I quoted ostripper over a week ago here and asked among many other things, how his giant NE5532 is outperforming the original chip with much longer vias? Other questions concerned the measured attributes of his designs with real parts (not from simulation). I'm particularly interested in the gain and phase margin measurements because they are highly indicative of the amps stability.
Thanks thimios!
When ostripper says the discrete circuits are embellished as compared to the chips, it suggests that the chips could incorporate these embellishments in a later design iteration. The point being that the Symasui is not just a larger version of the NE5532 even though the basic design is the same.
When ostripper says the discrete circuits are embellished as compared to the chips, it suggests that the chips could incorporate these embellishments in a later design iteration. The point being that the Symasui is not just a larger version of the NE5532 even though the basic design is the same.
Last edited:
Did anyone do a distortion analysis on the Symasui (or any of the other ostripper designs) with harmonic profiles?
You really should use the search feature to find some of the answers to your questions. That is what we would have to do at this point. Thimios built every IPS and posted his test results. He also posted links to each of his builds. OS has posted Sim results from each as well.
Is this(post#3200) what are you looking for?http://www.diyaudio.com/forums/solid-state/248105-slewmaster-cfa-vs-vfa-rumble-320.html
still4given
I looked at the links thimios provided before the one recently posted just above and they were scope shots only. We are approaching 1000 pages of content in this thread alone and the searches I've tried have not yielded results describing distortion harmonics. Given the background, I don't think it was unreasonable to ask for some guidance but your reprimand is noted.
thimios
Yes, the distortion plots are the most important and telling of the amplifier sonic character. It looks like the amp has very little harmonic preference. I tend to like the sound of slightly emphasized even order harmonics but the noise floor is pretty darn good. I saw another one of your charts showing a 1kHz plot with ~50dB of differential from the 50Hz and I am assuming this is the psrr across the whole band, is this correct? The scope shots give a good indication of the amp behavior but not much else. I can't do a search for PM and GM to look for the phase and gain margin specs and spelling it out hasn't yielded what I'm looking for, can you or anyone else point me to those measured specs?
I looked at the links thimios provided before the one recently posted just above and they were scope shots only. We are approaching 1000 pages of content in this thread alone and the searches I've tried have not yielded results describing distortion harmonics. Given the background, I don't think it was unreasonable to ask for some guidance but your reprimand is noted.
thimios
Yes, the distortion plots are the most important and telling of the amplifier sonic character. It looks like the amp has very little harmonic preference. I tend to like the sound of slightly emphasized even order harmonics but the noise floor is pretty darn good. I saw another one of your charts showing a 1kHz plot with ~50dB of differential from the 50Hz and I am assuming this is the psrr across the whole band, is this correct? The scope shots give a good indication of the amp behavior but not much else. I can't do a search for PM and GM to look for the phase and gain margin specs and spelling it out hasn't yielded what I'm looking for, can you or anyone else point me to those measured specs?
It seems like a reasonable request when talking about PPM level distortion from amplifiers. From what I understand, measuring distortion figured this low is quite difficult. Beyond that, you have the ever imposing question of whether or not any of this correlates with a preferred sound. Engineering wise, they seem to be a spectacular achievement, but have you ever heard an engineer sing.
Very predictable ....
All the Slewmaster simulations follow a simple rule.
- CCS based single LTP IPS's = nearly equal even/odd order cancellation.
Symasui and Wolverine are very similar (below1).
-Symmetrical IPS's , like K-ND and spooky (below 2) cancel the even's
out nearly completely.
Having more of the even's is said to be more musical. Perhaps with an amp
that is >100ppm , this can be discerned. The problem is , at <10ppm -
both (type)amps are so perfect , they are nearly indistinguishable.
OS
All the Slewmaster simulations follow a simple rule.
- CCS based single LTP IPS's = nearly equal even/odd order cancellation.
Symasui and Wolverine are very similar (below1).
-Symmetrical IPS's , like K-ND and spooky (below 2) cancel the even's
out nearly completely.
Having more of the even's is said to be more musical. Perhaps with an amp
that is >100ppm , this can be discerned. The problem is , at <10ppm -
both (type)amps are so perfect , they are nearly indistinguishable.
