[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[TCML] Re: comments on Finn's circuits



Hi DC,

We are getting closer now with some real numbers ;)

Do you see that the signal from the Predikter circuit's zero crossing is occurring 26.3 degrees ->"before"<- the actual Tesla coil primary current zero crossing occurs?

Do you see that the H-Bridge control circuit is slow and will take say 26.3 degrees "after the command" to 'actually switch' the IGBTs?

-26.3 degrees + 26.3 degrees = 0 degree switching...

At -26.3 degrees the Predikter circuit gives the command to switch by crossing over zero itself. 26.3 degrees later, the command is actually finally accomplished by the slow acting H-Bridge circuits right at zero degrees just like we want.

Don't worry about gate drive, GDT tuning, or signal noise. Those were never the real point.

If coil heating, vibration, current effects, streamer effects, etc. mess it up a degree or two, those effects don't really matter much. A +-3 degree error in the switching is still a whole lot better than constant 26.3 degree error.

We will try not to let our frustrations make us mad :) It is hard to explain these things over e-mails... But I think you will soon see what it is all about too and then everyone will be happy. :)

Terry



DC Cox wrote:
If I'm reading it correct ---- and I say if here because it's hard to see the scope divisions: The 5 volt drive signal is getting a 26.3 degree phase shift between voltage and current at the zero crossover point. This equates to 1.88 volts off of zero crossover point. Finn is switching at 26.3 phase angle --- that's nearly 1/3rd of his current is being switched at the cross over point. If he is running (guessing here) 30 amps, then thats nearly 10 Amps of current being switched at the zero cross over point. Look carefully at Finn's waveforms and you will see that I am right. am I right or wrong on this so far ????? If I'm missing some important point, hey, I'm open to criticism, not just open to cheap shot personal attacks that a few members have posted. The gate drive signal is a square wave rich in harmonics. If you do a fast Fourier transform on Finn's data you will see some nasty harmonics that are coming thru. Finn does filter out the harmonic using essentially a ham radio "trap" circuit. Nothing new here as Chris Craft has been building amateur radio broadband antennas using this exact same technique for over 40 years. I used one as a ham operator back in 1966. Finn is swamping out the bad harmonic using a trap filter, and, yes, he does get rid of the overshoot, but the real root of the problem is the gate drive transformer turns ratios are not correct. They can not just be 2:1, 3:1, or 4:1 if you want to have a drive without the harmonic overshoot. You have to make some minor adjustments to clear up the overshoot. Try it on an actual TC gate drive xmfr of an operating coil and you can see the changes in overshoot. Small adjustments of a few turns will clear up most of the gate drive harmonic problems. Or, as Finn has done ---- you can use a trap filter circuit. No problem with that ---- either one works. But, as I initially stated this WILL DEFINITELY NOT correct the zero point crossing problem. That's all I'm saying ---- my points are: (1) using trap filters is nothing new in EE circles --- been around for 40 yrs. (2) Finn's circuit will correct for a gate drive mismatch problem. He is swamping out the over drive, not correcting for it. (3) Finn's circuit will NOT correct the zero point crossing problem --- at least definitely not in the data he has posted. Thanks for your patience and comments. Dr. Resonance

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla