May be he want a cheap amplifier for mass production and sacrifice a little bit THD and slew rate?
All amplifiers are absolutely silent.
The 50Hz plot is a matter of measurement process
I concur on the output filter for stability, particularly with a fast amp. Not using one should be reserved for academic exercises only. The only time I would consider not using one would be on a class A design (continuous signal output). In this case the output filter may mask other defects that should be resolved earlier in the amp.
ostripper
Tomchr (the person who runs that website I linked above) has the AP equipment you specified and the knowledge to provide the accuracy. Try not to be so dismissive of his alternate grounding schemes. His results prove that the star ground is not always the optimal ground layout when the devices are switching so quickly.
thimios
They would probably be dead silent on my main speakers too (88dB) but maybe not on more sensitive drivers.
ostripper
Tomchr (the person who runs that website I linked above) has the AP equipment you specified and the knowledge to provide the accuracy. Try not to be so dismissive of his alternate grounding schemes. His results prove that the star ground is not always the optimal ground layout when the devices are switching so quickly.
thimios
They would probably be dead silent on my main speakers too (88dB) but maybe not on more sensitive drivers.
Last edited:
When i say death silent i'm sure for this.
I have test this with sensitive speakers and sensitive headphones also.
Look here a similar subject post #4799.http://www.diyaudio.com/forums/solid-state/164093-100w-ultimate-fidelity-amplifier-480.html
Last edited:
Thimios
I don't want to doubt you but the performance chart you posted earlier indicates to me that a sensitive speaker will make noise at full power output with that amp. There are amps that can do dead silence at three digit power on a sensitive driver but they are much rarer then you would imagine because of the many factors that can introduce it. Good PSRR is not all you need. Those living in a busy metropolitan city have a harder time with this ideal than those living in more relaxed environments. Do you agree with my choice of Symasui for my requirements (even harmonic bias or strong single H3 character and 65-75V rails) or do you believe I will like some other design better? The THD+N numbers on the ultimate fidelity amp you linked are not so impressive at full power output. Ideally an amp should specify its THD+N at just below clipping power or wherever the worst case THD+N is.
I don't want to doubt you but the performance chart you posted earlier indicates to me that a sensitive speaker will make noise at full power output with that amp. There are amps that can do dead silence at three digit power on a sensitive driver but they are much rarer then you would imagine because of the many factors that can introduce it. Good PSRR is not all you need. Those living in a busy metropolitan city have a harder time with this ideal than those living in more relaxed environments. Do you agree with my choice of Symasui for my requirements (even harmonic bias or strong single H3 character and 65-75V rails) or do you believe I will like some other design better? The THD+N numbers on the ultimate fidelity amp you linked are not so impressive at full power output. Ideally an amp should specify its THD+N at just below clipping power or wherever the worst case THD+N is.
Why in the world would you plan to run your amp at just below clipping ? That's just nuts. Build a bigger amp.
I don't know who else is running the symasui now , but I am.
Since you believe in "sonic character", let me elaborate. Very Very
hard to discern the Sym from the Kypton .... but it has real good
bass definition on techno/rock - kicks butt on movies. It is at least
many orders more quiet than my H/K680 at -107db S/N.
Objectively , it is nearly the "worst" at 120Vp-p 20K (40ppm). I
don't think that extra 30ppm is really audible.
The sym's are in my amp now - I see no reason to remove them.
They exceed my pioneer receiver , my harmon -kardon .... why
should I be anal and stress over 30ppm ?
Electrically , the symasui runs cool and should outlast your electrolytic's.
OS
Ok if you don't believe me you must built and test it using your ears.
The only reason this link posted ( to ax-14) is to shown difference when star gnd(psu gnd) used as speaker gnd and when using a separate star gnd for speaker return.
You must not see these measurements as accurate because a cheap sound card was used.
What IPS you like?
Really i don't know.
All these IPS are in very high level(to my opinion).
I don't necessarily plan to but I would like my amp to not degrade if I do. You are right, you can build a bigger amp but that introduces other negatives and possibly puts speakers at risk.still4given said:
thimios
Don't be offended by my skepticism. I think too many of us need more of it, including me 🙂
ostripper
That's quite an endorsement but its subjective nature makes me cringe slightly. I was attracted to the Symasui because I like the simplicity of the design, the flexibility in parts selection and the potential for layout optimizations to improve the performance further. I know you don't have the equipment to do the harmonic profiling but with RNMarsh MIA (missing in Asia), my earlier suggestion for a competent tester still stands as a good one. Have you reached out a feeler to gauge his interest?
Last edited:
Begun work on the "unislew" (all in one).
Even as I will eventually have a couple tested , I'm quite convinced
these designs trump most OEM's.
I not "looking back" -
New slews will have - (below)
- Full rail cancellation with alternating outputs (P/N/P/N .....)
- Non- inductive parallel emitter resistors and an Ayrton-perry output inductor.
- Kypton-ND , Symasui , or Spooky versions.
-Integrated DC protect with FET SS relays.
-3 pair TO-3P or MT-200 output and a smaller fast Toshiba driver stage.
Oh , and a book to go with them.
OS
Even as I will eventually have a couple tested , I'm quite convinced
these designs trump most OEM's.
I not "looking back" -
New slews will have - (below)
- Full rail cancellation with alternating outputs (P/N/P/N .....)
- Non- inductive parallel emitter resistors and an Ayrton-perry output inductor.
