100uR 1+W current sense resistor

[Simon]

I'm looking to do some high amperage current measurements (up 100A) with a shunt resistor, any ideas where I can get one. The ones i find don't have great specs often 5% and well over 100ppm. I can easily use a higher wattage value to try and keep in spec.

Note: i have tried a current clamp and I'm not happy so don't tell me to use one.

[German_EE]

I faced this problem a few months ago. Pull apart an old clock and extract the main spring. Three 5cm lengths of this in parallel clamped at both ends provided about 1,2 mV per amp, you then add a series resistor in your meter lead for calibration to bring this down to a 1 mV per amp ratio.

[Simon]

I'm considering using wire, but I can't calibrate it 5AWG cable has 0.001028R per metre so just under 100mm of that will do it and I can even do a twisted pair to get rid of inductive effects, or maybe a larger guage. But I don't know how accurate AWG wire cross sections are and it may depend on wire coatings etc and of course more heat more resistance as this is at 25C,

[krish2487]

Try a proper current shunt.
There are shunts available from ohmite.
0.5% tolerance and 30ppm tempco.
Look at the S and SH Series.

[rs20]

You shouldn't care about the tolerance at all, you only care about the tempco. You can calibrate any tolerance in the error out trivially (either in software if this is going via software, or with the voltage divider if you absolutely must have the correct analog voltage in-circuit, which I very much doubt you actually need). Dave regularly mentions this in multimeter reviews.

[Simon]

The problem is I have nothing to calibrate with, I have nothing to measure 100A with I can add up both fluctuating current readings from 2 power supplies that I hardly trust.

I need ideally to buy something that come warranted to work as mentioned above

[TMM]

The problem is I have nothing to calibrate with, I have nothing to measure 100A with

You don't need to calibrate it at 100A if you can ignore the self heating effects at 100A.

edit: might have been unclear - you can calibrate the resistance of the shunt at a lower current.

[Simon]

Yes i could flow 10A and use my meter as a reference

[Pillager]

Some of these sound quite promising:

http://www.isabellenhuette.de/en/precision-and-power-resistors/products/

And Mouser has a nice one:

http://www.mouser.at/ProductDetail/Vishay/WBPK600L0A00010001/?qs=kcL3wPdn6%2f2VZ5DGC96cKQ%3d%3d

[Simon]

hm not for 87 quid, I was hoping some something chuncky rather than surface mount although I suppose i can solder it to copper bar with a hole in for screw terminals.

[Zero999]

Why are you even considering a resistor?

A Hall effect transducer would probably be much more suitable.

[Simon]

hm thats true. I think i have a couple of those from Allegro

[rs20]

Why are you even considering a resistor?

A Hall effect transducer would probably be much more suitable.

Beware of Hall effect transducers; one that I got is very good at detecting its orientation w.r.t. the Earth's magnetic field (and yes, this is a current-measurement hall effect chip with pins for the current [datasheet]). I don't recommend that part, I was really confused while debugging it until I realised that tilting the device caused massive offsets in the measurement. You *might* get away with using it for AC though, but it's dodgy in my books -- I mean, if it picks up the Earth's magnetic field so well, surely some nearby 100A AC cables will gives if some pretty decent competing fields?

My v2 for the same circuit is ditching the hall effect transducer for a uCurrent-style arrangement instead -- ultra small resistance current shunt, with gain after that (I never needed the isolation in the hall effect transducer in the first place).

[HighVoltage]

May I ask, why you don't want to use a current clamp?

The Tektronix A622 (same as Agilent, AEMC and more, rebranded) goes to 100A and is pretty linear and reliable.

Otherwise, buy one of these cheap but large resistors from ebay:
http://www.ebay.de/itm/Nebenwiderstand-Shunt-100-A-60-mV-Mess-Shunt-zur-Strommessung-/121698634939?pt=LH_DefaultDomain_77&var=&hash=item1c55cdccbb
They are pretty accurate as well and can be re-calibrated at 10A

[Simon]

I tried a current clamp and was 5A out on 45A. I just don't trust them.

[HighVoltage]

I tried a current clamp and was 5A out on 45A. I just don't trust them.

Which model of clamp did you use?

The resistor in the ebay link I sent earlier will for sure work.
I have the same one and used it up to 100A for precision measurements.

[Simon]

cant remember, some no name thing, didn't even have a display you have to plug it into a multimeter in mV mode.

Mind you it was a motor driver I was measuring so plenty of emf floating around as spikes of 100's of amps get shoved in and out so might upset measurements. partly why i want a resistor and then I'll low pass filter it.

[tszaboo]

If you want that much resistance, without extra size and price, the best is to parallel smaller resistors. Now, I'm not talking about just paralleling 10 1mOhm resistor. You should use an amplifier next to each (current sense amlifier with fixed gain), and average the outputs. Use 4 wire resistors, a nice amp like INA199, 4-5 shunt each 1mOhm. It is not 100uR, but you dont really need that low, as all the cable and connection losses will be a lot already.

[Zero999]

I tried a current clamp and was 5A out on 45A. I just don't trust them.

Did you zero it first?

And this may seem silly but many current clams are current transformers and only measure AC: are you sure it was a Hall effect sensor, designed for DC use?