Thanks M. 
. It's what I suspected, but not as much for Cgs and Cdg.
I have another question about a fact that I can’t explain completely and of which I partially talk here during the discussion on dc-servos : Carrying out some tests on the input stage (fig 1) and the dc-servo of a power amp (in spite of the fact that I used a small value resistor for R after the integrator), I get much better results with the (fig 2) and a quite softer sound.
I thought that the problem could come from mixing 2 references, input ground and dc servo ground, then I tested the (fig 3) imagining that it could be a problem of ground currents and ground loops (I talk here about a power-amp), and in this case the sound was much more softer and more “live”. Do you think as I it suspect but I am not completely sure that it is a ground problem , have you another or better explanation?
Fig1-fig2-fig3
. It's what I suspected, but not as much for Cgs and Cdg.I have another question about a fact that I can’t explain completely and of which I partially talk here during the discussion on dc-servos : Carrying out some tests on the input stage (fig 1) and the dc-servo of a power amp (in spite of the fact that I used a small value resistor
for R after the integrator), I get much better results with the (fig 2) and a quite softer sound. I thought that the problem could come from mixing 2 references, input ground and dc servo ground, then I tested the (fig 3) imagining that it could be a problem of ground currents and ground loops (I talk here about a power-amp), and in this case the sound was much more softer and more “live”. Do you think as I it suspect but I am not completely sure that it is a ground problem , have you another or better explanation?
Fig1-fig2-fig3
Sorry the link is wrong! Here it is.
An externally hosted image should be here but it was not working when we last tested it.
Hi Upopa,
So you agree that a floating ground improves the naturalness, and the fidelity, as long as the ground reference for the integrator and the virtual ground op amp is the same and clean.
As you seem to know well the Borbely stuff, do you know the advantage of using his fet/fet cascode configuration even if the lower fet Vds is only about 1V.
So you agree that a floating ground improves the naturalness, and the fidelity, as long as the ground reference for the integrator and the virtual ground op amp is the same and clean.
As you seem to know well the Borbely stuff, do you know the advantage of using his fet/fet cascode configuration even if the lower fet Vds is only about 1V.
Hello, Upopa,
For sure that wizzards and gurus don't exist.... I built lots of amps and preamps, and also tested/upgraded some more for friends... Nothing to do with Black Magic, eh ! I don't think that my current scope and meter were on a vaudou influence too... But I can be wrong with these ones...
All I know is that John Curl NOT only is an engineer. He also collaborated to theory works... I have the works here at home. And I sometimes wonder what would audio world would be just juggling with uA709 & 741, and bad capacitors, Upopa. Seriously...
Was not so big a crowd of engineers advocating against poor opamp performances. Not defending special topologies... Still to see from other names you have in head associated with that.... NOW they are sharing data, but I didn't hear any in the business 25 years ago... Now, we have rather large access to patents and so on... Wasn't the case some years ago.
And it is the reward of sharing : Nelson Pass himself had to drop off his VERY interesting site. Guess why.
And because you seems to put me a lesson in testing here, I favourite Class A running, and experimented with almost all active parts available through years, from bipolars to IGBT with Mosfets in the middle and some others... Vfets as an exception ! And also experiences like Alexanders one ...
But sorry, Upopa, you can't read in my head nor put more in my words much that it is ! I am just advocating John Curl's work... What happened to him with time is out of my current thinking. And I already know that some guys just think him as a pretentious guy, guru like and very hard to deal with.... Perhaps. But I won't certainly judge nobody on a single word... On repeated facts, perhaps ...(Upopa, you don't know it, by I am formerly a lawyer... Hence my "special analysis" of things that happen... And please have my apologies for these rude words, John !!!). But that doesn't change anything in the full story...
Now, what am I doing here, in this forum ? For sure NOT having crusades against guys that just think different,, like you... But frankly, I am trying to see why some Curl's backworks are made for... And just to point any word against others "gurus" you have perhaps associated with my own mind. So, the debate is closed with you on my side, Upopa. Still to hear of you on a more cinstructive basis, and NOT a mind justice...
My current project ? Oh, yeah ! "The Flame Thrower" ! Guess why !
Jbaudiophile
For sure that wizzards and gurus don't exist.... I built lots of amps and preamps, and also tested/upgraded some more for friends... Nothing to do with Black Magic, eh ! I don't think that my current scope and meter were on a vaudou influence too... But I can be wrong with these ones...
All I know is that John Curl NOT only is an engineer. He also collaborated to theory works... I have the works here at home. And I sometimes wonder what would audio world would be just juggling with uA709 & 741, and bad capacitors, Upopa. Seriously...
Was not so big a crowd of engineers advocating against poor opamp performances. Not defending special topologies... Still to see from other names you have in head associated with that.... NOW they are sharing data, but I didn't hear any in the business 25 years ago... Now, we have rather large access to patents and so on... Wasn't the case some years ago.
And it is the reward of sharing : Nelson Pass himself had to drop off his VERY interesting site. Guess why.
