That is very true.
I'm using zero degeneration, local cascode feedback at the front end, and around 15dB of Global feedback.
I have not tried the more conventional local feedback loop.
Are you concerned about the gain and phase margins of the global feedback loop, shown in post #467? http://www.diyaudio.com/forums/pass-labs/300233-f4-beast-builders-47.html#post4956892
I don't have any concerns with those results, after quickly looking at them.
I would assemble them on protoboard for confirmation though.
Without seeing the circuit. I suspect they will work.
I would assemble them on protoboard for confirmation though.
Without seeing the circuit. I suspect they will work.
I'm not sure I would trust slew rate in ltspice.
It's not something I have experience with though. I've always just measured it the conventional way.
It wouldn't surprise me if the measured slew rate is better than the simulated result.
It's not something I have experience with though. I've always just measured it the conventional way.
It wouldn't surprise me if the measured slew rate is better than the simulated result.
Lynn if you run a distortion vs power test like it shows in the XA25 manual and also on the Baf 2016 video maybe it would show if the XA25 uses cascoding on the output.
I don't think that there is any need to run tests. Much available evidence show that it does not cascode the output fets. As a push-pull source-follower, with the proper choice of IXYS FETs there is not need for cascoding.
Well ok. I had just watched that Baf 2016 video and Mr. Pass seemed to be saying it was the cat's meow.
http://www.diyaudio.com/forums/pass-labs/299383-pass-xa25-4.html#post4891599 "Mostly you are wrong. The Villars photo showed a mock up of the circuit I used as an illustration of cascoded performance that I showed at BAF, and is not related to the XA25."
Have the three witches of this thread ever thought of using lateral fets that are in production for the second stage - i'dunno, maybe parts from semelab or exicon - instead of the nearly extinct toshis?
---(am i making myself a fool, again????)----
anyway, this is the most thrilling thread I've read here in the last years.
cheers & bravo!
A.
---(am i making myself a fool, again????)----
anyway, this is the most thrilling thread I've read here in the last years.
cheers & bravo!
A.
Yes, this has been already thought of numerous times.
Certainly worth trying if you already have those parts laying around.
Hi Attila,
we are all fools living in a yellow submarine......😀😀
I tried meanwhile the circuit without degeneration in the second stage. The gain augmented in my build around 4dB, so I had to readjust the feedback or the OLG.
And I tell you.... I can not really decide if it sounds much better.... I would need a second system of high and proofed quality to make real statements.
At the moment I have the feeling degeneration in the second stage or not...is secondary,
more important seems to me the right choice of the gain resistor, the load resistor and the feedback resistor.
Here I hear the dramatic changes of sound, dark, bright, involving, dull......
🙂
trip continues, hope my two fellows start hearing their builds soon!
we are all fools living in a yellow submarine......😀😀
I tried meanwhile the circuit without degeneration in the second stage. The gain augmented in my build around 4dB, so I had to readjust the feedback or the OLG.
And I tell you.... I can not really decide if it sounds much better.... I would need a second system of high and proofed quality to make real statements.
At the moment I have the feeling degeneration in the second stage or not...is secondary,
more important seems to me the right choice of the gain resistor, the load resistor and the feedback resistor.
Here I hear the dramatic changes of sound, dark, bright, involving, dull......
🙂
trip continues, hope my two fellows start hearing their builds soon!
What would the effect be in raising the bias of the toshiba's to .3 or.4 amps and using bigger heatsinks and then lowering that resistor connected to ground at their output.
I am still around with lower values of 50mA for the SK/SJ.
Wanting the fantastic low values of distortion the original XA25 shows in the manual, demands really no degeneration in the second stage.
Then gain in the second stage with the SK/SJ increases around 4dB.
Wanting the fantastic low values of distortion the original XA25 shows in the manual, demands really no degeneration in the second stage.
Then gain in the second stage with the SK/SJ increases around 4dB.
1) A little more open loop gain
2) Improved linearity
3) Reduced thermal distortion ie reduced to virtually zero
4) No thermal drift
I have mine around 120mA at the moment with the intent to experiment at higher bias levels.
Last edited:
If I go higher they will be going on the main heatsink.
At the moment I have 2 large to220 heatsinks connected back to back per device.
At the moment I have 2 large to220 heatsinks connected back to back per device.
By raising the bias to .4 amps and lowering the output in simulation it did squeeze out a little lower distortion. THD is down to 0.00089 at 1 watt. I wish the transistor spice models that I found were better matched. But maybe that doesn't matter I was mainly looking for a trend here. I see Juma uses the toshiba's in an amplifier biased higher than .4 amps.