After reading almost everything in this thread and admiring the great work of Mauro I have a question. According to:
http://sound.westhost.com/z-effects.htm
Using current source with high output impedance does not always produce good results. Can you guys comment on that article.
How does a topology like the one in this thread sound different/better than the classic version of the chip amp? Are there a specific type of speakers that would benefit from this topology. Maybe the most benefit would have the single driver speakers with no crossover filters...?
Thanks in advance for your feedback.
Greg
http://sound.westhost.com/z-effects.htm
Using current source with high output impedance does not always produce good results. Can you guys comment on that article.
How does a topology like the one in this thread sound different/better than the classic version of the chip amp? Are there a specific type of speakers that would benefit from this topology. Maybe the most benefit would have the single driver speakers with no crossover filters...?
Thanks in advance for your feedback.
Greg
Hi Greg
i did some simulations and it seems that this arrangement is not to sensitive to various speaker loads.
the sound can only be compare in the following way. the classic arrangement sounds "stressed" compared to this one. the sound you are getting here compare much better to larger solid state amplifiers. I got them sounding very much like my JLH class A amps and that says alot.
there are several tweaks that you can do to tailer it to your taste. if you like much more laid back sound use a 470uf in the feedback and then even more so if you replace R 13 with a 1M resistor, instead of the 100k. then you are almost in the valve playing field
cheers
Rudi
i did some simulations and it seems that this arrangement is not to sensitive to various speaker loads.
the sound can only be compare in the following way. the classic arrangement sounds "stressed" compared to this one. the sound you are getting here compare much better to larger solid state amplifiers. I got them sounding very much like my JLH class A amps and that says alot.
there are several tweaks that you can do to tailer it to your taste. if you like much more laid back sound use a 470uf in the feedback and then even more so if you replace R 13 with a 1M resistor, instead of the 100k. then you are almost in the valve playing field
cheers
Rudi
Hi all;
Greg,
I have explained all here:
http://www.diyaudio.com/forums/showthread.php?threadid=62483
Analog_SA,
A lot of more good (than me) planners use the zener-shunt-reg. My test confirms what you say.
Dr_H,
I understand thing say, but LM318 has not " low noise " in any case, for which the effects of a low noise PS are limited.
I have prefered use the shunt simply because from more guarantees in a structure as mine.
Much little high quality SS power amp uses 3 pin reg. PS for the diff-pair.
I have noticed that even great experts of electronics don't know the difference among the problems of "high energy" and that of " low energy". If I has to plan an RIAA pre surely I uses a very accurate stabilization, with a " virtual GND" autonomous by all the other circuits. An pre-amp is always an autonomous unity.
In the case of a power amp, LM318 is not an pre-amp, but does the work of " floating diff-pair". More LM318 is an integral part of LM3886, more the circuit works better...
Naturally there are other systems to do this, but my is a "very low cost " circuit, & "very synthetic".
Ciao
Mauro
Greg,
I have explained all here:
http://www.diyaudio.com/forums/showthread.php?threadid=62483
Analog_SA,
A lot of more good (than me) planners use the zener-shunt-reg. My test confirms what you say.
Dr_H,
I understand thing say, but LM318 has not " low noise " in any case, for which the effects of a low noise PS are limited.
I have prefered use the shunt simply because from more guarantees in a structure as mine.
Much little high quality SS power amp uses 3 pin reg. PS for the diff-pair.
I have noticed that even great experts of electronics don't know the difference among the problems of "high energy" and that of " low energy". If I has to plan an RIAA pre surely I uses a very accurate stabilization, with a " virtual GND" autonomous by all the other circuits. An pre-amp is always an autonomous unity.
In the case of a power amp, LM318 is not an pre-amp, but does the work of " floating diff-pair". More LM318 is an integral part of LM3886, more the circuit works better...
Naturally there are other systems to do this, but my is a "very low cost " circuit, & "very synthetic".
Ciao
Mauro
Hi all,
little precise statement on 220uF elcap:
The criterions of choice that I have proposeded am not "absolute" and it is have to always do of the verification test.
Cyril Bateman has found of the little ameliorations among a polar and a bi-polar Elcap. And the solution that he recommends is a " series bi-polar pair", what has resulted very similar to the film PET cap.
The sizing that I have done limits much the effects of the Elcap-THD on the audio signal ( <-100dB), but if you prefer a "absolute" solution is able use 2 470uF bi-polar 35V in series + 0.1 to 1uF film cap (not-inductive) bypass.
This solution has to guarantee about <-115dB THD...
Ciao
Mauro
little precise statement on 220uF elcap:
The criterions of choice that I have proposeded am not "absolute" and it is have to always do of the verification test.
