S.T.A.R. Instability hypothesis

[Steven Sesselmann]

Hi Guys,

The actual pleasure of physical experimentation is short lived in the fusion game. My 100 Kv power supply had to go back to Florida for repair..

In the mean time, I am trying to analyze and make sense of the little data I have.

In past experiments with the S.T.A.R. reactor I have found the system to be much more unstable than what the regular fusor seems to be, and at 60 Kv I found that it was behaving like a raging bull. As some of you have seen on the video's the cathode appears to be spitting flames.

After putting some thought into the theory of mean free path and gas pressure, I think I now understand what is happening.

The ion beam is heating the gas inside the cathode, which tries to escape through the apertures, but is held back by the ion beam itself, until the pressure eventually exceeds that of the ion beam, causing the cathode gas to discharge through the apertures and into the center of the accelerator tubes.

The sweet spot for best fusion must therefore be that point where the pressure reaches a maximum, ie. when the ion beam pressure equals the pressure inside the cathode.

The cathode pressure is therefore proportional to voltage * current both which are perfectly controllable.

By reducing the size of the aperture in the cathode, the pressure can be maintained at a high level without a large current.

As there is no theoretical limit to the voltage that can be applied to a S.T.A.R. cathode, the fusion rate in the cathode could in theory be increased without limits, using a very small current.

I think that is what we would like to achieve...isn't it

Thoughts....

Steven

http://www.beeresearch.com.au

[Starfire]

The hypothesis should be easy to test - increase the beam current and check if the sweet spot changes proportionately. ( but how to determine the sweet spot ?

[Richard Hull]

Modern supplies don't like current kick back bursts. Most are designed to a specific continuous maximum current and voltage, smoothly delivered. They have some protection in place of course, but the kind of stuff you are doing is just a 100% beatup on the circuitry.

You specifically need a robust, beefy, low impedance supply that has no electronics in the output legs if you feel you must go from no current to huge impulse levels. Electronics can be designed to handle this, but your supply doesn't possess such electronics.

Your device is not a smooth functioning system. Regular fusors are often not smooth functioning either, but I do believe they do not suffer the energy excursions that your system presents to its power supply.

Once repaired, prepare yourself for blowing it up again.

Richard Hull

[Steven Sesselmann]

Richard,

John Hendron has been telling me this right from the beginning, that I should have a more robust power supply.

I ponder if there is anything I can do or build, to protect the power supply I have?

I attach one frame from one of the tests, showing the "spitfire" effect I am referring to.

Steven

[Richard Hull]

I wish I could advise, but I have no schematics of your supply.

I would have worked my own supply up from scratch using an x-ray transformer and put a ballast resistor in it as most here have done and left the nice professional, super regulated supply for more kinder-gentler experiments.

You don't need a Ferrari here just a smelly, angry bull

Richard Hull

[John Futter]

Steven
Tell the supplier what you are doing link them to your results so far.

As to a cure (speculation get the Florida guys to give an opinion)
Near the supply ie minium coax mount series resistor equal to output impedance of supply. Also make up chain of Transils unidirectional equal to max voltage ( i think that the negative swing of the pulse with respect to polarity in your case negative so positive excursion {ringing}is mortally upsetting your supply.
use plenty of coax after the resistor to give more capacitance so that the fusor maximum demand in current pulses is satisfied by the storage in the coax.

Mount plenty of ferrites on both coaxes output to resistor and resistor to fusor to slow own the wave front that the PSU sees.

Will be interested in the replies especially from the makers of the supply

[Mike Beauford]

I know what they'll say. Don't use the power supply like that.

[DaveC]

Steven -

I don't recall how you have your HV output circuit configured. Is it connected to your fusor through some sort of ballasting resistance (or impedance), or do you have a constant current supply? The CC supply is the more sophisticated form of a ballasting circuit, using electronics to provide the current control. But if the control loop bandwidth is not large enough, the circuitry may not be able to stay on top of the current fluctuations.

