92.

Those aluminium stickers were used by the L&N lab at least until late 80's. And for one-off calibrations probably later than that.

If you post the serial number I can tell you the approximate year of manufacture.

[VintageNut]

Ok, here is the last set of measurements at 7A.
At least this SMU has reached its limit

Also interesting, the fan stays at full power for a long time, after the current has been reduced from 7A to 100 mA and the instrument is very loud.

This cable I assembled seems to work very well for this test.

Hello High Voltage

Are you performing an offset compensated measurement? If not, the multiple metal-to-metal contacts in the current path will induce random microvolts each causing significant error in your measurement.

There are multiple ways to accomplish the technique. In the end, you want two different currents forced, calculate the slope of the line, (v-v)/(i-i) and that is the true resistance of your device. The Y-axis intercept of the y=m*x+b equation is the lumped offset.

In a DMM, the highest current possible and lowest current possible are used to create the points to calculate the slope and thus perform the offset compensated ohms measurement.

For a SMU, you can use +7A and -7A which gives a much larger range and should be less error than a DMM all else being equal.

The old-school current reversal technique involves actually reversing the plugs at the current source. A more elegant solution is a polarity reversal switch between the current source and resistor.

The different methods have different sum-of-errors calculations.

For my 10 nano-ohm DIY tetrajunction, I used a KE2450 to force current. This allowed the use of TRIAX-to-alligator cables and placing the resistor and unshielded alligator clips into a shielded box which also prevented air flow over the device and alligator clips.

[Conrad Hoffman]

My only reference resistors (that I trust) are a 100 ohm L&N that I had measured some years back, and a Guildline 10 ohm that's lost its oil, like the one in the tear down thread. As far as I can tell, the L&N box is still within spec, but I don't have a good fractional ohm resistor I trust for comparison. NIST seems to be good at scaling things over large ranges, but I get nervous beyond 10X.

The L&N box has two metal plates that I thought were owner tags, but looking closer I think they're the L&N tags. The top one is two lines, "L.T." followed by "4171". Below that is a similar tag "Sp. Inst." The only other marking is "2377" stamped on the side of one of the taper keys, and a pencil notation inside the case I'm not sure about. I can also see the tape residue next to the cal sticker from a previous one, so it likely went back to L&N on a regular basis. I bought it at a hamfest decades ago when such things were being cleaned out of local businesses and universities. I seem to remember there were several of them, and they had trouble selling them at all. It's big and heavy. When I get a chance, I'll post some pictures of the insides. The 0.001 step section is basically a manganin bar with taps that's been filed slightly between taps to trim it. The rest is fairly conventional. I may also have a photo of a nice L&N knife-style reversing switch.

Another question- can an accurate ratio transformer be used in an AC measurement to compare resistors? Or are the big wound reference resistors so bad, even at 1 kHz or so, that it's not a good method?

[VintageNut]

Hello Conrad

I am curious how you connect to your L&N resistor box since the box has two connections. Do you use 4-wire ohms? What instrument do you use to measure the resistor box?

[Conrad Hoffman]

The giant binding posts are actually double nuts on a stud, so sort of a 4-wire connection, at least at that point. I put the device in series with a known resistor (my L&N 100 ohm) and use a 6.5 digit meter to measure the drop across that and the DUT. My resistor was certified about ten years ago, so it's as good as an L&N resistor is after ten years of drift!  I have a fairly large collection of them, plus a friend with more, so by intercomparison I know I'm not too far off, but no way would I guess better than 100 ppm or so without a fresh cert. An interesting aside, the resistor is quite old (chrome body) and when I first tried to have it certified, they rejected it as defective- too far out. Then we realized it was built before the ohm was adjusted (can't remember the year, '60s?) and it was actually almost dead on what it should have been.

Here's the insides, a bar for .001, loops for .01, coils for .1 and then wound and waxed coils for the rest. Also a not-great photo of the L&N switch that I believe was used for bridge reversals.

[VintageNut]

That is a very interesting box. I could see that used as a low volts, low offset, divider.

The solid copper bus bars are on every decade.

Looks like two correction/adjustment resistors?

[Conrad Hoffman]

The big resistors to the left are 100 ohms each. The box total is 300 ohms, the 100 ohm total decades plus the two 100 ohm coils that can be switched in and out. The odd value and "Sp. Inst." plate make me think this was a special for somebody, maybe based on a standard 100 ohm box. Alas, I have almost zero documentation on L&N, nor is there much on the web.

Proving anything resistance-wise requires a decent standard. I have tons of 0.1% parts, quite a few 0.01% parts, but only the one 4030 (or similar) L&N resistor that was measured to 6 places a decade ago. I think they charged me at least $100 back then, so I'm reluctant to do it again. Same problem with voltage- I have three standards that can be inter-compared that were calibrated a decade ago. They still compare closely, so I assume my volt hasn't drifted much. I just need somebody to ask for a traceable measurement and pay for a new round of calibration!

[manganin]

The L&N box has two metal plates that I thought were owner tags, but looking closer I think they're the L&N tags. The top one is two lines, "L.T." followed by "4171". Below that is a similar tag "Sp. Inst."

The secondary L.T. serial numbering system was for selected units better than the normal specifications. Used in L&N own labs and factories and possibly in some selected partner labs.

The number was engraved afterwards above the original serial number. It was usually not given at the time of manufacture, but based on years or sometimes decades of calibration history. For example I have a Rosa style standard resistor which has smaller L.T. serial number than a similar resistor manufactured almost 20 years earlier.

The Special Instrument tag could be the key for the missing main serial number. Maybe it was never built for sale.

[quarks]

Reading up on it, Leeds & Northrup did built a Thomas 4210 with one Ohm and they built it to 1 ppm specification.
I bought one of these 4210 a few weeks ago  but it has not arrived.

the LN 4210 is the very best 1 Ohm resistor I know of

more standards and details can be seen in zlymex post
https://www.eevblog.com/forum/metrology/teardown-standard-resistors/msg891744/#msg891744

[The Soulman]

An interesting aside, the resistor is quite old (chrome body) and when I first tried to have it certified, they rejected it as defective- too far out. Then we realized it was built before the ohm was adjusted (can't remember the year, '60s?) and it was actually almost dead on what it should have be.

It was on the 1st of January 1948 the standards were increased with 495 ppm.

I didn't even know it was changed at all, found it here:

https://www.nist.gov/sites/default/files/documents/calibrations/87mscohm.pdf

Thanks!

[Conrad Hoffman]

Good find- I saw that pub a decade ago but lost track of it. My resistor was certified 12/30/04, 12 years ago! Time flies. Back then it measured 100.05047 ohms, which seemed to upset the cal lab more than I would have thought. I should probably change the oil again, as it can become acidic over time and cause the value to drift.

I have various other standards, including a Fluke 5450A, but the 100 ohm L&N is the only thing with any calibration history, though I fully realize they're not the greatest standard.

I've experimented a bit with winding my own standards from both manganin and constantan. I've also got a spool of "800" alloy that's probably better than either of them, but you can't solder to it- has to be spot welded. Still, it all comes down to having a trusted standard to compare to.

IMO, it's best to have ancient standards with a known history of calibration so you can get some idea of long term drift! Or, being unable to afford much else, I keep telling myself that. 

[CalMachine]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

I would love to try

[VintageNut]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

I would love to try

Where are you located? What instruments do you have that will force 1A DC? Do you have a nanovoltmeter?

[CalMachine]

If you like low ohms, I have one of these in my office. It is about 4 nano-ohms. It is a challenge to measure.

I would love to try

Where are you located? What instruments do you have that will force 1A DC? Do you have a nanovoltmeter?

I am located in Ft. Wayne, IN.

As far as equipment, I have 2 Datron 4808s available, 2 Fluke 5520As, currently 1 Fluke 5700, as far as meters go, Keithley 181 or perhaps one of our 2002's or Agilent 3458As.  I also have a Fluke 752A Reference divider and Fluke 845AB Null detector if I want to get stupid with it

[e61_phil]

As far as equipment, I have 2 Datron 4808s available, 2 Fluke 5520As, currently 1 Fluke 5700, as far as meters go, Keithley 181 or perhaps one of our 2002's or Agilent 3458As.  I also have a Fluke 752A Reference divider and Fluke 845AB Null detector if I want to get stupid with it

4nR * 1A = 4nV

Only the Keithley 2002 in your list is capable to resolve 1nV. Isn't it?

[CalMachine]

As far as equipment, I have 2 Datron 4808s available, 2 Fluke 5520As, currently 1 Fluke 5700, as far as meters go, Keithley 181 or perhaps one of our 2002's or Agilent 3458As.  I also have a Fluke 752A Reference divider and Fluke 845AB Null detector if I want to get stupid with it

4nR * 1A = 4nV

Only the Keithley 2002 in your list is capable to resolve 1nV. Isn't it?

They keithley 181 is our nano voltmeter with resolution down to 1 nV.

[e61_phil]

They keithley 181 is our nano voltmeter with resolution down to 1 nV.

Sorry, I missed the 6.5 digit function. But even with 1nV of resolution it should be very hard to measure 4nV

Perhaps with some clever current reverse techniques.

[CalMachine]

They keithley 181 is our nano voltmeter with resolution down to 1 nV.

Sorry, I missed the 6.5 digit function. But even with 1nV of resolution it should be very hard to measure 4nV

Perhaps with some clever current reverse techniques.

Yeah it would be pretty difficult, but I would still love the opportunity to give it a go!  If not, I understand. 

What's the max current value?  Could always source more than an 1 amp to achieve a higher potential difference.

[VintageNut]

Max current is 5A. Even 20 nV is probably not going to happen for a KE181.

[chuckb]

My experience with measuring a 50u ohm current shunt. This is a large heavy multi-blade shunt that provides 50mV when 1000 amps passes through.

I used the Keithley 1801 nanovolt preamp to monitor the sense voltage. With 30 seconds of averaging the preamp has a 0.1nV pp noise floor. The K2001 DVM was set to high accuracy mode (100NPLC?), 20uV full scale and a 100 sample moving average.

As citizens of the planet we should all be concerned with not wasting energy. So for the first test I applied 100ua of current to the 50u ohm resistor and then I reversed the current. In the attached pdf you can clearly see the 5nV change in voltage along with the superimposed thermal EMF drift. It looks like the 1000 amp shunt is accurate! The 100ua was supplied by the 10v output of a F732A and a VHP202Z 100k ohm resistor. 

I tried to measure it with 10ua but there was to much voltage drift.

Later I applied 10ma and the voltage was rock steady at 0.5uV. With 2 amps from a bench supply the sense voltage was 100uV.
If the shunt is still accurate at 1000 amps (it should) it will have demonstrated it's linearity over a 10,000,000:1 range. Not to bad. A simple device is really the best.

With a 5 amp excitation current the K1801/K2001 combination should be able to discern 0.1n ohms. A battery powered A10 or A23 preamp from EM Electronics would work just as well.

[TiN]

chuckb
Did you also tried to use 2001's 4W function with preamp?
Interesting to see if 2001's current source stability is comparable for the purpose with separate current source.

[chuckb]

Yes I did but it did not work. I did not take the time to troubleshoot. This K2001 is a used unit that I use just for DCV with the K1801. I have not tested all it's functions. That's one more thing for the to-do list.

1 month to retirement!

[salbayeng]

Re Zero Ohm standard:
Couldn't one just short the two drive terminals together , and short the two sense wires together , then just connect a thin wire (or 100ohm resistor) between the two shorts?

[acbern]

Re Zero Ohm standard:
Couldn't one just short the two drive terminals together , and short the two sense wires together , then just connect a thin wire (or 100ohm resistor) between the two shorts?

Exactly

[HighVoltage]

If the shunt is still accurate at 1000 amps (it should) it will have demonstrated it's linearity over a 10,000,000:1 range. Not to bad. A simple device is really the best.

That is a pretty impressive resistor and demonstration of linearity.

What was this 1000A resistor used for?
Any specific application?
 

[chuckb]

The 1000 amp resistor is used to monitor current to the engine starter during development. We use it a few times a year.

[CalMachine]

If the shunt is still accurate at 1000 amps (it should) it will have demonstrated it's linearity over a 10,000,000:1 range. Not to bad. A simple device is really the best.

That is a pretty impressive resistor and demonstration of linearity.

What was this 1000A resistor used for?
Any specific application?

