Large load (solenoid, chiller, heater, etc) activation/deactivation kills LIFA USB communication. Why?

[jzzsxm]

Alright, here's the scenario:

I have Labview controlling the arduino just fine, no problems.  I have two 12V solenoids whose source is plugged into the same wall outlet as the computer/monitor/arduino power supply.  When the 12V solenoid is fired (opened) there are no negative side effects.  When the solenoid is unpowered and the rod falls back through the coil the RX and TX lights on the arduino go dark and the Labview program stops running. 

This is curious because the computer and the solenoids are in no way connected other than that they share an outlet.  We guessed that maybe the backwards voltage spike from the solenoid was making it all the way into the PC so we added a buffer cap to the solenoid and that solved the problem.

Now it gets weirder.

We have a chiller and a heater that were doing similar things to the PC/arduino communication so we plugged them into a completely different breaker but just now the chiller killed communication between the PC/arduino when it kicked on.  Again, it is on an entirely different breaker.

Does anybody know why this may be happening?  Solutions?  Explanation of the physics?

Thanks a ton!

[Nathan_B.]

How are these devices connected to Arduino?

[jzzsxm]

That's the thing, they aren't.

If you disconnect everything else, if you just connect the arduino to the computer and plug the computer into the wall and then plug the heater into the same wall, the heater will disrupt communications between the arduino and computer.  This behavior is repeatable. 

Less repeatable, but it has happened once, was when the heater and chiller, both turned on AND on a completely different breaker than the arduino/computer, interrupted communications.

It's weird.  Seriously, the only place they're connected is in the wall.

[jzzsxm]

Also, I'd be more than willing to shoot a little video to demonstrate.

[Nathan_B.]

That certainly is strange.  I wouldn't even know where to begin.

[Eric_BOB]

Hi

i have ever see something like that. In fact USB in pc are generaly protect from spike on their power lines, but some PC define those overshoots like default and cut definitively the USB connection and oblige you to reboot pc to restore it. The only solution in this case is to add galvanic isolation between PC and arduino. Something like that: http://electronicsshop.dk/?id=1038

[Kristoff2]

You are probably getting a ground loop which is causing a brown out.  This can be caused by some cheap hardware.  Once I was using my laptop to program a dsPIC on a robot via USB.  The robot was power by a cheap benchtop power supply connected to the same wall power line as my laptop.  When everything was connected the robot would brown out.  Unplugging my laptop (and using batter power) solved the problem.

My friend had the same problem with his Macbook and a projector on the exact same power line.  If the Macbook and projector were connected to the same wall power rail (the same one that cause problems with my robot) the projector would brown out.

-Sam K