I made the core out of a microwave transformer core..
IMHO this core is not really usable for this application, for the laminations are way too thick (as it's designed for 60Hz). You will get too much eddy currents at the proposed frequencies and therefore will have too many losses, which will result in a small q-factor.
As I already mentioned before: If you wanna take a laminated steel core, the laminations have to be thin. Just look at Stans steel-core: There the laminations were also thin. But I think also, that a Ferrite-core is probably a simpler solution. The main difference is, that the steel-core has much higher saturation values, but a smaller permeability, whereas for ferrite, this is usually the opposite (high permeability, low saturation)
How much energy in watts per hour (or joules) needed to spend to convert 1 liter of water in gas? I am interested in the efficiency of the cell.
Ahh, BTW. I just remembered, that Stan himself calculated the efficiency of his early cell in the independent evaluation report. There he calculated something like an efficiency factor of 300 times (input vs output).
I used 3 4007 diodes in parallel and work fine in the vic...
I absolutely wouldn't recommend using any 1N400X in any power switching supply. They are intended mainly as rectifier diodes for mains applications (60Hz). Their switching time is about 1000 times slower than that of a MUR.
Why not using UF400X diodes, if money is short. They are also quite cheap, and much faster. Or did you already use the UF-types?
EDIT:
I finally found now some time to have another short look at the original VIC-board.
A few things are interesting.
First, it is really exactly the circuit from the patent (WO9207861).
Second. The designer of the board made a layout error. The primary driving circuit, as it is in the patent is correct. But the board layout is wrong. The two resistors for the first transistor are connected the wrong way around. Therefore Stan had to wire the VCO-Out directly to the correct resistor.
I already thought it very strange, that Stan added these dividers for the PLL-Signal in the patent circuit. As this wouldn't make any sense, to divide the signal here, as the pickup-signal and the driving signal have the same frequency. But as can be seen on the original board, he didn't use them. He just used them to divide the signal down for display on the LED, for nothing else...
One thing is still strange. Namely that he wired the neg of the primary over an RC-damper to the comparator input. First I thought he wired it to the VCO-out, but it only looks like that on first sight (quite hard to differentiate the two blue wires).
Finally one could say, that today, one would really just replace almost the complete board by just one microcontroller. Only the Pickup-Signal, and some driving circuitry would be additionally needed. The driving of the primary one would most probably do with a FET and some special FET-driver-IC (much simpler, than this cascade of transistors).
