Couldn't find quite what I'm looking for re the D3A arrangement. I will have about 300V B+ for it.
I put the triode strapped D3A in TubeCad and modeled four standard grounded cathode configurations of about 36dB gain, better than +100V/-100V swing, 10 to 19mA current, plates 6k8 to 15k, Rk 100 to 180r, bypass cap probably 470uF or so.
I'm interested in knowing if this is a reasonable range to drive a pair of 2A3 in parallel, and also what would be a good LED to use on the D3A cathode as an alternative to the conventional way, or maybe a CCS on the plate would be better?
I put the triode strapped D3A in TubeCad and modeled four standard grounded cathode configurations of about 36dB gain, better than +100V/-100V swing, 10 to 19mA current, plates 6k8 to 15k, Rk 100 to 180r, bypass cap probably 470uF or so.
I'm interested in knowing if this is a reasonable range to drive a pair of 2A3 in parallel, and also what would be a good LED to use on the D3A cathode as an alternative to the conventional way, or maybe a CCS on the plate would be better?
Last edited:
Member
Joined 2009
Paid Member
Hi Bigun
Strange digging such an old post out of the wilderness.
The D3a is a nice valve triode strapped but watch out for mr. miller. IMHO D3a's should be kept aside to make very nice phono amplifiers instead.
Strange digging such an old post out of the wilderness.
The D3a is a nice valve triode strapped but watch out for mr. miller. IMHO D3a's should be kept aside to make very nice phono amplifiers instead.
As a triode wired D3a appears to be anything but wimpy, I suggest DC coupling an IRFBC20 source follower to the D3a's plate, to ensure O/P tube CMiller is a non-issue.
The matter of "current hogging" comes to the fore, when O/P tubes are connected in parallel. Either "fixed" or combination bias, with individual trim pots., disposes of the matter. However, DHTs are especially vulnerable to runaway, when "fixed" bias is employed. A MOSFET source follower allows the small grid to ground resistance safety requires to be used, without untoward effects.
The matter of "current hogging" comes to the fore, when O/P tubes are connected in parallel. Either "fixed" or combination bias, with individual trim pots., disposes of the matter. However, DHTs are especially vulnerable to runaway, when "fixed" bias is employed. A MOSFET source follower allows the small grid to ground resistance safety requires to be used, without untoward effects.
Eli, you can drive 2a3 without a source follower if you want too. No need for source follower with D3a.. Just use a good CCS and run the triode strapped pentode with at least 10 mA.
But I'm telling you that mr. miller will be there if you use a 100K ohm volume potentiometer. If you use a 50k ohm potentiometer you will probably be ok though. Is the schematic calling for parallel 470 Ohm resistors? much to high.. I know - been there, done that.
D3a has pretty high miller capacitance - at line-level input it will rear its ugly head. Better to keep D3a for phono, or to use as a pentode!
Parallel 2a3's is about as good an idea as parallel el34's. With all due respect to Mr. Ciuffoli, I don't like the power supply either. High dcr on single ended "power" amp is not my cup of tea.
Why use D3a? Not the cheapest or best pentode to drive 2a3.. Much better sound would be the E810F - but with that I would consider a source follower if I really must do 2a3 pse..
But I'm telling you that mr. miller will be there if you use a 100K ohm volume potentiometer. If you use a 50k ohm potentiometer you will probably be ok though. Is the schematic calling for parallel 470 Ohm resistors? much to high.. I know - been there, done that.
D3a has pretty high miller capacitance - at line-level input it will rear its ugly head. Better to keep D3a for phono, or to use as a pentode!
Parallel 2a3's is about as good an idea as parallel el34's. With all due respect to Mr. Ciuffoli, I don't like the power supply either. High dcr on single ended "power" amp is not my cup of tea.
Why use D3a? Not the cheapest or best pentode to drive 2a3.. Much better sound would be the E810F - but with that I would consider a source follower if I really must do 2a3 pse..
Last edited:
While we're at it.. those 2a3's must have individual grid stoppers too. 1k would be ok. without them it will oscillate. serious. I built amps like this in the past.. not anymore!
