I am using an ADG731 32 channel mux to monitor 28 points in a system and transfer the levels to a serial A to D. the duty cycle is about 2.3 seconds whilst thetest takes in the order of 164 ms. sometimes during the cycle the addressed channel pulls the applied input voltage to zero for the dwell time of the sample and hence the reading presented to the A to D is zero.
Are there any known issues of this kind and how can i correct the way we address the device, as it appears to be associated with the addressing and chip select.
What do you mean by:
Peteratcomss said:thetest takes in the order of 164 ms. sometimes during the cycle the addressed channel pulls the applied input voltage to zero for the dwell time of the sample and hence the reading presented to the A to D is zero.
Would you tie the addressed channel to GND or would you leave the channel floating?
Can you share a schematic or block diagram so I can understand the circuit better? What will be the load at A to D?
The mux O/P is connected to an A to D converter ADCS7476 (Nat Semi) via a 10 Ohm resistor with diodes to Gnd and +5V. The I/P to the mux is via a resistive potential divider chain reducing 15V to <5V in order to range the value properly for the A to D 36K/16K to Gnd. The mux channel is nominally at =3.2Vdc, but sometimes, when addressed to read, the voltage is pulled by the MUX to around 0V.
This circuit has been in production for over 6 years and we have seen a fault like this before and it was cured by changing the MUX chip, but this time the problem has not been fixed by the change and we have 2 units exhibiting the problem, and we are really struggling to get to the cause?
The SPI speed is quite slow (100Khz clock).
I will try and get a circuit to you.
If you can provide an email address I will send it to you directly.
You may send a private message to me here in engineerzone and I will reply my email address.
From here we can take your concern offline.
This thread has been taken offline.