for my first project I decided to make a scale using a load cell, the micro load cell 3133 from Phidgets. I also used an ICL7107 as A/D converter and the op amps are LM358.

I used a differential amp to get the voltage frome the cell and then I amplify and use another op amp to make the calibration of the cell. Then I use the ICL to show the load on a 7-segment display. I'm using 1.5 V piles to get 4.5V with which I'm powering my circuit.

I'm able to get the voltage from the load cell and to amplify it but the problem is that I seem to get very annoying noise from the last op amp because the display keep switching values (about 10 or 20 mV two or three times by second). I used decoupling caps (0.1 uF) for all the amps and for the IN of the ICL as for the ICL itself.

I find it kind of weird because when I first tested it without the ICL, I was reading values with my multimeter and they were right and weren't waving. But when I connected the last op amp with the A/D, I kept getting switching values. I tried many gains (from 5 to 800) but I always get this switching effect.

You can see the circuit on the following picture.

The decoupling caps aren't there but I can assure you I put it on the board.

I changed the last op amp configuration to make a straight line configuration for the calibration and we can't see it on the picture but I was getting the flickering before I changed it.

I made some tests when all was connected to the ICL and I realized that it was the output of the last op amp which was switching. I tried different op amp but I always got the same results.

I used the same ICL for a school project (a thermometer) and the display was fine. I don't know what to do right know and if anyone got a solution or have an idea I will be very grateful.

Thanks for your help.

  • \$\begingroup\$ The title of this question is a bit misleading. The ICL7017 inputs are "vacillating" too. So garbage in... garbage out... as someone would say. But the title is somewhat relevant in that the ICL is a non-trivial load and being a digital IC may induce noise in the rest of the circuit. \$\endgroup\$ – Fizz Oct 4 '15 at 20:53

You have quite a bit of gain there and the op-amp you're using is far from being low noise. There are also opportunities for the ~200mA LED currents to cause issues or mains hum (though most mains hum should be rejected if the integration clock is correct). You are only seeing maybe 10-25uV referred to the input.

Start by determining the nature of the noise by using an oscilloscope on the outputs of the op-amps and looking to see if it's random looking noise or stepped when the LEDs switch.

Edit: Based on the comment, and from your video, from which I can see that your resolution is about 1mV and noise is around 20mV you should use a lower noise amplifier. You want an amplifier with noise in the 0.1-10Hz range that's maybe 1uV (or better). For example, the (quite old) OP07 is typically 0.35uV. You could also consider a more modern 'zero-drift' amplifier or an instrumentation amplifier. It's not enough to replace the gain-of-820 amplifier- the differential amplifier also is contributing noise. The final amplifier in your circuit probably isn't contributing enough noise to bother about.

| improve this answer | |
  • \$\begingroup\$ I looked the output of the last op amp with my oscilloscope and there were no particular pattern in it. It looked just like noise. I recorded the oscilloscope while I was checking the output of the op-amp. There is also a short video of the display so you can see what I'm talking about. drive.google.com/file/d/0B7s-vzSOB78SMUVZVWF0VFVadFk/… \$\endgroup\$ – Charles Vandal May 22 '15 at 17:38
  • \$\begingroup\$ Sorry, here's the video of the oscilloscope. drive.google.com/file/d/0B7s-vzSOB78SWUFMUEl6a3h4OEE/… Also, I changed the configuration of the last op-amp (the one which the output goes in the ICL). Here's the schema of it. i.imgur.com/ZxOIpvI.png So the video of the oscilloscope correspond to the output of this op-amp. It keeps switching like you can see. \$\endgroup\$ – Charles Vandal May 22 '15 at 18:00
  • \$\begingroup\$ Ok, I will change the op-amp and use an instrumentation amplifier and I will come back with the results. Since I don't have these op-amps it will takes some days before I can make these changes. I will let you know if it works. \$\endgroup\$ – Charles Vandal May 22 '15 at 18:49
  • 2
    \$\begingroup\$ Good news. I received the op amps OP07DP and I replaced all the other op amps. I also used an instrumentation amplifier for the first op amp. So, now the waving has considerably reduced. It's below 5 mV and sometimes it don't change at all. But I don't know what else I can do to get better results. The scale is now usable but like I said, sometimes it don't change but in most cases it keep switching about 2 to 5 mV. Thanks for your help, it is really appreciate. \$\endgroup\$ – Charles Vandal May 30 '15 at 21:27

I'd suspect that your LCD drive frequency is getting into the offset input of your last op amp. You're only seeing a few millivolts, after all. Try replacing R17 with a pair of 3k resistors in series, with a 1 uF cap to ground from the connection of the two resistors.

| improve this answer | |

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.