Joe Khachan just sent me details on the power supply for the coils. They are designed to produce a brief high current pulse.
Joe says:
The part that took the most time to build was the power supply. I’ve attached a diagram that looks something like our circuit without dump resistor to dump charge of the capacitors when we need to service them. We used a hockey puck type of SCR (a type of thyristor) that can take 1000 A continuous or 10 kA pulsed. This may have been an overkill because we found that a maximum of 300 A was needed. However, we may need higher current as we increase the size of the polywell. The capacitor bank was make of 5 X 1500 microFarads electrolytic capacitors that can take a maximum of 450V all connected in parallel. You need some kind of transformer that can step up the voltage from the mains and be able to charge the capacitors within a couple of seconds. That means it shoud be a reasonably hefty transformer. We control the voltage output of the transformer with a Variac on the input side. You need to protect the SCR from back EMF with a diode across it. The diode should be able to take a few amps. The wire diameter about one millimeter and there were 10 turns per former.
Joe’s coil power supply looks like this:
Here is the preliminary bill of material (click for Mouser produce page):
1500 microFarads electrolytic capacitors max 450V
Silicon-controlled rectifier (SCR)
Diode Diode to protect the SCR from back EMF
Step-up Transformer Step-up Transformer
100 watt 2KΩ resistor for bleeding capacitor bank
Rack mountable chassis.
Please comment if you notice wrong parts.
Prometheus Fusion Perfection 
Hi, just looking at your circuit and materials.
A couple of points:
1) It seems you are building on wood? Unless you are planning only one shot every 10 minutes or so, double check on the thermal dissipation. Which means: take one shot and measure the transient temperature profile of the capacitors, diode, and SCR. Normally I would use my fingers; but that might not be a good idea at these voltages. Thermals are linear unless you commit a “thermodynamically irreversible” (burn it up) process; sorry about that but I have always loved that phrase. Linearity means you can separate and add individual heat pulses together to get the resulting temperature. If you want I can put together thermal and electrical (spice) models that run on one of the free programs. In the end the thermal model would require a thermistor/thermocouple time series reading though; or else rebuilding the unit with materials of know thermal characteristics.
2) The protection diode might be underrated for full protection. If inductive current is interrupted then the current continues to flow momentarily through protection circuits; at the same level. This means protection diodes and such must handle the peak current. One of the most common mistakes is to rate protection diodes for average current and not peak. During fault condition on the Polywell side you might get a reverse voltage surge back at the capacitors; causing problems. You might think about a reverse diode (full peak current) across your load.
3) To avoid to much grief during faults you might consider a fast acting hydraulic circuit breaker at the front end. The faster ones are fast enough to actually protect the components. If not fast protection; some fusing should be considered; regular fuses will typically protect the room if not the components.
Ray
Thanks for the feedback Ray.
The picture with the wood is from the Sydney team. I have not built this yet. Although I was planning on using a metal chassis, which would be better for heat dissipation.
I’ll spec a higher current diode.
If you use a lot of metal be carefull with the higher voltages. Teflon insulation and pads are fairly highly rated,flexible, and common. It’s been ages since I did work over 220 but…
Here is a list:http://www.allaboutcircuits.com/vol_1/chpt_12/8.html
Material* Dielectric strength (kV/inch)
===========================================
Vacuum ——————- 20
Air ———————- 20 to 75
Porcelain —————- 40 to 200
Paraffin Wax ————- 200 to 300
Transformer Oil ———- 400
Bakelite —————– 300 to 550
Rubber ——————- 450 to 700
Shellac —————— 900
Paper ——————– 1250
Teflon ——————- 1500
Glass ——————– 2000 to 3000
Mica ——————— 5000
I am sure all figures are for dry clean materials. I would use Teflon myself for insulation and barriers; but realize it “flows” if consistent pressure is applied. That is tightly wrapped around edges/corners.
Ray
Sorry, flows but is tougher than most plastics.
You need a reverse diode across the SCR. High speed, high current. Probably rated at 600V if you are going up to 450.
You need to build a voltage divider across the coil to watch the transients. A .002 ohm in series with the ground should do nicely for current. Get one of those flat metal jobs rated 2 W or 3 W. You don’t deed the power but it will help with large pulses.
And ramp your variac voltage up slowly. Energy storage goes up as the square of the voltage.
You can replace the relay with a large value resistor of suitable power. The reverse “kick” from the coil will probably turn off your SCR. Two diodes in series would probably guarantee turn off.
And if you are going to use a relay put the capacitor bank next to the diode bridge so the relay is not carrying high current when it closes. And add a little series resistance (50 to 100 ohms of suitable power). To limit the inrush current when you charge the capacitors.
OTOH 5K resistor in series with the bridge might be better from the standpoint of keeping the pulse leads short.
I also have a post up at Power and Control. And soon at CV.
Let me reiterate. You need a resistor in series with the relay to limit the current through the relay. The current should be limited to the maximum relay current for rated relay life. If you want to get longer life – nearer the mechanical limit vs contact limit – limit the current to 1/10th the contact rating. The current should be calculated from peak voltage. Although for this application such exactitude is probably not required.
Good on ya for adding the reverse diode. Them SCRs can be expensive. And I must add that I was not thinking clearly. A reverse diode across the coils (with a series resistor to absorb the energy) is also a very good idea. Send me an e-mail and we can discuss values etc.
BTW – nice transformer. It is probably vastly overrated for the application.
Safety: Consider a small (8 or 25A) SCR in Parallel with the capacitors to discharge them. A resistor in series with the SCR to absorb the energy is good. A 20 to 50 ohm wire wound rated at 25 watts (10 watts might do) should be adequate. If you want to get fancy calculate the RC time such that the voltage across the capacitance is reduced to less than 10 volts in 1/10th second. i.e. an RC time constant of 1/100th of a second will work.
[…] We received 2 of the 10 capacitors for the power supply. I got these cheap on ebay. They were bigger than I […]
[…] Parts for the coil power supply: […]
[…] Power Supply Made some progress on the coil power supply. Soldered an extension cord onto the 600V step up […]
“SCR (844-VSKH91/14P). Not sure how to wire this up.”
In case your not joking I downloaded the data sheet; the trigger coil goes between the small pins 4&5. To state the obvious make sure the SCR is mounted on metal with thermal grease or a phase change pad. Make sure the metal heat sink is large. If you sure it’s safe (measure) fire off a couple of realistic shots and check the temperature of the heat sink close to the SCR. I can’t make any more reasonable suggestions without knowing the load characteristics and PRR. Knowing that one can go backwards to the junction temperature. Since they cost a little money you should probably follow the recommendations for mounting torque and checking same on page 3 of the data sheet.
Ray
[…] I rearranged the equipment rack to make room for the coil power supply: […]
[…] Test Fire 18 04 2010 I successfully fired the coil power supply’s SCR today. Here’s how: I started by grounding the […]