SY, this is a difficult question for me to answer accurately. I just don't know. For more than a decade, before laying out this circuit, I used low rbb' complementary followers. They worked, but I always put a large electrolytic cap to ground AFTER the emitter, as well as in the base to ground. When I was forced to use only a 0.1 uf cap on the output of the regulator, I decided it was probably quieter to use a jfet follower. The input board was actually laid out to use 2SK146, 2SJ73 pairs for the series pass devices. Later, I switched to higher Idss j113, J271(p) parts to allow for more total current. It has remained that way for 30 years.
It would be interesting whether a good bipolar would be as quiet, WITH only a 0.1uF cap at the output of each follower. Perhaps one of the PhD's in engineering out there can give me a definitive opinion. By the way, is this the broken section in the Vendetta you are repairing. If so, check the 0.1uF cap. They sometimes short, which blows the follower.
Many thanks. Yes, those transistors were indeed bad, and I was convinced that they hadn't jumped but were likely pushed. Your hint may save me a great deal of time this weekend. This particular unit has J175- do you think that a J271 will work about the same here?
People (including you) have gotten excellent noise performance at low impedances from low noise bipolars, and of course, their transconductance will be higher, so that's what got me thinking.
John, what are the main preferences of his pcb layout design? Symmetrical layout, low L,C, or what? I`ve seen the pictures but i didn't look closely to the pcb layout properties. Thans, Borko.
Indeed.
I did all our T&M pcbs (analog with digital interfaces) taping 2:1 on a light box. Lots of Kelvin paths and guard rings so difficult to draw for others to do.
Then we switched to Orcad schematic capture and Racal Redac layout on a small HP mini. Mini tape cartridge and no true line width on the display!! Quite a shock in 1980(?) but it taught me how to hold the design "in my head". We finally switched to Pads-pcb on a 386 pc and 12" colour monitor. That was luxury indeed, but it still took 30 secs to refresh a full screen display!
My boards look simple and elegant, nothing to do directly with art just doing the job properly and taking care of every last aspect of the design, I could say they look beautiful because they do, but I still do not consider it art.
The art or main skill of PCB design is placement, get that right and its then just joining the dots.....
I disagree to an extent and when teaching PCB design I don't use the term artist. PCB design is all about pattern recognition, reading a schematic, understanding intimately how signals propagate from DC to high frequency, loop areas, dipoles etc. etc.
But then this is audio so normal situations don't apply.....
Oh look up the old Vutrax posters, you'll like them.
I am making a stand here as to much in audio reproduction is considered art, and I would never compare my PCB design skills with the art of the musician, or with the ART I employ in painting or photography, different worlds.
1st pcb layout attempt.
I'd think that experience far outweighs the engineering part, which makes it a craft.
In Germany, one would be bestowed with the Master title.
(Imo, the majority of 'artists' are no different. 90% craftmanship, 10% tuition, and if they're unlucky, zero talent. Those with talent are commonly completely bonkers)
sorry kasey, looks like ou've got some "unlearning" to do
which can be difficult - but read Cordell's site, book, negative feedback thread for a much better treatment
the reason DVV may not want to get into the theory is because it demonstrates the error of this formulation and his own prejudices that he admits he is unwilling to reexamine, has never given Cordell, Cambrell, Cherry a fair read on the subject
for the "speed" of the feedback - try looking at the time evolution of the error V, it is not limited by the "100 Hz pole" - it is also a function of the loop gain magnitude at the error frequency - which can easily be a Million at 100 Hz in a high loopgain audio power amp
and we are not limited to Miller Dominant pole compensation when designing high feedback amps
the Margan paper is particularly stupid for trying to draw conclusions from Class B output crossover dead band - feedback does indeed "not work" when there are dead bands == no loop gain, no way to influence the output from the amp's inputs
Don't expect feedback to fix deadbands - don't design, build amps with deadbands - most here suggest >100mA per output Q AB bias
and don't fall for the Otala "prescription", flawed feedback heuristics - he got a very few things right but with his "TIM" writing majorly set back (or permanently impaired) the understanding of many
Originally Posted by kasey197
which can be difficult - but read Cordell's site, book, negative feedback thread for a much better treatment
