10A 0.1V CC source for handheld DIY microohm-meter

[blubb]

Hi evereyone,

from time to time I need to measure DC resistances in the region of 1-500µOhm. At the moment I use a high resolution DMM (BM869s) and a lab power supply set to 5A in the usual 4-lead setup. It works quite well, I can measure down to about 1µOhm with an extra digit of resulution. Comparison with theoretical values showes me that my measurement errors are well below +-10%

Now I want to build a dedicated handheld setup with the flexibility to measure also at 1A and 10A. Power source could be batteries, a 12V wall plug or prefferably a USB power bank. I have build several projects with ADCs and shunt resistors. So I am not worried about measurement of current flowing and voltage drop. But at I am stuck trying to find a suitable current source. I looked for DC-DC converters but the high current ones only only work from 1.2V upwards in CC. Also, I do not need any voltages higher than 0.1V. In fact I prefer if the source is unable to supply anything above 0.1V. Looking at the power consumption, I do not think a CV-source with a power resistor in series is a good idea.
Ideally, I would look for a DC-DC converter with a spec of adjustable CC 0-10A and the ability to fix the voltage limit at below 1V.

Does anybody have a suggestion what search terms might help? Other ideas how to implement 1A, 5A, 10A CC capability without a full scale lab supply?

[inse]

That should not be too complicated to build.
As a power source, I would propose a capable LiPo or LiIon battery.
Then you only need a shunt, an OPAMP with RR input, a small DCDC converter to generate the supply for it and a suitable MOSFET.
Additionally some protection or timeout circuit as you are dissipating up to 36W.

[tszaboo]

Page 20 of the LT3080 datasheet shows a circuit that can be modified for 10A.
It's probably not cheap though. If you want cheapness, then transistors and opamps are the best bet.

[inse]

10 regulators in parallel? 🤔

[Boscoe]

Perhaps it’s best to use a CV DCDC intended for processors or DDR (low voltage, high current) to get you in the ballpark then use a linear ref to do the CC? It’ll be easier to get good noise performance and is simpler to design.

[inse]

Possibly running at 1MHz, not sure this is going to be easier.
Imho the analog approach is the simplest.

[Kleinstein]

I would start with a low battery voltage, like 1.2 V or 2.4 V and than use the more normal constant current sink circuit with a MOSFET and OP. For the OP part a second battery (like an extra 3-6 V) would be OK. Chances are one could use the same shunt resistor and just switch the set voltage to some 20 / 100 / 200 mV.

1 or 2 cells of high current subC size NiMH akkus may be enough for the main current.

The open circuit voltage may reach a little more than 1 V, but not that much.

[inse]

You could provide an open circuit voltage limit, using a shottky diode, you could limit to ~0,3V

[inse]

I wonder whether the initiator is going to return to all the suggestions being made…

[blubb]

First of all thank you very much to all of you who have replied with ideas. I think I will look at the opamp + MOSFET current source (or sink) ideas first. This is a type of arrangement I have never considered that may work for my purpose. Probably will have to order some parts as all I have are a handfull of LM358 which seem to be the wrong type.
The 0.1V open circuit is not a hard requirement, all I want is to reduce the chances of arcing and damage to other sensitive equipment in case of bad connections.
Thanks again for all your suggestions!

[2N2222A]

Is using direct transformer output for the 10A and measuring AC Voltage instead at all possible?

[Shay]

I have built a decent milliohm meter, however I used 100mA instead of 10A.
You have basically 4 blocks:
1. 100mA current source, using a precision voltage reference and opamp + p channel MOSFET. This is a very common and typical.
2. Op amp square wave oscillator set to 40Hz.
3. H bridge, to alternate the current (basically converting it to AC). This is to measure at ~40Hz. This allow you to only measure AC and get rid of any dc noise and improve noise rejection. (At least in theory).
4. AD620 instrumention amplifier is sensing across resistor under test, with a gain of 10. Output is the 1Vac / 1Ohm. This foes through a resisor (possibly increasing phase margin) to a common DMM.

This worked great for me. I was able to measure very low resistance with good accuracy

[magic]

For top efficiency look for a buck converter with synchronous rectification, otherwise diode loss will be ~0.3V and 3W even if you go with Schottky. That's still not as bad as a completely linear solution, but not optimal.

If you don't mind 12W of heat in exchange for a very simple circuit, build the linear current source powered by a few NiMH cells in parallel.

[Terry Bites]

 What you need here is a 1V regulated supply with a switchable current limit. You might want a skim of TI's SNVA829 (June 2018) ap note
Have you consdered a less brute force micro-ohm meter? Lower current better voltage sensing? www.maximintegrated.com/en/design/technical-documents/app-notes/1/106.html

[wishboneash]

I built a fairly inexpensive milliohm meter as an exercise to see how low I could measure with decent accuracy. I have some videos of it on Youtube. Main sources of error I would say are temperature variation and gain drift. Auto-zero works quite well.
https://youtu.be/BCJSTXyZ6oY