Markus,
I have read an artical (in russian). From the article this conclusion may be done: our ears don't hear 120db at the SAME moment of time, instantly: there is some period to adapt to "average current SPL". Are you agree?
BTW, my last post about external "game devices" is not ironic or joke. It is more logic and cheap to make neutral amps and such "colorizers" as different devices. Isn't it?
I have read an artical (in russian). From the article this conclusion may be done: our ears don't hear 120db at the SAME moment of time, instantly: there is some period to adapt to "average current SPL". Are you agree?
BTW, my last post about external "game devices" is not ironic or joke. It is more logic and cheap to make neutral amps and such "colorizers" as different devices. Isn't it?
No i haven't but i can see you are talking about what could be called artificial harmonics masking.
I don't have experiences with that at all.
Anli,
agree about external "game devices". You may think of it as a sort of sophisticated EQ. I see nothing wrong with the idea in principle, certainly better than universally colored amplifiers ;-)
120 dB: I have no exact reference, I just "know" this value is often quoted. It seems plausible that one would have to adapt to average SPL. Still , cocktail party effect points to significant potential for recovery of signals below noise level...
Re: the website of the article: on another page
MP3
they summarily claim that MP3 at 128 bitrate practically equals CD quality except for some loss of soundstage, judged too unimportant for listening to matter. Only classical music is sadi not to convert well. Strangely the author doesn't relate this to his soundstage comment 2 lines above, but to some electrical noise in his computer (?) or ambient noise in a concert hall.
Ah well.
agree about external "game devices". You may think of it as a sort of sophisticated EQ. I see nothing wrong with the idea in principle, certainly better than universally colored amplifiers ;-)
120 dB: I have no exact reference, I just "know" this value is often quoted. It seems plausible that one would have to adapt to average SPL. Still , cocktail party effect points to significant potential for recovery of signals below noise level...
Re: the website of the article: on another page
MP3
they summarily claim that MP3 at 128 bitrate practically equals CD quality except for some loss of soundstage, judged too unimportant for listening to matter. Only classical music is sadi not to convert well. Strangely the author doesn't relate this to his soundstage comment 2 lines above, but to some electrical noise in his computer (?) or ambient noise in a concert hall.
Ah well.
LARS CLAUSEN
About the sugggestion of a box generating controlled harmonics ahead of power amps :
Is it not the experience to do before saying amplifiers with 0.001% distorsion sound terrible ?
~~~~~~~~~~ Forr
§§§
About the sugggestion of a box generating controlled harmonics ahead of power amps :
Is it not the experience to do before saying amplifiers with 0.001% distorsion sound terrible ?
~~~~~~~~~~ Forr
§§§
To suit tastes
Take a "perfect" amplifier, that is one with say less than 0.01% THD at any frequency and the same about IM, able to drive within these specs any reactive load. Take the best available drivers, better yet cross over electronically (use as much perfect amplifiers as required).
This should be as close to a "wire with gain" system as measurable, and certainly will please some and will be found terrible by others.
NOW, feed it with program previously massaged by a fast 32 bit DSP. Invest some megabucks in characterizing the subtle (or not so subtle) nonlinearities and frequency response / slew rate arctifacts of all "ultimate" topologies, be they class A triode with 1.5 KV plate voltage, fet or bipolar class A or AB you name it.
Make now double blind tests with the best golden eared experts and see if they can reliably indentify the genuine "Brand X" unit from the emulated one.
My bet is they will not.
Rodolfo
PS I am not against tweaking listening conditions to suit tastes. What I should rather prefer is to have knowing control on the tweaking for each listening condition and source material.
Take a "perfect" amplifier, that is one with say less than 0.01% THD at any frequency and the same about IM, able to drive within these specs any reactive load. Take the best available drivers, better yet cross over electronically (use as much perfect amplifiers as required).
This should be as close to a "wire with gain" system as measurable, and certainly will please some and will be found terrible by others.
NOW, feed it with program previously massaged by a fast 32 bit DSP. Invest some megabucks in characterizing the subtle (or not so subtle) nonlinearities and frequency response / slew rate arctifacts of all "ultimate" topologies, be they class A triode with 1.5 KV plate voltage, fet or bipolar class A or AB you name it.
Make now double blind tests with the best golden eared experts and see if they can reliably indentify the genuine "Brand X" unit from the emulated one.
My bet is they will not.
