High-side measurement circuit.

[AvalancheEffect]

Hi.

I'm working on a high side current measurement for a DC motor. Currently (due the pandemic) the shipment of parts is taking forever so I got hands on into designing something of myself (and google). I read a TI design (SLAU502) and took the main idea of the circuit (the second stage).

I will be posting here the process and any commentary/help comes well received.

The idea of measuring in the high side is to avoid lifting the load(motor) from ground but this add the complexity of measuring the shunt voltage that now is floating. To do that the voltage measured is transformed into current and then refered to ground through a resistor. The one thing I noticed was the output in P2 is inverted. After probing P1 I noticed the signal was right (but way over ground). In this case I though maybe a buffer can be used in P1 and then an output capacitor in order to decouple the signal.

I'm not trying to measure bipolar currents so the is no problem in having the signal referenced by ground. Some pictures below show the actual cicuit but I'm not sure on how to proceed.

Other problem I had in the design was the input to the non inverting input of the opamp where a very high resistor was added. If that value is diminished the output is no longer valid as a voltage representing the current over the shunt resisitor, to be honest I don't know why this is happening in the circuit but a 80Meg resistor seems quite a hard value to find in the shelf. The input impedance of the opamp should be high but I think is in the band of 1-5 Meg.


Best Regards.

[Renate]

Mmm, I'm not sure what the idea here is.

Let's start with the basics.
In a Spice simulation it will try to make sense of whatever you give it, even if values are so far out of where its models work.

An 80 M resistor connected to a 0.01 ohm resistor?
For scale, that's like the difference between a 50,000 ton ship and a sheet of note paper.
Yes, there are 100 M resistors, but they are usually for high voltage probes.
Moreover, if you breathed heavily on the circuit you can probably make a 100 M short circuit yourself.

High side current measurement has to be differential across the shunt resistor.
You have to measure both sides and take the difference.
Texas Instruments makes tons of chips, both analog and digital that do a good job of this.
I'm mostly familiar with the digital ones, INA219, INA226.
Here's an analog one: https://www.ti.com/product/INA169

Yes, you can make your own circuit from scratch.
The tricky part is to make the parts matched well enough that having the average voltage of the two points of the shunt not affect the current measurement.
That's the common mode rejection ratio.

Edit: Ah, I see, that SLAU502 is a voltage to current converter, not a current measurement circuit.
Did you mean more like this? https://www.ti.com/lit/pdf/sboa310

[geggi1]

Get a Hall sensor.
I have found Hall sensors on Ebay for a few $.

[S. Petrukhin]

Get a Hall sensor.
I have found Hall sensors on Ebay for a few $.

Be careful with the Hall sensor - they are sensitive to magnetic fields.

[KT88]

The spice models for ADI or LTC amps often times have some Vos built in. If this is the reason for the observed effect, Ibias might add a voltage drop that compensates Vos.
Could you post your .asc file?
A similar circuit is mentoned in the DS - so it should work in reality...

Cheers

Andreas