That is a JE Labs 2A3 design with the SRPP 6SL7 drivers...
Now that I've progressed my 45 to 5842 choke-loaded + LED bias, and my 10/50/300B amp got some new soul with 5842 CCS-load + LED bias, I feel that my 2A3 with the expensive MagneQuest iron it got is being left behind sonically.
I think it's going to be boring to use my last pair of 5842 for this amp
My headphone amp is also using 5842, did I say I like 5842? 
Maybe I should graduate to another driver
I'm thinking of-
a) Pentode driver, I have a Western Electric 404A but it's just a pair!
b) DHT driver, I have 01A but I want to make a pre-amp from it. I have another 2A3 2A3 drives 2A3?
c) My friend bought about 80 pieces of C3m and I can get them really cheap. But I've read the C3m mostly for 300B applications...
Can the bright minds here suggest to this bored DIYer what to do with his 2A3?
Now that I've progressed my 45 to 5842 choke-loaded + LED bias, and my 10/50/300B amp got some new soul with 5842 CCS-load + LED bias, I feel that my 2A3 with the expensive MagneQuest iron it got is being left behind sonically.
I think it's going to be boring to use my last pair of 5842 for this amp
My headphone amp is also using 5842, did I say I like 5842? Maybe I should graduate to another driver
I'm thinking of-
a) Pentode driver, I have a Western Electric 404A but it's just a pair!
b) DHT driver, I have 01A but I want to make a pre-amp from it. I have another 2A3
2A3 drives 2A3?c) My friend bought about 80 pieces of C3m and I can get them really cheap. But I've read the C3m mostly for 300B applications...
Can the bright minds here suggest to this bored DIYer what to do with his 2A3?
Hello Arnoldc,
Your circuit is very close to the one I and my firends use with the 2A3. Differences are that the VR tube should return to the cathode not the ground connection. THe cap across the VR tube should be replaced with a zobel network after MArk Kelly's posting on AA on the subject.
We found the grid choke didn't add anything and size the grid leak resistor at 4-5 times the load resistor for the pentode. 3Cm is good in this application but we found C3g better - more life in the sound and better tone.
We also use the WE connection on the power tube. Other than these small tweaks it is spot on for a design we have proved works extremely well.
Good Luck
James
Your circuit is very close to the one I and my firends use with the 2A3. Differences are that the VR tube should return to the cathode not the ground connection. THe cap across the VR tube should be replaced with a zobel network after MArk Kelly's posting on AA on the subject.
We found the grid choke didn't add anything and size the grid leak resistor at 4-5 times the load resistor for the pentode. 3Cm is good in this application but we found C3g better - more life in the sound and better tone.
We also use the WE connection on the power tube. Other than these small tweaks it is spot on for a design we have proved works extremely well.
Good Luck
James
Triodes theoretically work the best when loaded with an infinite impedance load. Pentodes are supposed to work best into a constant resistance load.
In the real world these "ideal loads" do not exist. One of the best compromises for the triode case is the CCS load. They work best when the following stage has a high input impedance. In practice some triodes really like a CCS load (the 5842) and some work OK but not that great. In practice some pentodes work OK with a CCS load and some distort heavilly ( I have limited experience here). Triode strapped pentodes behave like triodes, and therefore should benefit from a CCS load. Every case is different and depends on the individual tube and the load conditions presented by the following stage.
If you present the tube with a perfect load for the DC condition (a CCS, or resistor) but the input impedance of the next stage is low, or non - constant (worse), the tube will still not be happy because the AC load conditions are not satisfied. Every circuit behaves differently in this regard and is highly dependent on the operating point.
In the real world these "ideal loads" do not exist. One of the best compromises for the triode case is the CCS load. They work best when the following stage has a high input impedance. In practice some triodes really like a CCS load (the 5842) and some work OK but not that great. In practice some pentodes work OK with a CCS load and some distort heavilly ( I have limited experience here). Triode strapped pentodes behave like triodes, and therefore should benefit from a CCS load. Every case is different and depends on the individual tube and the load conditions presented by the following stage.
If you present the tube with a perfect load for the DC condition (a CCS, or resistor) but the input impedance of the next stage is low, or non - constant (worse), the tube will still not be happy because the AC load conditions are not satisfied. Every circuit behaves differently in this regard and is highly dependent on the operating point.
