Silly workbench noise problem.

I have recently put together a xtal oven using a ad9850 and a pic12f629, it works well.

However if I pull a plug on my 6 way power socket on my bench the software in the pic reboots, I have a 100uF cap (with a 0.1uF in parallel) on the input to the 7805 voltage reg in the oven, and a 10uF on the o/p, theres 0.1's across the pic and ad9850.

I 'scoped the 24v supply to the oven, theres just a few mV of noise and it doesnt seem to change when i pull a plug, the oven itself has a grounded metal case.

I'm still thinking if noise isnt conducted and if radiated noise is screened, why is the pic rebooting.
I just tried my usb 'scope which is also in a ally box to see if i could see anything better and it too crashes whel i pull a plug, I'm starting to think theres something up with the bench's power supply.

Things that I would check are:

1 The mains earth at the wall socket is OK
2 The mains earth is OK on the 6 way power strip
3 The mains earth is OK on all the items on the bench. Beware of things with wall wart power supplies (including laptop PCs), the DC outputs, although properly isolated are sitting at 120v due to high impedance capacitive coupling.
4 Check that there are no floating inputs on the PIC, setting them as outputs sorts the problem.

That is all that I can think of for now.

JimB

Pepper: I had a weird one like that a couple months ago. Seemed to be noise related as improving the supply grounding reduced the prob a bit. Turned out to be a temperature PCL lookup table over run in the PIC code that happened sometimes based on a sample (thermistor) being out of range. If u have any of those asm lookups, disable them for a bit and see if the 'reboot' goes away.

I dont have a earth bond checker here, but my check plug says the earth is ok, and the extension has 100mohm resistance at 1 amp.
My old 'cope doesnt have the earth connected, but the probelm is still there with it unplugged.
The capacitive charge of wall warts is something to t hink of, but I'm not using any right now, unless something else in the house is coupling back through.
All of the pins on the pic are connected, its only an 8 pin, however 2 of the pins used for serial are connected to my usb to serial converter, and there are long wires unscreened running over the bench, maybe they are picking something up, also there is not protection to the pic on these lines, I thought about a resistor and a pair of diodes to the rail, if I can find room for them.

I do use lookups using the addwf pcl,f command, however I always check in code that the value doesnt jump off the table or the table goes off the page, I've been got by that before.

I'll try again, setup the o/p freq then i'll short the data i/o lines to ground then pull soem plugs, if it goes away its the data lines picking up junk, thanks for the advice on that one.