Spontaneous Reset after Setting Bit on IO Port

[Chris Ruff]

Oh- and one more thing- as soon as you touch a signal with your scope you have dampened or pulled down (lightly) the signal and you won't, therefore, be able to see what you are looking for

its the o-o-o-o-ld observer effect, kinda like heisenberg only different..

Chris

[DBrunermer]

Chris, are you suggesting that I try staking everything above the board and see if it makes a difference? Meaning, put stakes in all five holes locales, and see if it works with the relay way high?

[Chris Ruff]

No, as far as I can see, you have performed all of the detective work already. If I were you I would review all of the high impedance signals all around the relay coil and pull them one way or another. If the module itself is real close and there are no signals to dampen, you may need to either remote the relay off of the board or wind a steel shield around the coil to suck up the flux or put a steel box over the relay. You get the idea..

Chris

[pbreed]

Ok so the relay is ok as long as it does not switch the load, how is the load powered?
Is it powered by the same source or battery as the Netburner?

Do you have a snubber diode on the relay coil for when it turns off?

Does the relay turn on and stay on or is it a pulse?

IE is it possible that it is the relay turning off that's a problem...

Try replacing the relay with a mechanical switch in the exact same physical position and see if
switching that causes your problem...

[DBrunermer]

Sorry I didn't get back sooner, I was off work for a couple days. Back Problems. Age sucks. Anyway.

I decided to try the staking the relay high off the board on wire-wrap socket pins, even the load. The first time I engaged the relay, it worked perfectly, and I got excited. Unfortunately, it only worked once, and now it resets every time.

pbreed - The load is actually just the top of an h-bridge. It is an interposing relay, and it just turns a 100VDC signal on and off. The power itself is created in a separate linear supply. The board is powered by a +24V supply, going to DC/DC stepdown converters. The relay does have a coil snubbing diode, but it's not pulsed; it should stay on the whole time. I will try putting a simple switch in where the relay is and I'll tell you how that turns out.
Dan B.

-- EDIT --

I didn't have a switch, but I was able to simulate a switch with some wire and clips. Oddly, switching the voltage this way did not cause a controller reset. And, switching the relay coil without having the terminals connected still worked too. So, I guess I'll just have to give up on the cause, but I have found a workaround. I had a zero-ohm resistor inline with the high-voltage, so I pulled it out and put a ten ohm resistor in there. Now it seems fine.

Thanks for your help, everyone. I thought someone would see the 5ms reset delay and say, 'Oh yeah, I know what that is.' Clearly it's not a netburner issue. But thanks for hanging out anyway! Dan

[jking]

Try adding a resistor, 47 to 100 ohms, in series with the MOSFET gate to decouple the gate capacitance from the processor output. Connect it between the pulldown resistor and gate.

I have also seen the start-up step phenomenon you described. When driving an inductive load, like your relay coil, I saw volts of very high frequency ringing on the power supply that no combination of bypass caps would correct.