Nico Ras said:
Hi Nico,
Good investigation.
Maybe I'm missing something, but neither one of these square waves looks right. They are both asymmetrical. Is the amp clipping or slew rate limiting on these square waves? The initial overshoot on the first one is terrible.
As a starting point, we should look at small-signal square wave response comparisons, and these should be symmetrical.
Might the amps have been occasionally clipping in the listening tests? It happens more than we think on well-recorded music with good dynamic range. It is easy to believe that CCS vs bootstrap amplifier circuits would have different clipping characteristics and sound different.
Cheers,
Bob
I came to this conclusion after building too many CCS amps.
One can see/hear this by simming then building both types.
I found that one can moderate the "CCS effect" (overshoot)
by compensating the VAS side of the standard CCS.
A good analogy would be to visualize the CCS VAS as a spring,
as compared to the Bootstrap as a shock absorber.
I have found both to have their strengths and weaknesses.
The only topology I think can bring about the best of both
"camps" would be a balanced VAS or full comp.Diff.(leach)
(aksa says they sound bad but I guess I'll have to build one to find out).
These force the VAS linear (more like a tug of war) and give
a neutral sound (that is what we want,...for our amp to sound exactly like our soundcard/cd/preamp)I guess I'll find out in a few days .
OS
Hi Nico,
Your careful measurements revealed slightly lower levels of 5th, 7th and 9th harmonic. These are significantly bad sounding harmonics, are their origins in the overshoot you noted, which I've found also leads to offensive sibilance?
Nico, I wonder if, like me, you found that with the bootstrap you could relax the lag compensation slightly?
I'm learning LTSpice at present, finding it interesting, but am constantly reminded that correlations between reported distortion and the listening experience are thin at best......
I take my hat off to you for doing this important empirical research and sharing the results.
Cheers,
Hugh
Pete:
When you report me as saying full complementary sounds 'bad' you invite a volley of abuse from those that know better...... Frankly, I have listened to them at length, and they are clean and undistorted, but I prefer what I regard as the 'fuller' sound of a single LTP!
Your careful measurements revealed slightly lower levels of 5th, 7th and 9th harmonic. These are significantly bad sounding harmonics, are their origins in the overshoot you noted, which I've found also leads to offensive sibilance?
Nico, I wonder if, like me, you found that with the bootstrap you could relax the lag compensation slightly?
I'm learning LTSpice at present, finding it interesting, but am constantly reminded that correlations between reported distortion and the listening experience are thin at best......
I take my hat off to you for doing this important empirical research and sharing the results.
Cheers,
Hugh
Pete:
When you report me as saying full complementary sounds 'bad' you invite a volley of abuse from those that know better...... Frankly, I have listened to them at length, and they are clean and undistorted, but I prefer what I regard as the 'fuller' sound of a single LTP!
ostripper said:
I came to this conclusion after building too many CCS amps.
One can see/hear this by simming then building both types.
I found that one can moderate the "CCS effect" (overshoot)
by compensating the VAS side of the standard CCS.
A good analogy would be to visualize the CCS VAS as a spring,
as compared to the Bootstrap as a shock absorber.
I have found both to have their strengths and weaknesses.
The only topology I think can bring about the best of both
"camps" would be a balanced VAS or full comp.Diff.(leach)
(aksa says they sound bad but I guess I'll have to build one to find out).
These force the VAS linear (more like a tug of war) and give
a neutral sound (that is what we want,...for our amp to sound exactly like our soundcard/cd/preamp)I guess I'll find out in a few days .
OS
That analogy doesn't make a lot of sense. To be blunt, the waveform response of Nico's amp looks mediocre is either case.
This says more about the design as a whole, rather than anything absolute about the virtues and vices of a bootstrapped VAS or a CCS VAS.
With regards to your alledged "CCS effect" (overshoot), I've never seen this as an issue in any of the amps I've built and I certainly have never had to resort to "compensating" the CCS side of the VAS. If this is an issue then I suggest that there is something more complex going on.
