Push Pull 2A3

another option would be to follow the schematic I've got that uses fixed bias - i had thought it would be more complicated but since I've got all the details already worked out for me maybe it makes more sense?
 

paulbottlehead

Active Member
Fixed bias will give you a little more power. Cathode bias will be less fussy, you can just turn the amp on and use it.

Fixed bias will need about a 320V power supply with a -80V bias supply. Cathode bias will need about a 380-400V power supply.
 
Last edited:
I have attempted to draw up a diagram of the PSU I've been given that I've been told would work by someone that has built almost the same amplifier (they used an autoformer rather than an interstage/splitter). Its still incomplete as I am not quite sure how the fixed bias is implemented and haven't found any diagrams I've been able to get my head around (this one is a very similar amplifier but the PSU I've found very hard to understand - LINK )

the values are a bit lower than suggested by Paul above but maybe its still acceptable? the B++ works out about 300v ish (maybe a bit higher) and has a 50v bias supply

img014.jpg

T1 = dynakit P-782 ( 410-0-410 with a 50v bias tap - I have seen the same transformer listed with 55v and 60v)
T2 = Hammond 266M5

L1 = 10H choke
L2 = 1.5H choke

V3 = 5U4GB - this can change but I do have a few pairs already
V4 = 0D3 VR

R4 = hum pot?

? = i think i need something here so that C2 and C3 are seperated? another choke or a resistor?

* my understanding is that the bias tap would be connected to the center tap on the interstage transformer? via a variable resistor and test points or a meter dial as an alternative?
 

paulbottlehead

Active Member
While C3 isn't necessary, you can add a resistor where the ? is to reduce dissipation through the CCS a little bit.

If you look at the Dynaco ST-70 schematic, you'll see a diode (UF4007 is fine) connected to the red/black power transformer wire, then some caps and resistors. If you make the 47uF caps 220uF/160V caps instead, make the 1000 ohm resistor 220 ohms, and make the 18K resistor 22K, that supply will work for what you're doing. You will notice that the negative bias voltage goes to the junction of the grid leak resistors in the ST-70. In your circuit, this will be connected to the center tap of the output side of the interstage transformer.

If you want to put a meter on your amp to measure current, it will go between the wiper of your hum pot and ground.
 
While C3 isn't necessary, you can add a resistor where the ? is to reduce dissipation through the CCS a little bit.

If you look at the Dynaco ST-70 schematic, you'll see a diode (UF4007 is fine) connected to the red/black power transformer wire, then some caps and resistors. If you make the 47uF caps 220uF/160V caps instead, make the 1000 ohm resistor 220 ohms, and make the 18K resistor 22K, that supply will work for what you're doing. You will notice that the negative bias voltage goes to the junction of the grid leak resistors in the ST-70. In your circuit, this will be connected to the center tap of the output side of the interstage transformer.

If you want to put a meter on your amp to measure current, it will go between the wiper of your hum pot and ground.
Thanks guys, I'm off to look at the Dynaco ST-70 schematic!
 
I have found a few schematics for the dynaco and couldn't find a 1000 or 18K resistor, all i could find were several 10k resistors. THIS is the clearest one I have found and looks to be from a reputable source.

I have drawn up a revised PSU and have tried to include the bias circuit, i removed one of the bias points as the schematic was for stereo and im aiming to make monoblocks - Im not sure if Im correct to do this though:


img016.jpg

T1 = dynakit P-782 ( 410-0-410 with a 50v bias tap - I have seen the same transformer listed with 55v and 60v)
T2 = Hammond 266M5

L1 = 10H choke
L2 = 1.5H choke

V3 = 5U4GB - this can change but I do have a few pairs already
V4 = 0D3 VR

C1 = 50uf oil cap
C2 = 50uf oil cap
C2 = 50uf oil cap
C4 = 220uf 150v electrolytic
C5 = 220uf 150v electrolytic

R4 = ill work this out in PSUD but i think it will be something like 5.8k to get the B+ to a little over 155v, is that about what i want to be aiming for?
R6 = this was 10k in the dynaco st70 but Paul suggested changing the 1000R to 220R, is this the resistor you meant?
R7 = a 10K variable resistor (hum pot?) in the dynaco schematic
R8 = also 10K in the dynaco schematic

I can now identify the fixed bias circuit that I have been given for the 2a3 amp I mentioned previously and there are some different values given, what do you think of the following?:

C4 = 50uf
C5 = 50uf

R6 = 1kR
R7 = 25K variable
R8 = 18KR

I really appreciate the help
 

paulbottlehead

Active Member
Yeah, you'll want R6 to be a small value to keep the available voltage high, so 220 ohms. R8 can be increased significantly since you're not going to need much of the bias range. 22K-47K would be reasonable.

