12B4A SET - any critiques on this project?

[zobsky]

Last week, I actually built something cheap that works and sounds good from day 1 .

Any cheap mods on improving the performance ? All I have on my to-do list are CCS the driver and film-bypass the cathode and power supply electrolytics. I haven't bothered decoupling the power and driver stage (though each channel has a seperate CLC filter section).

Description at http://zobsky.blogspot.com/2008/06/12b4a-amplifier.html




Thanks

[oldeurope]

Originally posted by zobsky
...

Description at http://zobsky.blogspot.com/2008/06/12b4a-amplifier.html

...

Thanks

Hello,
I would recommend to replace the LED by a resistor and bypass cap like you did it in the second stage.
The LED is a non linear element and creates some kind of "LED sound".

Kind regards,
Darius

[Shoog]

"The LED is a non linear element and creates some kind of "LED sound"."

My experience is quite the opposite. I have found LED's tend to introduce a very clean detailed and punchy sound. Resistors are much softer and more blurred. Capacitor bypassed will sound like the capacitor which most people say is no good thing (though I find the problem overstated).
SY has done extensive measurements on LED biasing and finds it a very low distortion option.

Shoog

[oldeurope]

Originally posted by Shoog
"The LED is a non linear element and creates some kind of "LED sound". ...

My experience is quite the opposite. I have found LED's tend to introduce a very clean detailed and punchy sound. ...

Shoog

Thanks for describing LED sound.

Originally posted by Shoog
Resistors are much softer and more blurred.

Shoog

Resistors are making current feedback there.
This current feedback can be eliminated by a high
"well known factor".

Originally posted by Shoog
"The LED is a non linear element and creates some kind of "LED sound". ...
Capacitor bypassed will sound like the capacitor which most people say is no good thing (though I find the problem overstated). ...
Hello Shoog,

Shoog

The signal voltage drop at the ESR of the cap is much lower
than that dropping at the differential resistance of the LED.
This drop is what causes the influence in sound.

Originally posted by Shoog
"The LED is a non linear element and creates some kind of "LED sound". ...
SY has done extensive measurements on LED biasing and finds it a very low distortion option.

Shoog

I remember SY recommended to add some additional current
to reduce the differential resistance and it's influence in sound.

What about bypassing the LED with a cap?

Kind regards,
Darius

[EC8010]

LEDs are great for biasing valves, but this might not be the ideal option. Like any diode, their slope resistance is nonlinear so they have the potential for adding distortion. The important thing is to maximise the ratio of the slope resistance to the anode load resistance so as to minimise the feedback voltage developed across this nonlinear resistance. 22k is a bit low. If the stage was the input stage of an RIAA stage, you could get away with it because the signal levels are lower and the distortion proportionately lower. On the other hand, although an LED might increase distortion in this application, it won't cause blocking in the way that a resistor capacitor combination might. I'd say, "Try both." After all, it's an easy thing to change, and you can decide which you like best.

[alexg]

Is that really a 0.047uf coupling cap, or should it be 0.47uf?

I built a 6n1p driven 12B4 parafeed amp and it sounds excellent. Very neutral.

[Ciu]

Hello Alexq !

Would you share your schematics ?
I have some 6n1p and 12b4 on shelves...

Thanks !
R.C.

[Shoog]

What about bypassing the LED with a cap?

A bypass cap of about 47uf seems to reduce distortion a little. In order to fully bypass an LED would require a huge cap which is not desirable or practical.

I remember SY recommended to add some additional current
to reduce the differential resistance and it's influence in sound.

I wouldn't run an LED at less than about 6mA. In my preamp I used a little circuit off the heater winding to supply that supplementary current. It tightened up the sound. A simpler result can be achieved with a resistor from the +B to the top of the LED.

Shoog

[kenpeter]

I think some of the troubles we heard at highest gain could
have been the transformer saturating, rather than anything
clipping the power rail as we had originally speculated.

Might try parafeed experiment with a larger external choke?
Just to see if DC saturation of the SE transformer might have
been the limiting factor.

It sure sounded great up to a watt or two! Certainly we were
not using the most efficient loudspeakers available to us.
Need to bring both Karlsons to one venue at the same time
for another audition.

[alexg]

Originally posted by Ciu
Hello Alexq !

Would you share your schematics ?
I have some 6n1p and 12b4 on shelves...

Thanks !
R.C.

I got this schematic from diyasylum.com and I really don't know who designed it.

But if you have a very efficient (high sensitivity speakers), this will work (am using a 93db pair and it works).