Cyril Bateman has found of the little ameliorations among a polar and a bi-polar Elcap. And the solution that he recommends is a " series bi-polar pair", what has resulted very similar to the film PET cap.
The sizing that I have done limits much the effects of the Elcap-THD on the audio signal ( <-100dB), but if you prefer a "absolute" solution is able use 2 470uF bi-polar 35V in series + 0.1 to 1uF film cap (not-inductive) bypass.
This solution has to guarantee about <-115dB THD...
Ciao
Mauro
Feedback cap
Just becuase it was all I had that fit, a 470 ufd 50V Panasonic FM was used in mine. I was wondering about the voltage rating here, it seems a 10 volt should work.
Do not know about actual tests on "back to back" polar caps versus a real non-polar cap. In several applications I tried, a real non-polar worked better as a coupling cap. This was good quality caps, Elna Silmic as the polar, and BG NH for the non-polar. The nonpolar won every time. The comparison was listening, not measuring.
Looking at the circuit, the voltage on this cap should be close to output swing. Seems a lower voltage will be fine. Think I have some 220 ufd 25V Nichicon Muse non-polar. The big problem here is size, the Muse series are all large.
From a safety standpoint, what is the minimum voltage rating needed here? Think there are some 220 ufd 6.3v BG NHX in my closet. These would be the cat meow, if they will not be overvoltaged.
George
Just becuase it was all I had that fit, a 470 ufd 50V Panasonic FM was used in mine. I was wondering about the voltage rating here, it seems a 10 volt should work.
Do not know about actual tests on "back to back" polar caps versus a real non-polar cap. In several applications I tried, a real non-polar worked better as a coupling cap. This was good quality caps, Elna Silmic as the polar, and BG NH for the non-polar. The nonpolar won every time. The comparison was listening, not measuring.
Looking at the circuit, the voltage on this cap should be close to output swing. Seems a lower voltage will be fine. Think I have some 220 ufd 25V Nichicon Muse non-polar. The big problem here is size, the Muse series are all large.
From a safety standpoint, what is the minimum voltage rating needed here? Think there are some 220 ufd 6.3v BG NHX in my closet. These would be the cat meow, if they will not be overvoltaged.
George
Re: Feedback cap
Perhaps you are not understanding, or think as non-important, some of the points Mauro is raising for those caps.
He is talking about two non-polar caps in series, not back to back polar types. Those were found by Cyril Bateman as the best combination on the tests he made for his "Wireless World" article on capacitors. BTW a very good one.
Less 15dB in THD, as Mauro mentions for the two 470uF NP and 0.1uF film combo seems to me as a very interesting improvement and one that deserves listening to.
Such NP combination was also considered by Bateman as very good for input caps and feedback caps on power amplifiers, particularly the latter, where 'lytic caps are the most commonly used.
Perhaps Russ could find a way to provide for that option on Mauro's boards.
Carlos
Panelhead said:
Perhaps you are not understanding, or think as non-important, some of the points Mauro is raising for those caps.
He is talking about two non-polar caps in series, not back to back polar types. Those were found by Cyril Bateman as the best combination on the tests he made for his "Wireless World" article on capacitors. BTW a very good one.
Less 15dB in THD, as Mauro mentions for the two 470uF NP and 0.1uF film combo seems to me as a very interesting improvement and one that deserves listening to.
Such NP combination was also considered by Bateman as very good for input caps and feedback caps on power amplifiers, particularly the latter, where 'lytic caps are the most commonly used.
Perhaps Russ could find a way to provide for that option on Mauro's boards.
Carlos
A radical solution for people worried about this 220uF cap: why do not try to lift off it and short the pad? No cap, no dubt.
The feedback resistor netwok will amplify the max 4mV offset of Lm318 that will be added to the max 10mV of LM3886. This gives to a too large DC output offset, but it can be compensated by a trimmer on pin 1 and 5 of LM318.
I will use the cap, anyway.
The feedback resistor netwok will amplify the max 4mV offset of Lm318 that will be added to the max 10mV of LM3886. This gives to a too large DC output offset, but it can be compensated by a trimmer on pin 1 and 5 of LM318.
I will use the cap, anyway.
Re: Re: Feedback cap
Hi Carlos,
The noise floor and THD with a single Nichicon Muse non-polar is already low in the extreme. But yes, I would be glad to make room for two caps, let me see if I can manage it without too much fuss when I get back to the office. The biggest problem is size, most NPs are big, the muse especially so, and I am at the outer limits of what I can currently do on size.
