# Selecting current sensing chips for PWM applications

#### Gibson486

Joined Jul 20, 2012
327
I usually use an INA139 to measure current. Since it is for a PWM application, I was going to simply filter the output. Simple enough.

However, I also found the INA240 just from reading literature from TI. Aside from having built in gain and a huge common mode voltage spec, it has what is called Enhanced PWM rejection. It looks it has to do with having a step response at the extreme voltage swings, which causes a violation (or close to) of the common mode voltage because the voltage swing over time is just too big. Looking over the graphs, it seems that it is still effected, but it corrects the issue faster than say the INA181, which is about a fraction of the cost. However, it never fully corrects it. Am I thinking about this right?

It seems that the INA139 is still a good chip to use for my application (pulsing a high current LED @ 12V) and that the INA240 (or INA181) is more for space constrained applications that cannot afford the space for a filter and gain resistor or maybe they just need that 80V common mode spec (in the case of the INA240)?

Aside from that, the cost break down is kind of weird...

INA240 is the most expensive (not by much)
INA139 is in the middle
INA181 is the cheapest by a landslide

I find it sort of weird that the INA181 is cheaper than the INA139. Then again, I have learned that lots of this has to do with the fab process and what each vendor wants to try to push the industry to use. Or perhaps I am missing some detail that makes the INA181 unfavorable (aside from the internal gain, which some people may not like and the lower common mode spec).

#### TeeKay6

Joined Apr 20, 2019
572
I usually use an INA139 to measure current. Since it is for a PWM application, I was going to simply filter the output. Simple enough.

However, I also found the INA240 just from reading literature from TI. Aside from having built in gain and a huge common mode voltage spec, it has what is called Enhanced PWM rejection. It looks it has to do with having a step response at the extreme voltage swings, which causes a violation (or close to) of the common mode voltage because the voltage swing over time is just too big. Looking over the graphs, it seems that it is still effected, but it corrects the issue faster than say the INA181, which is about a fraction of the cost. However, it never fully corrects it. Am I thinking about this right?

It seems that the INA139 is still a good chip to use for my application (pulsing a high current LED @ 12V) and that the INA240 (or INA181) is more for space constrained applications that cannot afford the space for a filter and gain resistor or maybe they just need that 80V common mode spec (in the case of the INA240)?

Aside from that, the cost break down is kind of weird...

INA240 is the most expensive (not by much)
INA139 is in the middle
INA181 is the cheapest by a landslide

I find it sort of weird that the INA181 is cheaper than the INA139. Then again, I have learned that lots of this has to do with the fab process and what each vendor wants to try to push the industry to use. Or perhaps I am missing some detail that makes the INA181 unfavorable (aside from the internal gain, which some people may not like and the lower common mode spec).
Ignoring for the moment your actual question regarding "which is best", just what are you trying to do that requires an INA? A schematic would really help.

Joined Jul 18, 2013
20,875
In a PWM application that uses Mosfet drivers for e.g. A very low resistance is often used in the source connections of the mosfets and volt drop input to an op-amp used to monitor it.
There is also Mosfets with built in current sense outputs.
Max..

#### TeeKay6

Joined Apr 20, 2019
572
In a PWM application that uses Mosfet drivers for e.g. A very low resistance is often used in the source connections of the mosfets and volt drop input to an op-amp used to monitor it.
There is also Mosfets with built in current sense outputs.
Max..
I am acquainted with various current monitoring techniques but the use of an INA rather than a simpler opamp makes me curious what is being done (i.e. schematic).