Agentgates´s TPU setup with strange wavehill hump

[agentgates]

@GL

Thanks for the circuit diagram and the quick work. Some comments on it:
1. Those FB diodes will be an overkill for the PSU. Remember there is no flow of holes here, only electrons from the device. This is more a particle accelerator than a transformer. This is where most of you failed in the past watching the results.
2. BC337 cooks, there is a resistor missing in series between the +18V an 0V. Can somebody pls make me a correction with additional resistors and values. My primary winding is 4.75 ohms. Sry for the request but I have many things to document here and the 4017 is messing about.

On your device:
You use HUGE wire for primary. Forget that. I used to fire half a thinner wire with 1500W of IGBTs, not BC337...

Please make a comparison, look at my corrected design for the new BC337 circuit you made:
http://www.samsonium.org/sites/default/files/P1070270.JPG

More photos later on the separate primary winding.

@Stefan

1. I opened the topic because you requested me to start sharing ASAP if the 1W is correct.
2. I did and now I want to make sure what you will receive is not kilos of SM-like overheating demo tool + huge heatsinks, fans, starting caps, inductor piles but a reliable lightweight swiss watch that you can take to presentations with yourself and you'll not curse my name.
3. I wasn't prepared for crowd with ready to build schematics and competition run they urged me. If I know this I wouldn't come out yet until your device is done.
4. My health doesn't serve well and I agreed with you to disclose ASAP. I want to make sure if something goes wrong with me this thing will survive.
5. I am not looking to win the price or anger anybody. I can disclose it elsewhere if it doesn't work here.

So again to all, pls everybody, patient. Not a competition.

[szaxx]

hi all replicators,
to avoid the minute delay re 4017 resetting simply use 3 of em in a chain and get the output pulses from the carry out pin. if it is neccessary for total precision timing.
steve.   

[agentgates]

anthony said that the pulses were supposed to be very short, probably 10% or even 1% duty cycle would be enough.

Yes it is true but that was just a comment on the importance of the duty cycle. It works with 10% on all my designs, however that is the lowest edge.

The duty cycle makes difference when finally you want to lower and control the output current on the secondary to keep the voltage nearly constant. The best preformance ratio used to come on the 50% duty cycle.

[Groundloop]

Tony,

I did not know better at the point of coil making, so that is why the HUGE wires. LOL

Do you need one current limiting resistor at each transistor or do you need just one
resistor from the common plus point to the three coils?

Alex.

[Mannix]

Tony,

try high side switching!