No , but there is the finished "SYMasui" .
Finally have it done ! 😱
(below 1/2) is the "pretty " PCB.
(below 3) is the schema it came from.
Went back to the standard symasym , more stable. Added the capacitive
shunts like the original .... they bring the THD harmonics right in line with the
original project. These are all optional , of course ... the IPS will run
perfectly with just one compensation cap.
But ... this amp "buries" the original in every other way. 😎
Will release the whole package after "proofing" 😱 .
OS
Finally have it done ! 😱
(below 1/2) is the "pretty " PCB.
(below 3) is the schema it came from.
Went back to the standard symasym , more stable. Added the capacitive
shunts like the original .... they bring the THD harmonics right in line with the
original project. These are all optional , of course ... the IPS will run
perfectly with just one compensation cap.
But ... this amp "buries" the original in every other way. 😎
Will release the whole package after "proofing" 😱 .
OS
Attachments
found 3 mistakes already (I think that's it) D4 reversed and wrong/mislabeled
servo caps. Fixed (below).
I don't know .. Carl , I might take a break ... refine/debug these 3 main amps -
(gnome/spooky/symasui).
PS- .asc for error hunting below , as well.
OS
servo caps. Fixed (below).
I don't know .. Carl , I might take a break ... refine/debug these 3 main amps -
(gnome/spooky/symasui).
PS- .asc for error hunting below , as well.
OS
Attachments
Last edited:
Ah come on, you know that you want to! It would round out your collection. Ha, ha ...
OS, please take a much deserved break and let us builders catch up with you.....😀
i wish that you can have a place somewhere where the latest and mightiest revisions can be seen, all in one place....
i wish that you can have a place somewhere where the latest and mightiest revisions can be seen, all in one place....
Last edited:
Well ... here it is !
(attached ) is the "package" (debugged).
Let me explain it. M. Bittners "symasym" uses the Otala style VAS (CM +
differential) with a simple LTP input pair ...
he then adds adds R/C shunt compensation to the VAS's output.
I simulated the original side by side with my design.
Cvs1/Cvs2 (below 1 schema) , reduces total closed loop gain and lowers the
UG. The "side effect" of this reduces H3/5/7 even more ... but increases
H2/4. My design can compensate by adding a small amount of lead
compensation (Clc ... 2.2-2.7pF) , thus allowing the reduction of Cdom (C7).
This increases the UG point and slew rate back to where we started.
Bittner used 100K/330pF for his shunts. My design has 5X the impedance
(cascode/wilson)and drives an EF3 .... 330K/47pF, accomplishes the same task.
In the end , I simulated the original and mine - both had the same FFT. Should
sound almost the same.
What is different , most of my devices can be low voltage ... except
for Q3-4 /Q8,9,10 (regulated front end). THD is 1/5th the symasym (5ppm) at
60V p-p and 20 at 120V.
It can also go where the symasym can't ... just like the spooky , 150V p-p
is EASY for this IPS @ 25ppm.
NOT the same hawksford VAS , Not a symmetrical design (despite the name).
This should have a totally different "character" than the balanced IPS's.
The FFT says so .
-Balanced = cancellation of H2 /4 ... even harmonics.
-unbalanced (this amp)= almost no H3/5/7 ... odds are gone.
PSRR is also -110db or better across the audio spectrum.
PS - Cvs1/2 and Clc are totally OPTIONAL. IPS will run happily on just C7 (33-47pF).
OS
(attached ) is the "package" (debugged).
Let me explain it. M. Bittners "symasym" uses the Otala style VAS (CM +
differential) with a simple LTP input pair ...
he then adds adds R/C shunt compensation to the VAS's output.
I simulated the original side by side with my design.
Cvs1/Cvs2 (below 1 schema) , reduces total closed loop gain and lowers the
UG. The "side effect" of this reduces H3/5/7 even more ... but increases
H2/4. My design can compensate by adding a small amount of lead
compensation (Clc ... 2.2-2.7pF) , thus allowing the reduction of Cdom (C7).
This increases the UG point and slew rate back to where we started.