C4 and C5 aren't super critical, but I find that it's just easier to build with slightly larger parts that have fatter/longer leads, and to some degree with R6 being 220 ohms, bigger caps will reduce the increase in ripple that going to a smaller R6 will bring about.
 

Salectric

Senior Member
I would start out without R8. Just have the 25K pot connected between C5 and ground. This will give you the full range of the pot to dial in the proper bias. After you have it up and working, you can decide whether to add a resistor (R8). The advantage of the resistor is that it lets the pot cover a smaller range of voltages so you can fine tune the bias more easily. But if you use too big a value for R8 you may not have enough range to get the right voltage.
 
Yeah, you'll want R6 to be a small value to keep the available voltage high, so 220 ohms. R8 can be increased significantly since you're not going to need much of the bias range. 22K-47K would be reasonable.

C4 and C5 aren't super critical, but I find that it's just easier to build with slightly larger parts that have fatter/longer leads, and to some degree with R6 being 220 ohms, bigger caps will reduce the increase in ripple that going to a smaller R6 will bring about.
Thanks Paul, I really appreciate you taking the time to explain this to me. What rating for the resistors should I be looking for? I have some 100uf 500v jensen electrolytics that might do?
 
Guys can I ask what you think about the Psu in general? Does it seem like a good idea to proceed with it? I don't have the chokes and power transformers yet so It could be changed if its really less than ideal
 

paulbottlehead

Active Member
This will give you the full range of the pot to dial in the proper bias.
He's going to need the bias pot almost all the way fully negative, hence my recommendation to have R8 very high.

Resistor wattage isn't very critical here, there are a few mA flowing through them, so they won't get warm. 1/4 is just fine.

-PB
 

Salectric

Senior Member
Guys can I ask what you think about the Psu in general? Does it seem like a good idea to proceed with it? I don't have the chokes and power transformers yet so It could be changed if its really less than ideal
Since you asked, I am skeptical of the choke input. If I were building it, I would use cap input, which means a different power transformer, but the rest of the supply design looks fine. In the several projects where I tried choke input, I thought the sound was too soft and lacked impact. The one exception is my 46 SE amps where my PT allows me to use either choke input or cap input. With the 46 I prefer choke input.
 
Thanks guys for your thoughts!

Well I'm glad I've made it this far. After the help you guys have given me I'm starting to grasp some of this it's been quite fun.

I do have a power transformer that could do a capacitor input supply (I think) so maybe I should try and design a psu for that, I think I should be able to do that by myself now, and maybe it would be a good exercise. Would you do it fixed bias or with the cathode cap/resistor?
 
Last edited:
I'm a little late to this discussion, but I went down a similar path, starting a couple of years ago. One inspiration was a Dr Gizmo design, which looks similar to your starting point:
Interstage Coupled PP 6B4G Amplifier

A friend advised against the 417a because it can be prone to oscillation, so I decided to go old-school instead.
Page 44 of the pdf here, http://www.tubebooks.org/Books/Atwood/Robin & Lipman 1947 Practical Amplifier Diagrams.pdf

Page 149 here. 12 watts, class A.
http://www.tubebooks.org/tubedata/RC12.pdf

Both of those had cathode bias, but my available transformer was a little low on B+ so I had to decide between cap loading and cathode bias, or choke loading and fixed bias (so you don't lose the voltage across the bias resistor). I chose the latter (one version of the schematic attached). Notice that the original 2a3 spec sheets specify 3k for fixed bias but 5k for self bias; I ended up with 5k and fixed bias, which I think costs me some power, if I remember my load lines correctly.

