[fahque99]

What would be the best type of op amp for reading a shunt resistor on the high side of a ~300 volt load? Does the op amp see the 300 volts or only the mV of the shunt?


Dave.

[Al_The_Man]

If the Op amp input is across the resistor it will only see the volt drop, BUT if the 400v is referenced to ground at all then the op amp circuitry will have the same potential to ground.
There is an opto isolator H11F1 that has a variable resistance output for voltage input, but if mv it may need boosting.
I recall sometime ago reading about some analogue output devices that were fitted to Plasma supplies for AHC which were the same application.
Al.

[Zig]

Consider Norton current differencing op amp LM3900and its variants.

The op map sits at ground potential whereas its inputs are tied via high impedances to the signal to be measured.

Carefull selection of high value high voltage resistors in combination with lead lag compensation will allow you to happily set the input currents ( in the microamp region from memory ) into the inputs which are tied to ground at one Vbe drop.

A different way of thinking about op amps.

[fahque99]

This will be for a BLAC / BLDC drive. How do the other drives out there measure current?

[Zig]

Typically with the electronic driver modules the shunt is incorporated into the low side of the drive; that is the shunt is located between the motor and neutral or ground.

This topology makes it relatively easy to monitor and sample current on a pulse by pulse basis.
May be You ought to be a bit more specific about Your requirement.

[fahque99]

I don't have a strict requirement. I am just in the experimenting stage. Is a normal op amp used on the low side? Is low side safer for the op amp/circuitry ? I am better at the coding then the analog circuitry so I am trying to wrap my head around it.

Thanks

Dave.

[Zig]

Well..
Yes
and Yes

but at he same time it depends.

as in coding there are varoius ways of approaching a problem and various metrics ( big O say ) a coder may use to determine a better way NOT "the best" way.

So it is in hardware. It is hard to respond in a meaning full way to the question what is THE BEST? I t depends.

Often it is personal prefrence and experience.. Cost certainly plays a significant part in the design approach.

So lets start with some basic issues.

What is it you wish to implement?
What voltage and what power rating?
How do you want to control the "what it is you wish to implement" thing?
Control strategies as already mentioned pulse by pulse, average?
Control strategy: Current, voltage, torque, speed,..?
I think having a more complete spec can only help to come up with "the best" selection of solution.

[Al_The_Man]

Look up some of the manuf app notes for PWM/Power FET output stages, they show examples of what you may be looking for, IR, FreeScale, Allegro are a few.
Al.

[fahque99]

What I am looking for is a simple cheap way to measure current on a 20A 250V load with a microcontroller. I will need check out some app notes for sure. Can you recommend a good one to start with?

[Al_The_Man]

For starters you could check the HexSense devices http://www.irf.com/technical-info/anarchive.html
It has been a while since I visited these sites but also Motorola (Freescale) may have also.
Al.

[fahque99]

Would something like a INA201 from TI be a good option with some voltage dividers since it only has a max of 70V?

[bharbour]

If you are not in love with the shunt resistor approach, Allegro makes some Hall effect (magnetic) current sensors that work with AC or DC, have very low voltage drop and the measurement output is isolated from the signal being measured: http://www.allegromicro.com/en/Produ...0758/index.asp

I have not used these yet, so I don't know if they have any quirks, but they look ideal for this type of problem.

Good Luck,
BobH

[keebler303]

If you are not in love with the shunt resistor approach, Allegro makes some Hall effect (magnetic) current sensors that work with AC or DC, have very low voltage drop and the measurement output is isolated from the signal being measured: http://www.allegromicro.com/en/Produ...0758/index.asp

I have not used these yet, so I don't know if they have any quirks, but they look ideal for this type of problem.

Good Luck,
BobH

There are several companies who make them, I think they are much more expensive than a shunt resistor approach. You could also check out the IR2175, varies pulse width proportional to current. I think it can be used on high side, have to check the datasheet. Best thing is it is a digital signal so you can quite easily run it through an optocoupler for isolation.

Matt

[Zig]

Shunt resistors ( high power low inductance thick film types are not exactly cheap. anyway the hall devices provide You with isolation and solid analog output.
Overall cost difference with hall devices might in fact be lower than with shunt resistor and anciliary electronics

[jmelson]

What would be the best type of op amp for reading a shunt resistor on the high side of a ~300 volt load? Does the op amp see the 300 volts or only the mV of the shunt?


Dave.

There are some circuits for doing this that produce a current source proportional to the high voltage, and then you drop that current into a grounded resistor to produce a proportional voltage to the measured elevated voltage. I've seen these in the design tips in some EE magazines. It requires powering the op amp from the elevated supply with another current source, so there's a good bit of dissipation there.

A circuit I used was with the Analog Devices OP620 instrumentation amp, which has insanely good common mode rejection. Since it doesn't have a common mode range beyond the supply rails, I used resistive voltage dividers to bring the sensed voltage down to within that range. I matched these resistors with my DVM to get the best CMRR. In my case, I needed to sense the current from a full-bridge PWM driver running at 100 KHz. So, in addition to the 100 mV or so of sensed voltage, this was superimposed on a 100 KHz square wave at 80 V. The circuit performed flawlessly, and I'm sure the same game could be used at 300 V with suitable divider resistors.

Jon