Nano amp meter with HX711

[darksky]

I have been experimenting with the HX711 weighing scale chip.  I had two aims  - one to produce a milliohm meter and at the other end of the scale a Nano amp meter.

First project was a derivative of one posted here and now works repeatedly down to 10uOhms.

Second project which may be of interest here is a Nano amp meter to measure very low power consumption.

Project uses an Arduino Uno or Nano and the HX711 as the A/D.

With self-calibrating software I can measure to better than 2 nA.  Not bad for a £1 module and hand full of resistors, built on strip board.

Can measure the 6 Nano amps flowing from a 0.6 volt source through a 99Mohm resistor.

Max current is around 100 uA. Works in either direction plus or minus.

Accuracy at  10nA level appears to be about 20%. Rapidly improves as current increases.

Attached  is a plot of the % error in measuring 10nA from cold over 5 minutes. Standard deviation of 0.7nA.

If anyone is interested I will post code and details or if anyone can improve the idea.

[danadak]

I would be interested in project, code base.

Your project I assume has a pretty limited CM range for measurement ? Close to ground ?


Regards, Dana.

[darksky]

I use a floating shunt so must not share a common ground with DUT. Not fully tested the range  but think it is 100 uA max down to 2nA.  The max value can be  changed with different lower value shunt. Currently 100 R.  Could make it auto range up to amps with series of low value shunts. All calibrated in software via serial command. I am away until weekend but will attach  circuit and software and. I am back at the weekend.

[Kalvin]

darksky, very nice and interesting project! 

Placing two low-leakage diodes across the shunt resistor makes it possible to overload the current measurement and limit the maximum voltage drop to approximate 0.6V.   

[darksky]

Added two siganal diodes and 10 ohm limiting resistors accross the shunt to limit max voltage and to protect the diodes.
 
Measured max current and it is 0.670 uA resolution around 1 nA.

Will make it auto range with relay and one ohm shunt in parallel with existing shunt to increase max current to 60mA.

Will post circuit and code on Friday.

The HX711 is amazing.

[darksky]

So after much work here it is - my week end project to use a HX711 to measure down to nanoamps.  Having built it I was amazed how good it is.

I wanted a circuit to measure very low currents to test low power microprocessor circuits.

It has a LCD for local display and can stream data to a PC as a CSV format such that it can be processed in Excel.

Also added autorange so max can be up to 60 mA  - measurment uncertainty at low scale 1 nA.

The circuit was built on vero board.  Small relay is used to change range by switching in a 1R across the 100R.  Values arenot crfitical but you will need to calibrate.  Calibration can be achieved for low and high values via serial interface.  You must press d on first code upload to set defaults to EEPROM.

The circuit is an ugly bug build.  When finished I put the HX711 and circuirty in a metal box for added screening. Although it woked without.

The photos show measuring current from a 33 Mohm resistor sourced from 0.6 volt. Should read 18.2 nA. Photos show -18.5 nA and when reversed 18.05 nA.

The circuit is biased by 1M ohm resistors.  The 10k accross E- and E+ is to pull current such that the voltage is stable and 10 uF cap to smooth.

The 100k in series with the 1 Mohm are there because that is what I started with, so just left on the board. I built the code defaults with these values in series.  The 100k could be removed.

Code and library is attached, also schematic. The HX711 library is not mine  - so achnowledge the author.

Look forward to comments - hope it is of use to someone.

[darksky]

Ok done more work.

Cleaned up the code, added the ability to measure voltage and better calibration. Th enew design can measure current - two ranges 600uA to 60mA and 1nA to 600 uA plus voltage 600 vots max.

I also have ideas how to improve it.

I have attched the new code and ideas plus measurements.  Having calibrated aginst a voltage reference - it appears capable of better than 0.2% accuracy. If the input resistor is lowered to 100K it improves by a significant factor.

However, 100k is too low - so I intended to add a high Z buffer.

Also could add auto calibration.

I now need to better understand how to bias the inputs.

I believe this chip could measure voltage and current to better than 0.02%

Any comments welcome.

[darksky]

Tested after 24 hours and still accurate to better than 0.2% tested at 10, 7.5, 5 and 2.5 volts.

However, to get good results with these chips you need to supply from a good stable 5 volt source.

I have ordered some instrumentation amplifiers to use  at the front end and will make a Craddock  divider network.

[PointyOintment]

ICStation has 25% off their already cheap HX711 breakouts (and everything else) in their Tindie store for Cyber Monday (probably almost over by now—I don't know what time zone Tindie uses) (though it turns out you can get them a few cents cheaper normally on Amazon and Banggood with a slightly more cheap-Chinese-looking PCB), so I'm thinking about using one to make a combined short circuit-finding differential (probe vs. sample and hold or DAC) microvoltmeter with audio output and maybe also include this nanoammeter functionality (using both inputs, I guess).

Edit: ICStation's "free shipping" costs $3, maybe only today—see attached screenshots.

[vidalv]

Hello, very nice project! Congratulations. I know this project is a bit old but I just came across it while doing research on the HF711. Finally another use of the HF711 amplifier.
I have two small questions, the first concerns the frequency of conversion of the HF711 by default it is at 10Hz (100mS) would the measurement not be more stable at 80Hz (12.5mS)?
The second point concerns the ground, there is a common point between the GND link of the measurement and the one of HF711 circuit, wouldn't the use of a virtual ground be more judicious, using an isolation AOP with VCC/2?
Thank you very much !
Vincent.

[Edcali]

Hi, Wondering if you did this project, I just started to study it. If you can give me some hand help it would be great. I don't understand it yet. Just let me know. Thank in advance. Ed.