Operating a high voltage supply in constant current moded is common. Almost every Spellman HVPS is capable of stable operation in this mode. But the circuitry does need very robust design and a plethora of snubbers, of whatever configuration to work well.

A common problem is differences in bandwidth between the High Voltage Divider and current sensing circuitry. These need to be very fast and well matched, otherwise current control will occur at the wrong times, and possibly enhance rather than suppress the transients.

Were you able to determine what died on your HVPS?

This kind of trouble shooting sorts out the men from the boys.... hang in there.

Dave Cooper

[Steven Sesselmann]

John, Mike and Dave,

I am still a rookie at this HV stuff, but with friends like you, I am learning fast

I attach a picture of my power supply with the lid off, I love it when the damage is visible, that way I don't need to be an expert to find it.,,,, see there it it is!

An arch burnt a small hole through the silicone insulating material that covers the HV circuits, and went to chassis. the PSU still works, but it archs through the hole at 10 kv. as the unit is still under guarantee, I decided not to attempt any surgery on the circuit insulator.

The power supply is designed with an electronic current limiter, or I can switch to overload, which simply cuts out the supply at 6 ma.

I have around 2.0 meters HV coax going to a 400 K ballast resistor in a tub of oil, then around 1 meter of coax core without the shield, going from the resistor to the reactor.

Steven

[Frank Sanns]

Looks like a back EMF condition. This is the same principle that the ignition in a car works. Energy is stored either in a capacitor or inductor and when the circuit OPENS, a large voltage that by far exceeds the charging or original voltage can appear in the circuit. The voltage rise can be so rapid that the internal resistors or bleed down circuit can not deal with it an you see the result.

Frank S.

[Steven Sesselmann]

Frank,

My fusion reactor is essentially a spherical capacitor with cathode radius of 0.03 m and a anode radius of 0.15 m.

I found a formula on hyperphysics to calculate the capacitance;

http://hyperphysics.phy-astr.gsu.edu/hb ... apsph.html

putting the numbers into Google gave a capacitance of;

(4 * pi * 8.85418782e-12) / ((1 / 0.03) - (1 / 0.15)) = 4.17243771 × 10-12

I think that says 4 pico Farad, which at 100 kv it would store a charge of 4e-7 C.

...and I have no idea what to do with all that information or if it helps me at all

Steven

[Frank Sanns]

The inner and outer grids are only part of the capacitance but if you just look at a 4 pF sytem, it still can be problematic.

Energy= 1/2 Capacitance * Voltage squared so that is 0.02 Joules of energy. If this is switched over a 1 second interval then no problem because it is equivelent fo 0.02 watts. That is easily dissipated by circuitry. If however it is switchen in a millisecond then that same number of joules acts like 2 watts and in a micro second it is like 2,000 watts for that micro second of discharge. By Ohm's law, if the resistance is fixed, then the voltage will track up with the current.

I believe your capacitance is much higher in your system so you can see that you might have been bouncing a megawatt or so back into your sytem with voltages briefly spiking at 50% to 1000% or more above your operating voltage.

Frank S.

[Jon Rosenstiel]

Steven-

As I recall you said your HV cable was coax. If so that'll add a bunch more capacitance. For example, the capacitance of 100 kV x-ray cable is about 160 pF / meter.

Jon Rosenstiel

[Steven Sesselmann]

So if I read you guys right, should I try to reduce the capacitance, by using an unshielded cable?

Incidentally I have an old 100 kv x-ray cable (ebay scoop), it is a thick heavy rubber cable and I don't think it is shielded...any good?


Steven

[Mike Beauford]

I'm thinking the shortest possible cable you can make from the power supply to the fusor the better.

[Jon Rosenstiel]

Keep the lead as short as possible as Mike suggests, and ditch the shield if at all possible. X-ray cable (at least all x-ray cable I’ve seen) is shielded, but it’s easy to strip it off. If you do decide to use the x-ray cable be sure to solder all three of the center conductors together.

