I wonder if 4N36 would work as well in this design? I have a few tubes of them from many years ago that I have been saving for something special . This certainly qualifies as special! Thanks for sharing it Papa!
3300uF/16V sufficient
use 4N35 , proven and working
I didn't tried 4N36 , and I'm lazy to check datasheet for most important data - current transfer ratio and conducting treshold
however - if you get Iq in range of 1A5 to 2A , steady - no dramatic change between cold and hot , feel free to use it for good
use 4N35 , proven and working
I didn't tried 4N36 , and I'm lazy to check datasheet for most important data - current transfer ratio and conducting treshold
however - if you get Iq in range of 1A5 to 2A , steady - no dramatic change between cold and hot , feel free to use it for good
Last edited:
let introduce some thinking in thread , besides energy put solely in clonning
question which almost melted my non-electronic oriented brain , several yrs ago (I believe Papa is still laughing ) - what's role of 3 diodes in circuit ?
question which almost melted my non-electronic oriented brain , several yrs ago (I believe Papa is still laughing ) - what's role of 3 diodes in circuit ?
Thanks ZM.... Of course you are right. 4N35's are inexpensive. Just trying to jusfify keeping them all these years . The 36's are darlington phototransistor outputs if I remember correctly.
I really scratched my head on that one. My best guess is that they are steering diodes for the output coming back to the drive circuit . I don't pretend to understand the function - just guessing why they are there. The drop across the top two would be cLose to the upper reference V. Or perhaps some sort of clamp.
The other quesfion in my mind is if the input transformer is wired as an auto former?
other question - obviously yes
just draw it different , previously looking at xformer datasheet .
first question ..... clarify what you mean , and in which condition ?
just draw it different , previously looking at xformer datasheet .
first question ..... clarify what you mean , and in which condition ?
Well the opto isolator does just that.... There is no DC path back to the drive due to the isolation. I don't see any feedback per se either. The LED side of the bias circuit floats with the output swing but without some DC path back the bias/drive has no idea where the output is. Wow ... I really don't have any idea ! Still scratching my head ....
Last edited:
keep scratching ...
I said - I almost melted me few cells .....
Mine are going too!
I am puzzled why the need for the LM385's ? The opto led forward drop should be about 1.2 volts. So I don't see why they are needed for clamping.
Unless when the amp is first turned on and the circuit is in startup mode the LM385's and three diodes keep things under control until enough current flows in the output Mosfets to turn on the opto LED to stabilize the bias. Are you laughing yet??? I can't wait to see the real answer but it is fun to try to figure it out.
Last edited:
Well I ask myself when would these diodes become forward biased? The upper two would when the output voltage exceeds 1.2 volts more positive than Q1's gate. The lower diode would conduct when the output voltage is more than 0.6 volts negative with respect to Q2's gate. I can't yet see what conditions would make this happen unless driving an inductive load or something to do with startup.Anyone else want to jump in and shed some light on the subject?
Every part justifies itself by making it work. The 385's allow proper
operation into low impedances. The diodes protect the output Gates and
give symmetric clipping.
This used the Fairchild output devices and pot location assumes that P
channel Vgs is slightly higher than N. You can use other devices. If you want
to fool around, you can eliminate the cap coupling.
The Edcor transformer is steel core. If you want to use Jensen, then
set it up as autoformer and just drive the one tap above ground, letting
the other 3 coils create more voltage, for about X 4 (12 dB). If you
use other transformer, think about adjusting or eliminating frequency
compensation RC load to ground.
If you have a low impedance source you can eliminate the buffer altogether.
😎
operation into low impedances. The diodes protect the output Gates and
give symmetric clipping.
This used the Fairchild output devices and pot location assumes that P
channel Vgs is slightly higher than N. You can use other devices. If you want
to fool around, you can eliminate the cap coupling.
The Edcor transformer is steel core. If you want to use Jensen, then
set it up as autoformer and just drive the one tap above ground, letting
the other 3 coils create more voltage, for about X 4 (12 dB). If you
use other transformer, think about adjusting or eliminating frequency
compensation RC load to ground.
If you have a low impedance source you can eliminate the buffer altogether.
😎
and do not forget the Edcor was the winner…..
Nelson once wrote that the Edcor sounds best for reasons unknown.
I can confirm this, I first used Lundahl, but the Edcor had more juice!
Cat trying to understand the diodes in M2…….
Nelson once wrote that the Edcor sounds best for reasons unknown.
I can confirm this, I first used Lundahl, but the Edcor had more juice!
Cat trying to understand the diodes in M2…….
Attachments
Last edited:
I think the steel in the xfrmr core saturation figures vs the nickel alloy
were probably responsible for that. I liked both, and offhand I recall that
the nickel is preferred at lowest levels.
😎
were probably responsible for that. I liked both, and offhand I recall that
the nickel is preferred at lowest levels.
😎
I refer to this….😀
"So in those regards the M2 differs too. The only thing special about the autoformer is that it won the listening tests for reasons unknown."
6moons audio reviews: FirstWatt M2
"So in those regards the M2 differs too. The only thing special about the autoformer is that it won the listening tests for reasons unknown."
6moons audio reviews: FirstWatt M2