I am looking for a module or a design to see very low voltage using esp32, actually I am going to use a shunt resistor of 10 milliohm to observe 400 Amps DC current but the problem is how to detect voltage drop in millivolts, I know there is HIOKI multimeters but this is a project and I cannot use HIOKI multimeters coupled with ESP32, some Arduino voltage sensors exist but they are not better than 25 millivolts low limit, but the system I am designing needs something around 4 millivolts, any suggestion would be highly appreciated.

 

Have you considered using an INA219 (Or other devices in that family.) ? They have 12 bit resolution and the LSB is 10 uV.
Les.

 

as

Have you considered using an INA219 (Or other devices in that family.) ? They have 12 bit resolution and the LSB is 10 uV.
Les.

as far as I know, INA219 is a current sensor, and I don't think it can handle anything close to 400 Amps, it mentioned in the datasheet that it is a current/power sensor with internal shunt, I don't know any other module or IC for voltage sensing

 

I have used MAX4376 / 4377 / 4378 before with decent results. They do have a lower threshold to watch for meaning at no current flow they still output a small voltage. If you can work around that then they may work. I have a feeling based on your initial description that may be a problem. There are three output multipliers to consider also.

I haven't worked with the ESP32 ADC yet to know for sure, but I have read they are not known for accuracy. You may want to consider an secondary ADC also.

 

as far as I know, INA219 is a current sensor, and I don't think it can handle anything close to 400 Amps….

Your understanding is incorrect.
The INA219 only amplifies and reads the voltage drop across an external shunt resistor. This is mentioned in several places within the datasheet and is schematically shown in figure 13.
As long as you choose the correct shunt value, to maintain its voltage drop within the range specified in the datasheet (which varies depending on the PGA’s gain setting), you can measure any current level.

 

Ofcourse, for adequate accuracy and resolution, you will need to use isolated power sources for those amplifier stages and the A/D section. Plus opto-isolators in the digital lines. But by isolating the signal system you can avoid most of the common mode voltage issues completely.

 

You never talk about the voltage of you 400a system.
INA219 is limited to 26V

 

You never talk about the voltage of you 400a system.
INA219 is limited to 26V

Certainly a valid caution. Common mode voltage is ALWAYS a consideration when using shunt resistors. THAT is the reason I said to use isolated power supplies. Excess common mode voltages often lead to excess smoke issues.

 

You didn't ask about this, but make sure you consider the power dissipation in the resistor. P = i * i * R. Using your numbers, 400 * 400 * 0.01 = 1600 watts. A blow dryer dissipates close to that much heat by using a lot of forced air. If you are planning on a fanless design, you will need a very large heat sink.

 

I’m wondering if you could consider using a transformer core with a single turn of the 400A DC current, a linear hall sensor in the magnetic circuit and maybe a thousand turns of a ‘secondary’ with a controlled current up to 400mA to balance the hall sensor to zero? Or just calibrate the hall sensor output. That’s roughly how DC current clamp meters work

 

I considered that magnetic field scheme many years ago, when Hall effect sensors were much less linear. I never heard of anybody actually using it, but it certainly seemed reasonable at the time.

 

Absolutely, if you’re are measuring the current on a high/ voltage, high energy circuit, you DO REQUIRE insulation and a shunt resistor won’t provide it. Hall effect based sensors provide it and have improved vastly in its performance the last 15 years.

For the utmost precision, use a close-loop sensor, like the LEM model LF310-S, which is rated at 300 A RMS, or +/- 500 A DC, with an insulation rating of 3.8 kV RMS
But ( there is always a but) it is pricey.

The Tamura L37S400S05M is open loop and about 1/5 of the price, with reduced performance.

 

Probably adequately isolated power supplies for the instrument amplifier and processor module will be cheaper.

 

The hall sensor technique for measuring high current has been around for many years. A company called HEME, a small subsidiary of Pilkington Glass in the UK made transducers from around 50 years ago and made the first DC clamp meter. Around 40 years ago Mueller & Weigert in Germany made a 4 bit processor controlled AC/DC clamp meter with an additional 750V input with ranges including power and phase angle. The hall sensor was a Siemens component which was incredibly linear

 

400A X 0.010 ohms=4.00 volts. Do you really need that much shunt voltage?? Many shunts are rated at 100 Millivolts at the rated current. So a 0.001 ohm shunt will deliver 0.4 volts. That will be an excessive voltage drop in a 12 volt DC system.
So we still need to know the voltage in the system because the voltage drop does matter. So really, for this application , what is needed is a 100 millivolt shunt rated for 400 amps. That will be R=V/I =0.10/400= 0.00025 ohms. But still, the shunt will be dissipating 40watts.
It seems to me that this is an exercise to show that teaches about shunt resistor selection.

 

shunt voltage need to be small.... 50mV is common... 400A * 0.040V = 16W dissipation at peak current.

 

we still do not know what is the circuit voltage... so cannot propose concrete solution.
many products like INAxxx are meant for measuring high side. some of them can go to rather high voltages. and if common mode voltage is an issue, one option is to get rid of the issue and measure at the low side (like INA180).

personally, i like the isolation when dealing with high power circuits... measure whatever is needed in one circuit, report values in another circuit that is not galvanically connected.

 

Absolutely, if you’re are measuring the current on a high/ voltage, high energy circuit, you DO REQUIRE insulation and a shunt resistor won’t provide it. Hall effect based sensors provide it and have improved vastly in its performance the last 15 years.

For the utmost precision, use a close-loop sensor, like the LEM model LF310-S, which is rated at 300 A RMS, or +/- 500 A DC, with an insulation rating of 3.8 kV RMS
But ( there is always a but) it is pricey.

The Tamura L37S400S05M is open loop and about 1/5 of the price, with reduced performance.

My tests with Tamura sensors were not good for high precision current measurements but hall sensors in general are great.
https://forum.allaboutcircuits.com/...c-controlled-battery-array.32879/post-1496750

FM80 project 'dumpload' Hall current sensors.


Honeywell sensors.


For precision measurements with a shunt or Hall sensor a good ADC is needed. So a external ADC (16-bit or more) with a PGA will likely be needed.

 

OK, we do know that it is DC , and if it is a mobile application it might be in a vehicle with the body framework as a common negative, which could make low-side sensing a problem, or at least a challenge. Insulating a shunt resistor and the associated wiring is not a big deal, although isolating the attached processor modules could be a challenge. Many of the common DC supplies do not specify the output to ground voltage standoff rating. So never assume it is good for much beyond twice the mains supply rating.
And we still do not know if this is even in an EV, or the kilovolts plate supply of a really high power transmitter. A serious difference.

 

NSAspook;
Indeed Hall effect sensors don’t have the accuracy of a high-precision shunt.
But its errors can be calibrated in software. The zero-current offset is trivial. The full-load gain requires a way to simulate the high currents and thus not straightforward. That is the reason closed-loop are preferred for the highest accuracy.
https://www.lem.com/en/hall-effect-current-sensors