If i could find a good 35V 470uf NP that was ~10-13mm diameter it would be good. Does someone know of any?
Honestly I think a good 220uf low ESR or NP in that spot provides most excellent results.
I am just not sure how much I can add to the PCB. It becomes a question of practicality and space.
Cheers
Russ
carlmart said:
Hi Carlos,
The noise floor and THD with a single Nichicon Muse non-polar is already low in the extreme. But yes, I would be glad to make room for two caps, let me see if I can manage it without too much fuss when I get back to the office. The biggest problem is size, most NPs are big, the muse especially so, and I am at the outer limits of what I can currently do on size.
If i could find a good 35V 470uf NP that was ~10-13mm diameter it would be good. Does someone know of any?
Honestly I think a good 220uf low ESR or NP in that spot provides most excellent results.
I am just not sure how much I can add to the PCB. It becomes a question of practicality and space.
Cheers
Russ
Re: My mistake
I was also quite surprised when I read that. And the curves that showed the difference.
In fact I will be trying to implement that into a receiver I am modifying. Space is a problem though.
Many years ago, when Jung and Marsh modified a Hafler DH200 in The Audio Amateur, the feedback cap also got a special attention. They used a very large cap, I think it was a flash cap.
Carlos
Panelhead said:
I was also quite surprised when I read that. And the curves that showed the difference.
In fact I will be trying to implement that into a receiver I am modifying. Space is a problem though.
Many years ago, when Jung and Marsh modified a Hafler DH200 in The Audio Amateur, the feedback cap also got a special attention. They used a very large cap, I think it was a flash cap.
Carlos
Hi all;
The typical DIYaudio situation:
- choose the elcap...
- the change, am not sure...
- remove it and change the LM318 polarization...
- on the contrary, changes the front-end...
Calm, not there is need to arrive at this stage.
Use an Elcap polar or no-polar ( bi-polar ) of a type that know, if want use a film cap bypass, and is enough.
I have written the Bateman indications, to have to of chronicle, because I have spoken of he...
The Voltage regulation offset on LM318 it moves the internal points of work, and it may has resulted non goods.
LM318 in verC is compensated and linearized in very extreme way, and is not it may substitutes with any different chips.
In RevA yes, but not always with good result.
1M input: work, but it has an output offset voltage of 30- 100mV...
DC NFBs: They remember that here there are about 180 DB of open gain loop, don't play with the DC, is enough an anomalous condition to "melt" some component. The output offset voltage various in operation of the temperature and tolerances of the chips, and in my circuit is always " stopped " (AC coupled).
They don't use this techniques in the name of "probable" THD to -100 dB...
PS: The measurable THD is present near to the RC cut off frequency ( 1, 5Hz ) but over the 100 Hz the situation improves quickly... Are so worried of -100 dB @ 20Hz and -120@1Khz ?
The typical DIYaudio situation:
- choose the elcap...
- the change, am not sure...
- remove it and change the LM318 polarization...
- on the contrary, changes the front-end...
Calm, not there is need to arrive at this stage.
Use an Elcap polar or no-polar ( bi-polar ) of a type that know, if want use a film cap bypass, and is enough.
I have written the Bateman indications, to have to of chronicle, because I have spoken of he...
The Voltage regulation offset on LM318 it moves the internal points of work, and it may has resulted non goods.
LM318 in verC is compensated and linearized in very extreme way, and is not it may substitutes with any different chips.
In RevA yes, but not always with good result.
1M input: work, but it has an output offset voltage of 30- 100mV...
DC NFBs: They remember that here there are about 180 DB of open gain loop, don't play with the DC, is enough an anomalous condition to "melt" some component. The output offset voltage various in operation of the temperature and tolerances of the chips, and in my circuit is always " stopped " (AC coupled).
They don't use this techniques in the name of "probable" THD to -100 dB...
PS: The measurable THD is present near to the RC cut off frequency ( 1, 5Hz ) but over the 100 Hz the situation improves quickly... Are so worried of -100 dB @ 20Hz and -120@1Khz ?
mauropenasa said:
Calm, not there is need to arrive at this stage.
Use an Elcap polar or no-polar ( bi-polar ) of a type that know, if want use a film cap bypass, and is enough.
PS: The measurable THD is present near to the RC cut off frequency ( 1, 5Hz ) but over the 100 Hz the situation improves quickly... Are so worried of -100 dB @ 20Hz and -120@1Khz ?
I couldn't agree more. It is one thing to theorize, and quite another to test.
this amp (REV C) tests better than ANY amp I own, and I don't own any junk. Listen to it, and you will be sold at once.Cheers!
Russ