[AJinFLA]

Gished?  Whats that?

Jest  .

I dont listen to JayZ either but it was an example given to me by a younger person. It's what the next generation is listening to.

The point of the Cordell article is peak power demands of well recorded/dynamic music. Not what's popular with young folks.
I took issue with your premise that high peak power capacity is somehow frivolous/detrimental. It isn't, for dynamic music aficionados. For JayZ types, YMMV.

I agree class A is wasteful and not always the best sound.

Wasteful in the (general) sense of energy usage, in 2014. Especially for me, in warm FL.
I'll skip the whole (best) "sound" thing if you don't mind, thanks.

I meant to comment on the amp you mentioned. I looked at their website and found the idle power of 130 watts and output power of 40 watts but no statement if that's 40 per channel or 40 total watts.

It's 20watts, as you had asked about: http://www.stereophile.com/content/lindell-ampx-power-amplifier

cheers,

AJ

[Freo-1]

If the peak is 40 dB above the average then what level is your average SPL? If your average SPL is 80 dB then your peak is 120 dB SPL. What does that sound like?

Very loud and distorted.  Classical is  likely closer along the lines of 70 to 110 db.  Granted, 40 db is the extreme, but classical music variations often get to into the high 20s to low to mid 30s.  I think we can agree that many symphonies (think 1812 overture/Tchaikovsky/Rachmanioff, Mahler) have MUCH wider dynamic range than other types of music. 

So, the issue is, how to construct a system that supports this type of music.  The monitor/subwoofer approach (implemented correctly), seems to provide the ideal solution to reproduce this type of music (especially with tube amps). 

[G Georgopoulos]

hi guys, great reads thanks!!!

A problem with class-a and dynamic range
,basicly there are  limitations on the lower end of music (with hum)
and the higher end of music (with o/p watts)

cheers...

[AJinFLA]

A problem with class-a and dynamic range
,basicly there are  limitations on the lower end of music (with hum)

No, that is not a problem with "Class A". That would be component or system architecture/design.

....and the higher end of music (with o/p watts)

No, that isn't a problem either, unless the speakers are chosen unwisely. The "20 watt" amplifier I linked would drive a HE (>98db/watt) speaker system, in room, to very high levels, where the ears own distortion would most likely start to mask moderate clipping (>113db peaks).

cheers,

AJ

[Freo-1]

The "problem" is finding a HE speaker that provides reasonable linearity across the entire frequency band.   


There is a damn good reason many of the high end speakers that measure reasonably flat across their spectrum are in the mid 80's spl or so.

[Pneumonic]

The reason so few “true” class a amps are made today is because the days of having to bias heavy into class A, in order to get sufficiently low distortion levels, are long gone. Unlike back in the day, today's modern transistors have crazy good power capability and extraordinarily linear transfer functions which allows for the ability to design for (high) power outputs, with inaudible distortion levels, without the need to bias heavy (remember the point of bias is to reduce distortion by putting the amplifying device into its linear operating range). In fact, many of today's transistors have distortion level design specs that get them under 0.1% distortion without the need for much bias at all. Just add a little negative feedback to balance the distortion:heat ratio to manageable levels.

[opnly bafld]

[

[Roger A. Modjeski]

Great discussion going on here thanks to all...

AJ, thanks for the link to the Stereophile review. John Atkinson's measurements  and comments at the end are always a treat. I know many readers do not bother with the measurements section as it is just too techy. However I encourage all to read the last paragraph where JA gives his very honest opinion and often muses why a poor performing amps is so loved by the reviewer. That amp does look pretty good all in all. I am however surprised that it doesn't do better into low impedance loads given that it has 3 pairs or outputs per channel. It would have been a nice design feature to have it provide high current for those who need it. 


BTW I don't listen to JayZ and never heard of him before this test. I do think it is interesting in these days of the loudness wars that the producer or RLJ's album chose to record it so far below average level. A brave and appreciated move.

Pneunomic. I agree that transistors are better, not necessarily more linear as those physics have not changed.  They are faster and more rugged however. The physics and linearity for bipolar and Mosfets are quite different, MOSFETS being rather nonlinear (Voltage to current transfer) but because of their high transconductance make rather linear output stages when operated as followers.

As to class A bias: A few days ago I built up a little prototype of a MOSFET output stage driven by a tube just to look into crossover distortion with those devices. Using 18 volt rails I varied the bias from 0.1 amps to more than 1 amp. Distortion hit a low at around 0.2 amps and was worse at 1 amp, where bias approached class A for the power I was obtaining. This is what I have seen in the past, there is a sweet spot for bias that is not so high. It will not be the same for all configurations but of many configurations I have tried over the years I have found class A bias has no particular advantage.

