Member
Joined 2009
Paid Member
And where on this forum did you find this gem ? - kidding aside, there are folk who would disagree. What I learned from this forum was said by somebody else in a thread I can't remember but went along the lines of 'there is an amplifier for everybody amongst the different options'
Valery, agree with this. Not to be pedantic, it'll also depend on the LED used. Blue and white LEDs are noisier.
Doug, I haven't got the latest edition, but I have read about the Blameless amplifier on your web site. Your analysis of distortion mechanisms in standard amplifiers is, in my opinion, definitive.
But, why should we settle for standard amplifiers? More specifically, why should Class A never be built? I've never thought too much about the efficiency of Class A, I know it's less than Class B. But while its important to understand efficiency (you gotta get the right heat sinks), it is not something that should ever be used as a figure of merit for comparison purposes. I thought that the purpose of the blameless amplifier was to minimise distortion. This is an admirable objective. But i do not see any linkage between distortion and efficinecy. Is a Class A amp worse sounding merely because it efficiency is lower?
Anyways, I biamp my speakers, so I have 4 channels of Class A. Each channel draws 9 amps quiescent current. Yes, big heat sinks. Inefficent? Naturally! Should never be built? Get out of here!
Your comments about VAS are interesting. If push-pull VAS is inferior, is that actually casting doubt on all fully complemenatry amplifiers?
And what if you don't have a VAS stage? I'm not smart enough to design a VAS stage, my amplifier only has an input stage (long tail pair) which cascades into another long tail (ie the output stage also provides gain). Do you really need a VAS stage? My amp seems to work well enough without it. (BTW - feedback from the output stage goes to the source of input j-fets, so I guess that its CFA not VFA).
So in summary, more analysis of alternative topologies beyond standard Lin configuration would be good.
Last edited:
http://www.diyaudio.com/forums/parts/35821-some-noise-measurements-leds-zener-diodes.html
dalton95070 wrote in post#127
LED noise
Long ago (around 1981) someone that I respect in audio pointed out that a much-lower-noise reference could be made by attaching ordinary NPN transistors together base-emitter-base-emitter. With the emitter side pointed to the higher voltage, and a suitable resistor in series to produce about one milliamp of current, the equivalent of a 15V zener is formed. It is indeed far less noisy than any zener (especially low frequency noise, which is very hard to filter out), and I have used it as a reference in power supplies ever since.
I read about in the Analog Devices App Note #AN-211
"The Alexander Current-Feedback Power Amplifier"
Look at the end of page #5 where he talks about NPN Bipolar transistors used as Zener diodes. The transistors appear as Q14-Q17 on the schematic on page #6.
http://www.analog.com/static/importe...17334AN211.pdf
dalton95070 wrote in post#127
LED noise
Long ago (around 1981) someone that I respect in audio pointed out that a much-lower-noise reference could be made by attaching ordinary NPN transistors together base-emitter-base-emitter. With the emitter side pointed to the higher voltage, and a suitable resistor in series to produce about one milliamp of current, the equivalent of a 15V zener is formed. It is indeed far less noisy than any zener (especially low frequency noise, which is very hard to filter out), and I have used it as a reference in power supplies ever since.
I read about in the Analog Devices App Note #AN-211
"The Alexander Current-Feedback Power Amplifier"
Look at the end of page #5 where he talks about NPN Bipolar transistors used as Zener diodes. The transistors appear as Q14-Q17 on the schematic on page #6.
http://www.analog.com/static/importe...17334AN211.pdf
Last edited:
OK, good research. Quick summary of what it tells us:
1) Don't use IR and blue LEDs. All the rest @ around 5mA - no problem.
2) Don't use low power (e.g. 0.5W) Zeners below 12V. All the rest @ 10-20mA - no problem. Higher voltage ones are better anyway.
If you add a filter cap in parallel - it will be perfect.
Also, such references are often used in LTP CCS - in the end most of the common mode components are suppressed, so real influence on the noise floor is even lower.
Cheers,
Valery
I want a book that explore all amplifier topology. What the strengthen and weakness. And let reader decide which one topology is best for their need. Because how it sound is very subjective.
I learn from this forum, and found many point of view how to design an amplifier, what most importance specification is.