the reason DVV may not want to get into the theory is because it demonstrates the error of this formulation and his own prejudices that he admits he is unwilling to reexamine, has never given Cordell, Cambrell, Cherry a fair read on the subject
for the "speed" of the feedback - try looking at the time evolution of the error V, it is not limited by the "100 Hz pole" - it is also a function of the loop gain magnitude at the error frequency - which can easily be a Million at 100 Hz in a high loopgain audio power amp
and we are not limited to Miller Dominant pole compensation when designing high feedback amps
the Margan paper is particularly stupid for trying to draw conclusions from Class B output crossover dead band - feedback does indeed "not work" when there are dead bands == no loop gain, no way to influence the output from the amp's inputs
Don't expect feedback to fix deadbands - don't design, build amps with deadbands - most here suggest >100mA per output Q AB bias
and don't fall for the Otala "prescription", flawed feedback heuristics - he got a very few things right but with his "TIM" writing majorly set back (or permanently impaired) the understanding of many
Originally Posted by kasey197
Last edited:
Art is an attitude. A quest of harmony. A mysterious mix of skills, knowledge, rigor, and...love. You can apply in all and every of our human activities.
You look at some schematics, it is full of sweat. Few others are just luminous ;-)
"Simple and elegant" ;-)
marce, yes, I'm sure you are" too damn modest". Or you don't realize what happens in your brain, when you do something, PCB design or ...painting and photography.
"get the right notes at the right place and its then just joining them" ( in the rigorous context given by the laws of harmony and rhythm) for a musician ? Same attitude. Same pleasure when, mysteriously, things are turning suddenly perfectly "right". Working at its best, and beautiful to look-at.
Last edited:
The starting point is usually someone claiming "i can hear a difference" while others stating that he can´t because it is impossible (see 1.) in my short list) .
While i agree (as i´ve surely said numerous times before) that the jumping to conclusion "i heared something and that has to be the reason" is unfortunately common, it is otoh also common to negate any hearing event if a faulty explanation could be rejected. Which is basically the same jumping to conclusion, as Jay already pointed out.
Did i really say, that it isn´t the "same problem" ?
But, if you think a scientific approach is mandatory then it should be real science.
Yes, I understand that.
I'm simply saying that if something is measured that is within known thresholds, there's no requirement of controlled listening tests. If it's outside known thresholds, you're going to have to do more to establish that it's actually audible at least to someone.
se
Yes, an inductor in series. But they come with a a bunch of disadvantages. For nuclear science years ago the JFET-based charge preamps often had inductors in the input FET drain load. But they were not high inductance, as the bandwidth required did not extend to low frequencies.
EDIT: hadn't read 67587 when I posted this.
Last edited:
A way to reduce the impact of current source noise
Inspired by EUVL's use of floating supplies (for I-V converters), I've looked at how those can free us up in various design areas.
In the case of the simplified pre-preamp shown by John in post 67475, if a floating current source with sufficient voltage compliance is used in place of R2 and R8, except for certain circuit impedances that are not perfectly complementary, the noise in the current source nulls at the output!
Of course we need to have outstanding isolation if we're still attempting to power things from the mains, or use batteries---and the latter are going to run down in a hurry with tens of milliamperes being pulled all the time.
It's cumbersome, and extracting that last bit of noise performance may be deemed not worth the bother. But it can be done.
Brad
Inspired by EUVL's use of floating supplies (for I-V converters), I've looked at how those can free us up in various design areas.
In the case of the simplified pre-preamp shown by John in post 67475, if a floating current source with sufficient voltage compliance is used in place of R2 and R8, except for certain circuit impedances that are not perfectly complementary, the noise in the current source nulls at the output!
Of course we need to have outstanding isolation if we're still attempting to power things from the mains, or use batteries---and the latter are going to run down in a hurry with tens of milliamperes being pulled all the time.
It's cumbersome, and extracting that last bit of noise performance may be deemed not worth the bother. But it can be done.
Brad
- Status
- Not open for further replies.