Thermocouple amplifier circuit, the input becomes an oscillating waveform

jim_cheng

Thermocouple amplifier circuit, the input becomes an oscillating waveform [Copy link]

 

I made an instrument amplifier circuit with OP07, the power supply is ±5, and the voltage between v+ and GND is an oscillating waveform. Please help me analyze it!

topwon

20 milliseconds period, this is the 50Hz AC interference waveform. If you are using a DC power supply, first check whether the oscilloscope probe is fixed, especially the ground clip.

topwon

If you want to measure the waveform of V+ in the differential signal, then the ground clip of the oscilloscope should be connected to V- instead of GND, right?

topwon

Also, how come the gain I calculated is -210 instead of -230 as indicated on the graph? I don't know where I went wrong.

jim_cheng

topwon posted on 2018-8-21 16:21 If you want to measure the waveform of V+ in the differential signal, then the ground clip of the oscilloscope should be connected to V- instead of GND, right?

Thank you for your advice. I measured the voltage between v+ and v- and it is true that there is no oscillation waveform. I measured the wrong test point. I now have the following figure output from pin 6 of u8. Do you see it? It should be stable.

jim_cheng

topwon posted on 2018-8-21 16:34 Also, how come the gain I calculated is -210 instead of -230 as marked on the graph? I don't know where I calculated it wrong

It is -210.

maychang

The period is 20ms, so it is very likely that there is power frequency interference, not necessarily oscillation in the circuit. You can measure the output end to see.

maychang

jim_cheng posted on 2018-8-21 16:44 Thank you for your advice. I measured the voltage between v+ and v- and it is indeed no longer oscillating. I tested the wrong point. The following figure of my u8 pin 6 output...

"The following figure of my u8 pin 6 output" This is the high-frequency interference superimposed on the power frequency.

jim_cheng

maychang posted on 2018-8-21 16:52 The period is 20ms, and it is very likely that there is power frequency interference. It is not necessarily that the circuit has oscillated. You can measure the output end to see.

The output is indeed unstable. I don’t know if this is normal or caused by interference. The picture is the output picture on the 5th floor.

jim_cheng

maychang posted on 2018-8-21 16:54 "The following graph is output from pin 6 of my u8" This is high-frequency interference superimposed on the power frequency.

What should the normal output be like? Is it a stable straight line?

topwon

jim_cheng posted on 2018-8-21 16:44 Thank you for your advice. I measured the voltage between v+ and v- and it is indeed no longer an oscillating waveform. I tested the wrong point. The 6th pin of my u8 now outputs the following figure...

Because I don't know the situation of your thermocouple input, what should the detected signal be like? A fixed level? Does the thermocouple connection line pass through any interference source? Are the V+ and V- signal lines twisted pairs? Is the ground clamp of the oscilloscope consciously reduced inductive area, such as using a spring pin instead of the ground clamp to ground it near the 6th pin. Can a small capacitor be added between V+ and V- to try?

maychang

jim_cheng posted on 2018-8-21 16:55 The output is indeed unstable. I don’t know if this is normal or caused by interference. The picture is the output picture on the 5th floor

“I don’t know if this is normal or caused by interference” Judging from the amplitude, this is caused by interference.

jim_cheng

topwon posted on 2018-8-21 17:07 Because I don't know the situation of your thermocouple input, what should the detected signal be like? A fixed level? Does the thermocouple connection line go through any...

My thermocouple does not use twisted pair wire, and it is 1m long. I added a 1nf capacitor between v+ and v-, and I can't see any obvious change in interference

maychang

jim_cheng posted on 2018-8-21 16:55 The output is indeed unstable. I don’t know if this is normal or if it is interfered with. The picture is the output picture on the 5th floor

Since the input is a thermocouple, its change speed will not be very fast, that is to say, your amplifier circuit does not need a very high bandwidth. Why not take some measures to suppress high-frequency interference at the input, such as low-pass filtering?

maychang

jim_cheng posted on 2018-8-21 16:59 What should the normal output be like? Is it a stable straight line?

For normal output, since the thermocouple signal changes very slowly, the output should be close to a horizontal straight line, and no fluctuation should be seen.

jim_cheng

maychang posted on 2018-8-21 17:29 Normal output, because the thermocouple signal changes very slowly, the output should be close to a horizontal straight line, and no fluctuation should be seen.

OK, thank you, I will add a filter circuit

topwon

jim_cheng posted on 2018-8-21 17:26 My thermocouple does not use twisted pair wires, and it is 1m long. I added a 1nf capacitor between v+ and v-, but I can't see any obvious change in interference

For a 1m long weak signal differential trace, it is recommended to make it a twisted pair, so that the differential signal can play a role in resisting common mode interference. You can just twist the two wires yourself, and you don't have to consider differential impedance or anything, just make the twist distance as tight as possible.