The proper term for this is a current shunt. 

The 1000 amp resistor is used to monitor current to the engine starter during development. We use it a few times a year.

What he said!  They come in all kinds of ranges.  The smallest current shunt value I've seen is like 1 mA and there are some that go up to several kA.

[VintageNut]

Re Zero Ohm standard:
Couldn't one just short the two drive terminals together , and short the two sense wires together , then just connect a thin wire (or 100ohm resistor) between the two shorts?

That is essentially what the PCB cal short that Fluke and Keithley are. It is not a nano-ohm magnitude device. It is a micro-ohm magnitude device. These devices are good for setting the floor of a DMM. They are not a metrology-grade zero. The Ohm Labs 100-0 is a metrology-grade zero. The 100-0 in my possession is no grater than 4 nano ohms at room temperature. I have measured it at 4 nano ohms. It is quite the pain in the butt to measure. If I had a KE1801, it would be a breeze to measure. An A23 amplifier is on my list for future acquisition.

[zlymex]

I don't know other places, but here in China, there is an regulation(JJG 1069-2011) saying that shunt must be calibrated(and tested or measured) at rated current(10%, 20%, 60%, 80%, 100% of rated to be precise). This is mainly to avoid the changes result from TCR at the very high temperature rise usually associated with large current.
Here is the regulation(in Chinese): http://www.docin.com/p-284172614.html

[chuckb]

chuckb
Did you also tried to use 2001's 4W function with preamp?
Interesting to see if 2001's current source stability is comparable for the purpose with separate current source.

The correct operation became clear when I read the actual K1801 preamp manual. I needed to move the output wires of the preamp card to the other set of K2001 input terminals. The meter forces current out the normal voltage input terminals in 4 wire ohms mode.

All is good now. With a 9.6ma excitation current from the DVM (offset comp on), 10 sample filtering and SLOW filtering on the K1801 the current shunt measure 50.4 u ohms. The stability was about 0.1u ohms. It is graphed over 15 minutes in the attached pdf.

I really like the 2m ohm full scale feature.

 

[VintageNut]

chuckb
Did you also tried to use 2001's 4W function with preamp?
Interesting to see if 2001's current source stability is comparable for the purpose with separate current source.

Hello Tin

It is not a current source stability issue. It is a simple ohms law issue coupled with the uncertainty+noise of the 1801 voltage measure.

The maximum current forced by the 2001 is 9.2mA (per the factory spec). This develops 460 nano-volts across Chuck's 50 micro ohm shunt. The 1801 does not have the ability resolve 460.0000 nano-volts. My experience with the 1801 was that 0.1 nano-volts is stable but not below 0.1 nano-volts.

This agrees pretty well with Chuck's observation that the bottom 3 digits are not stable. The least significant digit is 10^-13 VDC. Anything 10^-11 and smaller is noise at room temperature from my experience.

[chuckb]

I connected the K2002 (without preamp) to a nice Julie Research Labs NB-102 4 wire, oil filled, 100 ohm wire wound resistor. This resistor was probably over 30 years old. Setting the K2002 up in 4 wire, offset comp, 30 sample avg I had less than 100u ohm of measurement noise over 20 minutes. This would indicate 1 ppm measurement noise or short term stability.
I will check the K2001 later.

[HighVoltage]

Original Leeds Northrup good quality calibration resistors seem to be very expensive.
But somehow I am lucky to find BURSTER Made in Germany, resistors.

Today I got a 200 milli Ohm version.
And look how accurate it shows up on my 3458A
It seems that Burster knew what they were doing, when these resistors were built

 

[e61_phil]

You were a bit faster than me

nice resistor

[HighVoltage]

You were a bit faster than me

nice resistor

Yes, indeed a very nice resistor.
Actually I was surprise that it was still there, when I looked.
There are some more on ebay Germany but far too expensive.

[e61_phil]

Yes, indeed a very nice resistor.
Actually I was surprise that it was still there, when I looked.
There are some more on ebay Germany but far too expensive.

The others on eBay are close to the new price with certificate. I ask Burster a while ago and the 200mR resistor was around 230€.
I think the 200mR is very nice, because it is the lowest resistor value with the good TC. 100mR is worse in TC.


BTW: The resistor should show 200,04mR at max, if it is still in 0,02% spec. The 3458A shows a bit high. But I don't think the 3458A is capable to measure this.

[HighVoltage]

The others on eBay are close to the new price with certificate. I ask Burster a while ago and the 200mR resistor was around 230€.
I think the 200mR is very nice, because it is the lowest resistor value with the good TC. 100mR is worse in TC.

Interesting... I did the same thing and got a written quote from Burster and decided that the remaining resistors on ebay are too expensive.

Model: 1240-0,1
Kalibrierwiderstand 0,1 Ohm
Toleranz 0,02 %
Euro 235

They also offer an official DAkkS calibration certificate.

Here is the datasheet:

[HighVoltage]

I had the newly acquired 200 mOhm resistor all day on my Keithley 2450 at 1.000 A source current and this is the result I got.
It seems this Burster resistor is right on the spot, even with 1A constant current for many hours.

 

[Vgkid]

Does anyone have any experience with the ex ussr milli-ohm standard resistors?

[HighVoltage]

Does anyone have any experience with the ex ussr milli-ohm standard resistors?

I think I found one and it is on its way from Russia, so I will post it here when I have it.

--------------
Just today I got another really nice low resistance calibration resistor.
This one has 1 mOhm with a 0.05% tolerance

What I find amazing is, how accurate the reading is with a modern instrument that was not
available 30 years ago, when this resistor was built.
The 2460 just sits at one value extremely stable.

[HighVoltage]

A forum member asked me for the resolution of the Keysight 34470A in the low Ohms reading.

So, I thought to make a comparison to measure this 1 mOhm resistor with a few different Ohm-Meters.
It seems that such low resistance is a challenge even for high end multimeters.


Keysight 3458A
100 Ohm range
Bias current: 10 mA
Measurement: 0.00100 Ohm (movement between 0.00099 to 0.00101)
10 uOhm resolution
7 1/2 digits available


Keysight 34470A
100 Ohm range
Bias current: 1 mA
Measurement: Seems to be out of calibration with a minus Ohm value
6 1/2 digits available


Keysight 34410A
100 Ohm range
Bias current: 1 mA
Measurement: 0.0010 Ohm
100 uOhm resolution
1 nOhm ! resolution shown in Average Statistic mode (very unstable in all uOhm and nOhm digits)
6 1/2 digits available


Keithley DMM7510
This is the 7 1/2 digit Keithley top of the line multimeter
1 Ohm range
Bias current: 10 mA
Measurement: 0.0010614 Ohm
100 nOhm resolution
100 pOhm ! resolution shown in Average Statistic mode
7 1/2 digits available



Keithley 2450 SMU
This is the 6 1/2 digit Keithley SMU

OHM-Mode
2 Ohm range
Bias current: 100 mA
Measurement: 0.001118 Ohm
100 uOhm resolution
5 1/2 digits available

SMU-Mode
20 mV range
Bias current: 1 A
Measurement: 0.001011 Ohm
1 uOhm resolution
6 1/2 digits available



Keithley 2460 SMU
This is the 6 1/2 digit Keithley SMU

OHM-Mode
2 Ohm range
Bias current: 100 mA
Measurement: 0.001167 Ohm
1 uOhm resolution
6 1/2 digits available

SMU-Mode
200 mV range
Bias current: 1 A
Measurement: 0.001000 Ohm
1 uOhm resolution
6 1/2 digits available

[The Soulman]

Is using single binding posts sufficient for measuring these low values? Well apparently it is. 

Edit: I missed the second set of terminals, why use 6 wires?

[HighVoltage]

There are 4 binding posts on the 1 mOhm resistor.

For the 2 high current posts I am using spade connectors and 2.5 mm^2 cable
All Banana plugs are gold plated from MultiContact

The extra spade to 4 mm connectors I used to measure the voltage drop separately.
But this was not hooked up for the measurements.



 

[CalMachine]

There are 4 binding posts on the 1 mOhm resistor.

For the 2 high current posts I am using spade connectors and 2.5 mm^2 cable
All Banana plugs are gold plated from MultiContact

The extra spade to 4 mm connectors I used to measure the voltage drop separately.
But this was not hooked up for the measurements.

Is that a Keithley supplied cable you're using?

[HighVoltage]

Is that a Keithley supplied cable you're using?

No, I did not like the cable that was supplied by Keithley SMU and made this one myself.
The cables are all lab grade LiFY, ultra flexible, ultra fine strain copper.
The banana plugs MultiContact gold plated.

For higher current measurements, I have not found a better cable setup.
 

[HighVoltage]

I calibrated the zero Ohm of the 34470A again new, to a short across the inputs.
And when the calibration is finished, it shows only zeros, as it should.

Also, this calibration is confirmed with the 100 mOhm and 10 mOhm reference resistors.
But the 1 mOhm resistor is not showing correctly on the 34470A at all. It moves between -10 uOhm and +6 mOhm.

May be the 34470A was not meant to be for low resistance value measurements.

[CalMachine]

Is that a Keithley supplied cable you're using?

No, I did not like the cable that was supplied by Keithley SMU and made this one myself.
The cables are all lab grade LiFY, ultra flexible, ultra fine strain copper.
The banana plugs MultiContact gold plated.

For higher current measurements, I have not found a better cable setup.

I love it!  Especially the braided sheath touch.  Very nice craftsmanship.      

[HighVoltage]

I love it!  Especially the braided sheath touch.  Very nice craftsmanship.   

Thanks.

Indeed, this cable works very well.
I see no difference in Low Ohm measurements, compared to a PTFE cable with only crimped gold plated copper spade connectors.
These MultiContact banana plugs are very smooth to insert in to the instrument.

[HighVoltage]

Now, this is funny...

I hooked up the 1 mOhm resistor to an older 34401A instrument and what a surprise, it sits at 0.0010 Ohm
The last digit is moving up and down by +/- 1 in short test.

So, one of my 34401A is better to use for the 1 mOhm range than the 34470A?

[CalMachine]

Now, this is funny...

I hooked up the 1 mOhm resistor to an older 34401A instrument and what a surprise, it sits at 0.0010 Ohm
The last digit is moving up and down by +/- 1 in short test.

So, one of my 34401A is better to use for the 1 mOhm range than the 34470A?

Have you tried with the offset comp on, on the 34470A?

[gamalot]

Now, this is funny...

I hooked up the 1 mOhm resistor to an older 34401A instrument and what a surprise, it sits at 0.0010 Ohm
The last digit is moving up and down by +/- 1 in short test.

So, one of my 34401A is better to use for the 1 mOhm range than the 34470A?

Have you tried with the offset comp on, on the 34470A?

I guess he didn't because I got much better result on my 34465A with the offset comp turned on. 

[CalMachine]

Now, this is funny...

I hooked up the 1 mOhm resistor to an older 34401A instrument and what a surprise, it sits at 0.0010 Ohm
The last digit is moving up and down by +/- 1 in short test.

So, one of my 34401A is better to use for the 1 mOhm range than the 34470A?

Offset comp always makes low ohms measurements more accurate! 

Have you tried with the offset comp on, on the 34470A?

I guess he didn't because I got much better result on my 34465A with the offset comp turned on. 

[onemilimeter]

May be the 34470A was not meant to be for low resistance value measurements.

Hi...

I wonder if Keysight states about this "low resistance value measurement" limitation in the datasheet... If it's not stated, can we assume that the 34470A, which shows 100.0000 at its 100-ohm range, must be able to measure a low resistance like 000.0010 ohm?

Thanks

[onemilimeter]

So, one of my 34401A is better to use for the 1 mOhm range than the 34470A?

That's really funny... 34401A can measure 1 mOhm but not the 34470A... 

[onemilimeter]

I guess he didn't because I got much better result on my 34465A with the offset comp turned on. 

Hi.. how much difference in the results with and without the offset comp turned on? Thanks.

[HighVoltage]

I guess he didn't because I got much better result on my 34465A with the offset comp turned on. 

Hi.. how much difference in the results with and without the offset comp turned on? Thanks.

Ok, it is true, Offset Comp was turned OFF yesterday.
So just now I made another test and here is the result: No difference!

[onemilimeter]

Ok, it is true, Offset Comp was turned OFF yesterday.
So just now I made another test and here is the result: No difference!

