Hello Alexiss
Why you don't try to understand more and learn about audio with free available diy project's for example :
M2 , Schade 50 watts, Vfet's and DEF amplifiers design decision etc ?
Reverse engineering discussion about new product protected by copy right's TM is your urgency ?
I think that front end using CRC filter with decent value of R enough to limit
unexpected current.
Output stage is limited to 10A, by electronic protection.
Opening of both front end mosfet with high current value, I think is possible only if amp fall in to oscillation. It should be analysed during development.
Obviously original XA25 passing that test.
For our Frankenstein XA25, there is no warranty
I think that front end using CRC filter with decent value of R enough to limit
unexpected current.
Output stage is limited to 10A, by electronic protection.
Opening of both front end mosfet with high current value, I think is possible only if amp fall in to oscillation. It should be analysed during development.
Obviously original XA25 passing that test.
For our Frankenstein XA25, there is no warranty
-----------------------------------------------------
Now this is a very interesting and a valid suggestion. This might very well be implemented, But still the capacitor closest to the source-pin can
apply the extra current....
I dont know, but this was a very smart suggestion.
I think the input JFets can actually conduct beyond Idss without oscillation, when the Vg is momentarily
higher than Vs, which might be the case when JFets are biased at Idss.
The input (gate) then pulls current.
Last edited:
-----------------------------------------------Hello Alexiss
Why you don't try to understand more and learn about audio with free available diy project's for example :
M2 , Schade 50 watts, Vfet's and DEF amplifiers design decision etc ?
Reverse engineering discussion about new product protected by copy right's TM is your urgency ?
The name of this thread is PASS XA25. Any other discussions are irrelevant and should be elsewhere.
We are here to discuss PASS XA25.
There is no ban against discussing commercial products here on diyaudio.
We are still living in the free west.
We are discussing possible circuit topologies, these iterations will not be equal to XA25.
Pass Labs personnel or moderators are more than welcome to contact us here if they find any parts of our discussions inappropriate.
Zen Mod, Sound Happey, you are obviously experts on Pass amplifiers and you don't need to be here.
Please leave the rest of us alone and let us get on with our discussions.
Thank you.
Last edited:
-----------------------------------------------------------------I think that front end using CRC filter with decent value of R enough to limit
unexpected current.
Output stage is limited to 10A, by electronic protection.
Opening of both front end mosfet with high current value, I think is possible only if amp fall in to oscillation. It should be analysed during development.
Obviously original XA25 passing that test.
For our Frankenstein XA25, there is no warranty
Radule, please take a look at the schematics below.
There are two resistors that are now bypassed by capacitacnces in the Source circuits of the MosFet VAS.
Is this how you mean? thanx
Attachments
In any case, it is a fascinating development.
Almost 90 percent are audio amplifiers are basic, Three-Stage amplifiers.
At Pass Labs, it has long been proclaimed that the fewer number of stages will lead to the best sound.
Pass Labs are now producing and selling Three-Stage amplifiers with really good measurement (see XA25)
Could it be that Three Stages is the best and minimum number of gain stages in an amplifiier? and anything
less will lead to compromises?
As long as I can remember has Mr Douglas Self claimed that the optimized Three Stage topology is the most suitable
number of gain stages in an amplifier.
And he has gone on with numerous books and measurements to probe and prove this.
My God.... could he be right???
Almost 90 percent are audio amplifiers are basic, Three-Stage amplifiers.
At Pass Labs, it has long been proclaimed that the fewer number of stages will lead to the best sound.
Pass Labs are now producing and selling Three-Stage amplifiers with really good measurement (see XA25)
Could it be that Three Stages is the best and minimum number of gain stages in an amplifiier? and anything
less will lead to compromises?
As long as I can remember has Mr Douglas Self claimed that the optimized Three Stage topology is the most suitable
number of gain stages in an amplifier.
And he has gone on with numerous books and measurements to probe and prove this.
My God.... could he be right???
Last edited:
Thank you Mr. Pass for your reply.
As you can see, Pass Labs comments about "not using degenerative feedback" has caused
quite a stir and much heated debate here on diyaudio.
We are still contemplating on how the removal of degenerative Source feedback in the 2nd complementary
stage of XA25, with its associated problems in a commercial product, has
been addressed.
Is the product scheduled for more reviews? There is almost no reviews to read about on the net.
-------------------------------------------------
PS:
I read your very informative article on feedback and distortion, where it is shown that the application of
feedback actually raises the amplitude of higher overtones.
However, I could not find any info on the gain stage used to produce those sets of measurements, and its Open Loop
characteristics, that is, the amount of available Gain vs Frequency.
Could you please elaborate?
As you can see, Pass Labs comments about "not using degenerative feedback" has caused
quite a stir and much heated debate here on diyaudio.
We are still contemplating on how the removal of degenerative Source feedback in the 2nd complementary
stage of XA25, with its associated problems in a commercial product, has
been addressed.
Is the product scheduled for more reviews? There is almost no reviews to read about on the net.
