allFET circlotron

Maybe you can post the latest schematic as you have so many variants. Whatever you think is the best option at this point and folks might be able to comment. It's hard to comment on so many parallel options.

I have done indeed some corrections, I post the latest one who did well and can be used, one with transformer and one with burr brown chip.

regards
 

Attachments

  • ScreenHunter_1198 Jun. 05 18.11.jpg
    ScreenHunter_1198 Jun. 05 18.11.jpg
    619.1 KB · Views: 339
  • ScreenHunter_1199 Jun. 05 18.12.jpg
    ScreenHunter_1199 Jun. 05 18.12.jpg
    127.7 KB · Views: 298
  • ScreenHunter_1200 Jun. 05 18.13.jpg
    ScreenHunter_1200 Jun. 05 18.13.jpg
    113.6 KB · Views: 290
  • ScreenHunter_1204 Jun. 05 20.50.jpg
    ScreenHunter_1204 Jun. 05 20.50.jpg
    573.7 KB · Views: 313
It would probably be wise to put a gate resistor on Q3/Q5 and what about an "external" input filter. Now you solely rely on the input C(iss) of Q8.(A/B)
Just a thought .....

That is I gonna do anyway, like the output mosfets, and a 100 ohms and 100pF on inputs, but when on transformer I do not now if this is needed. also on supplys and so. gate resistors cen be nicely used for P mosfets to get same respons with N mosfets using it as a corrector, but here not needed.

X I have order strafos, but go also order the sinc alloy ones.

and copper wire, how thick it was on primairy. wind the secondairy between the primary give good coupling. have it a bessel effect we can not do afcourse.

regards
 
Have done the advised corrections and put gate resistors and a low pass filter on input,

thanks Piersma for your advise.

I have seen that put two extra fets in the circlotron output stage itselfs ilding on 40 mA do make the amp much faster, even on 250 Khz the square looks good, (disconnected LP input who go to +- 100 Khz and is stolen from the alexander amp ;-) needs more drive voltage thought.



regards
 

Attachments

  • ScreenHunter_1214 Jun. 06 22.24.jpg
    ScreenHunter_1214 Jun. 06 22.24.jpg
    317.2 KB · Views: 267
Last edited:
X a compound stage output circlotron do amplify 2 volts in and 40 volts out, maybe just a transformer for it wil give a lot swing, and simpler it can not be done, extreme fast amp.. R13 and R 8 are local feedback resitors, making them smaller get more voltage output for same input voltage.

This was a old project i did not finish, but still interesting.

regards
 

Attachments

  • ScreenHunter_1212 Jun. 06 22.15.jpg
    ScreenHunter_1212 Jun. 06 22.15.jpg
    555.7 KB · Views: 141
Oke all.

I have today play some with combinations of topologies, but as you all now the analog amps are sowat out of design possibillities but can be combined, I think this is a nice one who need no transformer, so I can shift that to future plans.

Just see the schematic and let now there can be make things better as phillips says.

The schematic has no decoupling caps and such, it will be implemented later for sim this is not needed right away I have to play with circloron output to see it thing can better.

driver has enough drive, resistors there need 2 watts.

see pictures, for this also a constant voltage source is needed who can also be used as a multiplyer for prevent idle runaway, using a good negative voltage do make possible to use also a normally on sic j-fet. Output fets has not yet gate resistors, but she get a couple afcourse, it is just simming now..

amp go almost to 1 Mhz, square of 100 Khz looks oke, even 250 khz looks good. I think it is a nice fast amp, I can change it with a double feedback, now I have using source degeneration risistors as local feedback solution, so it have not a overal feedback, o afcouse almsot forgotten, it has current feedback, all of them.

regards
 

Attachments

  • ScreenHunter_1235 Jun. 07 21.55.jpg
    ScreenHunter_1235 Jun. 07 21.55.jpg
    105 KB · Views: 73
  • ScreenHunter_1234 Jun. 07 21.54.jpg
    ScreenHunter_1234 Jun. 07 21.54.jpg
    65.3 KB · Views: 68
  • ScreenHunter_1233 Jun. 07 21.53.jpg
    ScreenHunter_1233 Jun. 07 21.53.jpg
    65 KB · Views: 140
  • ScreenHunter_1232 Jun. 07 21.51.jpg
    ScreenHunter_1232 Jun. 07 21.51.jpg
    642 KB · Views: 164
  • ScreenHunter_1230 Jun. 07 21.27.jpg
    ScreenHunter_1230 Jun. 07 21.27.jpg
    132.1 KB · Views: 81
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
That looks very nice Kees. Looks like it will be the magic lucky 13 active devices once you put a Vgs multiplier in for temp control.

