user TURBO?s replication of Steven Mark?s TPU ?

[stevewal2]

I think I'll just go and howl at the moon. It's feels like a pretty sane thing to do after reading this thread.

Then I'll try to replicate Steven marks TPU by gathering some random bits of cable and parts, putting them in a box, shaking the thing up, then holding a 60w light bulb over it to see if it lights up.

That's about as close as I've got after reading all of this. Am I the only one who's feeling completely bemused?

[starcruiser]

Guys,

I really think we need to be looking at signal sources that can be driven to at least 1 to 2 Mhz. a LMC555 is capable or if the PWM or PIC's are capable then you might want to design them with this frequency capability. I say this since when testing my coils (TPU) the resonant frequency of the multiple control coils have a resonant frequency of around 560Khz. This was checked after the coils were wired together.

This is one of the reasons why I suggested to all to check your inductance and inter coil capacitance values since these will affect the resonant frequencies. Check them and calculate the resonant frequencies.

I have found while monitoring the collectors during multiple configurations of series and parallel connections, that this affects the output, I found a variance of 20Khz at least depending on the wiring configs.

If at all possible please use a sweep generator to check what the sweet spots are of the coils are and their wiring configurations, since this will impact what frequencies you choose to try with your oscillators.

[starcruiser]

I forgot to mention that with a 10vpp signal source (square wave) that the resonant point provides a 4 to 5 times increase in the collectors output voltage (40 to 60 volts AC). I haven't tried to drive anything with it just yet, still testing coils.

Nect step in the testing will be to add one additional frequency and determine the best frequency and then add the third.

One intersting note is that when the frequency is driven above 100Khz the pulse width seems to be irrelevant, I adjusted the duty cycle down from 50% and found no real change in the output wave form. I did however noticed if I drove it up above 65% it did affect the output.

[MeggerMan]

Hi Carl,
Thank you for the info its useful to know what the limits of testing need to be.

OK the UC28023 will go to 1Mhz, again, not sure about the lower limit.
UCC35706 will go to 4Mhz, but looks a bit different to the others I have mentioned so may or may not be suitable.

So the other option is to look at function generators couple the outputs to MOSFET drivers to drive the MOSFETs and thats it. They are more expensive though.
XR2206 up to 1Mhz, variable duty, needs a MOSFET driver though.

For those with very deep pockets the MAX0838, 0.1 to 20MHz, costs about 13 GBP each.
May be a process of suck it and see.

Regards

Rob

[MeggerMan]

The NE555 and 556 is rated to 100Khz.
On the other hand the CMOS version of the dual 555 timer: TS556 will go up to 2.7MHz.
So there is hope for the 555 timer yet.

Its cheap and chearful and everyone and their mother has got a circuit for it.
Using 2 of these you have 4 different outputs.
I will do some experiments a NE555 to see if I can fix this duty cycle issue.

Regards

Rob