Hi upupa !
Hmm, the 22pF is quite important for stability and avoiding
overshooting/ringing. I thought, i already do compensation different ?
Yes the jfet's are one thing more i wanted to try with this design,
i still have some sk170bl/sj74bl left. But not much place left on board...
And matching these is a pain !
Who said, i'm not satisfied with sound ? I think this one became very
good. Do you have any idea for making compensation different ?
Mike
Hmm, the 22pF is quite important for stability and avoiding
overshooting/ringing. I thought, i already do compensation different ?
Yes the jfet's are one thing more i wanted to try with this design,
i still have some sk170bl/sj74bl left. But not much place left on board...
And matching these is a pain !
Who said, i'm not satisfied with sound ? I think this one became very
good. Do you have any idea for making compensation different ?
Mike
carlos! the sound of jvc is soft with small volume .My amp is to
h! now i like my amp .Sense is depenp on my emotion state very much ..yes ,it is not perfec absolutely
carlos! you wrote so words so I often can't know you .
after I build my first amp ,i had just known that there are often some secrete on all schematics .I think your amp can work means you are quite lucky
My first amp could not work but I met a fairy .I don't know whether he be here Thanks my fairy very much
h! now i like my amp .Sense is depenp on my emotion state very much ..yes ,it is not perfec absolutely carlos! you wrote so words so I often can't know you .
after I build my first amp ,i had just known that there are often some secrete on all schematics .I think your amp can work means you are quite lucky
My first amp could not work but I met a fairy .I don't know whether he be here
Thanks my fairy very much
Hi upupa !
My pain with matching jfets is that they are so different, you need
at least measuring 10 for getting 1 pair. At their price the pain moves
into the portmonnaie.
My problem with feedback from VAS is that the outputstage get's less
feedback. But the outputstage is the distorting thing. Distortion in
the frontend+vas is far below than the distortions from output.
For example, checking with openloop at outputswing ~12volt, harmonics
at vasoutput are at ~7mv, after outputstage i have ~50mv.
But i will try, changes in feedbackcompensation have the most audible
effect. In fact i would say, feedbackcompensation is the thing that makes
an amp sound good or bad, not topology. This might be the mainreason
why amps without global feedback sound so good. They don't need
to be finetuned for correct feedback.
Mike
My pain with matching jfets is that they are so different, you need
at least measuring 10 for getting 1 pair. At their price the pain moves
into the portmonnaie.
My problem with feedback from VAS is that the outputstage get's less
feedback. But the outputstage is the distorting thing. Distortion in
the frontend+vas is far below than the distortions from output.
For example, checking with openloop at outputswing ~12volt, harmonics
at vasoutput are at ~7mv, after outputstage i have ~50mv.
But i will try, changes in feedbackcompensation have the most audible
effect. In fact i would say, feedbackcompensation is the thing that makes
an amp sound good or bad, not topology. This might be the mainreason
why amps without global feedback sound so good. They don't need
to be finetuned for correct feedback.
Mike
Hi Mike, matching is not so critical, if you have devices from one group ( ... bl for example ). About feedback : try to think differently : if you have precise stabil VAS, you can rise OLG, which " get down " distortion by closed loop. And remember, that " heart " of amp is input stage, which make the main differences in sound, not output stage .
.
Hi upupa !
My devices are all "bl", but still difference between devices is very
big. Not carefully matching give big DC and unbalanced diffamp.
I also have some sk389, but without complementary.
It's just that my bc546/556 are already very close, easy to match them.
And i have a bag full of them, hfe only differs +/-10%, vbe only 5%.
The thing with rising OLG is not a good idea, it seems that for low TIM
flat OLBW is very important. But i don't know how TIM behaves with
local feedback from vas. Need to check...
Mike
My devices are all "bl", but still difference between devices is very
big. Not carefully matching give big DC and unbalanced diffamp.
I also have some sk389, but without complementary.
It's just that my bc546/556 are already very close, easy to match them.
And i have a bag full of them, hfe only differs +/-10%, vbe only 5%.
The thing with rising OLG is not a good idea, it seems that for low TIM
flat OLBW is very important. But i don't know how TIM behaves with
local feedback from vas. Need to check...
Mike
Hi sajti !
I have built once a version without cascoded vas, but i must say:
cascoded vas rocks ! With cascoded vas you have 2 big advantages:
1, low openloopdistortion, means low thd with moderate feedback
2, a very detailed sounding amp, reproducing trebles without any effort
As i posted in another thread, i have no idea how to calculate RC-nets,
i get values by trying and checking with acsweep in sims...
My experience with RC-nets: They simply sound better !
Mike
I have built once a version without cascoded vas, but i must say:
cascoded vas rocks ! With cascoded vas you have 2 big advantages:
1, low openloopdistortion, means low thd with moderate feedback
2, a very detailed sounding amp, reproducing trebles without any effort
As i posted in another thread, i have no idea how to calculate RC-nets,
i get values by trying and checking with acsweep in sims...
My experience with RC-nets: They simply sound better !
Mike
MikeB said:
Hi mike,
I have some experiences with RC nets. The RC makes stepdown in the frequency range. With the C You can set the frequency where the stepdown starting. And the R set how big the stepdown will be. It works well if You set the R to reduce the gain from the amount of the open loop gain, to the amount of closed loop gain (this reduces the gain only in necessary amount). The RC network set the dominant pole, but makes less phase shift than the common VAS capacitor.
When I tried to use cascode stages, I found, that it makes the highs sharp
sajti
sajti said:
Have you compensated for the "sligthly" increased vas-gain ?
Without resistors to gnd inside the cascoded vas you get incredible
high gains ! My amp boosted up to OLG of 1:400.000, with max below
100hz. This generates lot of tim. I had another circuit where cascoding
boosted the gain to 1:8.000.000 ! It does not get unstable because
the gainbandwidthproduct keeps nearly the same.
If i remove the 100k in the vas, trebles get exaggerated.
About RC, i think i need to study them a lot more !
To upupa !
Can't i send you mails ?
Mike
MikeB said:
It also discuss the sharp high frequency sound. And the owner think that this sound caused by the low input impedance of the upper device (common base connected). This low impedance results overload for the lower transistor (common emitter connected). The overload makes distortion.
They made some tests with series base resistors for the upper device. This reduce the gain, and increase the input impedance of this stage. They found sweeter highs, as they increase the value of the resistor. But as the gain reduced, they finally remove all the cascode parts, and keep the common emitter stage only.
So I never tried this. It's more simple to forget all the cascode, and use VAS transistor with low BC capacitance, and local feedback.
sajti
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.