Rodolfo
PS I am not against tweaking listening conditions to suit tastes. What I should rather prefer is to have knowing control on the tweaking for each listening condition and source material.
About high priced amps (megabucks):
The person that bys a golden Rolex just wants to know what time is it?
The person that bys a golden Rolex just wants to know what time is it?
belated reply to MikeB
Sorry mike, I didn’t mean to leave you with the impression you’ve “proved” your point about vas load – I’ve just been too busy
http://www.diyaudio.com/forums/showthread.php?postid=710582#post710582
You comment that “…the only thing this schematic demonstrates to me is the benefit of higher feedbackfactor” – suggests we are on totally different wavelengths – OF COURSE my sim indicates the benefit of increased loop gain – and that for a vas with a given gain (= gm) that your vas load resistance throws away gain and reduces linearity in the closed loop amplifier
In your example sim
http://www.diyaudio.com/forums/showthread.php?postid=697807#post697807
you still seem to be under the impression that emitter degeneration is somehow equivalent/comparable to vas load – emitter degen addresses a completely different distortion mechanism than the dominant nonlinear output stage impedance in your example – linearizing the input gm has no effect on nonlinear Z load distortion from your darlington output stage’s finite and nonlinear current gain
You ignore the dual definition of circuit gain – matching open loop V gain but glossing over your boosting current gain by removing emitter degen in your loaded vas example – subcircuits with matching V gain and different output Z do not contribute the same loop transmission (Bode’s T) to the overall amplifier
Your circuit increases current gain and “throws away” the expected V gain by resistive loading and takes a lower output Z in compensation – I claim that you would see somewhat improved/similar closed loop distortion by omitting the vas R load while increasing input gm as you did with the advantage of the higher loop V gain reducing signal level at the diff pair input, the higher loop gain compensating for the reduced diff pair linearity compared to the emitter degen case
But better still is recognizing the dominant distortion mechanism in your sim circuit and addressing it directly – just add another follower stage to “buffer” the vas and the nonlinearity due to the nonlinear load from the output stage is reduced by the follower’s hfe ~100x
Sorry mike, I didn’t mean to leave you with the impression you’ve “proved” your point about vas load – I’ve just been too busy
http://www.diyaudio.com/forums/showthread.php?postid=710582#post710582
You comment that “…the only thing this schematic demonstrates to me is the benefit of higher feedbackfactor” – suggests we are on totally different wavelengths – OF COURSE my sim indicates the benefit of increased loop gain – and that for a vas with a given gain (= gm) that your vas load resistance throws away gain and reduces linearity in the closed loop amplifier
In your example sim
http://www.diyaudio.com/forums/showthread.php?postid=697807#post697807
you still seem to be under the impression that emitter degeneration is somehow equivalent/comparable to vas load – emitter degen addresses a completely different distortion mechanism than the dominant nonlinear output stage impedance in your example – linearizing the input gm has no effect on nonlinear Z load distortion from your darlington output stage’s finite and nonlinear current gain
You ignore the dual definition of circuit gain – matching open loop V gain but glossing over your boosting current gain by removing emitter degen in your loaded vas example – subcircuits with matching V gain and different output Z do not contribute the same loop transmission (Bode’s T) to the overall amplifier
Your circuit increases current gain and “throws away” the expected V gain by resistive loading and takes a lower output Z in compensation – I claim that you would see somewhat improved/similar closed loop distortion by omitting the vas R load while increasing input gm as you did with the advantage of the higher loop V gain reducing signal level at the diff pair input, the higher loop gain compensating for the reduced diff pair linearity compared to the emitter degen case
But better still is recognizing the dominant distortion mechanism in your sim circuit and addressing it directly – just add another follower stage to “buffer” the vas and the nonlinearity due to the nonlinear load from the output stage is reduced by the follower’s hfe ~100x
hi jcx,
obviously we have the one or other misunderstanding...
I agree with you that the benefits of degenaration are typically
lower than the loss of feedback resulting in less linearity for a closedloop.
But, if openloopgain is already too high, so that stabilizing gets
difficult/impossible, degeneration becomes a necessary step.
Maybe the combination of re AND vasload degeneration would be
the best choice.
Adding an extra followerstage and making a tripledarlington would
be the logical step, but i did not get it adequately stable, after it
was stable again distortion for 20khz had exploded.
This nonlinear load to the vas seems to be a bigger problem then
one would expect, especially in my case, as i use to92-bjts in vas
and have to keep currents low.