Aiii... 

Not sure Keysight knows about this...

[HighVoltage]

Aiii... 

Not sure Keysight knows about this...

I think something is wrong with my 34470A and I will send it to Keysight for repair and/or calibration.
Yes, the old 6 1/2 digit 34401A being more accurate for this measurement than the new 7 1/2 digit 34470A is kind of embarrassing.

[2N3055]

If I remember correctly member here had problems with noisy ohm measurements ( In kohms range ) I belive his DMM was replaced by Keysight (it was brand new)...

[onemilimeter]

Yes, the old 6 1/2 digit 34401A being more accurate for this measurement than the new 7 1/2 digit 34470A is kind of embarrassing.

100% agree...

[Alex Nikitin]

If I remember correctly member here had problems with noisy ohm measurements ( In kohms range ) I belive his DMM was replaced by Keysight (it was brand new)...

That was me  . I've checked the low ohm performance (4W) on the 34465A right now - with a short copper wire ~ 0.7 mOhm the span is about 0.5mOhm, either on NPLC 10 or NPLC 100. With the same Kelvin clips and the same bit of wire the Agilent 34401 I have here temporarily does show about 0.7-0.8 mOhm span.

Cheers

Alex

[Dr. Frank]

HighVoltage,

at NPLC10, the stddev of about 0.12 mOhm seems reasonable, my 34465A shows the same for a 10 mOHM resistor.
Both instruments resolve the same at those low values, i.e. 0.1 mOhm on the display, and 0.01 mOhm on the statistics.

At NPLC100, which you also might test, stddev is about 0.08 mOhm, span 0.4 mOhm, Min = 9.8 mOhm, Max = 10.1mOhm, Mean = 9.96 mOhm. So for a 1mOhm resistor, I wouldn't expect negative Min values as well. That makes no sense, as the Ohm circuit obviously is linear down to zero.


On the other hand, you shouldn't expect too good an accurcay at 1mOhm, these instruments are really not made for that, as the test current of 10mA is much too low.
The 100 Ohm range is specified worst, i.e. 30ppm of range (24h). That is +/- 3mOhm !!

Maybe your zero calibration did not result in a perfectly centered result, around 1mOhm, but in face of the specification, that's no wonder at all.

In the end I think, that your instrument is fine, if you look at the stddev values.


Anyhow, if you let KS calibrate and adjust your 34470A, for the 2nd time in its life, maybe, then the 100V and 1kV DC calibration then should be precise,  for the first time.
When I checked your 34470A, it also had this 15ppm bug in the HV ranges.
Frank

[gamalot]

I have done the test again 

[HighVoltage]

On the other hand, you shouldn't expect too good an accurcay at 1mOhm, these instruments are really not made for that, as the test current of 10mA is much too low.
The 100 Ohm range is specified worst, i.e. 30ppm of range (24h). That is +/- 3mOhm !!

Frank

Hello Dr.Frank
Thanks for this great explanation, this does make perfect sense what you are writing and I would totally agree.
But If I keep the 1mOhm resistor hooked up to the 34470A, the shown values drift significantly larger than the +/- 3mOhm
Sometimes I see even negative values.

And on the other side, ... the 34401A, 34410A and 34411A seem to be more stable for this measurement than the 34470A
And then you are also right, I should not expect the 34470A to perform so well in the low resistance values because that is not what it is made for.

Well, in a few weeks I will have my 34470A calibrated at Keysight Germany and then I will do another test with the 1 mOhm resistor.

[HighVoltage]

I have done the test again 

Hello gamalot,
Interesting, it seems you confirmed my finding, that there is no difference in Offset Comp ON or OFF

What is the real nominal value of your 1 mOhm resistor?
It seems you are also confirming with your 34465A, that this instrument really is not up for the 1 mOhm task.

[gamalot]

I have done the test again 

Hello gamalot,
Interesting, it seems you confirmed my finding, that there is no difference in Offset Comp ON or OFF

What is the real nominal value of your 1 mOhm resistor?
It seems you are also confirming with your 34465A, that this instrument really is not up for the 1 mOhm task.

No, don't get me wrong! 

Look at the two pictures carefully, you will find the difference between the two averages:

1.02 mOhm with Offset Comp turned on

0.73 mOhm with Offset Comp turned off

The resistor under test is 1 mOhm

[HighVoltage]

Thanks gamalot for the clarification, it makes sense now.

Your Hioki low resistance meter looks very good!
They are just sooo expensive.

[quarks]

because of the HIOKI RM3545 shown here,  I just updated my old "Resistance Comparison" chart, which I made several years ago, with some of the other gear shown here (and left in the FLUKE 8508A for reference)



Keithley 2460 is not fully shown here, because of the odd accuracy specs <2 Ohm and >20MOhm (which I did not calculate)


edit: found an error in HIOKI RM3545 graph and corrected it

[HighVoltage]

Thanks quarks for this nice chart.
And it becomes very clear now, why the 34470A is not suited for a 1 mOhm measurement as I have tried.

Also a surprise how much better the Fluke 8508A is in comparison to the Keysight 3458A for the full range.

[onemilimeter]

And it becomes very clear now, why the 34470A is not suited for a 1 mOhm measurement as I have tried.

Hi... quite sad to read this ... Is this "the 34470A is not suited for a 1mOhm measurement" mainly because of the low test current 1mA at 100-ohm range? Thanks.

[e61_phil]

Today I tried to measure a 1mR Isabellenhütte PBV 0.5% shunt on the 34470A. It showed 0.0010 Ohms.

[VintageNut]

Today I tried to measure a 1mR Isabellenhütte PBV 0.5% shunt on the 34470A. It showed 0.0010 Ohms.

What is the uncertainty of the measurement according to the datasheet?  You are only using 10 counts of a 10,000 count range.

[e61_phil]

Today I tried to measure a 1mR Isabellenhütte PBV 0.5% shunt on the 34470A. It showed 0.0010 Ohms.

What is the uncertainty of the measurement according to the datasheet?  You are only using 10 counts of a 10,000 count range.

I know it is nonesense to measure 1mR directly with the 34470A. I did it only because of the wired measurements HighVoltage took.

[HighVoltage]

Today I tried to measure a 1mR Isabellenhütte PBV 0.5% shunt on the 34470A. It showed 0.0010 Ohms.

What is the uncertainty of the measurement according to the datasheet?  You are only using 10 counts of a 10,000 count range.

I know it is nonesense to measure 1mR directly with the 34470A. I did it only because of the wired measurements HighVoltage took.

Thank you for making this measurement on your 34470A
It seems yours is at least better calibrated in this low ohm area than mine and I will have my meter definitely calibrated by Keysight Germany, in the next few weeks.

[HighVoltage]

Does anyone have any experience with the ex ussr milli-ohm standard resistors?

I got my first USSR precision resistor with a 1 Ohm nominal value.
No papers and no info was given by the seller.
So I can only go by what is written on the label

Model P3030 from 1986
It seems the 6 mm studs are pure silver or at least silver plated.

Measurements:
1.000068 Ohm - Keithley 2460 SMU
1.00005 Ohm - Agilent 3458A

I will keep this one hooked up to the Keithley 2460 SMU to take some Data over the next few days.
It will be interesting to see, how stable this one will be in comparison to the Burster resistors.

[e61_phil]

Can you tell me which uncertainty is achievable at 1R with the Keithley SMU?

I think a (or better two) 3458A can achieve 52ppm with an external current source (100mA) within 1 year specifications. Is the SMU better?

[manganin]

I got my first USSR precision resistor with a 1 Ohm nominal value. No papers and no info was given by the seller. So I can only go by what is written on the label

Soviet foil technology! Better initial adjustment accurary but less stable than the Krasnodar factory wire wound P321 (which unfortunately also suffers from the Soviet quality: there are very good examples, but you need to buy ten to find one.)

[HighVoltage]

Can you tell me which uncertainty is achievable at 1R with the Keithley SMU?

I think a (or better two) 3458A can achieve 52ppm with an external current source (100mA) within 1 year specifications. Is the SMU better?

I don't think the SMU is better than 52 ppm at 1 Ohm.
If I remember right, the Ohm uncertainty was something like 0.04% + 1 mOhm on the datasheet of the SMU. (I will have to look it up.)

But the 2460 SMU has 1 uOhm resolution in the 2 Ohm range, and the 3458A has 10 uOhm resolution in the 100 Ohm range.

[HighVoltage]

I got my first USSR precision resistor with a 1 Ohm nominal value. No papers and no info was given by the seller. So I can only go by what is written on the label

Soviet foil technology! Better initial adjustment accurary but less stable than the Krasnodar factory wire wound P321 (which unfortunately also suffers from the Soviet quality: there are very good examples, but you need to buy ten to find one.)

It seems you know these soviet resistors.
Thanks for this explanation.
Is the quality issue found in all ranges of the P321 or more in the lower or upper ranges.
May be I will get a 1 Ohm P321 and compare it to this P3030

[quarks]

Can you tell me which uncertainty is achievable at 1R with the Keithley SMU?

I think a (or better two) 3458A can achieve 52ppm with an external current source (100mA) within 1 year specifications. Is the SMU better?

if you look in my comparison chart above, you can see that 2460 accuracy is much worse than 3458

[e61_phil]

if you look in my comparison chart above, you can see that 2460 accuracy is much worse than 3458

Sorry, but I can't. The specifications for 1R are missing (best volts and amps combination for 1R).

And it seems to me, that you already calculated the best possible resistance uncertainty using volts and amps and not resistance measurement. Is that right? Because, the 3458A has about 10x15ppm+5ppm=155ppm uncertainty at 1R. If one combines 100mA (40ppm) and and 100mV (12ppm) you achieve 52ppm.

[quarks]

if you look in my comparison chart above, you can see that 2460 accuracy is much worse than 3458

Sorry, but I can't. The specifications for 1R are missing (best volts and amps combination for 1R).

And it seems to me, that you already calculated the best possible resistance uncertainty using volts and amps and not resistance measurement. Is that right? Because, the 3458A has about 10x15ppm+5ppm=155ppm uncertainty at 1R. If one combines 100mA (40ppm) and and 100mV (12ppm) you achieve 52ppm.

maybe I got your initial question wrong, but for me it is clear that the 3458A is much better than a 2460 for 1 Ohm

in my comparison chart I used resistance 1 year accuracy specs from the datasheets

[e61_phil]

maybe I got your initial question wrong, but for me it is clear that the 3458A is much better than a 2460 for 1 Ohm

It wasn't clear to me, that the Keithley SMU is in no range better than 100ppm. Therefore, 1R is also much worse than the 3458A.
I had a look into the datatsheet now.

in my comparison chart I used resistance 1 year accuracy specs from the datasheets

I read about 50ppm in your chart. The datasheet says 155pmm for 1R and 1year.

[quarks]

in my comparison chart I used resistance 1 year accuracy specs from the datasheets

I read about 50ppm in your chart. The datasheet says 155pmm for 1R and 1year.

3458A 1 year spec is 15ppm + 5ppm in the 10 Ohm range and for 1 Ohm this makes 65ppm

[e61_phil]

in my comparison chart I used resistance 1 year accuracy specs from the datasheets

I read about 50ppm in your chart. The datasheet says 155pmm for 1R and 1year.

3458A 1 year spec is 15ppm + 5ppm in the 10 Ohm range and for 1 Ohm this makes 65ppm

Yeyp, sorry. I interchanged ppm of reading and ppm of range.

[HighVoltage]

The Russian based Model P3030 1 Ohm resistor was now for almost 48 hours on the Keithley 2460 SMU at 100mA bias current.
And it shows a significant drift of 48 uOhm over 171.000 readings.
But may be that is normal?
And I dont know how much drift is contributed by the SMU.

I have also ordered a Burster 1 Ohm resistor for comparison purpose.

[2N3055]

The Russian based Model P3030 1 Ohm resistor was now for almost 48 hours on the Keithley 2460 SMU at 100mA bias current.
And it shows a significant drift of 48 uOhm over 171.000 readings.
But may be that is normal?
And I dont know how much drift is contributed by the SMU.

I have also ordered a Burster 1 Ohm resistor for comparison purpose.

