I successfully fired the coil power supply’s SCR today. Here’s how: I started by grounding the chassis:
Later inspection revealed less then fantasic electrical conductance between the ground cable and the aluminum chassis.
Wired up the SCR and polywell coils:
For the triggering circuit I improvised with available parts. I settled on this circuit:
Looks like this:
I tested it out and it works!
When the SCR discharges you can see the coils flex and hear a sound from the SCR… and the voltage across the capacitors drops sharply. I took it up to 400V with 5 caps.
Pretty major step forward for the coil power supply.
Nest steps include:
Making a front panel with master power switch, AC indicator, capacitor bleed switch, high voltage LED indicator.
On the back panel: IEC C14 AC connector, screw down terminals for coils power out, and D-sub interface for computer control (via DAQ). I want to computer control the AC power, bleed resistor and SCR triggering.
Stuart made a keen observation about a relay on the DC side: Relays have a much lower voltage rating for DC vs. AC. So a relay rated at 240VAC may only be rated at 30VDC. This is because DC sparks are harder to quench- they don’t naturally drop to 0V like AC does. So long story short – I’m going to put the relay on the AC side of the transformer.
I also want to fuse the AC and DC sides.
Prometheus Fusion Perfection 
If you put a zener or MOV across the Contacts you can use the relay contacts at greater than their DC rated voltage.
When you say “a relay rated at 240VAC may only be rated at 30VDC,” do you mean the one in img_2474.png under the writing “10uf@50V.” I don’t understand why it would have a problem quenching at most a 9 volt spark gap. Also, once the relay contact is closed, after the capacitor discharges, the voltage across the inductor is zero, and all 9V is across the 15kohm resistor, so when the relay contact opens it has 0V across it. How fast are your driving the switching of this? Or… what am I missing?
The relay does not go there. It would either go before or after the stepup transformer, to initiate charging the capacitors.
Further, while in general Stuart is right that DC sparks are harder to quench without a zero crossing, are you sure those specific ratings apply to the relay contact and not to the relay coil? They look to be about the right ratios for a coil. Under DC the coil has only resistance. Under AC the coil additionally has reactance, and so needs a higher driving voltage for the same hold-in current.
The AC/DC rating for a coil can be compared using the heat/power “P” dissipated for the steady state hold-in current. Actually, since heat is only dissipated in the resisitve part, it means the hold-in current is the same for AC and DC.
You can investigate this by holding the relay closed with driving voltage Vac and measuring the current, to find P=(Vac)(Iac). Then measure the DC resistance of the coil. Calculate the matching Vdc using P=(Vdc^2)/(R) (or Vdc=IR). Thus if a relay coil is rated for Vac, you can also drive it at the calculated Vdc.
My shop-mate Stuart says:
FYI, i read your latest blog post.
the 80/20 parts you’re using are anodized which isn’t conductive. you should sand down to raw metal where you’re trying to connect the ground.
Additional note: electrolytics are bad for pulse circuits because of high internal resistance. It slows the rise time. Fast gate risetime is essential to prevent SCR degradation or failure from hot spots. This is especially important in pulse applications although it is somewhat mitigated in your case due to the inductive nature of your load.
Use ceramic caps. Or parallel your electrolytic with a few .1uF ceramics. Disc ceramics with an X7R formula are good for the application.
I’ll put ceramic caps on my next mouser order.