Differential amp with current sense below or above op amp ground

[treez]

Hello,

I wish to measure 1 to 10Amps into LEDs.
I am using a 1mR sense resistor.

I am using a x100 differential op amp, but should I best have the current sense voltage going below the op amp ground, or staying above the op amp ground?

Current sense voltage going below op amp ground
https://i45.tinypic.com/2rxbk93.jpg


Current sense voltage staying above op amp ground
https://i48.tinypic.com/29qbps2.jpg

 

[crutschow]

You want it above the op amp ground (op amp negative terminal grounded). The other way you will also be measuring the current used to power the op amp.

 

[tpetar]

You can try something with LM3914 and opamp.

 

[treez]

thanks but i just have to know if its 1A, 2A....10A...so massive accuracy not required.....the opamp bias current will be negilgible in this case, and i do not worry about that.

 

[erikl]

As you want a positive VOUT, I think only the 2^nd circuit will work (sense voltage staying above op amp ground), if the current direction is such as shown by I1, not that suggested by V2 .

BTW: R2 is useless as it is in parallel to R5. If you want it because of accuracy reason (which I think wouldn't matter at the gain of 100), it should be connected to GND.

 

[treez]

thanks ericl....you are correct that the current is as the direction of I1 (the volt source is going to be a LED)

But i assure you that both circuits make the vout go positive.

You are right, i didnt spot that r2 is pllel to r5...........in fact, i realise that the differntial amplifier in this case, as actually just an inverting amplifier with R2//R5 added to equalise voltage drop due to input bias current......im glad you spotted this cuzz i didnt.

Anyway......is there any problem with either circuit?....does it depend how far below gnd the terminal of the sense resistor goes in the first circuit? (it will only go a max of 10mV below op amp gnd)

oh...hang on a minute....R2 and R5 must be separate resistors as shown...this is becasue a parallel 100n cap will go across R4 and a parallel 100n cap will go across R2.

..So these resistors must (after all) be kept separate so the filtering can thus be done.......(the current source is actually haversine at 100KHz and so the 100n's are needed.)

 

[crutschow]

There doesn't seem to be much difference which ground point you use, but the conventional way is to place the ground on the right in common with op amp common. Why would you want to do it the other way?

Filtering across R2 and R5 is only required if you are using the amp as a differential amp and you aren't. So you just need a cap across R4 for filtering.

 

[treez]

thanks, what about this connection here.....?

https://i49.tinypic.com/rm4dph.jpg

is this an improvement?

 

[erikl]

... i assure you that both circuits make the vout go positive.

Yes, you're right!

oh...hang on a minute....R2 and R5 must be separate resistors as shown...this is becasue a parallel 100n cap will go across R4 and a parallel 100n cap will go across R2.

..So these resistors must (after all) be kept separate so the filtering can thus be done.......(the current source is actually haversine at 100KHz and so the 100n's are needed.)

Again: R2 doesn't make sense, and you don't need any filtering across it, as crutschow suggested before.

what about this connection here.....?

https://i49.tinypic.com/rm4dph.jpg

is this an improvement?

Yes, this looks good. But again: R2 (and its possible parallel cap) is dispensable.

 

[treez]

Erikl

But again: R2 (and its possible parallel cap) is dispensable

...but that is a differential amplifier...you cannot dispense with R2, it wouldnt be a differential amplifier any more?

 

[FvM]

...but that is a differential amplifier...you cannot dispense with R2, it wouldnt be a differential amplifier any more?

That's right for the circuit from post #8, not for the circuits from post #1 (it's actually twice the same circuit with differently placed ground symbol), that are simple inverting amplifiers.

For an accurate current measurement, you would want to connect the shunt with separate sense terminals, then a you get a true differential amplifier topology in all cases. Usually, shunts below 10 mOhm will be connected this way.

 

[crutschow]

You don't need a differential amplifier if you are sensing current in the ground return line. You only need it if the current shunt is biased above ground (high side sense).

 

[FvM]

You don't need a differential amplifier if you are sensing current in the ground return line.

You don't need it if you can assume zero ohm wiring. Measuring with an 1 milliohm shunt to 1 % accuracy requires a trace resistance shared by current path and sense lines below 10 µohm. Usually the pad itself exposes higher resistance, so practically, you need to go for a four terminal resistor footprint and differential amplifier.

 

[treez]

Thanks, in the end i will go for this solution...........

https://i45.tinypic.com/72tvsi.jpg

...its because i want the op amp ground to be the same ground as the ground of circuitry to the right of the op amp....therefore i put the ground where it is , and let the current sense resistor voltage go below ground.

I dont need massive accuracy, i just need to be able to tell if its 1A, 2A, 3A, .....10A.

Do you think my configuration is ok?

The opamp that ill use in the circuit is the OPA335

OPA335 Dtasheet
https://www.ti.com/lit/ds/symlink/opa335.pdf



...i do apologise, i have certainly not ignored comments saying that R2 is unecessary.........i am afraid i do not understand it too well, and the configuration i give here is a typical, known diff amp configuration, so i am comfortable with it.................but i take your comments on board and will aspire to get a better understanding

 

[FvM]

...i do apologise, i have certainly not ignored comments saying that R2 is unecessary.........i am afraid i do not understand it too well, and the configuration i give here is a typical, known diff amp configuration, so i am comfortable with it.................but i take your comments on board and will aspire to get a better understanding

As said, R2 (and R5) will be needed, if R3 and R5 are 4-wire connected to the shunt, or in other words, if you want to supress possible voltage drops between shunt and circuit ground.