Urgent help with noise in a isolated dc voltage sensor

[joao.caracas]

Hi,

Can any one help me to identify and solve noise sources from this circuit?
It is a isolated DC voltage sensor, it uses a voltage divider to reduce 375V to 2V aplied to the input of a ACPL-C87A, the 2V output of the ACPL-C87A goes to a LM358 filter with gain to convert it to 3.3V applied to an ADC. There's approximately 100mVpp on the LM358 output.
How can I improve this?



- - - Updated - - -

 

[mrln55]

You might start with bypass capacitors on pin 1 of the ACPL-C87A and pin 8 of the LM358. Also, an image of the circuit layout would be helpful

 

[KlausST]

Hi,

show scope pictures of the noise you see. With timing and voltage range information.

**

Diode at OPAMP output.
It´snot good designe to short circuit an OPAMP output to limit output voltage.
* Either use a voltage limiting feedback circuit, (better solution, because there is no OPAMP saturation)
* or use a series resistor between output and diode.

Klaus

 

[andre_luis]

How did you measured that, by DVM or Scope meter ? It is important if you can determine if this noise has a fundamental frequency determinable, or if seems like a white noise. In addition, you could inform how the circuit was tested, if in a PCB properly coppered by ground plane, or if it was assembled at a breadboard.

 

[c_mitra]

First you need to identify the source. Please try the following the sequence:

1. Put a 0.1 or 0.01uF capacitor of sufficient voltage capacity across the terminals marked BAT. That will tell you whether the source itself is noisy.

2. Put similar bypass capacitors across the power supply (5V). Most of the modern ICs reject strongly power supply noise but still it is a good check.

3. Shift the diode after the resistor. Are you trying to clamp the output? The diode can also produce noise and test once without it.

4. Is 5V-OPT2 referenced to GND-OPT?

 

[schmitt trigger]

In addition to what everyone has mentioned...does the noise that appears in the output, is it independent from the high voltage level? In other words, if you have zero battery volts, does the noise change?

Could there be board leakage?