As to crossover distortion. That was a big problem in the early days (1970s) with Phase Linear amps and many others that had the output transistors biased off. In those amps the driver stage was used to power the very low level part of the signal and then the outputs would kick in at some 30 mA or so. This was very clever in some ways as it eliminated bias adjustment, bias drift and ran the outputs very cool. However it did create and make famous crossover distortion. The problem caught the attention of the Federal Trade Commission ( as I recall, correct me if you know otherwise) and around 1975 they instituted the current practice of specifying distortion from 0.25 watts to full power. Previously distortion was rated only at full power. Most of these amps were not so good at 0.25 watts though excellent at high power where the crossover distortion was still there but became a small percentage of the higher power. Even if the percentage was low, the order of the distortion was high which means it was very spiky when viewed on the scope. This made these amps sound "gritty" at low volume. At high volume driving a pair of  Magnepans they were just fine.

As to crossover distortion in today's world, I don't see much of it anymore and it doesn't take class A bias to eliminate it.

I know of no modern Hi Fi  amp (other than the QUAD 405) that has output transistors biased off. In the QUAD the crossover problem is solved in a different way. Of course its up to the listener if that worked out well. I do note that QUAD has sold several hundred thousand of these amps.

Perhaps this post should be changed from "who makes amps with Class A bias" to "who wants amps with Class A bias... and why"

[sonicboom]

Unlike back in the day, today's modern transistors have crazy good power capability and extraordinarily linear transfer functions which allows for the ability to design for (high) power outputs, with inaudible distortion levels

The physics underlying all amplifying devices has not changed. Transistors are still inherently the highest distortion devices out there, with the possible exception of Nelson Pass's custom made SIT's. The old direct heated triodes (300B, 2A3, 845 etc), are still the lowest distortion amplifying devices extant.

[Pneumonic]

Fellas. I can bias any modern transistor so that, just turning on, the measured distortion level will be say 0.5%. If I heavily bump up the bias current I can get #’s down to about 0.05%. But this would be a very wasteful design as it would be needlessly dissipating oodles of power. The last heavy bias amp I did it was disappointing around 50w. That's crazy! That same design, with the bias down to about 5w waste, gave me distortion #'s in the 0.3% range. Still way below audible amounts. Keep in mind this is all without any NFB! But we all love lower distortion #’s so, by properly compensating with minimal amounts of local feedback, the #’s can get down into the 0.001% range – a thousand times below our hearing limit.

[Freo-1]

Fellas. I can bias any modern transistor so that, just turning on, the measured distortion level will be say 0.5%. If I heavily bump up the bias current I can get #’s down to about 0.05%. But this would be a very wasteful design as it would be needlessly dissipating oodles of power. The last heavy bias amp I did it was disappointing around 50w. That's crazy! That same design, with the bias down to about 5w waste, gave me distortion #'s in the 0.3% range. Still way below audible amounts. Keep in mind this is all without any NFB! But we all love lower distortion #’s so, by properly compensating with minimal amounts of local feedback, the #’s can get down into the 0.001% range – a thousand times below our hearing limit.

And it will still sond like crap unless you design it correctly:







http://www.jockohomo.net/data/7470.pdf








Tubes are much easier to get it to sound right from a design principal. 

[AJinFLA]

The "problem" is finding a HE speaker that provides reasonable linearity across the entire frequency band.   

2 second search:
http://www.diysoundgroup.com/catalog/product/gallery/id/332/image/1042/
http://www.gedlee.com/summa_.htm
Of course if you're looking in all the wrong places, like glossy audiophile mags, that may indeed be a "problem"....
Now the goal posts can shift to size, looks, etc, in 3,2,1...

There is a damn good reason many of the high end speakers that measure reasonably flat across their spectrum are in the mid 80's spl or so.

Yes, it's called Hoffmans Iron Law...and a market based on fashion trends.

cheers,

AJ

[Roger A. Modjeski]

Fellas. I can bias any modern transistor so that, just turning on, the measured distortion level will be say 0.5%. If I heavily bump up the bias current I can get #’s down to about 0.05%. But this would be a very wasteful design as it would be needlessly dissipating oodles of power. The last heavy bias amp I did it was disappointing around 50w. That's crazy! That same design, with the bias down to about 5w waste, gave me distortion #'s in the 0.3% range. Still way below audible amounts. Keep in mind this is all without any NFB! But we all love lower distortion #’s so, by properly compensating with minimal amounts of local feedback, the #’s can get down into the 0.001% range – a thousand times below our hearing limit.

I have built several transistor amps of different configuration. Some behave as you have stated but not all. More often than not I find there is a sweet spot for bias. It may be different at different powers but there is usually a spot where the distortion is minimized and then goes up again. I have repaired and biased hundreds of amplifiers and have not found that ever increasing bias buys lowest distortion. Every amplifier design responds differently. When I was a warranty center for SAE I was required to buy a very expensive (at the time) Sound Technology distortion analyzer so that I could set bias by nulling distortion at 20 Khz. The reason for that rapidly became obvious.

