Burden resistors for DCCT transducers

[pansku]

Some background before I get to the topic. Many moons ago I made a thread on this section about a high current measurement setup. We ended up building one and already made interesting discoveries when comparing different suppliers in less than one month of usage

It is build from a few basic parts. There is a reference current transducer Danisense DL2000ID which is connected to their system interface unit. The output current is then turned into voltage within that box using their VOM1333-10 conversion module. The output voltage in 10V range is fed to a 3458A no1. The measured devices, be it open loop or closed loop hall effect (at the time of writing) are fed through a mux into 3458A no2 which is triggered synchronously with no1. When building the system I had no idea how noisy the environment would be, so I chose the 10V conversion module instead of 1V.

The weakest link in this is the voltage conversion. It is specced

7ppm linearity
13ppm initial offset with 1ppm/c tempco
30ppm initial maximum ratio error with 2ppm/c tempco

Since the company is throwing a lot of money around and I need a master's thesis project some time within next year or so, we were thinking if improving this could be a suitable project for that. I've already been looking into Z-foil resistors for the shunts of the DUT measurements of transducers with current output and the datasheets look very appealing.

1 ohm external burden resistor also could be calibrated in the company cal lab comparing it to a primary resistance standard. At maximum current we run, it would have around 1W of power dissipation and utilize the 1V measurement range of 3458A


When looking at the VPG catalog there are many nearly identical specced resistors for shunt usage. Do you have long term experience with those? Also any tips on heatsinking, mounting and actual build of high precision shunts? I know we could just buy a Fluke A40B and call it a day, but the point of this exercise is to be a learning experience. For characterizing the end product,  I also have access to weather cabinets.

Sorry if this discussion should've been kept to general resistor threads. I had a hard time finding DCCT related discussion

[ap]

Not sure about the A40, but it may drift quite a bid, since I think its not hermetic afaik and was essentially intended for AC measurements. If you want to get the best results, you should be using hermetic VHP resistors with low TC. Read the datasheets carefully, typ and max TC is important, aging as well. You need quite a few if you want to reduce drift at 1W total dissipation.
Also an option (have seen very little drift in my lab) would be the RUG-Z. Not cheap either.

[Kleinstein]

The Foil resistors are not really ideal when used at high power.  The construction with foil and "glue" to connect it to the substrate they can be subjet to drift and hystesis effect with changes in the power, even for the hermitic versions. With the low resistance range one could think about wire wound resistors of some kind.

There would also be the option to have more than 1 resistor in series to have a way to check for consistancy. With not that great overall performance there is no absolute need to use the full 1 V range. A current of 1 A is already more on the high side for an accurate measurement. More turns of a thinner wire and thus a smaller output current may be easier.
You may be lucky and find reference from CERN. The build the HPM7177 to work together with a DC current transformer - so they should have a really good one to call for that level of performce (though they may have access to superconductors - so they may play by different rules). 

[JohnG]

This may be of interest. No affliliation, etc., but I bought some of these for a power conversion efficiency measurement setup, where they are used as the burden resistors for some Ultrastab current transducers. I am not knowledgeable enough to comment on the actual performance in any rigorous sense, but they are much better than what we had been using.

https://www.ebay.com/itm/202568048586?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=20KLHoK2QDy&sssrc=2047675&ssuid=adCqyE2tTq6&widget_ver=artemis&media=COPY

Data sheet: http://www.goldeno.com/pictures/CH48T4HK.pdf

John

[Pip]

The Vishay VPR221Z with 4 leads (4 wire Kelvin connection) may be suitable. The only concern I have is that they are only rated at 8 W. This may however not be a problem.
My personal preference would be to use a 0.5 ohm or even better a 0.1 ohm shunt to reduce the power, however this does mean a lower output voltage. It's still within the range of the 3458A though.
Are you using this DCCT for both DC and AC? What is your range of measurement? 150 A to 1500 A?

[pansku]

The accuracy related measurements are DC with two power supplies. One has range of 0-40A and the other one 0-1200A. In the near future low current measurements are conducted only on open loop hall effect type devices so their accuracies are in the percents range. All though the big power supply does stay stable in the milliamps even at less than 10A current for minutes when measured at 100NPLC. I still need to do further evaluation with looped main conductor

There is also AC measurements from 50Hz to 1kHz up to 1200A RMS but those are just used to digitize the waveform. That is used for finding any abnormal performance

[mzzj]

The Vishay VPR221Z with 4 leads (4 wire Kelvin connection) may be suitable. The only concern I have is that they are only rated at 8 W. This may however not be a problem.
My personal preference would be to use a 0.5 ohm or even better a 0.1 ohm shunt to reduce the power, however this does mean a lower output voltage. It's still within the range of the 3458A though.
Are you using this DCCT for both DC and AC? What is your range of measurement? 150 A to 1500 A?

OP needs to plan for PhD thesis if using Vishay Z-foil's.   Even before the pandemic the delivery times and availability were sometimes insane.

"Thomas type" 1 ohm standard resistor would be the best bet for stability and accuracy but these are 1500...4000+ euros per piece. 
Heck, NIST webshop even has these for a mere 4600 USD https://shop.nist.gov/ccrz__ProductDetails?sku=51130C&cclcl=en_US
I'm not sure if the price includes the calibration with 0.04 ppm uncertainty.

[pansku]

Doesn't Vishay supposedly have these Precision centers around the world for quick prototyping parts? Paying a little extra for some resistors isn't exactly a concern compared to Keysight curve tracers or Lecroy 8-channel 12-bit scopes (yes, in plural)

I'm just thinking whether or not choosing and verifying a COTS and comparing it to the starting point will keep me busy for a few months. Just the hardware side of complete tester would've made for a great thesis had I not finished it before the university allowes starting the master's thesis process