Jon Rosenstiel

[DaveC]

Steven -

I recommend keeping the cable shielded, unless you are operating with a cable less than a meter in length. The shield properly terminated will help to limit the EMI that radiates when the plasma hiccups and burps... which it is prone to do.

But the cable terminations take probably up to a half meter for 100kV withstand and the possible surges, so a short cable won't allow being shielded. The shielding would raise capacitance it's true, but since the capacitance of a cylinder, varies according to the log of the shield to conductor ratio, it's not a huge increase, and what the shield does for your circuit, easily offsets the extra stored energy.

Be sure your ground system is complete, large conductance, and solidly connected... all bolted with soldered/crimped ring lug connectors in a star configuration.. where the central point is at the rear of your HVPS. This will do the most to reduce the induced voltages from ground loops when the transients happen.

I would have the ballast resistor between the HVPS and the HV Cable so as to make an HV low pass filter which will keep a lot of the transients out of the supply circuitry, and do something to help stabilize the plasma disturbances, somewhat.

Even better would be to have the ballast resistor in an insulated shielded enclosure, with the shields connected solidly to the cable ground on one side and to the supply on the other. This will tend to confine the EM surges to within the cable, resistor structures and not radiate so much of signal. But you have to keep the resistor's heat dissipation in mind.

An ideal design, would also match the characteristic impedances of the ballast resistor, cable, and fusor. This would eliminate reflected surges... and strongly limit the EMI. But that may not be feasible with your design.

Dave Cooper

[Steven Sesselmann]

Dave,

Thanks for advice..

The method used to connect my coaxial cable to the PSU, prevents me from simply stripping off the shielding. The PSU has a 200 mm deep hole at the back, into which the coax core with a simple banana plug, is inserted, a screw cap then connects shielding to chassis.

Connecting my multimeter to the PSU end of the cable, I measure 0.376 nf, significantly more than the theoretical 4 pf for the actual reactor.

I attach a simple diagram, and look forward to some suggestions

Steven

[UG!]

376nF!! thats HUGE! at 100 kV thats 1.9kJ! i suspect your meter is wrong. if its not, then no wonder stuff isn't working right

was the PSU connected when you made that measurement?

Oliver

[DaveC]

Steven -

Thanks for including a schematic of the HV output circuit.

I agree, you do not want to take the shielding off the HV cable from the HVPS. In fact and this applies to any fusor, if it were possible (practically speaking) to provide a full HV shielded cable lead all the way from HVPS to fusor, this would be best all around for stability, noise reduction and etc, etc. But the technical issues of building a shielded air to vacuum HV bushing are not trivial.

Your system would operate more stably with the resistor ahead of your hv cable, between the PSU and the line to the fusor. But unless the resistor can be put into a grounded enclosure, it's probably better left where it presently is.

As to the measured capacitance, Oliver's right, it does seem high. I would expect not more than 500 pF including the resistor's capacitance through the oil, and the fusor itself.

The DVM capacitance meters, I found to be either wildly off, or very close to measurements done with a capacitance bridge. Seems to depend entirely on the meter and its past life. Just check it with a few hundred pF ceramic cap. It should be within 10 or 20% of the printed value on the cap. Some have a zeroing adjustment, inside.

You can also get a good confirming estimate with a square wave generator, (even the calibration signal from your oscilloscope will do as a source). With a known resistance in series, (say 10 k ohms), the rise and fall times can be seen on the scope and you can work out the equivalent capacitance that way... It's a good way to check, if you have no other instruments. Just need a volt or two.. for the square wave signal.

One other possibility is trapped charge lurking in the HV components. Ground everything (short all HV conductors to ground) and warm things with a heat gun... Doesn't need more than about 10 or 20 C temp change... but more won't hurt. This will "empty" the trap sites..and eliminate any residual DC voltages, that the C meter circuit may be seeing. Of course you can also see this voltage (if it's there) with the meter in DC volts mode.

Testing each component separately will sort out where the strange readings are.