Well,
it looks like you have P-P 48 ppm drift+noise on 1 Ohm..
Noise is 9.686 ppm RMS, roughly it looks like about 20 ppm P-P.. So median drifted about 20 ppm.
Keithley 2460 has Extended specs +-40 ppm/°C of reading.... (0,1% /°C of +-0,04%)... on 20 Ohm/100mA range..
Keithley alone could have drifted that much with change in temp of less than 0.5°C.

But that noise looks high.. 100mA range on 2460 is specified at 100nA RMS noise..
So contribution to noise from SMU should be in 100nV range RMS on 1Ohm resistor.. You seem to have 100x more noise..

Maybe cabling, external disturbances?
Did I calculate wrong?

Regards,

Sinisa

[HighVoltage]

Maybe cabling, external disturbances?
Regards,

Sinisa

I also suspected the cable and made a new one.

Well, I also got a new resistor to play with.
This is a Russian P321 1 Ohm in good condition, was suppose to be "new".
But it came without any paperwork.

I made a cable, Silver plated copper with crimped lugs (Low EMF) and PTFE insulation.
The shield is grounded to the 2460 SMU

The results are much better already, with close to 100.000 readings.
Pk/Pk is only 26 uOhm.

[HighVoltage]

Finally I got a Burster 1240-1, a 1 Ohm calibration resistor, "Made in Germany"
Condition: looks like new, came in original box.
Thank you eBay !

This one was made on 12.05.2000
Has a listed tolerance of 0.025% and a stated Imax of 1.5A
I have tested this resistor on my Keithley 2460 with 100 mA for a few hours.

Here are the results:
- Measured average value: 1.000 081 Ohm
- Pk to Pk over 6841 readings: 19.6 uOhm
I think I can be very happy with this resistor and first results.

[alm]

Noise for the Burster looks very similar to the P321. Still, the 3 uVrms noise seems high compared to the SMU specifications, suggesting that improvements are still possible with better shielding.

I think it is hard to make any statements about short-term drift since it appears to be buried in the noise of the setup. Averaging might help, or just measure over larger intervals.

[HighVoltage]

I think it is hard to make any statements about short-term drift since it appears to be buried in the noise of the setup. Averaging might help, or just measure over larger intervals.

OK, I turned the "Filter" ON and had it run over night.
Now we can see a drift of 34 uOhm
The temperature did fall about 1.5 °C in the same time.

May be it is time for me to get a temperature chamber.

[alm]

If that drift is just from ambient temperature, then it would translate to about 20 ppm/K. Much higher than the 10 ppm/K specified by Burster. But if I read the specs right, the temperature coefficient of the SMU alone could easily be responsible for ten times that value. The specs state that the temperature coefficient is 0.1 * accuracy / K. Accuracy is about 2400 ppm (the 200 µV offset for the 200 mV voltage measurement range is responsible for most of that). Better put the SMU in a thermal chamber .

One thing you may be able to do is assume the Burster resistor meets its specs, and measure the Russian resistors relative to it. Either that or measure for long time periods and use regression to try to compensate for temperature fluctuations.

[HighVoltage]

This 1 Ohm Burster resistor was turned ON with the Keithley 2460, since I got it.

When I started out with this Burster 1 Ohm resistor, it had shown 1.000 080 Ohm on the SMU
A day later, I ended up with 1.000 040 Ohm
Again, two days later, it shows 1.000 013 Ohm
It has fallen about 55 ppm

So, there is clearly a drift that is no longer associated to only temperature variations.
Somehow I had expected it to be stable after a few hours with a straight line in the graph.
But may be equilibrium has not been reached inside the resistor from the current flow?

I would assume this drift is related to the resistor and not the Keithley SMU?

Any comments?

[HighVoltage]

I was gone over the weekend and left the instrument running with this 1 Ohm Buster resistor.

Now, after 5 days, it has settled to a perfect value of 1.000 001 Ohm on the SMU

Without a real temperature chamber, it will probably not be of any real value to keep it running, So, I will start looking for some good temperature regulation equipment.
 

[HighVoltage]

For a new project, I need to measure 80A DC very precise and accurately.
Current clamp probes would not be accurately enough.

So I got one of these "Made in Germany" WEIGEL 100A shuts with 60mV /100 A of voltage drop.
This is a 0.5% class shunt.

Today I did the first test for up to 5A and look at the result.
Almost too good to be true.
Interestingly, I see no marks of calibration on this resistor. It is like they built it and it is spot on.

In the next few days I will test it for up to 100A, to see how linear this resistor is.

Anyone here having experience with this Weigel type shunts for metrology type purpose?

[alm]

No experience with this type. A current shunt should be calibrated at full current to account for self-heating. But since the temperature coefficient may have a parabolic shape, it could be that 5 A (close to room temperature) and 100 A (max temperature) end up at almost the same resistance value, which could account for why the 5 A reading looks so good. See figure 6 in this paper for a curve from an older (manganin-type) shunt. Obviously the curve for Evanohm would be flatter.

[ap]

If it has to be very precisely and accurately (whatever that is  ), then Isabellenhütte are nice, 1ppm/K. Or Ultrastab. Not cheap.

[HighVoltage]

See figure 6 in this paper for a curve from an older (manganin-type) shunt. Obviously the curve for Evanohm would be flatter.

That was a nice paper to read, thank you for the link.

Today I hooked this shunt resistor up to 100A DC but of course, I can only read my PSU to one decimal place and using a current clamp also does not help to get better resolution.

It seems I need an additional calibrated shunt, to get any kind of current accuracy at 100A.

At 100A and running for about 30 min, the resistor elements heat up to about 70°C and the 35mm² cable heats up to about 40 °C

Removing the resistor at this temperature and measuring the resistance again at 5A, gives exactly 0.000600 Ohm again. That was a little surprise.

[alm]

I was going to propose heating the resistor up and measuring the resistance at a lower current if you do not have the facility to measure 100A accurately. Is obviously not a metrology-grade measurement because the temperature distribution could be different, but it should be pretty close.

Not sure what the proper way would be (apart from sending it off to a cal lab that can generate 100A). Is it possible to do a current transfer with a stable current source with the current equivalent of a KVD?

[HighVoltage]

Not sure what the proper way would be (apart from sending it off to a cal lab that can generate 100A). Is it possible to do a current transfer with a stable current source with the current equivalent of a KVD?

Good question, I do not know, may be someone else with precision high current measurement experience can explain this.

All I have read so far is to use a calibrated precision shunt resistor.
But they are very expensive.

Also, in the paper you posted yesterday, there was a nice explanation that the DUT has to settle over time to reach equilibrium.

Here is my test:
There is a fast immediate drop and then it slows down.
In this test, equilibrium has not been reached after about 9 min.

This means, we must wait many minutes before we take current and voltage reading, when measuring such resistor.

[alm]

But how do they calibrate the calibrated precision shunt resistor?

Would heating the resistor (ovenizing it) to just under 70°C reduce settling time? Obviously it would have to be calibrated like that.

[e61_phil]

But how do they calibrate the calibrated precision shunt resistor?

I think they use a direct current comparator. With such a unit you can compare currents with the ratio of turn over a core. The "secondary" winding is controlled in a way to null the flux in the core. Therefor, you can measure the secondary (lower) current with another shunt and transfer the reading.

By the way I did some measurements on such a shunt a while ago. But with 1A only: https://lowcurrent.wordpress.com/2017/02/10/kleine-widerstaende-und-offset-compensation/ (sorry, it is in german)

In the mikrocontroller.net forum was a discussion about a 100A source and a guy (from a german cal lab) said, that this current shunts which looks like yours are in no way good enough for measurements below .5%.

[HighVoltage]

But how do they calibrate the calibrated precision shunt resistor?

Good question, I think I will ask a few cal labs here in Germany.

Would heating the resistor (ovenizing it) to just under 70°C reduce settling time? Obviously it would have to be calibrated like that.

Probably yes.
I had expected this shunt to be calibrated at 100A and was surprised to see spot-on numbers at only 5A
The problem is that this resistor will have a different temperature for each current level.

I would really like to measure the TC curve of this resistor from 1A to 100A to get an idea of its temperature drift. But that requires a second better quality shunt. May be I should get a 1000A shunt for a second one that would not heat up too much and will stay in the more linear area.
Hmm, more and more questions.

Plus... if I have two shunts ... I need a third for comparison.
The good thing about shunts, they can all be put in a series connection without loosing accuracy.
   

[HighVoltage]

By the way I did some measurements on such a shunt a while ago. But with 1A only: https://lowcurrent.wordpress.com/2017/02/10/kleine-widerstaende-und-offset-compensation/ (sorry, it is in german)

In the mikrocontroller.net forum was a discussion about a 100A source and a guy (from a german cal lab) said, that this current shunts which looks like yours are in no way good enough for measurements below .5%.

Nice write up, thanks for the link.
Your picture is funny to see the little alligator clips on the large screws. But I guess as long as the current flows, it should not make much difference.

So far, I am under the impression that this resistor is much better than the .5% it is posted at. Otherwise I would not read 0.000,601 Ohm at 5.000,000A (This one was with a heated resistor).
Or may be I was lucky to just get a "good" one?

 

[HighVoltage]

It seems this setup requires about 1 full hour to settle and find equilibrium at 100A.
Here is the 1 hour graph...

It would be interesting to hear, how official cal labs do this.
Do they really wait 1 hour for each measurement point to settle?

[alm]

I think they use a direct current comparator. With such a unit you can compare currents with the ratio of turn over a core. The "secondary" winding is controlled in a way to null the flux in the core. Therefor, you can measure the secondary (lower) current with another shunt and transfer the reading.

Ah, so there is something like a voltage divider that lets you transfer between current ranges. Thanks!

The problem is that this resistor will have a different temperature for each current level.

True. What you could do is thermally couple a second resistor to it, and dissipate approximately (100A - expected current)R in it to keep the total dissipation approximately constant. You are not aiming for single degree C temperature stability here, the temperature coefficient is supposed to be good enough for fluctuations from 25°C to 70°C. Not sure if there is any precedence on this, or if this is a completely ridiculous idea. It will probably slightly degrade accuracy compared to waiting an hour, but should produce a much better reading after waiting a minute.

I would really like to measure the TC curve of this resistor from 1A to 100A to get an idea of its temperature drift. But that requires a second better quality shunt.

How about just heating it using some other method (oven, other resistor) and measure the resistance at 5 A, like you have been doing? Once you know the tempco as a function of temperature, you should be able to estimate the relation to current. Obviously this is not the cal lab method (temperature gradients will be different), but it sounds like you are not equipped like a cal lab either .

So far, I am under the impression that this resistor is much better than the .5% it is posted at. Otherwise I would not read 0.000,601 Ohm at 5.000,000A (This one was with a heated resistor).

What is the uncertainty on that resistance measurement based on the detailed specs for the SMU? It could just be that the SMU error happens to cancel the shunt error.

[Kleinstein]

It seems this setup requires about 1 full hour to settle and find equilibrium at 100A.
Here is the 1 hour graph...

It would be interesting to hear, how official cal labs do this.
Do they really wait 1 hour for each measurement point to settle?

There can also be quite some settling in the SMU. I would expect the shunt inside the SMU to be of about similar level of precision. In the SMU there is not only the dissipation of the shunt but also the power source part.
So to make sure it is the shunt and not the SMU, one would first measure a dummy or a short and only after something like an hour switch to the shunt.

[HighVoltage]

Thank you for your feedback, Kleinstein and alm.

I just received another shunt resistor and I want to use this one as a reference for the 100A one.

This new one has following markings
- 240A
- 5 Ohm
- 60 mV

It seems to have a manufactures logo and a part number on the side but I have no idea who made this one.
May be someone here recognizes that logo / stamp?

Also interesting, this one is marked 5 Ohm but in reality has 250 micro Ohm
Is that just a stamping mistake, hmmm?

And another curiosity came up during my research in to shunts.
Why is it that most of them have 60 mV at nominal current ?
May be in the past there was a volt meter specifically made for shuts with a maximum of 60 mV
But why just 60 mV?

[lukier]

I got this 150A shunt recently: http://www.ebay.co.uk/itm/322631256627

It is also for 60mV full scale (thus 0.4mOhm). I still need to do the testing under high load, but I was surprised how well my Keithley 2001 can measure it (well at least the resolution was there).

I guess the 60mV comes from old D'Arsonval movement panel meters.

[HighVoltage]

I got this 150A shunt recently: http://www.ebay.co.uk/itm/322631256627

It is also for 60mV full scale (thus 0.4mOhm). I still need to do the testing under high load, but I was surprised how well my Keithley 2001 can measure it (well at least the resolution was there).

