Moskido - Hybrid Aikido/MOSFET amplifier

[beau2317]

Albert

I emailed you a reply to your question about transformers, but the message bounced (unknown user). Can you email me again with your correct (or another) email address, and I will send the reply again.

The idea of the parallel output caps is (a) to get enough capacitance and (b) to bypass the electrolytic caps with high quality capacitors.

Because I had 2 x 2200uF and 2 x 20uF then the total capacitance was 4440uF (you add capacitance when you put them in parallel).

That was because I had a 4 ohm load. For an 8 ohm load you can get away with about 2000uF plus 20uF (So, no, 1000uF plus 10uF is not enough)

I used 2 x 2200uF instead of a single 4400uF as that was what I had in my parts box.

But in any case I would suggest that you use one large electrolytic, and bypass with a polypropylene or oil cap.

[tubee]

Originally posted by beau2317
Can you post the circuit diagram of your PSU?

Each mono channel:

120VA toroid (2X22V AC out) 44V AC to diy diode bridge of four HFA08TB60's on L shaped alu heatsink, isolated Center tap of 2x 22V to gnd.

0.12ohm 5W ceramic resistor in AC line to bridge, 0.1 ohm from diode in dc line to panasonic FC 10.000uF 50V lytic for the + and - rail each.

Lytics bypassed with Wima MKP 100N 630V.

Over +/- rail a 2.2k 2W carbon resistor, and a snubber 1 ohm/100N MKT.

At amp pcb bypass near Hitachi fets with 470uF panasonic/50V 100N Wima MKP and 12N styro each rail. Wired with solid core 2.5 mm (gauge unsure) copper lighting wire, hardwired, so no PCB for diodes/caps/snubber.

[marsupialx]

Beau2317,
Concerning the feedforward regulators on TubeCad.
What is your take on the Janus or the latest 12au7/opamp hybrid (for Aikido / Moskido use)? John posts so dang many circuits, I am only guessing at their suitability. Think I'll try the 12au7 / opamp hybrid, unless I shouldn't.
Stupid question time. Would it constitute the power supply in its entirety, or would I also need some CRC or CLC action?
Thanks - Jeffrey Hill (marsupialx)

[beau2317]

They look pretty cool. Not sure that aikidos need regulated PSU. I could not hear much (or any) difference between a tube regulated aikido and a just-nicely-filtered aikido.

But not everything has such a good PSRR so for those circuits that don't, his regulators would be ideal.

I don't think he intends every circuit he comes up with to be built, but it's more a case of learning how tubes work and how to use them.

He's said before that he's so dissappointed by the lack of new tube designs (like every design you ever see is either a common cathode, a cathode follower or SRPP) that his mission is to remedy this. With vengeance...

[marsupialx]

However, if the 12au7/opamp hybrid is all that is required for a power supply, it would be cheap to build, and it probably wouldn't suck. If, on the other hand, it is in addition to a crc power supply, then building it might be monetarily inefficient. Anytime he names a circuit, i.e. 'Janus', he would probably like to see someone build it. Just seems like it to me. - Jeffrey Hill (marsupialx)

[beau2317]

Go for it then. Be sure to tell us and JB how it turns out..

[Kashmire]

Lineup, I have been very interested in using a blocking capacitor on the amplifier output. See the thread here.

I have been running a Nelson Pass Zen v.4 "Penultimate" for the past year, and I really like it. My new speakers (Triangle Acoustics) seem to want more power/control, so I'm moving to a push-pull type. This will most likely be Broskie's version (named in this thread as the MOSKIDO) or the Pass F4.

Check out the Pass F4 thread here . The Pass amplifier reverses the orientation of the MOSFETs, and results in about twice the required gate current. Check out the schematic below and compare to the Broskie configuration. (Other differences are obviously the driver stage, whereas the Pass version uses two coupling capacitors to drive the MOSFET's gates).

So here's my questions:

1. Broskie points out that PSU capacitors are indeed in the signal path with push-pull amplifiers (see here ). Doug Self asserts that an output capacitor on an amplifier increases distortion by 5x. Self supplies actual measurements to back up his claim. So is a coupling capacitor in the MOSKITO generating 5x the distortion vs. using a bipolar power supply? So my question remains unanswered ... why do push-pull guys use bipolar power supplies (with dual PSU filter caps) instead of using a monopolar power supply (with one PSU filter cap and one output cap)? Is there a performance reason (i.e. distortion)?

2. The Broskie design and the Pass F4 design are similar, but hte F4 inverts the MOSFET transistors. The Pass design needs about twice the drive current at the MOSFET gates according to SPICE simulation. Is there advantages to the Pass configuration, that one should accept the drive requirements because the results are so much better?

3. Both Pass and Broskie and extremely accomplished designers, although I tend to listen more seriously to Pass for these reasons: (a) he is required to be excellent, because is employment is dependent on being the best, (b) it is apparently his full-time job build and evaluate designs for the best results, (c) he has an undisputed track record of winning Stereophile Class "A" recommendations with his designs. Many times Broskie tends to be "brainstorming" rather than trying to publish a serious design. I agree the Aikido is ground-breaking excellence, and I have several Aikido designs operating now.

[beau2317]

Some comments for what they are worth...

-- The amount of distortion that Self is claiming for the capacitor is quite small over most of the audio band. And he was using an un-bypassed capacitor. When you get down to these small levels of distortion the important thing is how it sounds imho

-- With class A amps, mostly the PSU is supplying a fixed current so there is no audio signal through the PSU caps.

-- The capacitor in the moskido is held at a large DC bias. From experiments I have performed, capacitors of all types sound way better when they are DC biased, so I am surprised at Self's findings about bias. It conflicts with something I can easily hear.

btw, I think of the moskido more accurately as "push harder/push less hard" rather than push/pull

I recently had a chance to listen to an amp of similar topology (tube voltage amp, mosfet output) but with no output cap. The moskido was better, in particular the bass was tighter and more tuneful (better than the one with no output cap), which did surprise me a little.

Perhaps a moskido with split rail supply would sound even better, but you'd need to have a servo scheme to zero the output. The other big advantage to an output cap is that in case of accident, your speakers are not toast.

I am puzzled by the comment about the F4 inverting the MOSFETs. They look the same to me? Please explain..what is inverted exactly?

[Kashmire]

You mean something like this .... ?

(Ignore some of the bias settings, etc. Still playing with simulations.)

[beau2317]

OK yes I see now. Why would you put them that way instead of the other way? I guess the drive is harder as the capacitance is higher that way (after all it's only the capacitance that you are driving really).

Would you get as good PSRR? That might be the downside. Not sure. Not very knowledgeable about mosfets.

I will be most interested if you build that to see how it sounds compared to your previous circuits. Also if you build the aikido front end. I have the feeling that the tubes dominate the sound of the amp, as it seems to have the same kind of sound (or transparency) as aikido line stages.