Square wave and clipping CRO shots of one such basic (CCS VAS) design starting here:
http://www.diyaudio.com/forums/showthread.php?postid=1488531#post1488531
Cheers,
Glen
MJL21193 said:
I've only have scope shots of my 10W "CCS VAS" amp readilly at hand ATM, but this alledged CCS VAS "overshoot" doesn't exist in any properly implemented amplifier of such topology, 10 W or 1000W. Maybe if you do something silly like running a high Cob transistor for the CCS, biased from a high impedance source.
WRT the bootstrap VAS, it has issues of its own with respect to driver saturation and rail sticking on clipping that the CCS VAS doesn't have.
Cheers,
Glen
G.Kleinschmidt said:
I believe you, but how much of an issue is this really? You singled out PSRR in my thread, and there is no evidence of the ill effects of this in the final product.
To be honest, overshoot was a problem with mine, as you know. I have since increased the emitter resistors in the LTP and further reduced it. It is just noticeable without the output inductor.
As for clipping, I'm taking a new tack. I am working on a bjt input version of roenders RMI amp that is powered with 72VDC rails, but the input sensitivity limits it to 150 watts into 8 ohms. This will give headroom for transients without clipping. That's my crazy thinking, anyway.
MJL21193 said:
The amp with better PSRR will measure better - the thread at the time, I thought, was about optimising the design. Who said anything about "ill effects of this in the final product"?
BTW, I'm currently doing 100% of my listening on a bootstrapped quasi comp amp. Just don't feed me BS that a properly (or otherwise) implemented boostrapped VAS has some kind of intrinsic technical advantage or magical sonic quality over a properly implemented CCS VAS.
Cheers,
Glen
Wow , "by the fanatics, for the fanatics" really rings true with
the right participants.
It seems we have a CCS proponent here (GK).I agree a misdesigned CCS VAS will not sound good (I did it). CCS
VAS's are just less tolerant of bad device choices and/or
just a poorly designed CCS itself.
The bootstrap VAS, on the
other hand just sees a frequency dependant, slightly out of
phase version of itself allowing for much more "slop"
as far as design goes.(no offense, DX or Aksa, you
have "done your homework")I actually ran my frugal with
bootstrap/ MJE350 to confirm this and it sounded
wonderful. (it will be a very nice HT sub amp soon,replacing a
"brother of quasi" w/mj15023/24 OP's)
One can refine both topologies to their limits and create
great sounding amps.
Yeah, improper devices (MJE- 350 is a high Cob dinosaur)...
thread is here..
http://www.diyaudio.com/forums/showthread/t-81150.html
The 350 is much better suited to
drive a hi- impedance CRT grid than a
audio VAS. Very surprising that it is used so widely in audio
amps as it absolutely sucks at the job at hand. In conclusion,
the standard CCS "blameless" in its many flavors, including
many projects currently being built in this forum WILL show
the "CCS effect" to varying degrees (nothing complex,just
bad device choices).
OS
the right participants.
It seems we have a CCS proponent here (GK).I agree a misdesigned CCS VAS will not sound good (I did it). CCS
VAS's are just less tolerant of bad device choices and/or
just a poorly designed CCS itself.
The bootstrap VAS, on the
other hand just sees a frequency dependant, slightly out of
phase version of itself allowing for much more "slop"
as far as design goes.(no offense, DX or Aksa, you
have "done your homework")I actually ran my frugal with
bootstrap/ MJE350 to confirm this and it sounded
wonderful. (it will be a very nice HT sub amp soon,replacing a
"brother of quasi" w/mj15023/24 OP's)
One can refine both topologies to their limits and create
great sounding amps.
Yeah, improper devices (MJE- 350 is a high Cob dinosaur)...
thread is here..
http://www.diyaudio.com/forums/showthread/t-81150.html
The 350 is much better suited to
drive a hi- impedance CRT grid than a
audio VAS. Very surprising that it is used so widely in audio
amps as it absolutely sucks at the job at hand. In conclusion,
the standard CCS "blameless" in its many flavors, including
many projects currently being built in this forum WILL show
the "CCS effect" to varying degrees (nothing complex,just
bad device choices).