I guess the 60mV comes from old D'Arsonval movement panel meters.

Yes, I totally agree.
There is something about Keithley and accuracy when it comes to low Ohm measurements.

Today I gave my 100A Weigel resistor 2 new Hirschmann binding posts for the sense wires.
This resistor also measures very accurately with the Keithley SMU.

These measurements were taken at 5A and the resistor is 1 uOhm off on the Keithley and the min max don't even change at 70nOhm standard deviation over 683 readings.

[tszaboo]

You can use the Kirchoff's law to calibrate these shunts. That is how we made accurate (0.03%) 100A at the lab. Basically we had a bunch of current sources, each with 0.03% accuracy, and we just summed the currents. We could produce up to 4000A with 0.03%. Each current source had it's own water cooled shunt (the entire thing was water cooled), kept at 23-ish degrees, to avoid the dew point. Each shunt individually was calibrated in house. Calibrating a 10mOhm shunt is "easy".

We were selling the current sources/sinks (aka, battery testing equipment). The proper calibration waited for the shunts to heat up. But that heating up happened fast, since they were directly on a temperature controlled cold plate. Seconds.

Now, that is how we did it. Then we sent our reference shunts (good up to 50A) to the cal lab, and magic happened and they came back calibrated. To be honest, I dont know what the magic was.

[Edwin G. Pettis]

I have some Tokyo Seiden 30A, class 0.5 current shunts, they are calibrated for 180mV full scale.

[HighVoltage]

I have some Tokyo Seiden 30A, class 0.5 current shunts, they are calibrated for 180mV full scale.

Interesting!
So, they are then 60mV / 10A
May be this was an old standard, to build everything around 60mV full scale?
But why?

Also, why not today build it always around 100mV / 100A to make easy to use a DVM for readout?

[HighVoltage]

Calibrating a 10mOhm shunt is "easy".

What is your procedure to calibrate a 10 mOhm resistor in your lab?
At how many points between a low current and full scale current do you check for linearity?


Very interesting with your 4000A of multiple current sources.
With my two Agilent 6V / 100A PSUs I just can reach 200A.

[TiN]

Perhaps if you really interested in high accuracy high current stuff, you'd be better off getting LEM Ultrastab DCCT.
They pop up on ebay from time to time to relatively cheap prices. I've got two with best linearity, Ultrastab IT-600S.
It is 1:1500 ratio CT, and with linearity spec <1ppm, offset error <20ppm and tempco <0.2ppm/K, so high chance that your measurements accuracy with these DCCTs will be limited by your burden resistor stability. There are also lower (and higher) current Ultrastab's with similar specs. Buying older Danfysik DCCT (model 867 is same as IT-600S) might save few $$.

I did a brief test before, with 5700A and 5205A as current source.

CERN using these beasts to calibrate and maintain LHC superconductor magnet currents to ppm levels stability. You can find related research paper, covering their transfer and calibration apparatus. 

[HighVoltage]

Hello TiN

You just opened a complete new door for me to look at currents in ppm level.
Thanks, I was not aware of these current sensors and ordered one.

[tszaboo]

Calibrating a 10mOhm shunt is "easy".

What is your procedure to calibrate a 10 mOhm resistor in your lab?
At how many points between a low current and full scale current do you check for linearity?


Very interesting with your 4000A of multiple current sources.
With my two Agilent 6V / 100A PSUs I just can reach 200A.

Well, we did characterization, and concluded that 2x2 point calibration is good enough for our specification.  I think we verified hundreds different scenarios on a new design. Different output voltages, currents, rise and fall times...

2x2 because sourcing and sinking current was separately calibrated. Keep in mind that we were selling machines, which had two to four ranges (50A the largest), and 80-120 channels, so calibrating these took hours. Any other measurement point would extend the calibration time by hours, which would mean increasing the lead time.

The procedure for the 50A range is the following. We had our reference 2 mOhm shunt. It was an Isabellenhütte RUGZ. 1ppm/K, 250W max power (we used 5W from this), on a big heatsink. Full scale was 100mV, this went into a 34410A. The two points were 10% and 90%. The reference shunt went for traceable calibration every year. I think it should have been more often, they drifted quite a lot.

We weren't just calibrating just the shunt of course, but the machine as a current source. Together with the AFE and the digital control values. But the principle is the same.
And I write past tense, because I'm not working there anymore. In fact, I think it was 2mOhm shunt built in, not 10... Memory.

In principle, if one can design a current source, then can calibrate any kind of shunt. The downside, you need a lot of equipment for it. For example, if we wanted to to (mis) use our stuff to calibrate a 0.2mOhm shunt, with 500A, it was just the matter of connecting them in parallel. Because the AFE is built in. But if you want to do it at home, you need 11 mulitmeters for it, simultaneously measuring all the 2mOhm shunts and the 0.2mOhm .  (Or do them sequentially, but then I'm not sure, how valid the measurements are then. )

The interesting part is, that the errors are not increasing significantly, because the voltage measurements are like a magnitude more accurate then current measurement. So if you have 10 pieces of 0.03% accurate current sources, and you use these to calibrate a shunt, then the shunt is something like 0.035% worst case.

[doktor pyta]

@Tin, HighVoltage

It seems that DCCT with higher nominal current have better parameters (noise, linearity etc related to nominal current). It may be considered as higher SNR situation- there is more magnetic flux with certain fluxgate sensitivity and earth's magnetic field.
Below photo of my first test setup. I'm building 1A, 10A, 100A precision current source using 900, 90 and 9 turns wound 900A DCCT. First tests using 90 turns are very promissing at 10A current.

[HighVoltage]

Perhaps if you really interested in high accuracy high current stuff, you'd be better off getting LEM Ultrastab DCCT.

Thank you all for the good input.

The Ultrastab IT 600-S or the IT 405-S are too expensive and I have not found a good low cost source source.

But for now I bought one of the DANFYSIK Current Transducer ULTRASTAB 867.
May be this one will work for my purpose.

Type: 867-700I
Nominal current: 700A DC
Ratio: 1:1750
Output +/- 400 mA

I am surprised how difficult it is to make some precision current measurements, even at only 1A

[Robert763]

I'm late to this thread but acouple of people have asked "why 60mV" value for shunts. This is not the only value used, 75mV seems quite common in the UK, 50, 100 and 150mV are also seen. It's related more to the impedance of the matching meter than anything else. Most panel meters for use with shunts have a 1mA movement. A coil resistance of 60 Ohm is not unusual for a 1mA movement, and a lower value will allow for a series calibration resistor.

[HighVoltage]

I'm late to this thread but acouple of people have asked "why 60mV" value for shunts. This is not the only value used, 75mV seems quite common in the UK, 50, 100 and 150mV are also seen. It's related more to the impedance of the matching meter than anything else. Most panel meters for use with shunts have a 1mA movement. A coil resistance of 60 Ohm is not unusual for a 1mA movement, and a lower value will allow for a series calibration resistor.

Thanks for this explanation.

I bought a couple more shunt resistors, one with 75mV and one with 150mV
Soon I will start measuring their linearity, when the Danfysik 867-700I will arrive.

I am expecting a really good and calibrated shunt to be a reliable source for a good current measurement, even at 100A.
But one big question is: How linear are these shunt resistors in real life?

[HighVoltage]

Perhaps if you really interested in high accuracy high current stuff, you'd be better off getting LEM Ultrastab DCCT.
They pop up on ebay from time to time to relatively cheap prices. I've got two with best linearity, Ultrastab IT-600S.
It is 1:1500 ratio CT, and with linearity spec <1ppm, offset error <20ppm and tempco <0.2ppm/K, so high chance that your measurements accuracy with these DCCTs will be limited by your burden resistor stability. There are also lower (and higher) current Ultrastab's with similar specs. Buying older Danfysik DCCT (model 867 is same as IT-600S) might save few $$.

CERN using these beasts to calibrate and maintain LHC superconductor magnet currents to ppm levels stability. You can find related research paper, covering their transfer and calibration apparatus. 

Hello TiN,

I am making progress .... to be reported.

On your site you are explaining that you are using the Burden shunt: 2.5 ? Vishay PG foil 1ppm/C 9W, CSNG
Why are you using the 74$/piece shunt resistor?
Why not directly hooking up the Ultrastab output to the DMM current input?

Thanks!

[tszaboo]

The CSNG shunt is better than the one built into the DMM. The Kethley 2002 doesnt have an excellent shunt built in, we are talking about 400 ppm accuracy for 1 year on the 200mA range. Voltage is much better of course. 3458A has 20ppm typical.
Though, I dont know how much sense the low current measurements make. I mean, he is measuring 2.5uV on a 200mV range??

But one big question is: How linear are these shunt resistors in real life?

Depends on cooling. Self heating is the biggest enemy. Look at the de-rating curves of a shunt, tells you a story. If the shunt is rated for 10W and it has a curve, which is flat to 80 degrees, and from 80 to 125 degrees it falls down to zero. That would mean, that in normal conditions, 10W will heat the shunt 45 degrees. Multiply that with the tempco...
Also, some heatsink mounted shunts will tell you the thermal resistance, but that value is usually better than this.
And a 4 terminal shunt will have a value, like 1 mOhm. Which often time means, there is 1mOhm between the sense terminal's connection, but not the entire shunt. So the total resistance is higher, leading to more self heating...

[Micke]

If anyone is interested in a high current resistor with super-performance (TC=±1ppm/°C  !!!) at great price, ELFA Distrelec have 2 m ±0.1% 250W TC=±1ppm/°C (Isabellenhütte RUG-Z-R002-0.1-TK1) at the great price 79.18 Euro (normal price 589 Euro, yikes!) you can put 350A continuously through this thing (with proper cooling of course...)
https://www.distrelec.biz/en/power-resistor-mohm-250-isabellenhuette-rug-r002-tk1/p/16057581?q=rug-z&page=1&origPos=1&origPageSize=25&simi=96.57

[HighVoltage]

Thank you Micke
I have placed an order, this will be a good comparison of linearity between a few shunt resistors and the LEM

Thank you NANDBlog for the explanation.
I will try a few current measurements with different DMMs, once I am setup for the next step.
May be I will get such a precision shunt later.

[TiN]

NANDBlog is right, as with high power systems PCR of the used resistors has big impact on accuracy. That is why I also chosen CSNG which is rated 9W to support maximum output current of the LEM's 400mA output.

That shunt is very interesting price, I wonder if this distributor can ship internationally to private customers..

[HighVoltage]

That shunt is very interesting price, I wonder if this distributor can ship internationally to private customers..

If you can not get it shipped, send me a PM and I will arrange it for you.

[2N3055]

I just ordered one RUG-Z-R002-0.1-TK1....... I needed to measure 50-100A currents many times and had to borrow shunt.. this time I will have mine.
I'll put it in case and stick it on CPU cooler...  Probably also a PT1000 for a good measure... Have to figure out mechanical package...
Anyways, thanks a lot Micke !!

[Micke]

Hi!
Always nice to help out fellow VoltNuts...    I also ordered one Isabellenhütte resistor, got it this Thursday, very nicely packed in a plastic box from factory (what else to expect...)
Heavy beast, weighs almost 300g...
It has a stamped number on the solid copper base, perhaps unique serial number?  or batch number at least...
Will put it in an aluminum box with four binding posts, high current binding posts will be 63A rated, Kelvin output will be a dual gold plated 19mm spaced version 
Getting like a 0,25W TC=1ppm/K resistor is expensive enough, but 250W... 

[HighVoltage]

Hi!
Always nice to help out fellow VoltNuts...    I also ordered one Isabellenhütte resistor, got it this Thursday, very nicely packed in a plastic box from factory (what else to expect...)
Heavy beast, weighs almost 300g...
It has a stamped number on the solid copper base, perhaps unique serial number?  or batch number at least...
Will put it in an aluminum box with four binding posts, high current binding posts will be 63A rated, Kelvin output will be a dual gold plated 19mm spaced version 
Getting like a 0,25W TC=1ppm/K resistor is expensive enough, but 250W... 

Thanks for the heads up on the 300 gramm weight, I did not see that.

I ordered two of these Isabellenhütte resistors, too good of a deal to not take advantage of.
And I also wanted to put one in a box with binding posts and separate Kelvin output to good quality binding posts.
What gold plated bindings posts will you be using?

