2
\$\begingroup\$

I have a differential op amp that I believe I have set up as a unity gain differential amplifier.

When I measure the potential difference across pin 6 and ground and adjust the 5K trimmer potentiometer, it doesn't change the output at all on pin 6. I expected it to be able to zero out the op amp when I have zero volts across the input pins.

I have removed the op amp and verified the 5k potentiometer is functional. I have verified the +15V input and -15V input and that there is 0 volts across the input signals, and I have replaced the differential op amp.

schematic

simulate this circuit – Schematic created using CircuitLab

Here is the datasheet.

Any ideas what I am doing wrong, have it wired wrong, incorrect reasoning, etc.?

\$\endgroup\$
5
  • \$\begingroup\$ What voltage are you seeing at the output? Is the output saturated? Is the input to the resistor to the non-inverting input floating? Or, is it tied to a fixed voltage? \$\endgroup\$
    – Math Keeps Me Busy
    Apr 15, 2022 at 20:11
  • \$\begingroup\$ @MathKeepsMeBusy So I thought I had the input pins grounded when testing but I did not. The output on pin 6 is at -0.52 volts when I let both pin 2 and 3 float. Then when I change the trimmer nothing happens to output on pin 6. If I ground pins 2 and 3 and do the same test pin 6 holds at 0 volts and changing the trimmer nothing happens to output on pin 6. \$\endgroup\$
    – axawire
    Apr 15, 2022 at 22:12
  • \$\begingroup\$ If your circuit is not soldered, try changing the gain to something like 1000, by changing the value of your feedback resistor, and the resistor between the non-inverting input to ground to 1 M\$\Omega\$. With the inputs tied to the same voltage (say ground) the output should continue to be 0V ideally. However, offset errors may make it substantially not 0V. Then try adjusting your offset null trimmer, to see if it adjusts the output, and see if it can adjust the output to 0. \$\endgroup\$
    – Math Keeps Me Busy
    Apr 15, 2022 at 22:58
  • \$\begingroup\$ For completeness, I thought I had been really thorough with checking my connections but I discovered when I replaced the chip LT1226 with LT1224 which is unity stable. I had R3 connected to the chip but R3 was not connected to ground. I also found the connection between pin 6 and R4 was severed and so my out was only ever connected to R4 but never made it to pin 6. \$\endgroup\$
    – axawire
    May 4, 2022 at 21:12
  • \$\begingroup\$ While I did switch op amps, I did discover that my strange readings actually ended up being from a poor ground connection. It was a 2 sided board and when measuring the voltages I was doing it on top because of ease of access and when checking ground connections ease of access was on the bottom... well gravity made the ground connection bad when I was taking readings on top and when flipping the board upside down made ground connection good :-/ \$\endgroup\$
    – axawire
    Sep 19, 2022 at 17:08

1 Answer 1

Reset to default
2
\$\begingroup\$

This is a very fast op-amp and it's not unity-gain stable: it won't work correctly at gain 1. It needs a minimum noise gain of 25x for stability. Such op-amps generally are incompatible with solderless breadboards. For add-hoc prototyping, dead-bug them on a solid ground plane.

The output on pin 6 is at -0.52 volts when I let both pin 2 and 3 float.

The output voltage in this condition is irrelevant, since the op-amp is not specified to work correctly in such circumstances.

changing the trimmer nothing happens to output on pin 6

Are you measuring with a voltmeter that has resolution of 10uV at least? If not, then you won't see much, because the entire trim adjust range will move the output a couple mV.

Recall that the trim terminals on op-amps are only to be used for trimming the op-amp's own input voltage. They must not be used for any other purpose.

Since these parts have small offset of just +/-1mV, to use the offset trim you should be able to measure the input offset with 1uV resolution - the voltmeter is across the input terminals while the op-amp is configured in the application circuit, and properly biased. Then tweak the trim potentiometer until this input voltage is down to 0uV+/-1uV. If you have a 10uV resolution voltmeter you may still do it, but that voltmeter better also have a comparable accuracy.

Misadjusting the trim potentiometer makes the op-amp perform worse. So, you must be able to measure initial offset and then be sure that you brought it down to zero.

The trim potentiometer cannot be used for trimming out any voltages external to the op-amp itself.

If your application really calls for a 1x gain difference amplifier, and not, say, a 25x difference amplifier, and you have to use the non-unity-gain stable LT1226, then you have to bump up the noise gain to 25x, by using the circuit below. The 1k resistors have to be matched to 0.01%. The 24 Ohm resistors have to be matched to 0.01%, but the match relative to 1k resistors can be 1% and is not critical, as long as the resistance is kept below 24.5 Ohms, to maintain the noise gain of 25x minimum needed for stability.

schematic

simulate this circuit – Schematic created using CircuitLab

Another way of compensating at low gains would be to use lag compensation, but that requires verification on a test bench.

\$\endgroup\$
5
  • \$\begingroup\$ Thank you, I had overlooked the 25x stable. I had been looking at the family of chips and had originally been on a stable one but didn't notice that had changed as I went with similar priced one with more bandwidth. I am going to replace it with LT1224. Would it be typical that the voltage adjustment be in uV range too in this case? My voltmeter is not that accurate. \$\endgroup\$
    – axawire
    Apr 17, 2022 at 20:33
  • \$\begingroup\$ I don't understand why you're using the trim terminals at all then? The maximum offset voltage of the part is +/-1mV. To trim this offset out, you must be able to measure it first and foremost, with a voltmeter across the input terminals of the amplifier, and then to get it down to zero. You should get the offset down to 0.000,00V, but ideally 0.000,000V. If you can't measure, the trim adjust potentiometer is pointless. You cannot use trim potentiometer for any other purpose but reducing the part's inherent input offset voltage. \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Apr 17, 2022 at 22:19
  • \$\begingroup\$ If you want to "trim out" any external DC offsets - those outside of the op-amp itself - you cannot use the trim potentiometer for that. It'd kill the offset tempco, increase differential input current, and you'll be unbalancing the input stage of the op-amp. To apply arbitrary DC offset, forget about the trim terminals. Proceed as if they weren't there. Because they are only for the op-amp's own offset. They are to make the op-amp better, not to make any application circuit better. \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Apr 17, 2022 at 22:21
  • \$\begingroup\$ I can do millivolt but not microvolt accuracy. \$\endgroup\$
    – axawire
    Apr 17, 2022 at 22:42
  • \$\begingroup\$ Then a maximum +/-1mV offset voltage is not your concern at all :) \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Jun 8, 2022 at 20:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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