Especially for higher freqs, but still below resonancefrequency of
the feedbackloop, the dynamic load starts to create wild phasehifts
into the vas (even with tripledarlington), creating instability.
This dynamic load is created by the very unlinear nature of the big
output-powerbjts and can create a resonance in the very last stage.
Sims showed this resonance at 3mhz, but realworld had it even
higher... (i "cured" that with the caps to gnd)
I will check your claim closer, i have the feeling that distortionlevels
would keep similar for same openloopgain if i omit the rload and use
re's in inputstage instead. But the topology i use for my symasym
is a bit crazy, with this heavy degeneration the tanh() curve of the
2nd diffamp should create massive compressive 3rd harmonics,
but it doesn't, for lower freqs it does the opposite, uncompressive...
Mike
obviously we have the one or other misunderstanding...
I agree with you that the benefits of degenaration are typically
lower than the loss of feedback resulting in less linearity for a closedloop.
But, if openloopgain is already too high, so that stabilizing gets
difficult/impossible, degeneration becomes a necessary step.
Maybe the combination of re AND vasload degeneration would be
the best choice.
Adding an extra followerstage and making a tripledarlington would
be the logical step, but i did not get it adequately stable, after it
was stable again distortion for 20khz had exploded.
This nonlinear load to the vas seems to be a bigger problem then
one would expect, especially in my case, as i use to92-bjts in vas
and have to keep currents low.
Especially for higher freqs, but still below resonancefrequency of
the feedbackloop, the dynamic load starts to create wild phasehifts
into the vas (even with tripledarlington), creating instability.
This dynamic load is created by the very unlinear nature of the big
output-powerbjts and can create a resonance in the very last stage.
Sims showed this resonance at 3mhz, but realworld had it even
higher... (i "cured" that with the caps to gnd)
I will check your claim closer, i have the feeling that distortionlevels
would keep similar for same openloopgain if i omit the rload and use
re's in inputstage instead. But the topology i use for my symasym
is a bit crazy, with this heavy degeneration the tanh() curve of the
2nd diffamp should create massive compressive 3rd harmonics,
but it doesn't, for lower freqs it does the opposite, uncompressive...
Mike
Hi jcx,
i checked this change, i made fft-plot of vasloaded amp (symasym5),
and amp without rloads, instead of them i used 8 ohms re's for first
diffamp. This gave same openloopgain/fb for 10khz.
Output is 12v peak into 4ohms 10khz.
Green is vasloaded, red is RE.
You're right, the difference is not big, but still, loaded vas has ~4db
less distortions. I should check this with complex signal...
Mike
i checked this change, i made fft-plot of vasloaded amp (symasym5),
and amp without rloads, instead of them i used 8 ohms re's for first
diffamp. This gave same openloopgain/fb for 10khz.
Output is 12v peak into 4ohms 10khz.
Green is vasloaded, red is RE.
You're right, the difference is not big, but still, loaded vas has ~4db
less distortions. I should check this with complex signal...
Mike
Attachments
Clarification
This may be a useless and pointless post for some of you but I just wanted to clarify something.
I noticed from a post related to my website http://www.biline.ca/critic4.htm the post contained "I guess you should read this content, knowing this is not a scientific or organizational website, but the homepage of a commercial speaker manufacturer."
I was never a commercial speaker manufacturer I mentioned my original interest in speaker building and my wish to try and start a business hence the research for a suitable name which ended up being 'biline'
This may be a small point but all of the website is completely free of any advertising and only contains the odd link to other sites I feel are of interest to others based on my exprience with the vender or website. To paint me as a commercial speaker vender undermines the whole point of my website. Which is to solely provide information that should read and taken into consideration before your next purchase.
thanks for reading
jeff@biline.ca
This may be a useless and pointless post for some of you but I just wanted to clarify something.
I noticed from a post related to my website http://www.biline.ca/critic4.htm the post contained "I guess you should read this content, knowing this is not a scientific or organizational website, but the homepage of a commercial speaker manufacturer."
I was never a commercial speaker manufacturer I mentioned my original interest in speaker building and my wish to try and start a business hence the research for a suitable name which ended up being 'biline'
This may be a small point but all of the website is completely free of any advertising and only contains the odd link to other sites I feel are of interest to others based on my exprience with the vender or website. To paint me as a commercial speaker vender undermines the whole point of my website. Which is to solely provide information that should read and taken into consideration before your next purchase.
thanks for reading
jeff@biline.ca
- Status
- Not open for further replies.