But why are you planning to have a "dual" output?
May be something to consider?

[TiN]

HighVoltage, thanks for the offer.

I'm trying to get account first, since:

Dear customer,
registration for customers outside EU has to be done manually. Simply send us your company details to export@distrelec.com and we will send you login details to our web for further purchase online. Thank you.
Your Export Team.

.

I've sent them an email, since I'm no company.

[0.01C]

Perhaps if you really interested in high accuracy high current stuff, you'd be better off getting LEM Ultrastab DCCT.

Thank you all for the good input.

The Ultrastab IT 600-S or the IT 405-S are too expensive and I have not found a good low cost source source.

But for now I bought one of the DANFYSIK Current Transducer ULTRASTAB 867.
May be this one will work for my purpose.

Type: 867-700I
Nominal current: 700A DC
Ratio: 1:1750
Output +/- 400 mA

I am surprised how difficult it is to make some precision current measurements, even at only 1A

Ultrastab IT 600-S=Ultrastab 866 I guess ,you can find it in eBay @$89 , I bought a lot.

[Micke]

...
Will put it in an aluminum box with four binding posts, high current binding posts will be 63A rated, Kelvin output will be a dual gold plated 19mm spaced version 

Thanks for the heads up on the 300 gramm weight, I did not see that.

I ordered two of these Isabellenhütte resistors, too good of a deal to not take advantage of.
And I also wanted to put one in a box with binding posts and separate Kelvin output to good quality binding posts.
What gold plated bindings posts will you be using?

But why are you planning to have a "dual" output?
May be something to consider?

I wrote a little bit confusing I realize now... I will only have one output, a couple of years when ELFA had a big sale, I bought very nice gold plated double binding posts very cheap (think they are USA made) 

[2N3055]

Sense connectors are not a problem. I have trouble finding what 63A to 100A posts to use.. I only found 63A from Hirschmann , but they have more contact resistance than shunt (they will heat up), and brass ones from Schützinger (3 models 63A/100A/200A). They have 0,4 mOhm, much better, and 63A ar not that big...
Did anyone find other type?

[HighVoltage]

My first Isabellenhütte 2 mOhm resistor arrived already.
RUG-Z-R002-0.1-TK1

I think I will use some M6 studs, may be out of copper for a direct connection through an aluminum box,
instead of the usual binding posts.

Similar to this:

[tszaboo]

I would order a pair. Unfortunately, I know how much it costs to calibrate these. I remember the price was some 350-500 EUR each. And I'm sure everyone sees the 1K hour drift value, .2% for the .1% resistors. And they drift quite a lot, even if you don't use it near full load. I think our resistors never went above 20W load (8mOhm 50A) and the drift was still significant. But of course these were used a lot, hours every day.

That being said, they are excellent resistors. Make sure to cool it well. Bolting down to an aluminum extruded heatsink works quite well.

[HighVoltage]

I would order a pair. Unfortunately, I know how much it costs to calibrate these. I remember the price was some 350-500 EUR each. And I'm sure everyone sees the 1K hour drift value, .2% for the .1% resistors. And they drift quite a lot, even if you don't use it near full load. I think our resistors never went above 20W load (8mOhm 50A) and the drift was still significant. But of course these were used a lot, hours every day.

That being said, they are excellent resistors. Make sure to cool it well. Bolting down to an aluminum extruded heatsink works quite well.

You are bringing up some good points.

Calibration.
Do you know how they are calibrated and what labs are capable of it to full current of 350A?
May be we can calibrate them ourselfs with the LEM / DANFYSIK sensors a few of bought?

Drift
Are you talking about a long therm drift?

Heaksink
I have been looking for a large heatsink already, that will fit with the resistor together, in to a nice aluminum box.
 
Linearity
I am really looking forward, testing these for linearity up to 100A, in comparison to the Danfysik / LEM. (may be 200A, if I hook my two agilent power supplies in parallel)

[tszaboo]

I would order a pair. Unfortunately, I know how much it costs to calibrate these. I remember the price was some 350-500 EUR each. And I'm sure everyone sees the 1K hour drift value, .2% for the .1% resistors. And they drift quite a lot, even if you don't use it near full load. I think our resistors never went above 20W load (8mOhm 50A) and the drift was still significant. But of course these were used a lot, hours every day.

That being said, they are excellent resistors. Make sure to cool it well. Bolting down to an aluminum extruded heatsink works quite well.

You are bringing up some good points.

Calibration.
Do you know how they are calibrated and what labs are capable of it to full current of 350A?
May be we can calibrate them ourselfs with the LEM / DANFYSIK sensors a few of bought?

Drift
Are you talking about a long therm drift?

Heaksink
I have been looking for a large heatsink already, that will fit with the resistor together, in to a nice aluminum box.
 
Linearity
I am really looking forward, testing these for linearity up to 100A, in comparison to the Danfysik / LEM. (may be 200A, if I hook my two agilent power supplies in parallel)

We were using a local Belgian cal lab, they were calibrated up to 50A, not to their full capability. I think I saw the invoice once, I'll try to recall what was the lab. They produced a very accurate reading, I think up to 7-8 digits. We sent our 3458A also there.

We were using these as a reference resistor, "gold standard", to calibrate the shunts built into the equipment we sold. The spec of our equipment was 0.03%, max current 50A. We were using the 2mOhm TK1 part, along with a bunch of other values for less current.

Drift is the long term drift. The datasheet specifies 0.2% drift @2000h, nominal load, 85 degrees. We were using it below maximum load (in fact below 2% load and almost room temperature) and it was causing issues with our spec. Like it drifted more than 0.03%, and we had to re-calibrate stuff. While the Arrhenius Equation would have suggested less drift, so careful, when trying to apply it. Also, these were used for 5+ years, so they had time to do initial drift. Might be, that some hysteresis or other physical phenomenon...

[HighVoltage]

We were using a local Belgian cal lab, they were calibrated up to 50A, not to their full capability. I think I saw the invoice once, I'll try to recall what was the lab. They produced a very accurate reading, I think up to 7-8 digits. We sent our 3458A also there.

We were using these as a reference resistor, "gold standard", to calibrate the shunts built into the equipment we sold. The spec of our equipment was 0.03%, max current 50A. We were using the 2mOhm TK1 part, along with a bunch of other values for less current.

Drift is the long term drift. The datasheet specifies 0.2% drift @2000h, nominal load, 85 degrees. We were using it below maximum load (in fact below 2% load and almost room temperature) and it was causing issues with our spec. Like it drifted more than 0.03%, and we had to re-calibrate stuff. While the Arrhenius Equation would have suggested less drift, so careful, when trying to apply it. Also, these were used for 5+ years, so they had time to do initial drift. Might be, that some hysteresis or other physical phenomenon...

Thanks for this great info.
Can you share, which lab in Belgium you are referring to, that makes 50A calibration?

[tszaboo]

I believe it was LHM calibrating our shunts.
https://www.lhm-instrumentation.eu/kalibratie.html

It took me some time to remember.

[HighVoltage]

I made a quick test with my first Isabellenhütte 2 mOhm resistor RUG-Z-R002-0.1-TK1
At least on my Keithley 2460 it is spot on.

Next I will do the same test with a switched polarity

[HighVoltage]

And are here the values with switched polarity to negative

Here are the calculated averages of the three voltmeters

3458A:      ( 2.00002 + 1.99875 ) / 2 = 1.999385
34470A:    ( 1.99945 + 2.00035 ) / 2 = 1.999900
7510:        ( 1.99632 + 2.00035 ) / 2 = 2.000004

[VintageNut]

And are here the values with switched polarity to negative

Here are the calculated averages of the three voltmeters

3458A:      ( 2.00002 + 1.99875 ) / 2 = 1.999385
34470A:    ( 1.99945 + 2.00035 ) / 2 = 1.999900
7510:        ( 1.99632 + 2.00035 ) / 2 = 2.000004

For the 7510 on the 100mV range, the uncertainty is 9ppm of range and 25 ppm of reading. The uncertainty of your measurement is 0.9uV for range. The reading uncertainty is about 50nV for a total uncertainty of 950nV.

For your 2.000004 milliohm calculation, drop the last three digits for being completely uncertain.

[chuckb]

The Isabellenhütte RUG series of resistors is physically very large. The factory recommends bolting it to a heat sink for high currents. The factory data sheet shows a 2000 hour test at 110 deg C with no detectable drift. I'm sure the 110 deg C test was in an oven with current only applied while testing it. The resistor is rated for 140 deg C.

NANDblog shared that they saw more drift than they were expecting with the RUG series at low power while bolted to a heat sink.

Could the large package be distorting when bolted to a heat sink and acting like a strain gauge? The package stress could come from the initial connection to the heat sink or a differential metal expansion as it's dissipating power.

[tszaboo]

The Isabellenhütte RUG series of resistors is physically very large. The factory recommends bolting it to a heat sink for high currents. The factory data sheet shows a 2000 hour test at 110 deg C with no detectable drift. I'm sure the 110 deg C test was in an oven with current only applied while testing it. The resistor is rated for 140 deg C.

NANDblog shared that they saw more drift than they were expecting with the RUG series at low power while bolted to a heat sink.

Could the large package be distorting when bolted to a heat sink and acting like a strain gauge? The package stress could come from the initial connection to the heat sink or a differential metal expansion as it's dissipating power.

The shunt was used for a long time. 5 years at least.
The datasheet 1st page specifies 0.2% The drift we read on the traceable calibration paper was less than this. It behaved according to the datasheet.
The drift was more than we expected.
The shunt was mounted on an extruded aluminium heatsink (about 8cm thick in total with fins). There was careful care take, so connecting the shunt assembly did not stress the current terminals of the shunt. Not directly anyway. The shunt had a big semiconductor fuse on it, and overload is not likely.

Again, the drift was within spec. I have high trust and confidence in the manufacturer.

[HighVoltage]

And are here the values with switched polarity to negative

Here are the calculated averages of the three voltmeters

3458A:      ( 2.00002 + 1.99875 ) / 2 = 1.999385
34470A:    ( 1.99945 + 2.00035 ) / 2 = 1.999900
7510:        ( 1.99632 + 2.00035 ) / 2 = 2.000004

For the 7510 on the 100mV range, the uncertainty is 9ppm of range and 25 ppm of reading. The uncertainty of your measurement is 0.9uV for range. The reading uncertainty is about 50nV for a total uncertainty of 950nV.

For your 2.000004 milliohm calculation, drop the last three digits for being completely uncertain.

Thank you VintageNut for pointing this out again about the uncertainty of the measurement.
I am aware of this but just wanted to show the raw data.

Somehow I find it impressive, that we have today instruments that can give us such a resolution of precision and the a 2 mOhm resistor shows up at 0.002000 Ohm.

Although, the absolute accuracy is a different subject, it still gives me the ability to see a trend within the last digits.
 

[JohnG]

Perhaps if you really interested in high accuracy high current stuff, you'd be better off getting LEM Ultrastab DCCT.
They pop up on ebay from time to time to relatively cheap prices. I've got two with best linearity, Ultrastab IT-600S.
It is 1:1500 ratio CT, and with linearity spec <1ppm, offset error <20ppm and tempco <0.2ppm/K, so high chance that your measurements accuracy with these DCCTs will be limited by your burden resistor stability. There are also lower (and higher) current Ultrastab's with similar specs. Buying older Danfysik DCCT (model 867 is same as IT-600S) might save few $$.

I did a brief test before, with 5700A and 5205A as current source.

CERN using these beasts to calibrate and maintain LHC superconductor magnet currents to ppm levels stability. You can find related research paper, covering their transfer and calibration apparatus. 

Just saw this and want to push for this idea as well. I have used this to calibrate shunts of different current ratings to use for DC-DC converter measurements. I used the Danfysik 866 version, with the following burden resistor: https://www.digikey.com/product-detail/en/vishay-foil-resistors-division-of-vishay-precision-group/Y169010R0000T9L/Y1690-10A-ND/3047016 .

I also use the standard shunts, and have found that by using heat sinks, I can minimize shunt temperature variation. I use a layer of self-adhesive thermal tape between the element and the heat sink, so that there very little pressure on the Manganin element. With fans on the heat sinks, I have found that I can keep the temp rise to < 10-15C at over half rated current.