Have fun....all this analysis, sharpens the wits, and builds experience. (We all say this to ourselves, as we groan and head off to solve yet one more problem... )

Dave Cooper

[Steven Sesselmann]

Dave and Oliver.

Oops, that was a bit sloppy of me, the capacitance measured was 0.376 nf and not 376 as I wrote.

You guys have some good ideas, building the ballast resistor into a metal can and grounding it should be pretty easy.

Onother possibility with the STAR reactor, is to eliminate the hv bushing and run the coax straight into the cathode and connect the shielding to the anode. I just need to figure out a way to make an oil seal.

The cable will then be shielded all the way.... neat!

Steven

[Chris Bradley]

Just search old threads for something else, and came to this.

Isn't this just a case of installing a suitable serial inductor in your output circuit?

You may well have to make your own so that the windings are insulated and/or well spaced apart - a big ferrite, 15AWG and with the actual loops having a reasonable gap between to avoid the blocked voltage transient from sparking across in cascade.

Someone else's input would be needed on the inductor characteristics/design, I don't know enough about this to advise.

best regards,

Chris MB.

[Steven Sesselmann]

Chris,

The setup has changed quite a bit since this old post.

My new reactor configuration with the smaller anode and the smaller oil tank has been commissioned and I now have the HV coax running straight through the tank and into the cathode, with no bushing.

Today I have also eliminated the ballast resistor and installed a new coaxial cable between the PSU and the reactor, because I didn't want to break the shielding. After I did this, I realized that I had a terrible earth loop, which caused arching at the reactor end. To solve that problem I removed the earth wire to the PSU and let it be earthed through the shielding on the hv coax.

I can now reach 95 kv without any arching.

X-rays have been an issue, so I also built a small box to fit over the reactor and lined it with 1.2 mm lead sheet.

So with all that in place the neutrons should be pouring out...

Noop, the results are poor, sometimes I get a couple of bubbles and sometimes non at all, the best runs gave 5 bubbles.

The reactor has four different modes, and I still haven't figured out which mode produces the best fusion. The different modes are not that easy to control and maintain.

Flashing Plasma Mode [50-80 micron/20-50Kv/4-5ma]
This mode is typical of higher pressures, blue/pink plasma oscillating rapidly in tubes.

Smooth Plasma Mode [10-20 micron/30-60Kv/2-3ma]
Typical at lower pressures, not very exciting to watch, slight pink glow in tubes, speckles on video screen. Soft x-rays above 30 kv, but occasionally no x-rays evident outside of shielding, even at 50-60 kv.

Hard X-Ray Mode [5-30 micron/40-60kv/1-2ma]
Occasionally the smooth plasma mode suddenly changes into hard x-ray mode, in this mode, my shielding appears transparent, even at a distance of 4 meters, the GM pancake detector goes off the 1x scale and almost off the 10x scale(0.5 to 5 mr/h). In this mode x-rays bounce around the room, forcing me to cut power.

Afterburn Mode [10-20 micron/PSU switched off]
This mode resembles flashing plasma mode, and happens after a long (2 minute or more) run, when the ion gun is hot. After switching off the PSU, the plasma continues to flash almost indefinitely. This mode activates both the He3 and B10 electronic neutron detectors, in fact the B10 detector goes off the scale. Occasionally x-rays continue in this mode, which falsely led me to believe I had activated my shield, after my 24th December run. I have kept this mode going for 4-5 minutes. I really don't have a clue what is going on in this mode, this shouldn't happen.

Time to do some thinking...., ideas and opinions welcome.

Steven

[Chris Bradley]

An opinion, yep...make sure you got your lead underpants on.!...

(and that's not really meant to be a joke)

[Steven Sesselmann]

Chris,

Where I stand and operate the reactor, I get 40-50 micro sieverts for a few seconds, when the reactor goes into hard x-ray mode, not a big health issue, unless I fall asleep at the helm .
More importantly, I would really like to find out why the S.T.A.R design has such an aggressive x-ray mode, compared to the standard fusor.

Steven