Precision resistor for calibration current shunt

Sep 13, 2022 Last reply: 3 years ago 13 Replies

(Sorry if this is a repost but it didn't show up on my end, the other day)


0.1 ohm 0.01% resistors seem a bit thin on the ground, so I was thinking:
formatting link

And:



formatting link

Idea is to take a square of copper-clad and Dremel a line down the middle, mount the resistors, and then drill out holes and solder on the banana jacks in the appropriate spots for a given piece of equipment's terminals, and shove the sumbitch on there when needed.



I don't have any mount-able male bananas on hand yet but I experimented with ten of the 1206 1 ohms in parallel as described, in theory the tolerance improves by sqrt(n) where n = 10.



Verifying if it actually does pushes the limits of my available equipment but my recently-calibrated 3478A in the 4 wire connection, after the copper-clad is thoroughly scrubbed clean before clipping the leads and after the HP warms up for a half-hour and settles down, seems to be telling me the resulting R is probably good enough for 0.03% rock and roll:



formatting link

But whether it's 0.0999, 0.1000, or 0.1001 this meter cannot decide.



Any construction tips for the next revision? the "prototype" is about an inch on a side with the Rs about evenly spaced, and kinda crudely solder-blobbed on each side bridging the gap but maybe it'd be better to mount them some other way or make the board they're attached to bigger or smaller


But only if the errors are random.

John

[about 0.1 ohm 0.01% sense resistors]

So, no improvement in temperature sensitivity nor aging, and... you'd get better performance (drop the 0.1 ohm voltage drop) sensing the magnetic field around a current-carrying wire. Old-school, that'd be a flip coil. Nowadays, Hall sensors and a thermostat, or maybe a fluxgate.

1-oz copper is 0.5 milliohms per square, so one square of copper at each end will blow your error budget by a factor of 100.

Cheers

Phil Hobbs

How much current?

1 oz copper is about 500 uohms/square, so be careful with that. You might do some clever 4-wire sort of thing using both sides.

Post pics!

There are genuine manganin 4-wire resistors.

I once visited the factory of an electricity meter manufacturer. They used manganin shunts with thick copper terminals electron beam welded to the ends so as to get a good 4-terminal connection.

John

Test

Sorry, my Usenet provider still seems to be dropping 75% of my replies so further discussion on my part may be delayed :(

Calibrating the meter on gear like the Agilent 3631A, etc. so on the fancy machines the calibration subroutine puts out a claimed 0.1 A and something like 5.6 A, and asks you to enter the actual value by measuring the drop across the shunt by measuring the voltage and shifting the decimal point.

It knows if your shunt/connections are too far out from 0.1 and will reject your input if it thinks it's BS.

I'd like to take another shot at it once I get more resistors in. Imagine a dog's dinner, tacking 1206 resistors across a gap on bare copper-clad isn't so easy for me..

I see resistance standards like these for sale, but how the hell would I even connect them to the machine's terminals?

formatting link

When we were in the NMR pulsed-gradient business, we made our own current shunts.

formatting link
We bought bulk manganin, had it rolled into sheets, punched it, folded, and annealed. The planar shunts were epoxied to constant-temperature aluminum blocks. We learned how to compensate eddy-current effects to get PPM-flat pulse response.

DC shunts are easy!

I use gold-plated copperclad. Solders magically.

formatting link
formatting link

On a sunny day (Tue, 13 Sep 2022 20:00:32 -0700) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:

Yes, but I have some nice HALL effect current transducers HX03-P en HX10.

Berylia has a pretty low thermal resistance - 285 W/mK - versus 235 for aluminium and a rather high electrical resistance. It is toxic,

Even easier if you know what you are doing.

We needed PPM flat and stable current pulses, which I suspect Hall sensors can't do.

But have no way of finding out. Hall sensors are temperature sensitive, and when I saw one used as a precision magnetic field sensor in a tandem mass spectrometer (designed to measure the atomic weight of protein fragments accurately enough to get the carbon/oxygen/nitrogen/hydrogen content out of the mass defect data) I was a bit surprised that the resistance of the Hall plate wasn't monitored to keep track of it's temperature. Somebody thought about patenting the idea, but it turned out to be well known, just not to the people who had developed the circuit.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required