It is important to make sure the heat sink covers the entire element, because Magnanin is a poor thermal conductor. If you look at one with a thermal camera, you will see temperature variation

I have not characterized this system thoroughly over the long term, sorry.

John

[JohnG]

Or, you could just use a Kelvin Bridge 

John

[TiN]

Got my greedy candies from DE Santa. Thank you, HighVoltage for your help 

These things are MONSTERs 



Quick and dirty check and comparison photo to other related stuff:

[HighVoltage]

Got my greedy candies from DE Santa. Thank you, HighVoltage for your help 

These things are MONSTERs 

That was fast.
Great that they arrived safely and in good condition.

Yes, they are much larger than I had expected them to be.

My Danfysik / LEM sensors also arrived and I am looking forward to some detailed testing in the next few weeks.

[TiN]

Well, I'll get to properly testing these only end of January, as I'll be OOC from this weekend for holidays.

Would be interesting to attempt PCR measurement of these resistors. I'd think first is to measure tempco using low current and sensitive nanovolt detector like Keighley 1801 (EM A10). Then it will be maybe possible to determine thermal part of PCR component. My best source so far is Time Electronics 9823 (10A range).

[VintageNut]

I would like to see the LEM sensor design working as DIY here. I would join in with one if they work.

[Pipelie]

There is a project call feshbach coils you might want to see.
https://www.physics.utoronto.ca/~astummer/pub/mirror/Projects/Archives/2007%20Feshbach%20Coils/Feshbach%20Coils.html

[HighVoltage]

There is a project call feshbach coils you might want to see.

Some nice information, thanks!

Also on that site, one can find a really nice article by Jim Williams from September 2011
Design a 100A load


https://www.physics.utoronto.ca/~astummer/pub/mirror/Projects/Archives/2011%20Fesh-MOT%20Coils%20redux/Docs/EDN%202011-9-22%20%27Design%20a%20100A%20Load%27%20article.pdf

[Pipelie]

yes, I read it before ,even printed out.
app note 133. I'll try to build it some time next year.
http://cds.linear.com/docs/en/application-note/an133f.pdf

[quarks]

because I found an error in HIOKI RM3545 graph I updated my comparison chart from Reply #101
https://www.eevblog.com/forum/metrology/low-ohm-precision-resistor-standard-and-testing/msg1215989/#msg1215989



edit: member onemilimeter found an error in 34470A, chart is now corrected

[e61_phil]

It doesn't make any sense, but it is impressive The 3458A justs says 0.000 06 Ohms.

The shunt is specified with 1000A 60mV class 0.05. But neither the DMM nor the setup is suitable for such a measurement. But it is pretty close.

[Vgkid]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

[tszaboo]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

Just got a cal certificate for a precision resistor (50 Ohm) and they used that Fluke for calibrating it.

[Zermalmer]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

Just got a cal certificate for a precision resistor (50 Ohm) and they used that Fluke for calibrating it.

They used only the multimeter?

[e61_phil]

They used only the multimeter?

8.5ppm isn't that bad for 50Ohms. Is it?

[tszaboo]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

Just got a cal certificate for a precision resistor (50 Ohm) and they used that Fluke for calibrating it.

They used only the multimeter?

Yes. I believe it was a few PPM uncertainty (8-ish), I needed some 50-100ppm. Just something better than what's inside the 34465A at work.
Strange thing, that technically, I could have used the 34465A to calibrate my resistor, and then use that as a transfer standard. But it is better to spend that money, than to re-call a production run.

[Zermalmer]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

Just got a cal certificate for a precision resistor (50 Ohm) and they used that Fluke for calibrating it.

They used only the multimeter?

Yes. I believe it was a few PPM uncertainty (8-ish), I needed some 50-100ppm. Just something better than what's inside the 34465A at work.
Strange thing, that technically, I could have used the 34465A to calibrate my resistor, and then use that as a transfer standard. But it is better to spend that money, than to re-call a production run.

I know that this 'baby' is really nice and the 8.xx ppm is a also a very nice number.
(I didn't want to criticize the device or the lab)

And naturally it is good invested money instead of throwing away one or more  production runs.

I commonly run it more the "old way" with the DC current resitance, but this is more historical based.

[TiN]

I took some pretty photo while ago too.
Short resistance of Fluke 5700A in 4-wire mode, using 5440B-700x cable.

[tszaboo]

That is impressive. One thing I have noticed is that the 8508a is in a lot of labs that specialize in resistance measurements.

Just got a cal certificate for a precision resistor (50 Ohm) and they used that Fluke for calibrating it.

They used only the multimeter?

Yes. I believe it was a few PPM uncertainty (8-ish), I needed some 50-100ppm. Just something better than what's inside the 34465A at work.
Strange thing, that technically, I could have used the 34465A to calibrate my resistor, and then use that as a transfer standard. But it is better to spend that money, than to re-call a production run.

I know that this 'baby' is really nice and the 8.xx ppm is a also a very nice number.
(I didn't want to criticize the device or the lab)

And naturally it is good invested money instead of throwing away one or more  production runs.

I commonly run it more the "old way" with the DC current resitance, but this is more historical based.

OK, I looked up the cal certificate, they wrote 12 PPM uncertainty on it forr a 50 Ohm resistor, with EA 4/02
http://www.european-accreditation.org/publication/ea-4-02-m-rev01--september-2013

[e61_phil]

I commonly run it more the "old way" with the DC current resitance, but this is more historical based.

Can you explain that a bit further?

I took some pretty photo while ago too.
Short resistance of Fluke 5700A in 4-wire mode, using 5440B-700x cable.

Nice    The 8508A showed 35µ Ohms on my Fluke 5450A.

OK, I looked up the cal certificate, they wrote 12 PPM uncertainty on it forr a 50 Ohm resistor, with EA 4/02
http://www.european-accreditation.org/publication/ea-4-02-m-rev01--september-2013

Das anyone know how they calculated the 12ppm? Is it the 99% Tcal +/-5K spec? That would give 11.2ppm

[tszaboo]

I commonly run it more the "old way" with the DC current resitance, but this is more historical based.

Can you explain that a bit further?

I took some pretty photo while ago too.
Short resistance of Fluke 5700A in 4-wire mode, using 5440B-700x cable.

Nice    The 8508A showed 35µ Ohms on my Fluke 5450A.

OK, I looked up the cal certificate, they wrote 12 PPM uncertainty on it forr a 50 Ohm resistor, with EA 4/02
http://www.european-accreditation.org/publication/ea-4-02-m-rev01--september-2013

Das anyone know how they calculated the 12ppm? Is it the 99% Tcal +/-5K spec? That would give 11.2ppm

"The measurements have been calibrated using measurement standards traceable towards international and national standards, and meets the requirements of ISO/IEC 17025:2005."

"Based on a standard uncertainty multiplied by a coverage factor of K=2, which provides a confidence level of
approximately 95%. The standard uncertainty has been determined in accordance with EA-4/02."

[Zermalmer]

I commonly run it more the "old way" with the DC current resitance, but this is more historical based.

Can you explain that a bit further?

Hi Phil,
will try it.

You supply the resistor with a constant current and measure the voltage drop over the resistor and simply calculate the resistance.
R=U/I

And according to the fact that a resistance in DC usage isn't equal in both directions you have to do the same with switched polarity.
(If you use the resistor always in the same polarity that isn't really needed, but who can guarantee that?)

The MU of this configuration is for sure higher than the accuracy of the 8508A, but the results of this method give normally very trustfully results.


ADD:plesa already mentioned this method in the 2nd post of the thread ... looks like I have to do some reading in the thread 

[e61_phil]

Hi Zermalmer,

thanks for your explaination.

Normally it is much harder to measure current very accurate than comparing resistors.

I also doubt that a resitor behave in one direction other than in the other direction at DC. The reason why you change the direction of the current is to eleminate offset voltages (thermal EMF). Most of these voltages stay constant no matter in which direction the current flows. Therefore, they cancel each other out if you measure in both directions.


@NANDBlog: I wonder why they stated 12ppm. The 95% uncertainty 1 year specification of the 8508A for 50 Ohm is better than that.

[Zermalmer]

Normally it is much harder to measure current very accurate than comparing resistors.

That's the reason why you use accurate current source and measure voltage 

@NANDBlog: I wonder why they stated 12ppm. The 95% uncertainty 1 year specification of the 8508A for 50 Ohm is better than that.

You need a closer look for what the numbers are and what they mean.
The specified uncertainity in the manual is the accuracy of the device itself under referenced conditions.
The number from the calibration lab is the measurement uncertainty.

For the measurment uncertainty you have to calculate based on all known accuracies,  conditions and specifications (and sometimes to assume from experiences).

For that the „Guide to the Expression of Uncertainty in Measurement“ ... in short form GUM is the key.

If you want to know more about GUM
german wikipedia : https://de.wikipedia.org/wiki/GUM_(Norm)
english wikipedia : https://en.wikipedia.org/wiki/Measurement_uncertainty

You can find a lot of info about GUM in the web and direct at the source... https://www.bipm.org/en/publications/guides/
(if link doesn't work correct, go to main page > publications > Guides in Metrology )

But attention... lot of math is involved  and very very dry 

[e61_phil]

Normally it is much harder to measure current very accurate than comparing resistors.

That's the reason why you use accurate current source and measure voltage 

It doesn't matter much if you create a current or measure it very accurate. The source also needs a feedback which ist in principale the same as a measurement. I don't know any current source outside a NMI which gives you such an accurate absolute current to measure resistors better than the 8508A does. Even the best Fluke 5720A 24h current specification is 25+x ppm.

You need a closer look for what the numbers are and what they mean.
The specified uncertainity in the manual is the accuracy of the device itself under referenced conditions.
The number from the calibration lab is the measurement uncertainty.

For the measurment uncertainty you have to calculate based on all known accuracies,  conditions and specifications (and sometimes to assume from experiences).

And what is there to add? The Fluke 8508A specification already includes the instrument drift, TC, noise and so on?

[CalMachine]

Normally it is much harder to measure current very accurate than comparing resistors.

That's the reason why you use accurate current source and measure voltage 

It doesn't matter much if you create a current or measure it very accurate. The source also needs a feedback which ist in principale the same as a measurement. I don't know any current source outside a NMI which gives you such an accurate absolute current to measure resistors better than the 8508A does. Even the best Fluke 5720A 24h current specification is 25+x ppm.

The idea is to use a stable current source and 'characterize' your nominal current you plan on using by measuring the voltage drop across a known resistance.

Put your known resistance in series with your DUT resistance, connected to your current source (IE: 5700/5720).  Measure the voltage drop across your known resistance and DUT resistance.  Swap polarity of your sense connections and measure again.  Average your +/- results to null out any thermal emf errors.  Compare your results of your known resistance to your DUT resistance.

This procedure is further documented in 5700/5720A performance verification.

[e61_phil]

Normally it is much harder to measure current very accurate than comparing resistors.

That's the reason why you use accurate current source and measure voltage 

It doesn't matter much if you create a current or measure it very accurate. The source also needs a feedback which ist in principale the same as a measurement. I don't know any current source outside a NMI which gives you such an accurate absolute current to measure resistors better than the 8508A does. Even the best Fluke 5720A 24h current specification is 25+x ppm.

The idea is to use a stable current source and 'characterize' your nominal current you plan on using by measuring the voltage drop across a known resistance.

Put your known resistance in series with your DUT resistance, connected to your current source (IE: 5700/5720).  Measure the voltage drop across your known resistance and DUT resistance.  Swap polarity of your sense connections and measure again.  Average your +/- results to null out any thermal emf errors.  Compare your results of your known resistance to your DUT resistance.

This procedure is further documented in 5700/5720A performance verification.

Yes, but that is a direct resistor comparison like the 8508A does it.
The absolute values of voltage and current doesn't matter. The current has to be stable only and the voltage measurement have to be linear. No need for correct gain.

I think that wasn't meant by Zermalmer. Or was that the way you meant by R=U/I?

[Zermalmer]

Normally it is much harder to measure current very accurate than comparing resistors.

That's the reason why you use accurate current source and measure voltage 

It doesn't matter much if you create a current or measure it very accurate. The source also needs a feedback which ist in principale the same as a measurement. I don't know any current source outside a NMI which gives you such an accurate absolute current to measure resistors better than the 8508A does. Even the best Fluke 5720A 24h current specification is 25+x ppm.

