Hello, colleagues! There is INA125 powered by plus 24V and -10V, created by ICL7660AIBA.
For some reason, the output signal without load is quite noisy, about 100mV, due to which the zero point is shifted. I don’t know what to assume, so I’m asking you for advice and recommendations.
The sensitivity of the strain gauge is 3mV/V, the gain factor is G = 337 (R = 180 Ohm.) I also attach a diagram.

 

Hi Max.
Welcome to AAC.
Please post your LTS asc file.
E

 

Never mind. Made an inaccurate statement.

 

What is the appearance of the output noise on a scope? Does it look periocic, digital or just random? The most vulnerable part of your circuit to EMI is at the input of the amplifier. How long is the wiring between the strain gauge and the IC? Are you using shielded or twisted wires?

 

What is the appearance of the output noise on a scope? Does it look periocic, digital or just random? The most vulnerable part of your circuit to EMI is at the input of the amplifier. How long is the wiring between the strain gauge and the IC? Are you using shielded or twisted wires?

I'll try to take an oscillogram tomorrow and show it. The board is in a prototype and yes, the wires from the sensor to the IC are about 10 - 15 cm. The wire is a regular copper multi-core (it came with the sensor). There is no screen.

 

Hi Max.
Welcome to AAC.
Please post your LTS asc file.
E

Do you mean LTspice model?

 

I'll try to take an oscillogram tomorrow and show it. The board is in a prototype and yes, the wires from the sensor to the IC are about 10 - 15 cm. The wire is a regular copper multi-core (it came with the sensor). There is no screen.

After you have made a record of the noise, try twisting the input wires, D+ and D- together and also the +10V and ground, making two twisted pairs. Then check to see if the noise has reduced.

 

Hi Max,
I already have the INA125 model I need your asc file.
E

 

Hello everyone, I was finally able to return. In general, yes, I figured out that it is most likely the sensor that makes such noise, since without it (with shorted terminals D- and D+) the noise is much less (something around 4-5 mV). Here I had another question, what is the best way to connect the strain gauge screen to the common Ground or just to the case? When I get around to the next prototype, I will write.

 

That noise is from the power supply. ICL7660 is not renowned for being quiet.

 

Use the scope’s cursors to determine the noise’s period.
To improve the measurement resolution, speed up the timebase at least 10X.
If it is around 10khz, then it is the ICL7660.

 

Why don't you find an instrumentation amplifier with an output that goes to the negative rail, then eliminate the negative supply?