diyAudio (
-   Solid State (
-   -   Circlotron amp using N-channel mosfets (

Circlotron 4th June 2002 04:39 AM

Circlotron amp using N-channel mosfets
Hi friends. This my very first post.

First off, I decided to name myself after what I think is the most fascinating amplifier topology I have ever seen - the circlotron. It is a model of simplicity and symmetry. I particularly like it because after drawing the basic circuit about 50 different ways to try and make it look the most logical, I noticed that if you place 2 loudspeaker loads on it, 1 between the sources and the other between the drains, it can be drawn as a perfect tetrahedron! As far as I was concerned that was the icing on the cake for a most aesthetically pleasing ampifier.

Now that I have firmly established my reputation as a nutcase, uh, sorry, *enthusiast*, Let me talk amplifiers. I discovered this circuit only a few months ago, and as it turns out I have quite a few large collection of hexfet switching type mosfets laying around at home. The last mosfet type amp I built was the Electronics Today one, back in about March '81 I think. It used complementary fets and I never felt comfortable with P-channel devices. They always seemed the poor cousin (or brother/sister?) to the N-channel. Then I found the circlotron cct that would let me use both N-channel devices. Coolarama! I wound up a driver transformer, picked a matched pair and away it went.

I started a thread on Google Groups that I will part copy and paste to here, telling of my adventures with this cct if anyone is interested. One unusual thing I did was to use a ferrite core tranny normally used in a switchmode supply for the driver tranny. With sufficient turns on the core it works really well, contrary to the advice I was given by others.

Anyway, I think the topology is the greatest thing since sliced bread. I hope there are others out there that are perceptive enough ;-) to also see what is so great about this cct.

Regards. GP.

mrfeedback 4th June 2002 01:36 PM

The Circlotron Circle
Hello Aussie,
Do you have links for FET versions ?
This might even get Harry in a spin !.

Regards, Eric.

Joe Berry 4th June 2002 08:09 PM

GP, in case you are not aware, there is another thread related to this subject in which you may want to get involved:

Circlotron 5th June 2002 08:48 AM

Re: The Circlotron Circle
2 Attachment(s)

Originally posted by mrfeedback
Hello Aussie,
Do you have links for FET versions ?
This might even get Harry in a spin !.

Regards, Eric.

What do you mean FET version? Audio specific FETs or something else? Attached is the current version of the o/p schematic drawn for optimum clarity of seeing how it works. FETs are IRF540, rails 20v, resistors to earth = 100R, broken wire ends go to 3.5 volt bias supply, driver tranny is wound on ferrite 1:1:1 ratio as previously mentioned. Not that the circuit really needed posting or anything, I just love it that's all.

Who is Harry that he might get in a spin?


mrfeedback 5th June 2002 08:35 PM

Harry Who ?
Hi mate,
Quick Google search gave mainly tube circuits.
Maybe yous is the afterburner circuit that I have been looking for.
I agree that tranny coupled driver stage ought to be sonically decent.


Who is Harry that he might get in a spin?
Harry seems to have a thing about FETS - and I'm happy to listen to his advice - go search H.H and you are sure to find plenty of good advice, and unfortunately much flak also - ignore this stuff, for his experience and advice is GOOD.

Regards, Eric.

djk 6th June 2002 09:56 AM


yves 6th June 2002 10:10 AM

re: de palma
Did he get his labeling of the battery polarity wrong, or
are his fets labeled with the arrow in the wrong direction?


Circlotron 6th June 2002 02:40 PM

I'd say they are meant to be pnp bipolars. It did mention them in the text as being such, at least for one version of the circuit.

Well it's a Circlotron all right. I remember seeing that one about a year ago and it did not occur to me that there was something special about that topology.

Did you get a load of that cct number three?? I'll bet he never actually made it. What makes me say that is the diodes in the source legs of the driver fets. He says they are for temp compensation but higher temp means lower diode drop so the source is pulled lower (equals gate being pulled higher) and so both the drivers are turned on harder, the output fet gates are pulled further apart, quiescient current rises, and whoa! you got thermal runaway. :eek:

Bruce DePalma's N-machine was always surrounded with controversy. I have my doubts about this thirs amplifier cct too.


Circlotron 18th June 2002 04:52 AM

2 Attachment(s)
Hi everybody. This is how the project is progressing. As you can see I have made the output into a Sziklai pair on the suggestion of AKSA, (thanks heaps, mate:) ). This makes the output LOTS more linear. As an example, before I had to use ~200mA quiescent current to eliminate crossover distortion, now I can do it with 75mA. Actually I am running 100mA right now and may return to 200 and see if it sounds and better.

The fets are now tricked into thinking they are P-channel devices and so the whole circuit is upside down. It looks a bit confusing at first, but take my word for it, the things works. To me it sounds just so smooth and clean. Even through that 3C90-cored input tranny!

Next thing to do is figure out some feedback from output to input because the output impedance is 0.47 ohms courtesy of the "source" resistors (that are connected to the drains).

I have not cranked the volume way up yet because of lack of drive voltage avaliable. I gotta make a preamp for it yet. Doesn't that symmetry of the schematic make your mouth water?;)


mrfeedback 18th June 2002 07:40 PM

Hello Circlotron

Doesn't that symmetry of the schematic make your mouth water?
Yep, nearly as much as the Cantonese Chicken I'm enjoying right now !.

More seriously, this setup looks pretty ideal in terms of total symmetry.
Have you tried a step up transformer as the driver transformer to increase input sensitivity ?
Are you finding the 0.47 ohms output impedence to be a problem ?.
This has to be pretty much minimum component count output stage, and well suited to tri-amp cabinets ?.
Do you have any distortion/performance figures ?.
Please keep us posted on your results.

Regards, Eric.

All times are GMT. The time now is 02:51 AM.

vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright 1999-2017 diyAudio

Content Relevant URLs by vBSEO 3.3.2