Baxandall oscillator vs Royer

[dreamth]

For a week i am gathering data and shared experiences about this ascillators that most people call Royer converter , even the Mazzilli converter was called a Royer one ... but the thing is that the only places i found  that clearly claimed a  difference between them are these ones:
https://en.wikipedia.org/wiki/Royer_oscillator
http://www.sophia-electronica.com/Baxandall1959JM.pdf
https://patentimages.storage.googleapis.com/33/30/c3/37a4e97853a657/US2783384.pdf

From what i can see None of them are the actual  circuit used in usual CFL or CCFL inverters as described in Jim Wallace Linear technology app notes.. The Royers were using PNP transistors and split feedback coils while the original Baxandall had split feedback coils for the bipolar transistor bases too, LC resonating tank and . From what i could find up until now it looks like the actual Baxandall-Royer circuit is actually the creation of Jim Williams as described in LT app notes and the book of Bob Dobkin and Jim Williams, but i might be wrong too...

So i created this subject just because google might send the curious people right at the beginning of these converters and possibly set the right pace for develloping a project. I am investigating myself the Baxandall inverter as i need it for some tube preamps and already experimentally built about 5 inverters by now in order to get the unregullated right voltage and the current i need.For the moment i couldn't exceed 10w of power from my gapped  cores although i tried huge nonsaturable cores and 150Mhz 100watts  transistors....and i just found this new forum yesterday so i'll probably need to read and build a lot more transformers to understand exactly how i can get higher power from a Baxandall inverter.
For the moment i am really glad i found you and hopefully i'll be able to get more advice on how to build a better Baxandall inverter.I was first considering the Mazzilli version as it's used to deliver huge powers, but Baxandall inverter looks bullet proof as compared to Mazzilli, the transistors are stone cold while working at any load  and i measured and visualized the waveform of a baxandall in many variations and it never blows the transistors no matter what happens with its 12v supply .
I am not looking at more efficient ways of building a high voltage resonating smps as i established my strategy already .
A tube amp needs a 6.3 ...12v fillament supply so i'll probably be building a 12v resonating supply based on a dedicated IRS chip and from that voltage i'll get also the high voltage through these Baxandall inverters as the high voltage doesn't need regulation. Unfortunately my contry's power grid is 220v ac 50hz and i don't own transistors with higher voltage than 1000V CE to try an off line baxandall  or Mazzilli  converter as i saw that some of you did from the 110v ac 60 hz US power grid.Also, if i'd need to controll the high voltage, i can simply use some LT1180  to drive the Baxandall-Royer circuit as JW indicated in his app notes.

For the moment i am really glad i found you and hopefully i'll be able to get more advice on how to build a better transformer for these Baxandall inverters. I only tried litz wire 7+7 turns primary and lots of variations of secondary turns, just no separate windings for the secondary yet , with various cores and gaps, but , as i said i couldn't exceed 10w of power  while the sinus had some better, some worse secondary resonating peaks.

[dreamth]

Just some more info on the subject:
https://patentimages.storage.googleapis.com/a9/e9/2d/bf729556dae9cb/US4517472.pdf
http://www.sophia-electronica.com/Baxandall_parallel-resonant_Class-D_oscillator1.htm

[dreamth]

so the  transformer quality is paramount...now i can look to improve the rest of the circuit.

[dreamth]

Better transistors ztx1051a, better efficiency and higher power output too.I'm improving incrementaly the system cause i have no idea what its limits are.The transformer should allow 5 times higher power though...as i get no theoretical help from you guys...bear with me!

[petespaco]

Hello Dreamth.

I'm not the brightest bulb in the chandelier, but I had trouble figuring out why you are going to all that trouble to:

A tube amp needs a 6.3 ...12v fillament supply

.
  Back in the vacuum tube days, we got away with mains AC filament supplies by heating a cathode most of the time.

For instance: I had trouble reconciling these two comments:

i couldn't exceed 10w of power from my gapped  cores

the transistors are stone cold while working at any load

Ten watts doesn't sound like much load to me.

Personally, I don't have any trouble with the Royer/Mazilli oscillators failing.

So, I was just waiting for the smart guys to chime in.

Pete Stanaitis
---------------

[dreamth]

It seems that getting higher power out of this thing isn't too difficult, but opted to optimize it for  just the amount of power i need : 390v/25ma for both phono preamp channels .When using a ztx 1055a and somehow a too thick secondary wire i get a slightly modified sinus .The final circuit consists of a 70v ac output at 40khz and a quadrupler .The secondary wire is a bit too thick for my needs which also puts the transistors in danger in case of an output short, but i'll leave it like that for a while.Now i just need to make a better power supply  for the 12v input as the 50watts output of my sony playstation PS2 smps is a bit at the limit for my whole preamp including tubes fillaments.

[dreamth]

lots of other trials but for low voltage output and higher power it looks like the mazzilli converter was better than the baxandall and the most precious information i could find on the differences before building a mazzilli was this one:
https://www.eevblog.com/forum/projects/royer-converter-with-mosfets-for-low-power-dcdc/

[shrad]

I'm also interested in this, I was looking to do the same kind of power supply some time ago... looking forward to see you developments Getting rid of the big iron is a graal for some people