OS
Attachments
No, not unreasonable and not a reprimand. When a thread gets this long and we have built so many of the IPS sections, the only way we can find the answers to questions is by using the search feature. For folks who use the forum a lot, learning how to use the search feature is important. I was trying to be helpful, not critical. Sorry if it came off that way.
I have done listening tests on each of these IPS through an A/B setup with good speakers. They all sound the same. Absolutely accurate. If you are seeking a colored sound I would suggest an ASKA or maybe an SKA GB150. Every one of the Slewmaster typologies are wire with gain.
Blessings, Terry
Think it had been announced that RNMarsh by occasion should real world pro benchmark various builds this thread, when or if that happen wouldn't it answer terrences request.
terrences,
Make advanced search this thread for "User Name" "5th element", then you will find good build documentation layout/wire ring/alternative input pair and real world distortion measurements.
terrences,
Make advanced search this thread for "User Name" "5th element", then you will find good build documentation layout/wire ring/alternative input pair and real world distortion measurements.
Last edited:
An NE5532 might match the symasui at a few volts , but they are indeed
different at the 100V p-p level.
A dual op-amp is on a 3mm sq. die , both amps are powered from one bus
internally. PSRR is fully dependent on the amps design. You can't individually
decouple amp 1 from 2.
Some OPAxxxx single op-amps might beat the symasui and a few LMExxxx
headphone amps can match a Kypton-ND or Wolverine in raw THD numbers.
Still , not at 100V p-p.
Don't be afraid of the slew series IPS's. Worst case , they are overcompensated.
Wolverine and symasui , I've gone down to <33pF .... still no ringing
into a piezo tweeter or cap.
The symmetricals (K-ND/spooky) , can nearly run on the main VAS Cob
without compensation. The 33-47p specified gives quite a lot of "slop" ...
gain and local degeneration can be changed with stability.
The only oscillators you can create are if you degenerate the input pairs
with 10R ... or some other obscene value.
I've done this by misreading a few resistors on a few of my builds.
Or , input pairs that have obscene beta (>500Hfe). Here you can
increase degeneration 100-150R - they too will work.
You can screw up an output stage (and have it oscillate). Thimios used
low Ft outputs (MJL21193 types) - had an oscillator.
Best is a very fast prediver/driver and standard 35mhz BJT outputs - rock
stable.
OS
Not exactly a 5532 but his Glasui should allow very good perfs with some rework and the inherent proper compensations, the below is what i got with some slight modifications.
Attachments
You mean the sansui with the single ended second stage ? And the
level shifter (for the VAS) ?
I progressed to the fully symmetrical 2'nd stage with the kypton-V , but
it is basically the same as the old sansui (3900Z receiver) . Post
your changes , I might bring back the "V" in either form 😎 (as an IPS).
Wahab - you show 1K ?? I can easily get the 2'nd harmonic down >120db on
all the slewmaster designs @ 1K !
OS
Last edited:
ostripper
Agreed, the small die will limit the power capability but for smaller signals, the shorter vias should trump most any layout of a discreet design. I've been on this webpage and his alternative grounding example might come in handy to try on your very fast amps.
BYRTT
Thanks for the guidance. I looked at 5th elements posts and he does provide the harmonic distortion numbers but they are only for his CFA and not the SYMSUI. Also, he is using a very low voltage in order to compare Class A operation. I am not interested in a Class A amplifier and have a choice of 65V and 78V rails for a small bias class AB build. I've noted in his measurements the elevation of the odd harmonics the larger the voltage swings. Is this attributable only to his CFA or do all the ostripper designs share this characteristic? As I mentioned before, I prefer the sound of de-emphasized odd harmonics. The only odd harmonic that can sound right to me is H3 and that is only when it is dominant over all the other harmonics. A dominant H3 adds a certain richness that is not as sweet and reverberant as a dominant H2. The right amounts of H2 and H4 sounds similar to a dominant H3 on my speakers (Ohm F3 updated). Given what I've said about my preferences, Is my choice of a SYMSUI build a good one or should I be looking at one of the others?
For those interested, I've recently asked RNMarsh about his offer to measure these amps and he said he is still overseas without any plans to return home anytime soon. Well, taht's good for him abut not so good for us. The person I linked above has some very good measurement equipment and teh skills to produce accurate data. He may be persuaded to help if asked.
