CatSkin Class A2 SET with Secondary Flux Cancellation

[BinaryMike]

This project is a follow-up to previous experiments with the cheap MCM 555-7125 speaker line matching transformer. The goals I had in mind were to prove that it's possible to make a decent single-ended power amp with this transformer and to produce a listening experience dominated by the characteristics of a single power triode -- free of any possible benefit or drawback from the driver section. My power output goal was 5W, mainly because that's about all I expected from the "30W" OPT at very low frequencies.

I feel that I've succeeded, to the point that I now plan to transplant the breadboard circuit into a two-channel amplifier chassis with it's own PSU. Midrange power output approaches 10W at clipping with regulated lab power supplies. Grid current commences at 5W. There is no visible change in the THD analyzer residual waveform as power output increases beyond that point. The 3.3nF feedback cap was needed only to insure no-load stability. It has no obvious effect on 10KHz squarewave response. The 100Hz squarewave shows a small linear tilt. At 1KHz, there is no visible deviation from perfect.

Regarding the voltage amplifier section, I owe a nod to Frank Blöhbaum for his articles on Multiplied Transconductance Amplifiers in Linear Audio.

The secondary flux cancellation feature produced a surprise: LF distortion at higher power output levels can be reduced significantly by flux overcompensation. I haven't decided to pursue this finding any further at present, but the ability to control core flux independently of tube bias current is a unique feature of this architecture that could facilitate research.

I've attached the amplifier breadboard schematic and a proposed PSU schematic for two channels. Comments are invited.

 

[ballpencil]

Very clever. Thanks for sharing. Thread subscribed.

 

[ratbagp]

That is really interesting. I might try doing the same thing with my 6C33C amp later in the year when I get into amp building mode.

Did you do a PCB or just ptp?

ray

 

[ballpencil]

How about DC offset on the speaker output?

 

[Wavebourn]

Several years ago we discussed here a concept of a SE sound from PP transformers, by applying a DC current to the opposite shoulder to cancel magnetization. I designed then an amp that was called Alligator, with triode on one shoulder, and MOSFET modulated CCS on the other one. This time I am going to repeat the trick, but using sweep tubes. Stay tuned! ;-)

 

[TonyTecson]

How about DC offset on the speaker output?

what about it? with such low dc resistance of the secondary windings and the lowish current supplied by the ccs, this is not a problem...

 

[ballpencil]

On the schematic, secondary DCR is said to be 2R02 (not sure if it's correct or why it's so high). Multiply that with 400mA standing current, that amounts to be 800mV on the 25V tap, not an insignificant number. Connecting the speaker to the 70V tap confuses me but i think there still should be some offset? Do correct me if i'm thinking incorrectly.

 

[TonyTecson]

if the speaker diaphragm veers too much away from dead center, that could be a problem,
but let us see what the OP has to say....

 

[20to20]

Why did you choose to use a 6P36S for this project?

 

[BinaryMike]

How about DC offset on the speaker output?

I get about 0.3VDC across an 8R load, which amounts to 12mW in the voice coil.

 

[BinaryMike]

Why did you choose to use a 6P36S for this project?

I started out using a 13GB5 because it was handy and about the right size and smoking-amp published some attractive triode curves for it. I tried the 6P36S later, just out of curiosity, and found that it produced slightly less distortion. The difference wasn't big and I didn't try many samples. I suspect either one will be fine for construction.

 

[BinaryMike]

Did you do a PCB or just ptp?

It's just an ugly breadboard so far. The circuit complexity is probably enough to justify a PCB design, especially for two channels, so I'll probably tackle the PCB artwork next. My new PC won't run my old PCB tool, however. I've pretty much decided to learn DesignSpark PCB instead of paying for another upgrade, but these things are always a struggle for me. I think I can...

 

[TonyTecson]

nice job there BinaryMike, how are the silicons holding up in terms of flange temperatures and can you tell us what your ambient temperature at your beck was? thanks

btw, i got a lot of 13gb5's from Stans' for a dollar each....

 

[BinaryMike]

I haven't measured transistor temperatures, but I did calculate power dissipation for all that matter and annotated the schematic linked in my first post. The numbers are well under 1W, which might not demand heatsinking under ideal conditions. In a warm climate, inside a tube power amp chassis, small radiators on the TO-220 devices are de rigueur. The diff-amp cathode resistor was added to reduce dissipation in the CCS device to a low enough value that long life is assured without heatsinking. That CCS still has plenty of compliance range and collector voltage for good performance.

Who is Stan?

 

[TonyTecson]

thanks for the reassurance...it is not everyday that we come across posts such as yours....


Summer Dollar Days - Vacuum Tube Sale - $1.00 Vacuum Tubes

but the 13gb5's seems to have been gobbled up, this is the downside of
mentioning such tubes here...

 

[ballpencil]

Why do you need so much headroom for the secondary flux cancellation? The MJW3281A goes to -40V and as annotated there it dissipates 18W. 5W@8R of output on the 70V tap translates to 6.32Vrms of voltage output. On the 25V tap (where flux cancellation is done), this means 6.32 * (25/70) = 2.26Vrms or 3.2Vpeak. Wouldn't it be more reasonable if the MJW3281A goes to something less like -10V? This would substantially reduce the dissipation (and heatsink size). Am i missing something?

 

[TonyTecson]

again it would be interesting to find out the rationale for it,
it may boil down to the cost of a real SET traffo vs his contraption...
i am all ears....

 

[jazbo8]

I think BM stated that "this project is a follow-up to previous experiments with the cheap MCM 555-7125 speaker line matching transformer...", i.e., just an alternative to deal with the OPT's saturation problem.

 

[ballpencil]

Why do you need so much headroom for the secondary flux cancellation?

Nevermind. I got it. Number of turns on the 25V tap is actually more than the 70V tap because i need to refer to the voltage to the primary side.. So to convert 6.32Vrms (for 5W@8R) to 70V needs 11:1 ratio while to the 25V only needs 4:1 ratio.

6.32Vrms * 70/25 = 17.6Vrms = 25Vpeak.. which makes your 40V headroom more reasonable though you can reduce it to 30V which would cut the device dissipation by 25%.

Bonus cool points: drive the flux cancelling device as well to obtain some funky push pull and more output power.

 

[316a]

I once did a breadboard using SE 6B4G and a 50VA 230:9+9V toroidal mains transformer . One secondary used for the speaker and the other as offset winding . The offset winding was placed in series with the DC filament supply of the 6B4G and a 280mH series choke (A CCS could also be used with a much higher voltage filament/offset supply) to obtain approx. 40mA offset against 1A of filament current . It worked fairly well but with the added complexity and cost , not really worth persuing . Dave Slagle of Intact Audio was the inspiration for this , he advocated using a reduced gap (from what I remember) and a tertiary offset winding to provide a wideband transformer suitable for the higher Ra tx types such as 75TL

316A