Well that square wave performance is very impressive indeed. Are you sure we can drive a single IRFP240 at 65v?

Thanks for all the simulations - let's get some feedback from folks and we move to layout next. :cheers:
 
That looks very nice Kees. Looks like it will be the magic lucky 13 active devices once you put a Vgs multiplier in for temp control.

Well that square wave performance is very impressive indeed. Are you sure we can drive a single IRFP240 at 65v?

Thanks for all the simulations - let's get some feedback from folks and we move to layout next. :cheers:

I can put two active devices extra and have Q3 and Q4 cascaded, cascade is good for sound, it has big free swing capability, like a spekaer in a horn so to say who get then much lower cone movement.

I can give the irfp 240 also lower voltages, it is what you need, can put some paralell also, as it are source followers capacitance and miller is not a issue, so it why i use smal mosfets where possible. if you look to distortion graph you see a very vriendly amp.

today is the last day for sims, after this I go make a pcb.

very small parts is not possible for good respons, however do not cound CCS and multiplyer as she are not in signal path, making it 10 parts.

regards
 
Cascode the drivers has no benefit, I have put in a multiplyer constant voltage source, however i am not a mathboy, so it need calculated such that idle running is corrected.

maybe I can find a tool on the internet to do this.

see schematic, last one. I can change input so it accept also balanced input with transformer, it is what I go do testing on if transformers are here, I make a stepper motor
wind machine, some software in a pic can make a nice machine, with extra small motor I can do also a way to let wind it very beautifull, I do not now how the name of that is, but it follow the wire when winding.

regards
 

Attachments

  • ScreenHunter_1236 Jun. 08 12.25.jpg
    ScreenHunter_1236 Jun. 08 12.25.jpg
    659.2 KB · Views: 271
Founder of XSA-Labs
Joined 2012
Paid Member
Nice work Kees. The design is clean and the topology is really different than other Circlotrons that I have seen.

The device you refer to I believe (same as on fishing reel) called an oscillating line layer for winding. Perhaps a mechanical line counter (distance) can be used vs a complicated stepper motor?

https://www.amazon.com/gp/aw/d/B005OTYW2E/
 
Nice work Kees. The design is clean and the topology is really different than other Circlotrons that I have seen.

The device you refer to I believe (same as on fishing reel) called an oscillating line layer for winding. Perhaps a mechanical line counter (distance) can be used vs a complicated stepper motor?

https://www.amazon.com/gp/aw/d/B005OTYW2E/

Hi X

no need for a counter, just a stepper and a pic with a display, so I can set the windings and it stops when reached. But is for later, there a much steppermotor drivers and such, maybe even diy winding machines op internet so is not a issue.

The last part of de design is people who look at it, you now I have never be on school to learn amps designing, I did learn it by experience and the nice sim software of the last years, so explane topologies I am not as good as our colleges who have learn it on a school, I am for shure not a mathboy, but can combine and sometimes get nicely result.

I have two kinds of output stages, X if you like overal feedback I can do that.

regards


regards
 
What you have is good - would global feedback make for lower distortion but at expense of bandwidth and slew emirate?

This is very cool and looking forward to seeing the layout. Please squeeze into a 10cm X 10cm PCB if possible. :)

Feedback is current, so not limiting bandwidth and slew rate. feedback do widen the amp bandwidth as i now, special current feedback is ecept is less stable for offset it have low tim etc etc and give very wide bandwidth.

Have somebody models for the SIC J-fets, now infinion? who works in multisim.

regards
 
a a mistake, not option global feedback, yes this is different, need this because of high open loop, well for this one it was not that bad needing just low ohmage risistors for that.

I mean overall current feedback, that is the last one i go try, but be aware, the global version has nice harmonics, and had 1 % for full power.

I thought about two idle pots, remove two D1 and D2 zeners and put these in there, use the corresponding mosfets as stabilization idle current, it can work, I think.
 
Last edited: