Class A2 Direct MOSFET Coupled SE
Well, I finally got power applied to the breadboard amp I've been
working on. The purpose is to experiment with driving power grid
tubes with the MOSFET coupling circuit alluded to in these discussions:
Here is the schematic as currently operating:
The main idea is to provide a fixed reference voltage for the driver
anode and power tube grid (fixed bias) while coupling the driver
anode into a mu-follower to isolate the driver from the sudden onset
of grid currrent.
Q2 and R3 provide a DC reference with high AC impedance. I made R9
adjustable to allow dialing in the bias voltage. Good thing, too...
Q1 provides a high impedance AC load with fixed voltage to the
driver anode, and provides a low impedance mu-follower output
to drive the power grid.
Q1 is supplied by a separate stacked power supply (ultra simple)
which is returned to the cathode of the power tube such that the
grid current loop does not include the driver power supply.
As shown below, it works quite well, with the 5842 driving up to
300V P-P to the 4-65A grid (to +70V in this trace) while supplying 50mA
of grid current:
The 4-65A triode load line I am working with needs about 260V P-P
drive, going about +40 maximum, so I have some headroom in this
I'm also thinking about smaller screen grid resistors.
I am getting about 16W output, which the 10 Watt Transcendar
OPT seems to handle remarkably well. I was worried looking at the
waveforms on the scope but the low end is quite nice sounding.
I haven't done an A-B comparison with my 16W push-pull amp but
this one seems to play quite loud before the distortion gets
The distortion spectra is all 2f up to ~1 watt, then some 3f + 4f,
etc in a waterfall pattern increasing up to about 10 watts, when
the higher order harmonics start filling in. I just made some brief
measurements and have been listening to it on and off over the
weekend. It has a very neutral sound, flat frequency response, and
sounds great with some drum parts I usually find challenging for
Unfortunately I can not or will comment your intesresting circuit.
Because when comes to tubes I am still freshly beginner in knowledge.
At least you have there a very unusual hybrid type design.
Valve-> Mosfet-> Valve-> Output transformer.
Wow! Nice amp-porn!
One question: why the 2N7000 to bias the 5842? Don't like LED's or small signal diodes?
Thanks for that paper from John Curl! The stories of Matti Otala and
TIM distortion brought back some memories of my own.
In the late 1970s I worked in an audio store in Alaska and we sold
mostly Sansui. I was into DIY audio design back then and working
on my own versions of comp-symm output circuits, blowing up a lot
of transistors but I was tired of my day job as an audio tech consisting
of mainly replacing from 4 to a dozen transistors in an amp and
hoping you get them all so the whole thing doesn't smoke again...
Sansui back then was in the habit of sending their design engineers
out to collect customer feedback from reps. So one day who should
show up but Susumu Takahashi, wanting to talk about TIM and their
new DD/DC output circuit!
I was still trying to get my head around this new concept but to me
it explained a lot of what I and others knew, that harmonic and IM
distortion measurements were not predictors of sound quality. I
knew this because I was still listening to my old Fisher, HK, or
Scott tube receivers.
So I spent the afternoon talking with Susumu about Sansui's
approach to solving the TIM problem. As I recall it was a direct
coupled diffamp output stage with huge peak current capability
and local current feedback to boost the power bandwidth above
100 KHz. Sansui claimed a slew rate of 200V/uS for those amps.
But at the time I wasn't sure about the TIM claim; in the back of my
mind it had the ring of some snake oil claims. So I kept pestering
Susumu about it until I realized he must get that from a lot of
people. He finally explained quite well why you needed 100 KHz
power bandwidth for 20 KHz audio using the spectral mirror. Of more
precisely why the power bandwidth needs to equal the small signal
Anyway, enough of that. I only read the first few pages and need
to turn my attantion elsewhere now.
I like all the above. The LEDs I've tried have dynamic resistance of
about 10 ohms at 12mA. The 2N7000 connected like this has a dynamic
resistance of about 7 ohms (1/gfs I guess) and it's cheap and easy.
It gets me close to the 3V needed, using a single device.
Nice use of the 4-65A tubes. I have a handful of them in my junkbox just waiting for an application like this. Your schematic shows Q1 as a 1N60 but thats an ancient germanium diode.
This circuit kinda look like a mix between a common tube driver and tubelab´s powerdrive.
Nice circuit! Any sound difference or spectra diff. between Q1 in Mu follower or anti-triode?
(1N60 -> FQP1N60 likely)
what about a LED between base and emitter? I mean tempco.
Can you be more specific, Wavebourn?
What is obvious to you, might need a bit of explaining to others, professor!
And all different level in audio would be able to get something from my text.
This means adjusting to
basic english understanding & basic audio electronics understanding.
Demo - Krati .... by the people, for the people
and justice and equal rights to every woman and man regardless of gender, religion and status in society
/lineup Sweden May 2009
|All times are GMT. The time now is 05:43 PM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2015 DragonByte Technologies Ltd.
Copyright ©1999-2015 diyAudio