op amp offset

[Simon]

I just spotted an older blog I had not seen yet about "chopper amps" very informative and an eyeopener (thanks dave). It was an eye opener on what I might have been getting wrong !

now I have this current regulating circuit and I am using a 0.001 ohm resistor as the sensing resitor, I know VERY low, well it passes 20 amps and works in a confined space so power dissipation is an issue, so I have up to a 20 mV signal.

i used one part of a TL084 opamp to make a 100 gain differential amp and feed the output to a pic, things went a bit haywire although the circuit did work but it was reading less than it should have. having seen the blog it jumped to mind (as I am currently working on an upgrade to the project), now the TL084 has a 15 mV offset, so does this mean that potentially I was getting a 5 mV signal instead of 20 mV ? how do I counteract this ? use an op amp with a small offset voltage ? high precision is not required but repeatability is.

[migsantiago]

There's a new low offset ampop from Microchip, the MCP651.

http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en540381

It has an auto-calibration pin which trims the offset up to 200uV.

[Simon]

right but is my theory correct ? to be honest I am limited to what RS and farnell supply, they usually have a part that meets my needs but I don't get to choose the specific one

[jahonen]

If you want repeatibility, then look also for the offset drift. And remember that in a differential amplifier, your offset will be amplified by the amplifier, so if you have 15 mV input offset, it will produce a 1.5 volts offset with 100x gain at the output.

Regards,
Janne

[Simon]

yea well that looks like the cause of the odd behaviour, I was using a 100 gain so the 15 mV became 1.5 V. its not the effect of the gain that is the issue it's that I start with a 0-20 mV signal on my input so a 15 mV offset is like an overwelming part of that. so is the offset normally positive or negative ? in my case the output was lower than expected so I'd say negative but I'm new to all this so probably overlooked something.

[jahonen]

Offset can be to either direction. Basically that will ruin your otherwise carefully designed circuit. BTW, are you using a single supply for TL084? If you look at the output swing specifications, you'll notice that output will not go to either rail. That is often showstopper for single supply application unless you can bias everything in middle of the supply voltage range.

If that is the case, something like OP747 would be more appropriate.

Regards,
Janne

[Simon]

yea I'm using single supply but really I'm just waiting for 2 V to appear on the opamp output to trigger a comparator, I'm not interested in finding that I'm passing 1 amp because 0.1 volt is coming out of the op amp

[jahonen]

I forgot to mention that input also has common mode range limitation, TL084 input common mode range is -12 to 15 volts with ±15V supplies. That means you can't use it to reliably amplify signals near the ground (or negative supply rail, whatever that is). If negative supply rail is at zero volts, then both inputs must be at least at three volts for the amplifier to meet its specs.

Regards,
Janne

[Simon]

well the inputs are over 12 volts ! thats why i use it as a differential amplifier to "extract" or isolate the voltage on the sensing resistor (that is on the positive rail) and convert it to an amplified voltage with reference to ground, while I had not taken what you mention into consideration it seems to be a happy coincidence that I chose the right operating conditions (save for the horrible offset)

[jimmc]

Why not use chip designed for exactly this job?

 eg http://datasheets.maxim-ic.com/en/ds/MAX4172.pdf

Jim

[Simon]

Thats interesting, although I have a PCB layout already done and my method seems to be ok but for the actual part choice and i may well find a drop in pin for pin replacement. I'd also need to use this on a 0.001 ohm resistor and I can't change that much maybe 0.002 as I'm passing 20 amps but may make this up to 50 later

Edit: having read the data sheet it appears that there is no issue with resistor values, I just have to watch that the output is not over 1.2 V (doable). oh well looks like finding a supplier and relaying my PCB, hm exspensive chip ! and it looks like I'd better get some SMD soldering skills QUICK !