The idea is to use a stable current source and 'characterize' your nominal current you plan on using by measuring the voltage drop across a known resistance.

Put your known resistance in series with your DUT resistance, connected to your current source (IE: 5700/5720).  Measure the voltage drop across your known resistance and DUT resistance.  Swap polarity of your sense connections and measure again.  Average your +/- results to null out any thermal emf errors.  Compare your results of your known resistance to your DUT resistance.

This procedure is further documented in 5700/5720A performance verification.

Yes, but that is a direct resistor comparison like the 8508A does it.
The absolute values of voltage and current doesn't matter. The current has to be stable only and the voltage measurement have to be linear. No need for correct gain.

I think that wasn't meant by Zermalmer. Or was that the way you meant by R=U/I?

Ah sorry... yes...totally forgot the reference resistance

[e61_phil]

Ahh, ok. That will work of course. But that is exactly what the 8508A does in True Ohms mode with the build in reference resistors.

[Zermalmer]

The important thing to understand is, that the specification of a device is only part of the measurement uncertainity.
In this case of such high quality devices the manufacturer give us a lot more informations, which makes it more easy to decide and calculate the measurement uncertainity.

[e61_phil]

Do you have an example what is to add to the manufacturer specification?

[Zermalmer]

Hi Phil,
that depends on all involved components and if their uncertainity affects the whole calculation or if you can neglect the value.

One simple thing is for example the aircondition of the lab.
Commonly this one will be within the needed range, but this measuring of this also have an uncertainity.
If this one is small enough you can neglect this value... if you have sometimes more worse conditions you have to include this into the calculation.

You can breakdown a measurment uncertainity to the used test leads... if it is necessary or usefull...this have to be decided.

As told before... measurement uncertainity is build on math, experiences and theories and is a very dry stuff.

greetings
Andreas

[e61_phil]

If Fluke stated a specification for 1 year with a temperature range of 23°C +/- 5°C I would expect this specification cover the described enviroment. They also say what the maximum lead resistance in your 4-Wire measurement should be. Therefore, I would expect even that is included.

[CalMachine]

If Fluke stated a specification for 1 year with a temperature range of 23°C +/- 5°C I would expect this specification cover the described enviroment. They also say what the maximum lead resistance in your 4-Wire measurement should be. Therefore, I would expect even that is included.

You still have to take into account all of the uncertainties that are inherent to the DUT and measurement system, as a whole.  TCR of resistor, uncertainty of ambient temp, stability of ambient, uncertainty of temp measurement of resistor

[e61_phil]

Thanks CalMachine!
But how do you know these parameters?

Lets say I send you a DIY Box (a 10k reference for example like many people here have build). How do you know the TC and so on?

The certificate can only include short term influences anyhow. Therefore, I thought the uncertanty only describes the DUT as found during the measurement and no DUT properties. Another story might be a noisy measurement due to noise of the DUT. But this will affect the measurement itself.

[Zermalmer]

That's the reason why in the certificates always is stated that the result can't make states about the long term stability.

Manufacturers invest a lot of time and money in alaysis of their products and work with aging, stress test aso.

Otherwise there where no possibility of such detailed data which for example is given with high end devices.

For your DIY resistance box you have basically the specifications of the manufacturer of the resistor.
All additional parts, like plugs, wire, case aso. you have to calculate for yourself.
For the case you can't test it for your own you send it for example over a time of 5 or 10 years to a calibration lab.
After this time you can make, based on the uncertainity of the lab and the results, a long term probability calculation... but only based on this time
... but be aware... even good and well handled reference resistors can change their drift direction 

That's why a lot of people at the german PTB have a whole lot of stuff to explore.
And have a look to the new VIM... after a long period the definition of some SI elemantary components are defined in another way, based on new experiences over the last years.

Actually I have no english link to latest changes...  but sure there is some in the internet
here is short description from PTB in german: https://www.ptb.de/cms/fileadmin/internet/presse_aktuelles/broschueren/intern_einheitensystem/Das_neue_Internationale_Einheitensystem_V2.pdf
Some of the used constants are known here... some for light for example... and for Ampere and others.
But does this change everything for us? Commonly not, except you work in a very high level of research or production level... and if you work their you already know about the changes for sure 

[TiN]

I wonder where these come from..





 

[Vgkid]

I own some of those(30 now). I should have bought the rest  
Looking forward to your test.

[babysitter]

These are so cool you should have some to... well, have some.

[doktor pyta]

Can anyone tell me the value of Rth (bulk foil- case) for above mentioned Vishay resistors ?
In the datasheet of VHP-3, VHP-4 and VPR247 I couldn't find such crucial info.
Some of my older resistors from Sfernice times had Rth 7K/W.

[HighVoltage]

I wonder where these come from..

TiN, you really do find great deals!

I bought a few Vishay VHP-4 100 Ohm resistors, but have not taken detailed measurements.

[Vgkid]

Gravedigging this with a current shunt
JRL CS-1R-100 . Yes it is bolted to a very heavy fan cooled AL heatsink
Last measured value 5.00017mOhm
rough measured value using a 34420A calibrated most likely after it was made. only has a cal count in the low 30's
4.9994354m avg with a std dev of 71,665nOhm measured across 303 readings.

[HighVoltage]

That resistor has some serious wattage behind it with such heat sink.
Any idea what the power rating is?

I have never seen such resistor.
 

[Vgkid]

Taking a guess since the cs-100-1 is rated for 100w, maybe 50, but could be more. I emailed ohm labs to see if they have any old docs.

[HighVoltage]

This is my newest edition to the precision low value resistor current shunts.
After years of searching, finally one showed up on the used market.

This one is a Burster, Made in Germany

- Type: 1281-Q
- Built: 1990
- R: 100 mOhm
- Tol. : 0.02 %
- Tk : <10ppm/K
- Pmax: 10 Watt

The resistor seems to be spot on, when measured with the Keithley 2450 SMU.
Here are some pictures and the Datasheet (German from 1990)

[HighVoltage]

And here are the measurements at 5A with the Keithley 2460

[gamalot]

Congrats on your new toy! 

I don't have any Keithley sourcemeter and I don't really understand why it shows 0.100015 ohms instead of 100.015x milliohms.

---

Please ignore my stupid question, just read the 2450 datasheet ...... 

[coppice]

This is my newest edition to the precision low value resistor current shunts.
After years of searching, finally one showed up on the used market.

This one is a Burster, Made in Germany

- Type: 1281-Q
- Built: 1990
- R: 100 mOhm
- Tol. : 0.02 %
- Tk : <10ppm/K
- Pmax: 10 Watt

The resistor seems to be spot on, when measured with the Keithley 2450 SMU.
Here are some pictures and the Datasheet (German from 1990)

I'm surprised that such an elaborately constructed resistor doesn't use material with a better temperature coefficient.

Very few low value resistors, especially ones in the micro-ohm range, are all that precise. The market is usually for an approximate resistor with a first class temperature coefficient, and the error is just calibrated away at the system level.

[HighVoltage]

I don't have any Keithley sourcemeter and I don't really understand why it shows 0.100015 ohms instead of 100.015x milliohms.

Well, this is the default setting of the 2450.
You could apply an mx+b function to the measurement to show mOhm instead of Ohm

[HighVoltage]

I'm surprised that such an elaborately constructed resistor doesn't use material with a better temperature coefficient.

This Burster 1281 is made of standard MANGANIN (CuMn12N) and has a T.C. 10ppm/K.

There was an older type Burster 1284 (Not in production anymore), that was made of ZERANIN (CuMn7Sn) and has a T.C. <3ppm/K. Allegedly, ZERANIN is no longer available from the German manufacturer, Isabellenhuette.

See datasheets enclosed for both materials.

[coppice]

I'm surprised that such an elaborately constructed resistor doesn't use material with a better temperature coefficient.

This Burster 1281 is made of standard MANGANIN (CuMn12N) and has a T.C. 10ppm/K.

There was an older type Burster 1284 (Not in production anymore), that was made of ZERANIN (CuMn7Sn) and has a T.C. <3ppm/K. Allegedly, ZERANIN is no longer available from the German manufacturer, Isabellenhuette.

See datasheets enclosed for both materials.

People tend to refer to any low coefficient material as manganin, so its often hard to tell exactly what things are made of. However, there are a lot of shunts with a demonstrable coefficient of around 4ppm/K across a number of samples. These materials don't have a specific temperature coefficient, as there are tolerances on the mix of metals in the alloy. You need to check if 10ppm/K is typical or maximum.

[Conrad Hoffman]

Interesting. Most of the better-than-manganin wires I've seen are not solderable and have to be spot welded. Too bad they don't make that anymore.

[HighVoltage]

Interesting. Most of the better-than-manganin wires I've seen are not solderable and have to be spot welded. Too bad they don't make that anymore.

The ZERANIN datasheet claims that the material can be soldered.

Based on their datasheet:
The Burster models from 25 uOhm to 100 mOhm use a MANGANIN sheet metal
The Burster models from 200 mOhm to 100 Ohm use a ZERANIN wire

I do not know, how they are internally connecting the sheet metal or the wire but I would assume that it is mounted directly to the 4mm binding posts without welding or soldering.

Maybe one day I get a bad one and can take it apart.

[coppice]

Interesting. Most of the better-than-manganin wires I've seen are not solderable and have to be spot welded. Too bad they don't make that anymore.

Why is solderability a big issue? most shunts I've encountered are welded to copper supports. Either butt welded between 2 chunks of copper, or flat welded to the surface of copper. Why would anyone spot weld these things? That doesn't sound like a route to stable results. Solderability may be an issue on parts of the shunt where the sensing pair of a 4 wire attachment occurs. Most seem to solder to the shunt material, but there might well be plating there, to create solder compatibility.

[Conrad Hoffman]

For shunts, sure, but with smaller wire spot welding is easy and practical. I've never seen the wire I'm thinking of (NiCr "800" series like Kanthal Nikrothal LX) in a format suited to shunts- ribbon or flat stock.

[HighVoltage]

Newest addition to the Burster family of low ohm resistors, the BURSTER 1284

- BURSTER Type: 1284-0,1
- Built: 1996
- R: 100 mOhm
- Tol. : 0.01 %
- Tk : <3ppm/K
- Pmax: 45 Watt
- Imax: 200A
- Material: ZERANIN


This should be a true metrology grade low ohm resistor.
It seems like the ZERANIN is a sheet metal, squeezed between the two heat sinks.
I will use it for some high current measurements for up to 100A.

This thing is huge in comparison to the small Burster 1240 or even the 1281

[HighVoltage]

And here are some first measurements at 1A

[Smokey]

You guys were talking about measuring the resistors at max current operating temperature earlier. 
If you had an accurate thermal oven and let the thing settle between readings, would it be valid to characterize the tempco over the full temperature range but at one current much lower currents than maximum?  Or is there something special about actually pushing the real current through the shunt?

[TizianoHV]

I would expect a different temperature distribution. A thermal chamber will heat up uniformly. The nominal current might create temperature gradients. And those will for sure influence thermal EMF. For most of the cases it will give you a reasonable value.

One trick I use to evaluate thermal EMF in function of the current is to use AC test currents to heat up the shunt and a DC voltmeter to measure EMF (the voltmeter will filter all the AC component and measure only the offset). That's good when you don't have a DC high current supply since you can use any cheap transformer rewound for high currents (I used a free microwave oven transformer, good up to 400A).
Then I measure the resistance at lower DC currents and temperatures.

The photo shows some tests of a DIY low value resistor (I made it to measure pulsed currents, thats why I'm using a BNC). In the first plot there are temperature and EMF in function of AC current. In the second plot there's DC resistance (at just one temperature).


Next step would be to use both DC and AC current to heat up the shunt and measure its resistance at the same time, but I didn't had time for that, jet.

[Kleinstein]

The thermal EMF part depends on the diretion of the current, as it adds a voltage that is little (low thermal EMF materials also have a low Peltier coefficient, as they are linked) effected by the direction of the current.
The difference from heating from the outside and heat from the resistor itself is more from mechanical stress caused by temperature gradients.

[mendip_discovery]

DC EMF can be dealt with by reversing the connections. But we have spoken about that before.

I have a few shunt resistors and I wonder if they could be improved by adding some heatsinks, using thermal pads to insulate the electrical part from the heatsink but keep the thermal connection. It's also a handy way to also apply a temperature sensor.