And because you seems to put me a lesson in testing here, I favourite Class A running, and experimented with almost all active parts available through years, from bipolars to IGBT with Mosfets in the middle and some others... Vfets as an exception ! And also experiences like Alexanders one ...
But sorry, Upopa, you can't read in my head nor put more in my words much that it is ! I am just advocating John Curl's work... What happened to him with time is out of my current thinking. And I already know that some guys just think him as a pretentious guy, guru like and very hard to deal with.... Perhaps. But I won't certainly judge nobody on a single word... On repeated facts, perhaps ...(Upopa, you don't know it, by I am formerly a lawyer... Hence my "special analysis" of things that happen... And please have my apologies for these rude words, John !!!). But that doesn't change anything in the full story...
Now, what am I doing here, in this forum ? For sure NOT having crusades against guys that just think different,, like you... But frankly, I am trying to see why some Curl's backworks are made for... And just to point any word against others "gurus" you have perhaps associated with my own mind. So, the debate is closed with you on my side, Upopa. Still to hear of you on a more cinstructive basis, and NOT a mind justice...
My current project ? Oh, yeah ! "The Flame Thrower" ! Guess why !
Jbaudiophile
A self referenced cascode is a good compromise design component. This is where the cascode is made by a fet with its gate tied back to the input fet's source.
Usually, we use 2 different fet types for this. A short gate, low Vg part (hi transconductance) for the input part and a long gate (low transconductance) part for the cascode part. Identical parts may be an interesting solution, but I tend to avoid it. The input capacitance will be very high, BUT the change in drain-gate voltage will be very small, so the non-linear capacitance will be essentially invisible. This is not true with a non cascode fet that has to swing significant voltage at its output. Here, operating voltage across the fet is very important. Of course the best is to operate the fet with a reasonable voltage, and then cascode it as well. Best of both approaches.
Usually, we use 2 different fet types for this. A short gate, low Vg part (hi transconductance) for the input part and a long gate (low transconductance) part for the cascode part. Identical parts may be an interesting solution, but I tend to avoid it. The input capacitance will be very high, BUT the change in drain-gate voltage will be very small, so the non-linear capacitance will be essentially invisible. This is not true with a non cascode fet that has to swing significant voltage at its output. Here, operating voltage across the fet is very important. Of course the best is to operate the fet with a reasonable voltage, and then cascode it as well. Best of both approaches.
Oh yes his cascode configuration is very attractive! I agree, but as I said, in this case the “active“ device only sees something like 1V or so, if it where a BJT no problem, but in the saturation region? A lot of parameters seem to be less interesting, Cgd Cgs Yfs, but I guess that as the active device doesn't sees the signal, it doesn't matter...
Inpact of DC servo to sound is more complicated and isn't based only on ground " problem " . You must imagine, that this output's resistor is connected in parallel with feedback resistor and output of integrator substitute true ground for AC signal. If integrator will be forced, thanks to uncorrect condition ( for example too high output current ), to work at his own parameter border, will injected into feedback loop own distortion ( virtual floating ground will be not perfect ). So by another words spoken : we must be very careful with value of components and type of this circuit at this position. By my opinion most of objections against this solution are cause by this reason.
Hi, JustCallMeDad
If your schematics refers to the Blowtorch, please consider that it works BOTH BALANCED and UNBALANCED... So, it is just much more complex that the usual interator way.. . That is ways off the real problem...
And here we are not in the books guruing (as I ever heard from somes), but in real engineering... An answer to this differential problem, anyone ?
Still to hear from your own studies, Upopa... Sorry for, but pressure will increase A LOT from my side... Want to see your proposal first... Mine second only... And guess who is the lessons on that I took from ? Any Guriu ? NO !
Go for the challenge, Upopa ???? Good .... But if your question was off thsi topic, please consider I haven't said a single word...
Jbaudiophile
If your schematics refers to the Blowtorch, please consider that it works BOTH BALANCED and UNBALANCED... So, it is just much more complex that the usual interator way.. . That is ways off the real problem...
And here we are not in the books guruing (as I ever heard from somes), but in real engineering... An answer to this differential problem, anyone ?
Still to hear from your own studies, Upopa... Sorry for, but pressure will increase A LOT from my side... Want to see your proposal first... Mine second only... And guess who is the lessons on that I took from ? Any Guriu ? NO !
Go for the challenge, Upopa ???? Good .... But if your question was off thsi topic, please consider I haven't said a single word...
Jbaudiophile
Looking at those servos, I see a BIG problem. You should not create an artificial ground and then run audio through it. It will map itself into the audio signal itself, i.e. any impurities generated by the servo IC will be part of the audio path. No wonder you can hear these servos so easily. Always separate the servo from the audio thru-path. This can be done by adding a low value resistor to ground, and the servo resistor high valued. Then a small imperfection in the servo will be attenuated many times before it is added to the audio path. 
- Status
- Not open for further replies.