Ultimately, I decided that the amp didn't have enough LF or gain for my needs, but better IT's might have fixed that. But I couldn't afford the iron to do that (running unbalanced current through the IT makes it difficult to wind one with 20-20khz response) so I changed my topology. (27-->6a6 LTP phase splitter-->IT-->PP2a3-->5k PP OPT). It delivers about 14w before clipping, which is adequate for Altec speakers.

I will say that even with AC filaments, and only small capacitors in the PS, hum is negligible. The required chokes make it heavy to move, though! No balance pot, just four bias pots. I'm cheating and using a diode with electrolytic capacitors to provide the bias voltage; I might change it to a tube bias rectifier at some point.

mhardy6647 has heard both the original and new versions; Redboy has heard the new version. I've been listening to it in a breadboarded format since the spring, so no photos, but I'm hoping to make a proper chassis this winter.
 

Attachments

Last edited:
I'm a little late to this discussion, but I went down a similar path, starting a couple of years ago. One inspiration was a Dr Gizmo design, which looks similar to your starting point:
Interstage Coupled PP 6B4G Amplifier

A friend advised against the 417a because it can be prone to oscillation, so I decided to go old-school instead.
Page 44 of the pdf here, http://www.tubebooks.org/Books/Atwood/Robin & Lipman 1947 Practical Amplifier Diagrams.pdf

Page 149 here. 12 watts, class A.
http://www.tubebooks.org/tubedata/RC12.pdf

Both of those had cathode bias, but my available transformer was a little low on B+ so I had to decide between cap loading and cathode bias, or choke loading and fixed bias (so you don't lose the voltage across the bias resistor). I chose the latter (one version of the schematic attached). Notice that the original 2a3 spec sheets specify 3k for fixed bias but 5k for self bias; I ended up with 5k and fixed bias, which I think costs me some power, if I remember my load lines correctly.

Ultimately, I decided that the amp didn't have enough LF or gain for my needs, but better IT's might have fixed that. But I couldn't afford the iron to do that (running unbalanced current through the IT makes it difficult to wind one with 20-20khz response) so I changed my topology. (27-->6a6 LTP phase splitter-->IT-->PP2a3-->5k PP OPT). It delivers about 14w before clipping, which is adequate for Altec speakers.

I will say that even with AC filaments, and only small capacitors in the PS, hum is negligible. The required chokes make it heavy to move, though! No balance pot, just four bias pots. I'm cheating and using a diode with electrolytic capacitors to provide the bias voltage; I might change it to a tube bias rectifier at some point.

mhardy6647 has heard both the original and new versions; Redboy has heard the new version. I've been listening to it in a breadboarded format since the spring, so no photos, but I'm hoping to make a proper chassis this winter.
Thanks for your thoughts and schematics, very interesting.

I have come across several people on hifihaven now who have built push pull 2a3 amplifiers.

I'm not totally sure what I'll end up doing and I can do anything as I don't yet have the Psu parts ordered, but Salcritic has given me some great advice already so I'm keen come up with a capacitor input with cathode bias psu as he thinks it should be better.
 

paulbottlehead

Active Member
Yeah, the #56 driving a 2A3 doesn't exactly have lots of gain, but the 417A would ;)

I like that the interstage in your design has a split secondary, as you can balance the idle current of each tube. That of course means a heater winding for each tube, but that's not too rough all things considered.
 
So I have tried to design a capacitor input PSU, this time with cathode/cap bias - I'm not sure that I was going in the right direction as B+ was coming out at 380v. After comparing the PSU with another PSU I saw that using a small value first cap brought the B+value down considerably, seems like a good way to control the B+ but is that a good idea?

here is what I managed to get from PSUD

capinputPSU.jpg

and here is my drawing and parts

img017.jpg

T1 = Hammond 374BX (375 - 0 -375) - this was a recommended PT for push pull monoblocks
T2 = Hammond 266M5

L1 = 10H choke
L2 = 1.5H choke

C1 = 1.5 - 1.8uf
C2 = 50uf
C3 = 50uf
C4 = not sure this is necessary?
C5 = 100uf

R4 = 550R ish?

if this is good to go(?) that would be great as luck would have it the hammond transformers are in stock in the UK and thats a great incentive to use this type of PSU over the choke input which would take quite a while to have the parts made etc

cheers!
 
Top