OS
ostripper said:
Oh no, how awfull.
ostripper said:Yeah, improper devices (MJE- 350 is a high Cob dinosaur)...
thread is here..
http://www.diyaudio.com/forums/showthread/t-81150.html
The 350 is much better suited to
drive a hi- impedance CRT grid than a
audio VAS. Very surprising that it is used so widely in audio
amps as it absolutely sucks at the job at hand. In conclusion,
the standard CCS "blameless" in its many flavors, including
many projects currently being built in this forum WILL show
the "CCS effect" to varying degrees (nothing complex,just
bad device choices).
Yeah, but I havent had this "CCS effect" (In terms of the horrible squarewave / clipping response as detailed in this thread) with MJE340/MJE350 CCS and VAS transistors either. So yeah again, I'd say that in such cases there is something more complex going on (stuffed up) than just a possibility of "improper" devices.
Cheers,
Glen
EDIT:
I haven't found a set of SPICE models for the MJE340/MJE350 that were not screwed up, thus predicting far worse than real-life performance. But I have not looked that hard either.
NO **** sherlock, but, I didn't want to stress anyone out
with a long dissertation as to exactly why. (I do know)
The ONLY point I was making was that the wrong components
in a "schoolbook" amp (as Lineup put it) will cause overshoot
as was seen on this thread. It can be corrected somewhat
by OLG,Cdom, or lower imp. , but switching to a better device
is by far the best solution.
As far not knowing, I do comprehend the effects of
transconductance vs linearity, Cob, etc and the basic physics of
a transistor junction.... So feel free to throw some valuable
"over my head" insight my way. today I will scratch my head
but tommorrow I will know.
wrong smiley thats better.OS
ostripper said:
NO **** sherlock, but, I didn't want to stress anyone out
with a long dissertation as to exactly why. (I do know)
The ONLY point I was making was that the wrong components
in a "schoolbook" amp (as Lineup put it) will cause overshoot
as was seen on this thread. It can be corrected somewhat
by OLG,Cdom, or lower imp. , but switching to a better device
is by far the best solution.
As far not knowing, I do comprehend the effects of
transconductance vs linearity, Cob, etc and the basic physics of
a transistor junction.... So feel free to throw some valuable
"over my head" insight my way. today I will scratch my head
but tommorrow I will know.
OS
Wow.
I think the best solution by far is to get the "schoolbook" amp right in the first place - then contemplate better devices.
I hope that doesn’t cause any further distress.
yes , I realize this and should have kept to the schoolbook
(Mr. self does not use 340/350's for the CCS/VAS, only as drivers.)
At least I did not follow the crowd and do this...
[this amp was run for awhile then "freaked out"]
Not a bad design ,but a high Cob non EF'ed 340/350
with the constructor using junkbox Cdom of wrong value
is the route I chose not to take (I used 68p silver mica and
improved the circuit) so my stress was less.
If I blew one I would post it so everyone could "gloat"
Most definitely!!!
The same circuit with better devices has stability
at a much lower Cdom. no over shoot. performs as the "schoolbook" says it should(even better).
Not to satisfy the bootstrap crowd , it is interesting to note
that a bootstrapped mje350's sound/stability matches the
CCS with better components.Guess I just lucked out
making the right choices even before I knew the technical
aspects..
OS
(Mr. self does not use 340/350's for the CCS/VAS, only as drivers.)
At least I did not follow the crowd and do this...
[this amp was run for awhile then "freaked out"]
Not a bad design ,but a high Cob non EF'ed 340/350
with the constructor using junkbox Cdom of wrong value
is the route I chose not to take (I used 68p silver mica and
improved the circuit) so my stress was less.
If I blew one I would post it so everyone could "gloat"
Most definitely!!!
The same circuit with better devices has stability
at a much lower Cdom. no over shoot. performs as the "schoolbook" says it should(even better).
Not to satisfy the bootstrap crowd , it is interesting to note
that a bootstrapped mje350's sound/stability matches the
CCS with better components.Guess I just lucked out
making the right choices even before I knew the technical
aspects..
OS
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.