Kapanadze Cousin - DALLY FREE ENERGY

[verpies]

Trying to recall the document you posted a while back, would you use a series resistor on the front side and another resister at the end?

There are three ways to terminate but in this case it would be the easiest to connect a non-inductive resistor at the end.

The value of the resistor at the end of the cable should be chosen in such manner that a scope connected at the beginning of the cable (the driven side) does not see any pulse reflections coming back, like in this video or this video.

I don't remember what Dally did to the end of the coax, but if I were experimenting I would start with a properly terminated end and if that did not work then I would try an open end and shorted end, because I would be afraid that the reflected pulse would come back and blow up my pulser ...especially the inverted reflected pulse.

[Dog-One]

I don't remember what Dally did to the end of the coax, but if I were experimenting I would start with a properly terminated end and if that did not work then I would try an open end and shorted end, because I would be afraid that the reflected pulse would come back and blow up my pulser ...especially the inverted reflected pulse.

Yes, not all diodes can react in less than 10 nanoseconds and clamp 1400 volts at 28 amps.  I hope Itsu gets it pretty close to correct on the first try.

[Jeg]

I finally was able to  make a new nano-pulser.

I knew that you were cooking something good. Nice job Itsu.! I am curious on how your coils will react on applying this pulse.

D1. Thanks for the link. Eric is a good school by himself alone.

[Void]

... I don't remember what Dally did to the end of the coax ...

Daly stated on his blog that the end of the 50 ohm coax was shorted (at least that's how Google Translate translates it):

"Далее обмотка коаксиалом 50 ом (ранее применялся в компьютерных сетях) получается бифиляр - конец обмотки закорочен - нагружен на генератор наносекундных импульсов."
Google Translation:
"Next, the winding 50 ohm coax (previously used in computer networks) obtained bifilar - end shorted winding - loaded to the generator of nanosecond pulses."

From:  http://www.liveinternet.ru/users/edward_lee/post235233940/

[itsu]

Very fine Litz wire would help with wire overheating.
Evenly distributing the winding over the entire circumference of the toroid will eliminate localization of the heat buildup in the toroid's ferrite.

Larger toroid will help with the ferrite heating problem also, but too large toroid will prevent it from saturating which is essential for this type of DSRD driver.

I choose teflon wire back in the Dally days because it was able to withstand the temperatures created, but now without any cooling the temperature jumps to over 90°C in seconds, so good cooling is a must even with litz wire i guess.

Also this teflon wire is thin and i am able to wind the 6:12 turns just on this small toroid; with litz wire i probably have to find a bigger toroid, but this i probably need to do anyway.

I tried to evenly distribute the 6:12 turns as much as possible, but still there seems to be a temperature buildup in the middle of the 6 turns primary on the inner side of the toroid.

That's 39kW of power for 10ns if you are measuring this voltage across a 50Ω resistive load.

Correct, a 50Ω resistive load

The frequency content of 10ns pulse with low pulse repetition frequency (PRF) will contain components from kHz to 100s of MHz.
This is Mr.Fourier, not noise !

Ok, good to know, it does show different then the signals when i still had the MOSFET oscillating.

Wouldn't you get a better performance with a 600V MOSFET than 600V IGBT ?  MOSFETs are much faster than IGBTs although they they do have problems with R_{DS_ON} at higher voltages.
What is the peak voltage appearing between the drain-source of a MOSFET or between  the collector-emitter of the IGBT, anyway ?

The use of the IGBT comes from the oscillations i had with MOSFETs used, my idea was that the lower capacities of the IGBT would minimize these oscillations, but that did not turn out to be thru, its for 90% the (PCB) layout that matters, but this IGBT stayed (till now).

I did measure the drain/collector signals, but they vary much with the voltage used (the higher the better/cleaner/smaller) but there seems to be a sharp peak (nano-pulse feedback?) at the shutdown side. 
I will take some screenshots.

Will it be opened, shorted or terminated at coax's characteristic impedance at the far end?

Anyway, AFAIR last time you did that, I think the nanopulses were detectable outside of the coax and I was surprised because they should've been confined to the coax's innards.

P.S.
Please remind us the schematic of your nanopulser ...with updated component values.

The Dally setup used a shortend coax, so i am planning to do again the same.
The idea will be that then the pulse will be reflected back along the coax shield where it can do its "work".
I will start on low drain voltage (125V).

A schematic will be upoaded later today.

Itsu