When the fusor’s plasma becomes unstable and sparks, it makes my DAQ crash. This is likely due to transient voltage spikes in the wires, or electromagnetic interference through the air.
I’ve been exchanging emails with Raymond R. about the details of voltage suppression. He suggest using TVS diodes to protect each channel of the DAQ from voltage spikes in the wire by clamping them just above the channel’s operating range.
I’ve ordered these TVS diodes:
Raymond suggests the following:
1) Buy the NI USB-6000 Series Prototyping Accessory if you have the money and want to avoid messiness; otherwise you can kludge the wiring. (I’ve done this).
2) 16 (+ 4,5 spares) of LCE10A-ND (I bought corresponding parts at mouser).
3) Go to differential mode on NI USB 6008 (good idea).
4) Connect the TVSs from each input/output line to frame ground.
5) Take the cable shield and do not connect it at the USB board; connect back to the computer frame. The TVS diodes go to the cable shield not to the USB box grounds. Connect the USB grounds together and to the shield ground through a 1 meg resistor (alternately a .1 uf ceramic cap).
Prometheus Fusion Perfection 
Have you checked your building ground?
If you are not sure where your power grounds or what its quality is, look into driving a 25 mm galvanised pipe 4-5 feet into the ground outside (or two in locations separated by more than 20 meters) to be sure you have a rock solid earthing ground for the instrumentation. High voltage applications like these play havoc with instruments when the ground is less than perfect. Sounds agricultural but eliminates lots of head scratching and debugging later on.
Dude I’m in New York fucking City. I don’t think there is a patch of ground not covered with concrete for a mile!
But I’ll double check that.
Is there a way to test the quality of the ground you already have ?
See if you can trace your building ground first then. Do you even have a building ground wire in the lab there?
Are your cable shields, computer frame, etc connected to building ground?
If not, I’d recommend it … if you are sure building ground is of good quality. (hence my first question).
Have you considered using an opto-isolator to remove your test equipment from the circuit? Here’s an example: http://dkc3.digikey.com/PDF/US2010/P2602.pdf
They have ratings in the low kV range rms – which will probably handle most spikes. This in combo with your existing tvs’s would probably be about as good as you can do.
I wouldn’t think the ‘quality’ of your earth ground is all that important as long as everything in your lab is bonded to the same ground. You definitely don’t want multiple paths to earth ground on your equipment. Everything needs to be bonded to the same ground or isolated and never connected. I would suggest running a large ground conductor around the baseboard of your lab and bonding it to the main power entry point’s ground (i.e., at your breaker panel). Then all equipment cases would be grounded to this to minimize ground potential differences, and give transients a path to ground.
Good luck!
Brad
Can these opto isolators transferee analog voltage levels? I thought opto interconnects were mostly for digital.
[…] voltages spikes or EMFs crashing the data acquisition (DAQ) card. Today was a big test for the new transient voltage suppression system . It FAILED big time. But I learned something in the […]
BTW: On
http://media.photobucket.com/image/iec%20fusion%20polywell/Torulf2/IEC_Polywell.jpg?o=1
they have a chicken wire, probably copper, cage around the reaction chamber and magnets.