Bittner used 100K/330pF for his shunts. My design has 5X the impedance
(cascode/wilson)and drives an EF3 .... 330K/47pF, accomplishes the same task.
In the end , I simulated the original and mine - both had the same FFT. Should
sound almost the same.
What is different , most of my devices can be low voltage ... except
for Q3-4 /Q8,9,10 (regulated front end). THD is 1/5th the symasym (5ppm) at
60V p-p and 20 at 120V.
It can also go where the symasym can't ... just like the spooky , 150V p-p
is EASY for this IPS @ 25ppm.
NOT the same hawksford VAS , Not a symmetrical design (despite the name).
This should have a totally different "character" than the balanced IPS's.
The FFT says so .
-Balanced = cancellation of H2 /4 ... even harmonics.
-unbalanced (this amp)= almost no H3/5/7 ... odds are gone.
PSRR is also -110db or better across the audio spectrum.
PS - Cvs1/2 and Clc are totally OPTIONAL. IPS will run happily on just C7 (33-47pF).
OS
Attachments
Last edited:
I have heard a lot of people talking about badly designed DC servos in audio circuits, does anyone know of a source of information on these problems, so I can get up to speed with everyone else?
Who are "lots of people" ...and what servo ?
Most servo's I've seen are 30 years running. 😕
I would not use them if I thought they would fail ...
PS - they are just "bombed out' ... and are too scared to add some 21'st century additions.
50c chip don't scare me !
OS
Last edited:
Hi OS
How necessary is to put the servo in such topology if match well the LTP and VAS ?
can we put ksa1015=bc560 and ksa992 ksc1845 =2n5401-2n5551 ?
what is the substitutability transistor?
Nikos
How necessary is to put the servo in such topology if match well the LTP and VAS ?
can we put ksa1015=bc560 and ksa992 ksc1845 =2n5401-2n5551 ?
what is the substitutability transistor?
Nikos
Last edited:
Many years ago I had built the Titan 2000 with LTP diamond topology match very good without servo 5-10mv with servo 1-6mv offset , but the sound badly .
Attachments
Last edited:
Must be something wrong with servo. In my latest prototype I've got less that 0.1 mV offset and no influence on sound with servo.
Hi,
I assume that Keantoken rather referred to the sonic implication of servos and not the possibility of failure.
The typical first-order-filter inverting servo still puts out some signal at as high as 1kHz and above.
In a configuration like in #1624 its corrective voltage is fed as error signal into the most sensitive point of the amplifier.
Imho no wonder that it may influence on sonic character.
I see two possible solutions.
First is the use of more filtering, for example post LP-filtering the servo, to get rid of artefacts within the audible range.
Or second to find a less sensitive point to feed the error signal into.
This point could e.g be a source/emitter of a current source or current mirror.
jauu
Calvin
I assume that Keantoken rather referred to the sonic implication of servos and not the possibility of failure.
The typical first-order-filter inverting servo still puts out some signal at as high as 1kHz and above.
In a configuration like in #1624 its corrective voltage is fed as error signal into the most sensitive point of the amplifier.
Imho no wonder that it may influence on sonic character.
I see two possible solutions.
First is the use of more filtering, for example post LP-filtering the servo, to get rid of artefacts within the audible range.
Or second to find a less sensitive point to feed the error signal into.
This point could e.g be a source/emitter of a current source or current mirror.
jauu
Calvin
vzaichenko
The sound quality does not change with or without the servo.I m talking for Titan 2000.
The sound quality does not change with or without the servo.I m talking for Titan 2000.
Last edited:
Interesting. I deliberately tried to see some artifacts at DC servo output in my CFA prototype (non-inverting servo). I fed a square wave through the amplifier, having about 20W output power @ 4ohm resistuve load. 100 Hz, 500 Hz, 1000 Hz. Nothing. Just very low level HF noise, not influenced by the signal.
I will run some tests at the lower end (around 20 Hz) over the weekend.
Nikos, sorry, then I did not understand your initial post about Titan correctly.
No problem .
when you say 0.1 mV offset on prototype - which amp ?
and without servo - offset?