Hello to All,
i am trying to put general rules together in the hope to understand something
Well i can tell you ... it is not that easy for sure.
Topic ... Mosfets.
I have read that mosfets have good performance when operated in class A
So my trivial question is ... why so many AB amps use mosfet in output stages ?
As an aside i have never heard a class AB mosfet power amp that i like
They lack punch and are not transparent.
But if you know one i am all ears.
Thanks and regards, gino
i am trying to put general rules together in the hope to understand something
Well i can tell you ... it is not that easy for sure.
Topic ... Mosfets.
I have read that mosfets have good performance when operated in class A
So my trivial question is ... why so many AB amps use mosfet in output stages ?
As an aside i have never heard a class AB mosfet power amp that i like
They lack punch and are not transparent.
But if you know one i am all ears.
Thanks and regards, gino
That is not the way to understanding, which may be why you find it hard. You still seem to be searching for silver bullets - recipes for success. Are you a manager in your day job? In my experience managers often search for silver bullets; sadly (for their staff) they sometimes think they have found some.ginetto61 said:
Learn about power MOSFETs, and amplifier classes. Then you may understand why they can be useful in Class A, and Class AB too (and Class D (and C and E etc. - for RF purposes)).
This is your question: If A equals B, then why does C equal A? See the logical fallacy?
You questions are mutually exclusive. MOSFET can have good performance in class A or class AB.
Hi and thanks for the reply.
I have clearly misunderstood the all thing.
It is a fact that the amps i prefer use bjts and i did not like some mosfet amps.
But i have not listened all of them of course
I see now that is all in the design and execution
Thanks again. Gino
I have clearly misunderstood the all thing.
It is a fact that the amps i prefer use bjts and i did not like some mosfet amps.
But i have not listened all of them of course
I see now that is all in the design and execution
Thanks again. Gino
vertical mosFETs are usually used by the masses because they are cheap.
1 pair of To220 150W Vertical mosFETs (very cheap) can give an advertised output of 150W into 4r0
That means the amplifier is actually rated for 75W into 8ohms speaker loading and capable of driving a 4r0 dummy load for as long as the tiny heatsink stays cool.
But the advertise gets even more ambitious, let's change the "4r0" to read 4ohms and hope the customer does not notice that the amplifiers regularly blow up when driving a reactive load.
Last edited:
Hi and thanks a lot for the interesting advice.
But i got confused by the statement about class A being the best operating mode for mosfet. I translated this as the only recommendable mode, clearly wrong.
Now i understand that they are perfectly fine also for AB operations.
For instance the much praised F5 have a mosfet output and the sound is excellent.
It is all in the design and construction then.
Thanks a lot, gino
But i got confused by the statement about class A being the best operating mode for mosfet. I translated this as the only recommendable mode, clearly wrong.
Now i understand that they are perfectly fine also for AB operations.
For instance the much praised F5 have a mosfet output and the sound is excellent.
It is all in the design and construction then.
Thanks a lot, gino
The Arcam 8/9/10P amps all used Mos-Fets in Class AB and, although not stunning, were really good amps.
A lot of the Pass amps, use Mos-Fets in Class A but only up to a point, then they change to Class AB.
As you said, it's all in the design.
A lot of the Pass amps, use Mos-Fets in Class A but only up to a point, then they change to Class AB.
As you said, it's all in the design.
Last edited:
The confusion comes from only reading part of the information.
A BJT output stage can be in two forms, single ended or push pull.
The single ended is only ClassA.
The push pull can be ClassA or ClassB or ClassAB. (there are other operational classes).
The ClassAB push pull BJT version can be high bias or low bias.
In between these extremes there is a special case called optimal ClassAB bias. This is where the crossover distortion of a ClassAB amplifier is minimised. Unfortunately D.Self has renamed optimal ClassAB as his ClassB and that causes more confusion.
A mosFET output stage can follow exactly the same topology as BJT stage.
Single ended, push pull, ClassA, ClassAB
BUT !!!!!
there is no optimal ClassAB for a push pull mosFET output stage.
The crossover distortion gets less as the bias current is increased, from zero bias current all the way up.
Where does all the way up take us?
It becomes ClassA.
The push pull mosFET stage when biased such that the output current never exceeds (twice the quiescent bias current) remains in ClassA. It has no (zero) crossover distortion.
That leads to the conclusion that mosFETs make a good ClassA output stage.
That conclusion ignores all the other topologies.
There's the problem:
you read a conclusion without reading the conditions that applied to that conclusion.
You really must learn to research.
A BJT output stage can be in two forms, single ended or push pull.
The single ended is only ClassA.
The push pull can be ClassA or ClassB or ClassAB. (there are other operational classes).