-------------------------------------------------
PS:
I read your very informative article on feedback and distortion, where it is shown that the application of
feedback actually raises the amplitude of higher overtones.
However, I could not find any info on the gain stage used to produce those sets of measurements, and its Open Loop
characteristics, that is, the amount of available Gain vs Frequency.
Could you please elaborate?
Last edited:
The open-loop gain will increase dramatically with the caps. That must be considered.
If you are referring to figure 10, it is a reprint of the graphic offered by
Peter Baxandall many years ago. I recall that it was a bipolar device, but his
full paper is available on the net, and was also incorporated into John Linsley
Hood's book on linear audio circuits.
If figure 11, it was as stated - a Mosfet operated single-ended Class A at
1 watt into 8 ohms with varied Source resistance. It is only intended as
an example of the altered distortion complexity with degeneration.
------------------------------------------
Fascinating....
This is exactly what I experienced myself when I bypassed the
Source degeneration resistors with a pair of 470 uF Rubycon ZA series caps.
The sound slightly lost its clarity and air.
I tried many other caps and values and
anything smaller than 50uF dramatically altered the sound.
I preferred the sonics with the degeneration in place. Strange....
Last edited:
--------------------------------
You are absolutely right of-course.
However it is very much device-dependent.
The gain of Vertical MosFets vary considerably from sample to sample.
Without degeneration you will have - exactly as you said - unpredictable Open Loop Gain variations.
Everything from low to very high. Specially without a stabilizing bias arrangement.
But the increase will not always be as dramatic.
Particularly in the case of the XA circuit in which I suspect Cascode-Feedback is used to linearize the Aol to an optimum value.
Nevertheless it is an interesting design decision, from one of the leaders in
Audio amplifier industry. It will require a lot of adjustment and matching in production.
But one can equally question why God created the Heavens and the Earth....
...He certainly had less headache without them.
Anyhow, the bypassed Source resistors was not a (sonically) good idea.
So I think my version of the XA25 concept will feature Source degeneration.
I am going to try to build one without degeneration altogether and see what happens.
My problem is that I am using slightly older Harris parts with massive Gain. They runaway instantly, making them
difficult to use in a real circuit with no degeneration. Mounting them on a huge heat-sink did not help either.
There was a reference to the Bias Generator earlier, that it might be a split-supply.
I have had difficulty realizing the advantages of a split-supply Bias Generator, but I don't know...
Last edited:
Very interesting result.
Applying series-degenerative-feedback in a Single Stage Fet Amplifier with an 8 ohm load would lead to
alteration of Vds across the Gain device and the output impedance.
That might be the cause of the increase in higher harmonics.
Have you had the chance to perform THD tests with the above-mentioned parameters kept constant?
Last edited:
Distortion & Feedback
-------------------------
https://passlabs.com/sites/default/files/distortion_and_feedback.pdf
A very interesting paper, but I am sure you have all seen it before.
An important point in the test procedure was how the feedback was applied for making these sets of measurements.
Thankfully Mr. Pass informed us about that in the previous post.
This might explain Pass Labs recent disinclination towards using degenerative feedback.
Cheers.
-------------------------
https://passlabs.com/sites/default/files/distortion_and_feedback.pdf
A very interesting paper, but I am sure you have all seen it before.
An important point in the test procedure was how the feedback was applied for making these sets of measurements.
Thankfully Mr. Pass informed us about that in the previous post.
This might explain Pass Labs recent disinclination towards using degenerative feedback.
Cheers.
I have been thinking and thinking about this article.
....something is not right. There must be something wrong in the production of these curves...
This should not happen in a Single Stage, linear, series degenerative feedback, into a resistor load...
Feedback, specially as done with a resistive degeneration, into a resistive load,
in such a simple circuit, should not give rise to higher amplitude of higher order harmonics...
....something is not right. There must be something wrong in the production of these curves...
This should not happen in a Single Stage, linear, series degenerative feedback, into a resistor load...
Feedback, specially as done with a resistive degeneration, into a resistive load,
in such a simple circuit, should not give rise to higher amplitude of higher order harmonics...
Last edited:
Simulations
Let's run some simulations on this.
Let's create a circuit as follows:
A single stage, Single Ended, MosFet (IRFP240) biased by an 8 ohm resistor, having 11 Vds and 2 Adc through it.
Then we create two versions of it:
- One with no degenerative feedback applied,
- The other with degeneration feedback amounting to approxiamtely 15 dB.
According to Mr. Pass paper, there should be higher amplitude on higher harmonics components on the FFT
analysis of the output signal.
Let's see:
Let's run some simulations on this.
Let's create a circuit as follows:
A single stage, Single Ended, MosFet (IRFP240) biased by an 8 ohm resistor, having 11 Vds and 2 Adc through it.
Then we create two versions of it:
- One with no degenerative feedback applied,
- The other with degeneration feedback amounting to approxiamtely 15 dB.
According to Mr. Pass paper, there should be higher amplitude on higher harmonics components on the FFT
analysis of the output signal.
Let's see:
