EMF/Noise Resetting my ZX24p! Do I need filters?

[everest]

Hi everyone,

I've got a fairly simple circuit set-up as indicated in the attached schematic. While it's performed well in most cases, I have a user that is operating some heavy duty AC relays with the smaller 12v relays on the schematic, which are in turn operated by a ULN 2803A transistor array, which is in turn activated by an I2C I/O chip, which is in turn operated by the ZX24p processor. Whew! So here's the problem:

When the user cycles the AC relay, the entire board locks up, or sometimes just resets itself. We believe the culprit is likely noise from the AC relays which in turn are driving some heavy duty A/C motors. The user has installed MOV noise suppressors across the coil of the A/C relays which helped a LOT, but he's still occasionally seeing his ZX24p reset itself.

There are number of inputs also running around the observatory, that are read by another I2C I/O chip. Those input lines are in shielded cable, but do run in parallel to the A/C line that drives the coils to the larger relays.

I'm just at a loss in terms of helping resolve this problem. Do I need to build filters for all the input lines coming into my board? Is that best practice? Any suggestions? I'm stumped!

The schematic is attached if it helps. . .thanks!

-Jeff

[dkinzer]

When the user cycles the AC relay, the entire board locks up, or sometimes just resets itself.

Until you know what is causing the problem you're just grasping at straws. It could be ground noise or it could be noise on the supply rail or it could be induced noise on signals unrelated to the relays. To get started understanding the problems and some solutions, you might want to study EMI Design Techniques for Microcontrollers in Automotive Applications.

It would probably help to add optoisolators between the ZX-24p and the ULN2803. That would electrically isolate all of the relay circuitry from the other logic circuitry, eliminating the possibility of ground or Vcc noise due to switching. In an electrically noisy environment, you may find that you need to enclose the logic circuitry in a metal can thus shielding it from the induced noise. Tom Becker may have some additional ideas since he has used ZXs in a marine environment, at least as noisy (if no more so) than an automotive environment.