The ClassAB push pull BJT version can be high bias or low bias.
In between these extremes there is a special case called optimal ClassAB bias. This is where the crossover distortion of a ClassAB amplifier is minimised. Unfortunately D.Self has renamed optimal ClassAB as his ClassB and that causes more confusion.
A mosFET output stage can follow exactly the same topology as BJT stage.
Single ended, push pull, ClassA, ClassAB
BUT !!!!!
there is no optimal ClassAB for a push pull mosFET output stage.
The crossover distortion gets less as the bias current is increased, from zero bias current all the way up.
Where does all the way up take us?
It becomes ClassA.
The push pull mosFET stage when biased such that the output current never exceeds (twice the quiescent bias current) remains in ClassA. It has no (zero) crossover distortion.
That leads to the conclusion that mosFETs make a good ClassA output stage.
That conclusion ignores all the other topologies.
There's the problem:
you read a conclusion without reading the conditions that applied to that conclusion.
You really must learn to research.
Research includes asking questions.
But those questions should be based on the knowledge gained so far.
You need some knowledge to formulate the question.
There is a saying that takes this advice to the extreme:
You should know the answer before you ask the question ! That allows one to assess the answer/s and decide which parts of those answers one believes, or does not believe.
But that is an extreme view.
You just need sufficient knowledge to assess the answers, to be able to "Sort the wheat from the chaff".
keep asking.
But those questions should be based on the knowledge gained so far.
You need some knowledge to formulate the question.
There is a saying that takes this advice to the extreme:
You should know the answer before you ask the question ! That allows one to assess the answer/s and decide which parts of those answers one believes, or does not believe.
But that is an extreme view.
You just need sufficient knowledge to assess the answers, to be able to "Sort the wheat from the chaff".
keep asking.
A perfect example of what AndrewT is talking about is my F4.
It's currently running double the specified bias.
A normal F4 will run up to about 10W in Class A and the rest of the way to 25W in Class AB.
My F4 will run Class A up to about 20W and the last 5W or so is AB (which you don't notice).
It sounds absolutely fantastic if you can keep your heatsinks cool but even at the standard bias point it was still a superb amplifier.
It's currently running double the specified bias.
A normal F4 will run up to about 10W in Class A and the rest of the way to 25W in Class AB.
My F4 will run Class A up to about 20W and the last 5W or so is AB (which you don't notice).
It sounds absolutely fantastic if you can keep your heatsinks cool but even at the standard bias point it was still a superb amplifier.
If you double the bias current, then the ClassA output current is also doubled.
That doubled output current results in a quadrupled maximum ClassA power.
P = IV = I²R = V²/R
That doubled output current results in a quadrupled maximum ClassA power.
P = IV = I²R = V²/R
Member
Joined 2009
Paid Member
Bias is just one aspect, I suspect that the input impedance of the output stage as 'seen' by the input stage plays a role and BJTs look quite different than MOSFETs in this case.
Hi and thanks again ... last question or so
Let's take this F5
Let's say that when it runs in class A mode it sounds superb (and thanks for sharing this, very interesting)
When it runs instead in class AB, at higher levels, does it still sound superb ? This is the question
Maybe when pushed with current hungry speakers it sounds less convincing ?
It could be that for class AB operations bjts are better ?
Actually i see much more bjts amps than mosfets amps.
Maybe there is a reason.
So mosfet amps have splendid sound also when run in class AB ?
If the answer is YES i am done. No more doubts.
Thanks again, gino
P.S. For instance a class A mosfet amp could be the very best solution for high eff speakers that do not need high power. The best solution for a specific application.
Let's take this F5
Let's say that when it runs in class A mode it sounds superb (and thanks for sharing this, very interesting)
When it runs instead in class AB, at higher levels, does it still sound superb ? This is the question
Maybe when pushed with current hungry speakers it sounds less convincing ?
It could be that for class AB operations bjts are better ?
Actually i see much more bjts amps than mosfets amps.
Maybe there is a reason.
So mosfet amps have splendid sound also when run in class AB ?
If the answer is YES i am done. No more doubts.
Thanks again, gino
P.S. For instance a class A mosfet amp could be the very best solution for high eff speakers that do not need high power. The best solution for a specific application.
Last edited:
Don't become overly dependant on which class your amplifier is operating in. Class A is a great marketing tool for wheting the appetite of customers because of the lore that surounds the title, but it is no indication of sound quality. One of the persistent myths surrounding class A is that they have the lowest distortion. This is utter hogwash. They do have the complete absense of crossover distortion at all levels, including levels far lower than we could listen to music, but it ends at that point and their claims omit mention of any other types of distortion. Most of those "audiophile" class A amps exhibit high distortion. There are so many other areas to pay attention to than biasing alone, and crossover distortion constitutes a small piece of the pie.