Nothing is absolute in the tube world. Of course, you can load a pentode with a CCS
http://www.tubecad.com/2006/06/14/variable_ul aikido_cf_2.png
http://www.tubecad.com/2006/06/14/variable_ul aikido_cf_2.png
Hi James,
So, this thing will actually work I hope it sounds better than my 6SL7-2A3
I checked Thorsten's Legacy again and it seems the 12K was designed for some 400V B+ and with a little bit over 300V for the 2A3, I think the plate resistor should be calculated accordingly... or should I say I have to make a compromise because I only have 8k2 and 7k5 big Kiwame resistors
Can you tell me why you connected the VR to the cathode?
No improvement with grid choke? Hmmm, I have some 270K Riken I could put to use for the meantime. But I'll compare them anyway.
Zobel
resistor + cap... I don't know how to compute for those I'll search the AA for that post.
WE connection! Yey! I have two huge Obbligato 30uF from Brian Cherry which I haven't used since I bought them
Is this close to what you described?
ps.
If the VR returns via the cathode resistor, then I can't use LED biasing, can I? I plan to use LED soon to bias the C3m...
agent.5, good to know that, but with a low B+ for this 2A3 amp, CCS load may not just make it, don't you think?
So, this thing will actually work
I hope it sounds better than my 6SL7-2A3 I checked Thorsten's Legacy again and it seems the 12K was designed for some 400V B+ and with a little bit over 300V for the 2A3, I think the plate resistor should be calculated accordingly... or should I say I have to make a compromise because I only have 8k2 and 7k5 big Kiwame resistors
Can you tell me why you connected the VR to the cathode?
No improvement with grid choke? Hmmm, I have some 270K Riken I could put to use for the meantime. But I'll compare them anyway.
Zobel
resistor + cap... I don't know how to compute for those I'll search the AA for that post.WE connection! Yey! I have two huge Obbligato 30uF from Brian Cherry which I haven't used since I bought them
Is this close to what you described?
ps.
If the VR returns via the cathode resistor, then I can't use LED biasing, can I? I plan to use LED soon to bias the C3m...
agent.5, good to know that, but with a low B+ for this 2A3 amp, CCS load may not just make it, don't you think?
Attachments
I've done a 717 to px4 amp ala yamamoto
Some things I learned
- plate chokes don't work, better to use them on the screen.
Pentodes love resistive loads
-grid chokes don't work either. I transferred my S&B grid chokes to the input grid of the 717.
-WE connection is a matter of taste and takes some time to break-in
-I use separate filament trans both for driver and pwr tubes for low noise given my spkrs are 100db
-Overall, I'm quite satisfied with the sound a good mix of pentode aggresiveness and triode sweetness
-I've done a lot of iteration over the last four years on my amps. 6sl7-300b, 6s45-px25, 6sn7-2a3, 6j5-px4
Next would be dht drivers and dht output!
Some things I learned
- plate chokes don't work, better to use them on the screen.
Pentodes love resistive loads
-grid chokes don't work either. I transferred my S&B grid chokes to the input grid of the 717.
-WE connection is a matter of taste and takes some time to break-in
-I use separate filament trans both for driver and pwr tubes for low noise given my spkrs are 100db
-Overall, I'm quite satisfied with the sound a good mix of pentode aggresiveness and triode sweetness
-I've done a lot of iteration over the last four years on my amps. 6sl7-300b, 6s45-px25, 6sn7-2a3, 6j5-px4
Next would be dht drivers and dht output!
arnoldc said:I'm thinking of-
a) Pentode driver, I have a Western Electric 404A but it's just a pair!
b) DHT driver, I have 01A but I want to make a pre-amp from it. I have another 2A3
c) My friend bought about 80 pieces of C3m and I can get them really cheap. But I've read the C3m mostly for 300B applications...
Can the bright minds here suggest to this bored DIYer what to do with his 2A3?
How about driving the 2A3 with a Kimmel? This could either use a pentode on top or make it a hybrid with a power MOSFET on top. If you DC couple, you avoid all the problems that coupling with a capacitor can present. Add a few db's of gNFB and you'd really have something.
^^^^
Here you go: Kimmel Mu Stage.
It's like an SRPP or a regular Mu stage, but better and more flexible. It basically adds an external load resistor so that it may be used with completely different types of tubes since they don't have to run on the same currents. Thus, you can use a small signal, high gain triode for the voltage gain, and a high g(m) pentode as a follower for current sourcing capability. It also adapts quite well to solid state, making for an even beefier driver if you're going to push the 2A3 into Class A2. Even if you don't go that extreme, it will improve transient overdrive behaviour.
