Ward's Build of Bob Boyce's TPU

[HumblePie]

Earl,
Thanks for the advice.  Im trying to understand how just resistance would work... like air gap? 

EDIT Nov 5th:  OK it is about balancing impedences so that the energy yield flows only through the load.  Eldarion's Simulations in his thread Nov 4th, with high Vs low load impedence says a thousand words.  It reminds me of how the Dirac Delta Function High-pass extra load trimmer-pot' tuned the ring out after the impulse that was created from the square wave... via impedence matching in that case.  In this case it seems we desire the correct impedence mismatching to encourage directioanal flow of energy.  I will try it before trying a large battery charger secondary as Bob does.

BTW - I am amazed at your post proving just a resistor between Q and Preset on an ACT' FFactually induces delay.  Your Scope and probes are superior to mine I found when comparing that pulser with ACT Vs HCT... my 100MHz Tek' 465B with 100MHz Tek P6109B probe made it look like only HCT could reach Vcc, nor could I detect any delay. 

Edit: Now that Eldarion has discovered pulse-train interference problems, I am beathing easier because there is an explanation.  (I have always found the most difficult problems occur when something believed to be true is not really true.) While Eldarion was wrestling with that, I went tangent about the ringing and 'perfect edge speed'.   I have heard ringing is our friend... with this HV potential 'biasing it all positive'.   Then again SM says not and Tao's TPU End Game[2] shows a big 'no symbol' through the ringing after the overshoot.  SM's, BB's.  I can see driving open-ended does not see any DC offset from square wave per no Gnd reference, and my scope shows AC voltage with respect to Gnd... and DVM can measure Amps only in AC, none in DC.  (BTW -  I was trying that while driving just one Primary).  I hear Eldarian warning us not to Scope Secondary while driving.  Anyway, Bob never said SM's rules did not apply to his toroid design... more like the opposite with all the similarities.  So maybe ringing is really bad in this app' too.  IGBT's are so fast they must ring though.

If ringing is bad AND we need very fast edge speeds, then the answer is in 'Delta Functions' type 'high-pass' as Grumpy has posted about.  When I tested it, I included a pot to Gnd from the stimulated end of the coil I was driving.   The cleanliness and purity of the pulse surprised me.  Perfect!  FF pulser, square wave input, no difference.  Pure, without ringing, but only 3V out after 5V in.   I have not replicated Jason's XFRMR method of boosting pulse amplitude.  That might affect the Delta magic of yielding all harmonics at nearly same amplitude as fundamental... as a true impulse approaching infinity is supposed to.  With pulses so pure, maybe 3V is enough.  With 15V drivers, I expect 9V pulses out.  I still have not connected my HP Spectrum Analyzer to check the harmonic population or amlitudes because I was believing ringing in this application was AOK. 

Lastly, I really worry about hand wired PCB and will order a vendor PCB after concept proven.  I will finish this version and make another with the high-pass filters and diodes for pure frequency pulses.  I am certain these are needed for the other TPU applications at least.     
Ward

[HumblePie]

I have seen Copper pad developer board work and do not like it as it is hard to change.  You can wire wrap around Resistor or Cap leads.  When finished fiddling with design, solder every post and lead cut off close to Vector Board.  Practice soldering wire wraps elsewhere, not on your fully wrapped card.  30AWG insulation melts like butter!  Get the cheaper WW tool with stripped inside.  This one costs too much at $25 and the stripper scratches the wire.  Use a good wire stripper. 

These 'BI' brand 1K Trimmers shafts are so stiff I warn you now... if you follow this lead, find another brand!  Trimmers should turn easily or you will hate it quick.  Desperate as I am, I am going with these.  Cermet does not make 1K pots.  I can break off the JB Weld later if needed.  I plan on soldering wraps and cutting the posts down eventually, but not before I get better 1K Trimmer pot's. 

With RC phase delay, lowering the frequency makes the delay longer.  There is no way to sweep frequency while maintaining phase delays as well as timers, counters, or gate delays can provide.  I would use the DS1023, but I want to keep this model basic.   Once I know if best phase is same for all frequencies, (I hope), I will build it in with gate delays, looped internally inside GAL's to keep the wires down.

Ward

[eldarion]

Hi Ward,

Would you mind posting a schematic of your high-pass "delta function" filter?  I had not thought of doing anything like that, and it is quite intriguing.  Might help my new controller quite a bit! 

Thanks,

Eldarion

[HumblePie]

Hi Ward,

Would you mind posting a schematic of your high-pass "delta function" filter?  I had not thought of doing anything like that, and it is quite intriguing.  Might help my new controller quite a bit! 

Thanks,

Eldarion

Sure.  It is called DeltaFunctions.doc.  Do not miss anything else posted here.  It is all very relevent to what we want to accomplish..  For 2nd approximation, try the 'Pole-Zero Cancelation" also posted here.  EDIT:  This uses a resistor in parallel with the CR  follwed by an RC to more perfectly shape the back side / falling edge of the pulse. 
http://www.overunity.com/index.php/topic,1825.msg32161.html#msg32161

Go straight to Pg 18, then 16.  Then LOOK at pictures of spectrum analyzer showing square wave harmonics Vs 'Delta Function' and tell me this is not important.  It is excellent background on harmonic components of sine, square, and triangle waves that should be fully read and understood after you get to Pg 18 meat of matter.  1N5711 is the equivlent diode. 

Ward

[eldarion]

Do not miss anything else posted here.  It is all very relevent to what we want to accomplish..

I'll try not to!   I've had my nose buried in the test bench for a while, and therefore have not poked around the forum as much as I would like.

Thanks!

Eldarion