Mysterious Reset

[skimask]

skimask,

A quick look at the datasheet would have shown that the ULN2004 does have Output Clamp Diodes. Datasheet boy.

Joel,

Maybe you need a snubber on the relay driving the contactor.
I use .022uF 400V and a 47ohm 1/2w resistor in series across the output, but then I'm using SSR's

Works good for all the contactors I've used.
<br>

Ya I know, I just downloaded it and looked That'll learn me for not learnin' me...
But still, what if the '2004 isn't right next to the relays? What if it's on the PCB and the relay is X feet down the road on a cable.
Maybe an opti-iso would be the ticket in this case...still doesn't resolve the relay spike but it might isolate it a bit better...
How 'bout a decent cap across MCLR? If there's a pullup on MCLR, maybe it's being dropped out on occassion?

[Darrel Taylor]

Just wanted to add that skimask's idea of Solid State Relay's can also reduce that spike, by waiting for the Zero-Crossing point of the AC wave before turning on.

A relay can't do that, and may turn on at the peak of the AC, causing a very large spark and spike.
<br>

[keithdoxey]

Do you have MCLR tied to +5V directly or have you used a resistor ?

Or have you done what I did on a protoype board and forgotten it. My board had been in used for months as I was developing code and was working fine. I kept adding bits of code and other circuitry and then one day started getting spurious random resets.

I had forgotten a pull up on the MCLR pin and it was now triggering on noise picked up from the ratsnest of prototype wiring. fitted a 1K resistor and it has been fine since

[JoelMurphy]

OK OK OK

The 5V relay is about 1.5 inches from the ULN2004.
Darrel, what is the cap-resistor across the relay coil supposed to do?

I have a 10K R pulling MCLR high
I have a 1000uF cap across +5 and GND right at the chip, along with a .1uF there too.

I'm going in today to place diodes across the relay coils to see if that could solve the problem...

I imagine that some kind of back fluxing is the issue... but I don't understand how the spike could travel through the 'unconnected' parts of the relay, and how it could effect my battery powersupply...

Joel

[Darrel Taylor]

what is the cap-resistor across the relay coil supposed to do?

From what I understand, it limits the rate of change in the voltage across the contacts.

Here's a thread where someone had a similar problem. Snubber seems to have helped.
http://www.picbasic.co.uk/forum/show...hlight=snubber

[Jerson]

Joel

I am not sure if I can be of any help here. But anyway, here it comes.

- Have you looked at the Brown Out settings? Try to disable it for a while to see if that is the cause.

- another item you can miss out on is the layout of your board. May happen that the heavy current is flowing past your PIC and lifting the ground. I would expect the layout to follow this route if possible or else have a star point grounding
MCU -> ULN -> relay -> PS (ground / VCC) This keeps the heavy load near the Power Supply (PS)
In one case I had a problem where the layout was causing problems.

I dont know if you understand what I am saying. What has to happen is that the lowest ohmic path should be from PS to relay and then to the lighter loads.

Excuse me if these are lame suggestions and you have them sorted out

Jerson

[mister_e]

it's common to place a few hundred nF cap in parallel with the relay coils. For the contact, if you switch DC load, the link Darrel provide is nice.

If you switch AC load, you really want to switch them on the AC zero cross. This have tons of advantages.