Fusor@40kV+Electronics=Danger

[Jon Rosenstiel]

I believe I have killed about as much stuff with my fusor as I did with a Tesla coil. Latest victim was my Canberra mca. OUCH! (Luckily, I've got a spare).

With 40kV applied to my fusor any small hv “hiccup” creates a large, high frequency pulse of EMI.

After the mca went south I decided to do some snooping. My scope indicated over 500V peak to peak centered around 30mHz on my well grounded fusor shell. YIKES! At the input to my mca the signal was around 8V peak to peak. (With no power applied to the preamp or the pmt).

I’m open to suggestions on how to tackle this problem.

Jon Rosenstiel

[3l]

Hi Jon:

Sounds like what I do in pulsing.
I use torriods wrapped to protect scopes and sensitive stuff.
It's the fast risetime that destroys most stuff.
A ferrite torriod will block fast hv stuff while letting signal pass.
My first Ebay Oscope burned a preamp before I discovered torriods.
All caps blown....lots of magic smoke!

Happy Fusoring!
Larry Leins
Fusor Tech

[Frank Sanns]

Jon,

I do not know if you are using a filter capacitor in your HV feed or not but there can even be quite a bit of capacitance in a high voltage transformer. When the fusor demands more current, the capacitance supplies it very quickly giving EMP or brief RF oscillations. This is still another reason why I use a ballast resistor just before the fusor. I know you are shooting for higher powers so you may add a choke just before your fusor. This way all the DC current you need will flow but no pulses will pass. Just a few turns of wire should solve the problem. It will also damp some of the oscillations that start within the fusor.

Frank S.

[Jon Rosenstiel]

Larry, Frank:

I don't use a filter cap, but I do use 10 feet of 100kV x-ray cable. At 40pF/foot that's 400pF. Plenty of capacitance to cause trouble.

I think I'll attack this from both ends, ballast resistor or choke at the feedthrough and ferrites on the coax cables.

Thanx for the ideas.

Jon Rosenstiel

[Richard Hull]

I ran into this with the 100kiloamp water arc gun work. I remoted the sensors from the gear with a single fiber optic line and moved the expensive stuff way back from the source. Inverse square law gets 'em every time.

Unfortunately, the data rate out of a Scintillator is high. I don't know if it could be easily optically coupled from a preamp and driver to a fast enough LED. Maybe a PIN diode setup would cut it.

Richard Hull

[Richard Hester]

If it were a simple digital setup, there would be no real problem with fiber optic transmission. However, I would worry some about linearity when transmitting fast analog signals. One possibility for isolation would be a high quality pulse transformer (ferrite is good enough for NaI output pulses), with a shield outside and between windings that would help to block common mode stuff, expecially in conjunction with a common mode filter on the signal cable at the MCA. As Larry suggests, this can be as simple as a a large ferrite with a few turns of the signal cable through it.

[DaveC]

Protecting sensitive instrumentation around HV circuits is not a trivial problem.

With the 40 kV Xray cable as an "HV filter", you have inadvertently made yourself a big transient problem. Under HV fault or breakdown conditions, possibly arising from within the fusor, travelling waves are launched in the cable, which then are often doubled at the impedance mismatch between the cable and the fusor, whose characteristic impedance might be several times the 40 -50 ohms of the Xray cable.

The current limiting resistance will do a lot to tame this beast. A common value used with Xray equipment is about 1ohm per volt. So a 40 kV supply would have 40 k Ohms series resistance in the HV output lead. It needs to be placed between the HV cable and the fusor, so it can damp out the voltage surges. The resistor, however, needs to be capable of withstanding at least the normal output voltage of the supply without flashover. A safety margin of 2 times the normal voltage is wise, since you have a low surge impedance line feeding the fusor (load).

DIgi-key sells a ceramic non- inductive resistor with about 7 kV surge rating (when potted). Make a string of these to achieve your needed withstand voltage and either pot the entire assembly in silicone rubber or epoxy or put it in an oil bath.

This should do the trick to greatly reduce the induced voltages everywhere. Vind = L di/dt and you have just reduced di/dt by about 1000x. That has to help.

Happy damping...!!

Dave Cooper

[Jon Rosenstiel]

I've fixed my problem by doing the following:

1. Shortened the hv supply cable from 10' to 5'.
2. Removed the hv cable outer shield.
3. Re-routed the mca cables away from my power controller. (Big help)
4. Installed clamp-on ferrites on all mca cables. (Another big help)

The best news is…. My mca is back in action! All I had to do was to pull the cpu board and reset the memory with the shorting jumper.

My thanks to all for the great ideas.

Jon Rosenstiel

[jst]

Think vacuum tubes

A common practice where intense radiation or EMI is a problem ( and fusors certainly rate here ) is to use vacuum tube circuits in the area and transmit to solidstate equipment further away in faraday boxes.