Agreed, the small die will limit the power capability but for smaller signals, the shorter vias should trump most any layout of a discreet design. I've been on this webpage and his alternative grounding example might come in handy to try on your very fast amps.
BYRTT
Thanks for the guidance. I looked at 5th elements posts and he does provide the harmonic distortion numbers but they are only for his CFA and not the SYMSUI. Also, he is using a very low voltage in order to compare Class A operation. I am not interested in a Class A amplifier and have a choice of 65V and 78V rails for a small bias class AB build. I've noted in his measurements the elevation of the odd harmonics the larger the voltage swings. Is this attributable only to his CFA or do all the ostripper designs share this characteristic? As I mentioned before, I prefer the sound of de-emphasized odd harmonics. The only odd harmonic that can sound right to me is H3 and that is only when it is dominant over all the other harmonics. A dominant H3 adds a certain richness that is not as sweet and reverberant as a dominant H2. The right amounts of H2 and H4 sounds similar to a dominant H3 on my speakers (Ohm F3 updated). Given what I've said about my preferences, Is my choice of a SYMSUI build a good one or should I be looking at one of the others?
For those interested, I've recently asked RNMarsh about his offer to measure these amps and he said he is still overseas without any plans to return home anytime soon. Well, taht's good for him abut not so good for us. The person I linked above has some very good measurement equipment and teh skills to produce accurate data. He may be persuaded to help if asked.
Last edited:
Yes that s the Sansui 3900Z, as for the 2nd harmonics it is at -98dBV while the fundamental is at +30.37dBV, at 10KHz H2 s at -75dBV.
With decent stability it s difficult to do better, and when i say decent it means up to 1uF loading without LR network, of course this imply drasticaly reduced slew rates, but as aknowledge by the IMD graph very high speed is not necessary to get very low intermodulation.
I looked at that chip amp website. We (I) was at that "level" back in the
pre-badger days !!
Grounding for a simple LM3886 is not the same as a 400W OPS/IPS. I
have "G1/2" and a lifted ground reference , multiple stages of decoupling
and cap multipliers.
I DO have a spare OPS and one each of the wolverine/K-ND that I could
ship (somewhere) ??
Who ... is the question. I can't measure most of these amps on a PC based
FFT. The "someone /somewhere " - would need to have a 20K$ AP.
Short of that , I won't do it.
This is why I would of sent one to RN - he has one (two?).
OS
pre-badger days !!
Grounding for a simple LM3886 is not the same as a 400W OPS/IPS. I
have "G1/2" and a lifted ground reference , multiple stages of decoupling
and cap multipliers.
I DO have a spare OPS and one each of the wolverine/K-ND that I could
ship (somewhere) ??
Who ... is the question. I can't measure most of these amps on a PC based
FFT. The "someone /somewhere " - would need to have a 20K$ AP.
Short of that , I won't do it.
This is why I would of sent one to RN - he has one (two?).
OS
LM3886 vs a slewmaster !!
I've owned various 3886 logitech's , and surround receivers with 3886's.
THEY ARE NOISY ! I can hear them (Hsassssss...)
I just had to buy a -120db Asus Xonar to enjoy the silence of my slews.
Now - PC on or off , I hear NOTHING (ear on tweeter).
Slewmaster's are forcing me to upgrade my sources and even my
loudspeakers. None of my OEM's made these weaknesses quite as
apparent.
Even the Badger builders express dismay at the jump in performance
over their store bought equipment. Many start threads to ask how they
can either improve these replaced OEM's or just want to sell them outright.
Slew's are even a little better than Badger's ..... !
OS
I've owned various 3886 logitech's , and surround receivers with 3886's.
THEY ARE NOISY ! I can hear them (Hsassssss...)
I just had to buy a -120db Asus Xonar to enjoy the silence of my slews.
Now - PC on or off , I hear NOTHING (ear on tweeter).
Slewmaster's are forcing me to upgrade my sources and even my
loudspeakers. None of my OEM's made these weaknesses quite as
apparent.
Even the Badger builders express dismay at the jump in performance
over their store bought equipment. Many start threads to ask how they
can either improve these replaced OEM's or just want to sell them outright.
Slew's are even a little better than Badger's ..... !
OS
Why would be necessary that an amp should be stable with no LR network at the output with 1uF loading, if just 0.5 uH//2R2 will make it rock stable up to 2-4 uF??