As I stated in my current fooling around with a simple MOSFET follower increasing bias does not decrease distortion.

[AJinFLA]

And it will still sond like crap unless you design it correctly:
http://www.jockohomo.net/data/7470.pdf

http://www.cordellaudio.com/papers/another_view_of_tim.shtml
Audiophile myths never die...

[AJinFLA]

The physics underlying all amplifying devices has not changed. Transistors are still inherently the highest distortion devices out there, with the possible exception of Nelson Pass's custom made SIT's. The old direct heated triodes (300B, 2A3, 845 etc), are still the lowest distortion amplifying devices extant.

Commodity parts of amplifiers are just that. The only thing that matters is the electrical output at the terminals. Mind showing that lower distortion of triodes, full bandwidth into loads? TIA.

cheers,

AJ

[AJinFLA]

John Atkinson's measurements  and comments at the end are always a treat. I know many readers do not bother with the measurements section as it is just too techy.
However I encourage all to read the last paragraph where JA gives his very honest opinion and often muses why a poor performing amps is so loved by the reviewer. That amp does look pretty good all in all. I am however surprised that it doesn't do better into low impedance loads given that it has 3 pairs or outputs per channel. It would have been a nice design feature to have it provide high current for those who need it. 

Yes, as he states the obvious, it's not the class/type etc. that matters most.
Yes, 2 ohm compatibility would have been nice, but keep in mind the intended market is nearfield passive monitoring, whose impedance's are usually benign, unlike "audiophile" injurneering stuff.

cheers,

AJ

[Freo-1]

http://www.cordellaudio.com/papers/another_view_of_tim.shtml
Audiophile myths never die...

What a shock, engineers disagree to some extent about the best overall design.   I think Matti Otala is closer to overall circuit design characteristics that Mr. Cordell.  Cordell's arguements center around negative feedback in general, which I have no quarrel with.  However, the paper rather oversimplifies the issue, and kind of overlooks the fact that music is not a simple sine wave, but a rather complex waveform. 


This is closer to the truth:


" "Traditionally, designers increased feedback to make a larger portion of the output signal control the amplifier’s response. Our listening tests showed us that simply applying more feedback was not the answer. In fact, as one kind of distortion went down, other parameters would be adversely affected, leading to an overall degradation of sound quality. We knew that the other conventional design approach of eliminating feedback completely was not the answer either, because this would cause high distortion levels, and as a result would produce a “woolly” sound.
The answer to the dilemma was found in a novel approach to feedback theory. We developed a feedback concept that allowed local feedback to be applied around individual stages of the amplifier circuit. This approach allowed us to avoid the sonic disadvantages of overall feedback from output to input. The concept was further developed to reduce phase and interphase distortion between stages of the amplifier as well. We were able to concentrate the loop feedback on the stages of the amplifier where it resulted in audible improvement.
Stability margins were also expanded because feedback no longer affected the frequency response. The use of this concept of individual gain blocks—complex in design but simple in function—allowed us to reduce distortion to minute values in all the products. The amplifier is divided into two separate sections or gain blocks. The input block is a transconductance amplifier without overall feedback. This avoids large output current being fed back to the input, and mixed with the minute input signal. The output block is a transresistance amplifier with parallel feedback. This is done to prevent higher frequencies than the feedback loop can handle from entering the loop. An approach like this will prevent Transient Intermodulation Distortion (TIM) and Slewing Induced Distortion (SID), eliminating the need for an extremely wide bandwidth.”


Anyone who has actually listened to equipment with these design principals incorporated will hear the results.

[Freo-1]

2 second search:
http://www.diysoundgroup.com/catalog/product/gallery/id/332/image/1042/
http://www.gedlee.com/summa_.htm
Of course if you're looking in all the wrong places, like glossy audiophile mags, that may indeed be a "problem"....
Now the goal posts can shift to size, looks, etc, in 3,2,1...
Yes, it's called Hoffmans Iron Law...and a market based on fashion trends.

cheers,

AJ

The claims made on a website have little to do with actual performance.  There are also space restrictions in many cases.

[AJinFLA]

What a shock, engineers disagree to some extent about the best overall design.

Cordell debunked the NFB>TIM myth. That's a bit more than a "disagreement".

[/color][/i]Anyone who has actually listened to equipment with these design principals incorporated will hear the results.

Interesting. Please provide the links so we can see the specifics, controls used, statistical analysis of the positives, etc. TIA.

[AJinFLA]

The "problem" is finding a HE speaker that provides reasonable linearity across the entire frequency band.   

The claims made on a website have little to do with actual performance.

What claims?

Now the goal posts can shift to size, looks, etc, in 3,2,1...

There are also space restrictions in many cases.

Thank you sir.

cheers,

AJ