1. THD profile is most importance
2. Lowest as posible THD value is most importance
3. Slew rate is most importance
4. IMD is most importance
And what minimum value of each specification should achieve?
How can you have four things that are all 'most important'??
Here is the correct link to Mark Alexander's paper:
The Alexander Current-Feedback Power Amplifier
In fact, he uses the same sort of thing - PN junction under reverse bias. It may be the case, that reverse-biased transistor's junction works based on a "true Zener" effect (tunneling), unlike most of "so called Zener" diodes, that are actually avalanche-effect based (presumably more noisy), but as we just saw from the measurements above, some diodes are more than acceptable.
Here is the link on short description of the physics behind both reverse bias effect options:
PN junction under reverse bias
Cheers,
Valery
The Alexander Current-Feedback Power Amplifier
In fact, he uses the same sort of thing - PN junction under reverse bias. It may be the case, that reverse-biased transistor's junction works based on a "true Zener" effect (tunneling), unlike most of "so called Zener" diodes, that are actually avalanche-effect based (presumably more noisy), but as we just saw from the measurements above, some diodes are more than acceptable.
Here is the link on short description of the physics behind both reverse bias effect options:
PN junction under reverse bias
Cheers,
Valery
Not to criticise, because I have built similar energy sponges over the time I've been a DIY, it's clear that we can have different objectives and priorities to professional designers like the OP.
If a manufacturer were to market full class A active array speakers, they would have a hard time making profits with any but a small and choosy group of audiophiles with the extra cash to spend. As DIY projects are all about individuality in the quest for our personal audio nirvana, we do lots of crazy things no one would attempt commercially. As a professional designer, to say that you think an idea, design or principle is poor, inefficient or should not be built, is to look first at the prospects of selling the idea or products in tomorrow's markets - globally.
There are also the awkward social conscience issues, likely peer and community sentiment about power consumption, heat emission, cost effectiveness, safety and even mooted legislation to think about. Watch the news and energy prices for a hint of where that's going.
We may have to open our eyes a bit wider to see more of Douglas Self's perspectives.
I believe, Bimo mentions them as possible options and questions which one is more "correct"...
No direct answer. Always a compromise. Combination. And the final result - listening impression - is very subjective...
I am sorry, it is not what I mean. I ask what most importance specification? I resume what people think about it.
On efficiency and Class A. How can you call it a "POWER AMP" with an output of only a few Watt?
Efficiency must be one of the primary design goals of a "power amp".
No sane person would want a 100W amp with an efficiency of only 1%. Just think of the PSU and that for one channel only, think multi channel and it only gets worse.
Efficiency must be one of the primary design goals of a "power amp".
No sane person would want a 100W amp with an efficiency of only 1%. Just think of the PSU and that for one channel only, think multi channel and it only gets worse.
The efficiency at any given moment is a function of the power delivered to the load. Its not a constant factor. Even a pure Class B (as in zero quiescent current) has extremely poor efficiency if its just providing low level background music and zero efficiency if switched on and the volume turned down. So whether its a Class A dissipating 100 watts at idle or a Class B dissipating 5 watts, both are extremely poor when the user is listening to say 50 milliwatts.
Class A though is not a topology I would ever consider for an audio power amp. I don't want the equivalent of a radiator sitting in the corner of the room on a hot day.
Class A though is not a topology I would ever consider for an audio power amp. I don't want the equivalent of a radiator sitting in the corner of the room on a hot day.
A reverse-bias transistor zener is extremely noisy and is routinely used as a noise source...
Jan
But should you expect to find such an amp in a book on "power amplifier design"? The word "power" does exclude all wimpy amps.
So, what do you think about the most importance specification?
You are one of the experts here, I will consider your opinion.
Whenever you single out ONE parameter as most important, then you will not give enough attention to other important parameters. To be a successful designer you must realise that a great design is the result of carefully considered compromises between different factors. It's never black and white.
Some people are hung up big time about slew rate and seem unable to consider anything else.
Or they are absorbed by THD only and fight for the last ppm or even ppb.
Yet others do not fully grasp feedback and thus champion anything as long as it has no feedback.
A good amp is a rather complex system of many interlocked issues.
Jan
Thank you, Jan.
I hope someone else can give their opinion about it, too.
I think, Mr. Self fight for lowest THD
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.