0
\$\begingroup\$

Circuit below amplifies 0-5mV signals to 0 5V level. But at zero input there is like 0.5V offset. Is there an easy way to offset null?

enter image description here

edit: enter image description here

edit2:

enter image description here

\$\endgroup\$

3 Answers 3

3
\$\begingroup\$

You could add a pot as shown. The LM324 is not a good op-amp to use for mV DC signals- offset can be as much as +/-20mV and this will only adjust for +/-5mV offset (you have observed 0.5mV).

schematic

simulate this circuit – Schematic created using CircuitLab

If you don't like having to come up with +/- supplies (or having it drift around by maybe 5-10mV/°C at the output), use a better single-supply op-amp.

Edit: Since you are using a thermocouple, and assuming it is floating, you can (and should) offset the input by adding a voltage divider to a reference on the (-) side of the thermocouple. Otherwise you'll never be able to measure below the temperature of the instrument terminal block. It's possible to combine the voltage divider with a temperature sensitive element to both offset and cold-junction compensate the input but that's way beyond the scope of this question.

\$\endgroup\$
4
  • \$\begingroup\$ I dont use a dual power supply \$\endgroup\$
    – user16307
    May 8, 2016 at 17:39
  • \$\begingroup\$ what opamp would you suggest for this range? i also have 741 \$\endgroup\$
    – user16307
    May 8, 2016 at 17:41
  • \$\begingroup\$ 741 is worse than LM234. \$\endgroup\$
    – Transistor
    May 8, 2016 at 18:08
  • \$\begingroup\$ LT1013 is an older part (dual). 741s are not single-supply so they won't work at all. There are many parts- more if you have lower voltage supply. \$\endgroup\$ May 8, 2016 at 18:09
0
\$\begingroup\$

The LM324 datasheet says

Large output swing: 0 V to V\$_+\$ - 1.5 V.

However it's not very good at close to zero. See my answer to How can LM324 work in dual supply mode when it only has 2 pins for the supply? where this is discussed in some detail.

Try connecting the op-amp negative supply pin to a -2.5 volts supply and run the simulation again.

\$\endgroup\$
10
  • \$\begingroup\$ I only have single supply not dual \$\endgroup\$
    – user16307
    May 8, 2016 at 17:37
  • \$\begingroup\$ On your simulator you can, I presume, have dual. Add one and try it. If that fixes the problem then you are gaining some understanding and I suspect that you will find that it's not an offset problem. \$\endgroup\$
    – Transistor
    May 8, 2016 at 18:05
  • \$\begingroup\$ I mean in real life I dont have dual power supply. But when I reduced the rail voltage to 5V the offset reduced from 600mV to 130mV \$\endgroup\$
    – user16307
    May 8, 2016 at 18:08
  • \$\begingroup\$ iNSTEAD OF NULLING Cant I just remove that offset after collecting the output voltage data instead? I mean if the output is 3.5V can I treat it like (3.5 - Offset) and calibrate input versus output? output will go to a data acq program \$\endgroup\$
    – user16307
    May 8, 2016 at 18:09
  • \$\begingroup\$ You can if the offset is constant through the range of interest and you don't mind it drifting with temperature. If it's clamping at 600 mV as you approach zero then you can't. Why won't you run the simulation with a negative supply rail, as suggested, and find out if that's the real problem? It would probably cost you a minute. \$\endgroup\$
    – Transistor
    May 8, 2016 at 18:11
0
\$\begingroup\$

You may not need to, and I suspect it would be the wrong approach, upsetting the behaviour during normal operation by trimming the input for an output-related problem.

The datasheet (Table 6.5 on page 5) shows that the LM324's output stage can source 20-40mA but only sink as many microamps under similar conditions.

Given this information, you can improve the performance of the LM324's output stage in certain applications by forcing the output stage into Class-A, so that it only ever sources current and never sinks it.

To do this, simply connect a resistive load from the output o V- (0V in your example). 1 kilohm if you can afford a few mA extra current consumption, up to 10 kilohms if you can't. Start with 1k, and if that fixes the problem, experiment with higher values if you need to.

One benefit is the elimination of crossover distortion (to which the LM324 is prone in AC applications such as audio).

But here, the benefit you're looking for is improved output swing when the output voltage is close to 0V.

\$\endgroup\$
6
  • \$\begingroup\$ I tried ur advice please see: i.stack.imgur.com/BL9rX.png But still having 600mV offset in smulation \$\endgroup\$
    – user16307
    May 8, 2016 at 17:48
  • \$\begingroup\$ ahh sorry I think u meant from -v PIN to GND \$\endgroup\$
    – user16307
    May 8, 2016 at 17:54
  • \$\begingroup\$ I meant from output to ground. But that was before I noticed the insane closed loop gain, with a 1-2mv input offset voltage (1-2V output offset). Even if you do trim that out, I expect it will drift a lot with temperature. Your best bet is a low offset opamp instead of the 324. \$\endgroup\$
    – user16324
    May 8, 2016 at 17:58
  • \$\begingroup\$ Cant I just remove that offset after collecting the output voltage data instead? \$\endgroup\$
    – user16307
    May 8, 2016 at 18:01
  • \$\begingroup\$ I mean if the output is 3.5V can I treat it like (3.5 - Offset) and calibrate input versus output? output will go to a data acq program \$\endgroup\$
    – user16307
    May 8, 2016 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.