- Kypton-ND , Symasui , or Spooky versions.
-Integrated DC protect with FET SS relays.
-3 pair TO-3P or MT-200 output and a smaller fast Toshiba driver stage.
Oh , and a book to go with them.
OS
Attachments
Building , listening , and tweaking these present creations is "school".
VERY hard to actually improve these base designs.
I have seen the current Accuphase , Pass creations - they have gone
the "twisted triplet" route right on to the main output stage ....
full cancellation right down to the last BJT.
Adding the special output inductor and paralleled emitter resistors is a
sourcing thing. I'm actually looking at current cheap mouser stock
to make these design decisions.
If I combine the current designs with these "anal" layout techniques ,
I believe we will have true "high end" in DIY form.
Also , a careful choice of the standard Nichicon caps ,
Vishay non-inductives and our standard Sanken/ON semi's will augment
this "next generation" .
OS
VERY hard to actually improve these base designs.
I have seen the current Accuphase , Pass creations - they have gone
the "twisted triplet" route right on to the main output stage ....
full cancellation right down to the last BJT.
Adding the special output inductor and paralleled emitter resistors is a
sourcing thing. I'm actually looking at current cheap mouser stock
to make these design decisions.
If I combine the current designs with these "anal" layout techniques ,
I believe we will have true "high end" in DIY form.
Also , a careful choice of the standard Nichicon caps ,
Vishay non-inductives and our standard Sanken/ON semi's will augment
this "next generation" .
OS
os, if you keep cranking these out like this, it's very difficult to settle down on a design and build it. Sure, they are all so good that the differences are inaudible, but for those who cannot accept that there is something better that could've been built, this is hell.
I still have a SlewMonster 5 pair board and CFA-XH IPS I still haven't built. Mainly because by the time I built a BOM to order stuff, you had started cranking out the Kryptons.
But I think I like this version of hell. 🙂 Keep up the good work man.
I still have a SlewMonster 5 pair board and CFA-XH IPS I still haven't built. Mainly because by the time I built a BOM to order stuff, you had started cranking out the Kryptons.
But I think I like this version of hell. 🙂 Keep up the good work man.
ostripper
Going forward in your designs, I will remind you that there are already amps with <5ppm worst case THD+N distortion at full power (Halcro, etc.) and there is a growing sentiment that they all seem to exhibit a certain lifelessness compared to simpler designs. John Curl is not the only prominent designer to espouse the notion that more line stages/parts degrade the original signal. There are many reasons for this, not the least of which is the imperfections of the parts. I am a bigger fan of refining an excellent simple topology with layout and parts selection/matching than adding more stages in order to achieve distortion crushing PSRR and phase cancellations which can also crush some of the hairiness in the original signal in ways that don't show up on distortion measurements as we do them today. I would also take a close look at the output stage of the krill amp. Steve Dunlop (Dunlap?) is another audio design giant with a highly underrated accomplishment in my opinion. His output stage addresses crossover distortion more effectively than any other I know of and may be the puzzle piece to one of the slew input stages that makes the greatest DIY amp ever.
Going forward in your designs, I will remind you that there are already amps with <5ppm worst case THD+N distortion at full power (Halcro, etc.) and there is a growing sentiment that they all seem to exhibit a certain lifelessness compared to simpler designs. John Curl is not the only prominent designer to espouse the notion that more line stages/parts degrade the original signal. There are many reasons for this, not the least of which is the imperfections of the parts. I am a bigger fan of refining an excellent simple topology with layout and parts selection/matching than adding more stages in order to achieve distortion crushing PSRR and phase cancellations which can also crush some of the hairiness in the original signal in ways that don't show up on distortion measurements as we do them today. I would also take a close look at the output stage of the krill amp. Steve Dunlop (Dunlap?) is another audio design giant with a highly underrated accomplishment in my opinion. His output stage addresses crossover distortion more effectively than any other I know of and may be the puzzle piece to one of the slew input stages that makes the greatest DIY amp ever.
Last edited:
I haven't had much time to listen to the Kypton-ND yet, but I have listened to the Kypton-C for extended periods of time. It is not "lifeless". At the same time it does not colour the sound.
These amplifier designs are very inexpensive and easy to source parts for. Maybe build some and try them. Repeating the same discussion that has already been done in loads of other threads doesn't really lead to any advancements in design. Designing, building, and testing does.
These amplifier designs are very inexpensive and easy to source parts for. Maybe build some and try them. Repeating the same discussion that has already been done in loads of other threads doesn't really lead to any advancements in design. Designing, building, and testing does.
I was the one who first simulated steves "krill" OPS.
Built a krill even ...
Also owned a "Super A" JVC. And had the greatest Class A "genesis" to
audition for a month.
Was I impressed ? The Class A impressed me - yes. JVC and Krill
were good , as well.
But this EF3 , based on the harmon kardon either matches or exceeds
with some of the better input stages.
Believe me , I would go with an error correcting OPS (hyperbolic - whatever),
if I thought that all this would bring any substantial improvement to the table.
The EF3 is the perfect balance of simplicity to performance - I don't have to
"babysit" builders with assorted problems !!!
It really don't - A properly designed EF3 with the right "suckout" cap ,
local decoupling and layout matches any of those less stable , "fancy"
OPS designs.
Technically , adding local FB loops to the OPS reduces slew and stability.
Adds nearly nothing that a little more global NFB can't negotiate.
OS
Last edited: