Please critique my fault tester circuit

wayneh

Joined Sep 9, 2010
17,496
Adding a little hysteresis to your comparator might help with the chatter. Something like a 1 to 5MΩ resistor from the output back to the input.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
I tried 2Mohm to + input, and I tried 1Mohm each to - inputs, and did not notice a change in relay chatter.
 

wayneh

Joined Sep 9, 2010
17,496
The hysteresis resistor goes back to the non-inverting, + input. The idea is it adds a bit of the output - either state - to the input, but at high impedance. That requires the input signal to overcome the feedback before it can trigger a flip to the opposite state. Try a 100k and see if you can tell a change.

Another source of chatter that I've fought with was due to wimpy ground wires for the load, like maybe a shared ground wire for the load and for the PCB. So every time the load would start to kick on, "ground" would shift up a few mV. Doesn't sound like much, but adding a simple wire fixed it. Always good to isolate the brain from the load as much as possible.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
The hysteresis resistor goes back to the non-inverting, + input. The idea is it adds a bit of the output - either state - to the input, but at high impedance. That requires the input signal to overcome the feedback before it can trigger a flip to the opposite state. Try a 100k and see if you can tell a change.
Will do.
Another source of chatter that I've fought with was due to wimpy ground wires for the load, like maybe a shared ground wire for the load and for the PCB. So every time the load would start to kick on, "ground" would shift up a few mV. Doesn't sound like much, but adding a simple wire fixed it. Always good to isolate the brain from the load as much as possible.
That's probably it. The whole thing is powered with 30-something AWG wires. Will upgrade tomorrow

Thanks for all the help BTW.
 

Thread Starter

strantor

Joined Oct 3, 2010
6,782
I put my scope between the GND of my relay switching transistor and the GND of my power supply, and noticed a 1.2V noise while the oscillations were occuring. I replaced the ground wire with a better one, and got it down to 1V. I replaced the V+ wire and they went down to .5V, but still oscillating. Tried the 100Kohm hysterisis resistor, no difference. I put .1μF caps between my opto output and +V and oscillations went away. Funny, now I remember having the same problem with my arduino model; when I was watching the input from the optos, it would swing wildly (maybe 100 points out of 255) before I put the caps in; afterwards I got a nice steady output. Not sure why I decided to omit those caps in the new design. So it's working good now, just need to draw up & (try to) make a PCB for it.

Thanks again!:)
 
Top