Here you go: Kimmel Mu Stage.
It's like an SRPP or a regular Mu stage, but better and more flexible. It basically adds an external load resistor so that it may be used with completely different types of tubes since they don't have to run on the same currents. Thus, you can use a small signal, high gain triode for the voltage gain, and a high g(m) pentode as a follower for current sourcing capability. It also adapts quite well to solid state, making for an even beefier driver if you're going to push the 2A3 into Class A2. Even if you don't go that extreme, it will improve transient overdrive behaviour.
Hello Arnold,
I replaced a cascaded 6SN7 circuit with this and this one was more transparent and open with more extended bass response and much better bass tone and articulation, My son liked the amp so much he now has it and won't let me try any further mods in it
To your questions:
a) With the 300V B+ and value of 7k5 load will work nicely. I use a 7K there.
b) The pentode references the screen voltage to it's cathode so if you decouple to ground you have the low frequency pole that this generates distorting the bass response of the pentode (hopefully nelow the audio range but having phase effects through the bottom end of the audio range). You also have any cathode resistor/bypass variations imposed on the signal. This applies at the top end as well - again hopefully above the audio range but not necessarily. Referencing the sg to the cathode avoids these effects. So really it is just putting the sg reference where it should be anyway...
c) Grid choke is definitly a try it and see option - it add anything for us and was slightly worse - this might just be a reflection of pentodes not liking a reactive load (in general) i.e. producing a simpler distortion spectrum when resistively loaded.
d) The zobel I use is 100nF and a 390R - it is to neutralise the VR tubes inductance and reduces the small amount of rf hash that the tube produces.
e) WE with the 30uF Obbligatos would be very nice with a 100-140uF cathode bypass
f) Led biasing could still be used but you need to allow for the VR tube static current as well - so you would need a small LED array to handle that as the VR tube reaches minimum noise at about 14mA for a 40mA rated tube (It gets seriously noisy below about 3-4mA)
g) I wouldn't even think of ccs loading a pentode in this application - it works well for high loads on the system and give maximum voltage swing but it doesn't like lower loads and/or varying current draw (i.e. reactive load) - like a DHT tube with it's Millar cap on the input...
Something I'm increasing doing to get the cathode resistor heat above the chassis is to use a mains light bulb as the cathode resistor. You need to experiment with values as the resistance changes a bit with applied voltage - in Europe a 60W bulb makes a nice cathode load for a 2A3 for instance.
Hope that helps
James
I replaced a cascaded 6SN7 circuit with this and this one was more transparent and open with more extended bass response and much better bass tone and articulation, My son liked the amp so much he now has it and won't let me try any further mods in it
To your questions:
a) With the 300V B+ and value of 7k5 load will work nicely. I use a 7K there.
b) The pentode references the screen voltage to it's cathode so if you decouple to ground you have the low frequency pole that this generates distorting the bass response of the pentode (hopefully nelow the audio range but having phase effects through the bottom end of the audio range). You also have any cathode resistor/bypass variations imposed on the signal. This applies at the top end as well - again hopefully above the audio range but not necessarily. Referencing the sg to the cathode avoids these effects. So really it is just putting the sg reference where it should be anyway...
c) Grid choke is definitly a try it and see option - it add anything for us and was slightly worse - this might just be a reflection of pentodes not liking a reactive load (in general) i.e. producing a simpler distortion spectrum when resistively loaded.
d) The zobel I use is 100nF and a 390R - it is to neutralise the VR tubes inductance and reduces the small amount of rf hash that the tube produces.
e) WE with the 30uF Obbligatos would be very nice with a 100-140uF cathode bypass
f) Led biasing could still be used but you need to allow for the VR tube static current as well - so you would need a small LED array to handle that as the VR tube reaches minimum noise at about 14mA for a 40mA rated tube (It gets seriously noisy below about 3-4mA)
g) I wouldn't even think of ccs loading a pentode in this application - it works well for high loads on the system and give maximum voltage swing but it doesn't like lower loads and/or varying current draw (i.e. reactive load) - like a DHT tube with it's Millar cap on the input...
Something I'm increasing doing to get the cathode resistor heat above the chassis is to use a mains light bulb as the cathode resistor. You need to experiment with values as the resistance changes a bit with applied voltage - in Europe a 60W bulb makes a nice cathode load for a 2A3 for instance.
Hope that helps
James
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.