How to solve the signal fluctuation from the sensor?

sfcsdc

How to solve the signal fluctuation from the sensor? [Copy link]

 

The signal from the sensor is very weak, and the circuit behind it has been amplified and filtered. When the detection quantity remains unchanged, the amplified signal has certain fluctuations. Looking at the waveform, it is found that the signal from the sensor is fluctuating, as shown in the following figure The yellow is the original signal, and the blue is the amplified signal. How can I suppress this fluctuation? Is it possible to solve it in the circuit, or can I only do software compensation?

maychang

The sensor must convert a physical quantity (such as pressure or temperature...) into an electrical quantity (such as voltage or current or frequency...). What physical quantity does your sensor convert into what electrical quantity? If the result of the conversion is frequency, then you can shape the fluctuating signal. If the result of the conversion is voltage, then you have to take other measures and consider why the converted voltage is unstable.

topwon

It is recommended to focus on checking the power supply of the sensor chip, such as power supply ripple, decoupling design, etc., and also check the grounding method and position of the sensor chip to avoid interference caused by external ground current. Finally, you can consider adding some shielding measures.

sfcsdc

maychang posted on 2019-6-3 09:11 The sensor must convert a physical quantity (such as pressure or temperature...) into an electrical quantity (such as voltage or current or frequency...). Your sensor...

The sensor is a thermopile, which converts the infrared light intensity of a certain band into a voltage signal. It may be that the optical path is too long and the air flow interferes with the signal. Can the original signal be stabilized through the circuit?

sfcsdc

topwon posted on 2019-6-3 09:36 It is recommended to focus on checking the power supply of the sensor chip, such as power supply ripple, decoupling design, etc., and also check the grounding method and position of the sensor chip, ...

Well, there may also be electromagnetic interference. Thanks for your reply

topwon

sfcsdc posted on 2019-6-3 09:57 The sensor is a thermopile, which converts the infrared light intensity of a certain band into a voltage signal. It may be that the optical path is too long, and the air flow interferes with the signal. May I ask...

It may be that the optical path is too long, and the air flow interferes with the signal------If the input signal source to be collected is unstable, then there is no way for the back stage. I think it is relatively simple to use algorithms to process it, such as averaging, filtering and other processing.

topwon

Try to extend the time to see if this fluctuation is cyclical or is it decreasing steadily? Is it caused by the environment or the temperature rise of the chip? In addition, is there a problem with your amplifier circuit? It seems that the center line is moving downward. Is there no DC bias? Or try changing the op amp circuit to one with positive and negative power supply?

sfcsdc

topwon posted on 2019-6-3 10:12 It may be that the optical path is too long, and the air flow interferes with the signal------If the input signal source to be collected is unstable, then the post-stage can't do anything about it...

Yes, I just fanned the air next to it, and the signal dropped immediately. Of course, electromagnetic interference will also affect it, but the electromagnetic interference circuit can still be solved. If it doesn't work, add a shielding cover. But air flow will also affect it, there is really no way to solve this problem.

maychang

sfcsdc posted on 2019-6-3 10:28 Yes, I just fanned the air next to it, and the signal dropped immediately. Of course, electromagnetic interference will also have an impact, but the electromagnetic interference circuit can still be solved, no...

If this phenomenon is caused by air flow, there is no good way. Note the reply from topwon on the 7th floor, "Fanning the air next to it, the signal dropped immediately", this phenomenon may be caused by the increase in chip (sensor) temperature.

sfcsdc

topwon posted on 2019-6-3 10:20 If you extend the time and see if this fluctuation is cyclical or has been decreasing steadily? Is it caused by the environment or the chip's operating temperature rise? In addition, your...

I think I found the reason.

Due to structural reasons, the sensor is not directly soldered to the board, but is lifted up a bit before being soldered to the PCB .

This leads to two problems:

One is that it is very susceptible to electromagnetic interference because the input signal loop is very large. No matter how short the signal path on the PCB is, it is useless because there is a large loop outside the board.

Also, I mentioned before that air flow affects the signal. I found that the signal fluctuates when I blow on these pins. I don’t quite understand why the signal fluctuates when I blow on these metal pins? There are charges in the air, and the rapid flow generates voltage on them?

maychang

sfcsdc posted on 2019-6-4 08:47 I think I found the reason. Due to structural reasons, the sensor is not directly soldered on the board, but is lifted up a section before being soldered to the PCB. ...

The pin and solder are two kinds of metals. When the temperature is different (there is a temperature difference), thermoelectric potential will be generated (forming a thermocouple). This may be the cause of the fluctuation. For precise measurement, the temperature must be the same everywhere.

sfcsdc

maychang posted on 2019-6-4 10:13 The pin and solder are two kinds of metals. When the temperature is different (there is a temperature difference), thermoelectric potential will be generated (forming a thermocouple). This may be the cause of the fluctuation. Precision...

Oh, that's it, I understand. I was still looking at thermocouples and thermopiles before, but I didn't think about this. Suddenly I see the light

maychang

sfcsdc posted on 2019-6-4 11:41 Oh, that's it, I see. I was looking at thermocouples and thermopiles before, but I didn't think about this. It suddenly dawned on me

This is just a possibility, not a certainty. To confirm that this fluctuation is caused by the change in temperature difference, further experiments are needed.

叶春勇

Is this noise caused by DCDC?

sfcsdc

Ye Chunyong posted on 2019-6-5 12:32 Is this noise caused by DCDC?

No, the power supply has been checked.

jack160801

OP, I recently encountered the problem of unstable pyroelectric output signal, which kept jumping. I think it was because I did not handle the power supply and output signal of the pyroelectric properly. How do you handle the power supply and output signal? Thank you!!!