DIY 2A3 by mred

MrEd

Senior Nobody
This is an amplifier project that's been in the works for some time.. 4 years maybe
It has been through several iterations on the board and I am ready to make two chassis and build them.
20181209_084956.jpg 20181211_191752.jpg 20181215_110305.jpg
I started the lay out of components and have my sheet metal ready to start forming the chassis.
I hope to have these done in the next few months.... life permting... I am getting exited about it, it has been a long time coming.
 
Are you using the Lundahl for phase splitting? Which model?

Can you post a schematic of your intended circuit?

I am curious because I have mulled over building a PP 2A3 for at least 10 years but haven’t started one yet.
 

MrEd

Senior Nobody
Are you using the Lundahl for phase splitting? Which model?

Can you post a schematic of your intended circuit?

I am curious because I have mulled over building a PP 2A3 for at least 10 years but haven’t started one yet.

Interstage is lindahl 1692A.
Power supply is clclc , first cap is small and adjusts the voltage.
The driver stage gets another rc filter to set voltage.
No caps in the signal only in ps
The driver is fixed bias, even though the photo shows caps there I like it best without
The 2a3 are auto bias by a K&K ccs.
 

marantzfan

Administrator
Staff member
I remember this project! Glad to see it’s gaining traction again as I’m really interested to hear what you think of it.
 
Interstage is lindahl 1692A.
Power supply is clclc , first cap is small and adjusts the voltage.
The driver stage gets another rc filter to set voltage.
No caps in the signal only in ps
The driver is fixed bias, even though the photo shows caps there I like it best without
The 2a3 are auto bias by a K&K ccs.

Looks like a cool build: have been sorely tempted to try a pp 2A3 amp...

Looks like your output stage biased by a constant current sink bypassed with capacitor. FWIW, that bypass capacitor is in the current loop (also know to some as the signal path).

How are you biasing the input/driver?
 

paulbottlehead

Active Member
The 2a3 are auto bias by a K&K ccs.
What's the operating current and the bias voltage? I have some experience with that particular heatsink and with PP 2A3 amps, and if we presume 80mA with 62V of bias for the push-pull operating point in the book, that's 5W of dissipation. If that is the 2" version of that heatsink (which it looks like), then you'll see a temp rise of 37C on the heatsink and another 10C from the device itself. This will be just barely OK operating in an amp that's reasonably well ventilated (from my experience at least), but you aren't really guaranteed that the grid bias voltage will be that low.

If at all possible, I would put a resistor in series between the CCS and ground to dissipate some of that heat. A 250 ohm 5W resistor (assuming 80mA) will eat up about half the heat and it would buy you a ton of thermal margin on the CCS.
 

MrEd

Senior Nobody
I have changed it to a resistor and cap common bias to both tubes. Sounds better to me than the ccs. 295v on the plate @45 ma. 50 v bias
I have been experimenting with the one on a board now for some time.
It has changed quite a bit.
I have try multiple drivers, operating points, bias methods, caps etc.
 

paulbottlehead

Active Member
With a single filament supply, fixed bias will let you balance the DC current through your output transformer. You can also clip into class B with some grace.
 
Top