For even minimally biased modern devices, 100mA of bias bestowed upon a class AB Mosfet amplifier will reduce crossover distortion to far below audiblity. No one can hear below -90dB; it isn't possible because the mass of the tympanic membrane and cochlea damp it from moving. The actual devices and their requirements are pertinent - not claims like X mA bias. The level of bias does not improve nor enhance the amplifier's current delivery capabilities with "demanding speakers". The current capability is dependant on the output stage impedance and the power supply. A low impedance is derived from higher feedback, and from there the available current is determined by the power supply. I want to also add that contrary to marketing and magazine rhetoric, a large power supply and current doubling does not neccesarily dictate the sonic performance of the amplifier.
For even minimally biased modern devices, 100mA of bias bestowed upon a class AB Mosfet amplifier will reduce crossover distortion to far below audiblity. No one can hear below -90dB; it isn't possible because the mass of the tympanic membrane and cochlea damp it from moving. The actual devices and their requirements are pertinent - not claims like X mA bias. The level of bias does not improve nor enhance the amplifier's current delivery capabilities with "demanding speakers". The current capability is dependant on the output stage impedance and the power supply. A low impedance is derived from higher feedback, and from there the available current is determined by the power supply. I want to also add that contrary to marketing and magazine rhetoric, a large power supply and current doubling does not neccesarily dictate the sonic performance of the amplifier.
Last edited:
Hello and thanks a lot indeed for the very valuable advice
I have noticed that also many class AB units have very low distortion
I have a friend whose ears i respect that has moved from a class A amp to an AB amp because more transparent (i.e. lee distorting).
I have to say one thing. Sometimes i read about class AB mosfet power amps not able to provide the same "punch" of bjts. They say that the sound remains soft and a little confused, less transparent that with bjts.
Is this true ?
I see. Then may i ask what dictate the sonic performance in your opinion ?
Low distortion maybe ?
Thanks a lot again. Regards, gino
Try listening to a Quad 606/909. I was pleasantly surprised, it gives my Aleph 4 a run for its money.
The Quad only runs in Class A at the crossover point, the rest of the time it is in fact in Class C.
It's only just in Class C, it might be better to say that it's slightly outside Class B. (Same thing)
The Quad only runs in Class A at the crossover point, the rest of the time it is in fact in Class C.
It's only just in Class C, it might be better to say that it's slightly outside Class B. (Same thing)
Actually, the stock F4 runs Class A to about 25 watts into 8 ohms (50 watts peak).
You might find the following helpful:
www.firstwatt.com/pdf/art_leave_classa.pdf
😎
You might find the following helpful:
www.firstwatt.com/pdf/art_leave_classa.pdf
😎
That sounds to me like inadequate power supply problems. I haven't researched it but possibly mosfet output based circuits have poorer rejection of voltage rail modulation than bjt, as a generic behaviour. Which should mean, make sure the unit has a 'heavy duty' supply - I did this very thing to an 80's mosfet amp, which gave it far more "bite".
Good morning Mr. Pass and thank you very much for your kind and valuable reply.
I am not educated in electronics but i read with great interest your articles.
And i take this opportunity to thank you sincerely.
I am very interested in power amps with low voltage gain.
Usually line preamps have no difficulties to give out from 5 to 10 good clean Volts. Nevertheless i see power amps with 0.5V of sensitivity for max power.
But to stick to the topic, my question is if this value of up to 25W into 8 ohms within which the amp runs in Class A can be recommended as the optimum working range for the amplifier.
Or instead mosftes are perfectly ok also when operated in class AB ?
With a right speakers selection this value could be more than enough for many home listening set-ups.
I think that i usually listen within 1 Watt
Moreover i have decided to move to high efficiency speakers lately.
Thank you very much again.
Kind regards, gino
Member
Joined 2009
Paid Member
MOSFETs, LATFETS, BJTs, Triodes, Tetrodes, Pentodes, JFETs, ..... they all work just fine in Class A and Class AB as already said.
Mosfets are perfectly OK also when operated in Class AB 🙂
Of course, it maybe that you have a preference for BJT type of amplifier just because of the limitations of particular amplifiers you have heard and you would not be alone. Many folk have a preference but I suspect there preference is because they hear an amplifier that had some limitations which had nothing to do with the output devices.
Mosfets are perfectly OK also when operated in Class AB 🙂
Of course, it maybe that you have a preference for BJT type of amplifier just because of the limitations of particular amplifiers you have heard and you would not be alone. Many folk have a preference but I suspect there preference is because they hear an amplifier that had some limitations which had nothing to do with the output devices.
- Status
- Not open for further replies.