For "simple", single tube, SE, you'll get no argument from me. However the OP did state PSE and the problems associated with that sort of O/P stage, "current hogging" in particular, make (IMO) the FET buffer a good idea.
Hi Eli
I'm not against the source follower, but with 18mA on the plate, the D3a won't have any problem driving parallel 2a3's. but I re-iterate - I would not use D3a.. nope.. there are better and cheaper candidates. 🙂
IMHO, its a worth investigating if you can drive without any follower just to see if works. Triode-strapped pentodes are prime candidates for this.
Again, I would not pick D3a for this duty. It does work, but it's very "forward" sounding. It is not the top of the contenders, and as you will likely agree, there is much choice here.
I've probably said too much by suggesting the E180F - but prices are still wonderfully low on that valve. A nice implementation can include CCS cascode, grid stoppers close to the pin, and moderate heater elevation.. then the user will be quite pleasantly surprised. Again - if they really want to do pse with 2a3, then they will need the source follower when using the E180F. I would rather build a push-pull if the user needs more power out of 2a3 but that's another story.
If you look at Mr. Ciuffoli's page, you will see that he tried Mu follower schematics including the one of Alan Kimmel, which is in my opinion a pretty darn good circuit. Cascode would be the clear way to do it today. Use a gyrator.. etc.. but then he throws it all away because the simple design has a more "natural sound". Huh? what??????
To be frank and honest, I think the problem he faced was that the Mu-follower designs were likely better and more clearly exposed mr. miller. If he had used a 10K POT instead, he might have noticed something significantly nicer... something perhaps even a bit magical. My own crude measurements put Cmiller for D3a at around 200pF (!)
Or it might have been that the higher miller capacitance rolled off enough high end to make the D3a sound a little less "forward".. I can't speculate much more since I just don't pick D3a for this implementation! I save D3a for phono, where signals are low enough that the high Cmiller won't much matter.
Of course getting rid of that second 2a3 will make it FAR superior too. Not to say paralleling output valves can't work.. its just that 2a3 already has two (likely dissimilar) triodes in the envelope...
Ian
I'm not against the source follower, but with 18mA on the plate, the D3a won't have any problem driving parallel 2a3's. but I re-iterate - I would not use D3a.. nope.. there are better and cheaper candidates. 🙂
IMHO, its a worth investigating if you can drive without any follower just to see if works. Triode-strapped pentodes are prime candidates for this.
Again, I would not pick D3a for this duty. It does work, but it's very "forward" sounding. It is not the top of the contenders, and as you will likely agree, there is much choice here.
I've probably said too much by suggesting the E180F - but prices are still wonderfully low on that valve. A nice implementation can include CCS cascode, grid stoppers close to the pin, and moderate heater elevation.. then the user will be quite pleasantly surprised. Again - if they really want to do pse with 2a3, then they will need the source follower when using the E180F. I would rather build a push-pull if the user needs more power out of 2a3 but that's another story.
If you look at Mr. Ciuffoli's page, you will see that he tried Mu follower schematics including the one of Alan Kimmel, which is in my opinion a pretty darn good circuit. Cascode would be the clear way to do it today. Use a gyrator.. etc.. but then he throws it all away because the simple design has a more "natural sound". Huh? what??????
To be frank and honest, I think the problem he faced was that the Mu-follower designs were likely better and more clearly exposed mr. miller. If he had used a 10K POT instead, he might have noticed something significantly nicer... something perhaps even a bit magical. My own crude measurements put Cmiller for D3a at around 200pF (!)
Or it might have been that the higher miller capacitance rolled off enough high end to make the D3a sound a little less "forward".. I can't speculate much more since I just don't pick D3a for this implementation! I save D3a for phono, where signals are low enough that the high Cmiller won't much matter.
Of course getting rid of that second 2a3 will make it FAR superior too. Not to say paralleling output valves can't work.. its just that 2a3 already has two (likely dissimilar) triodes in the envelope...
Ian
Last edited:
Anyway, the original poster is likely long gone. this is another thread pulled out of the murky depths by Bugun...
Member
Joined 2009
Paid Member
I have a friend with a small audio shop. He has lots of amplifiers, including some real boat anchors and indeed the AN Neiro. He doesn't use any of that stuff in his daily listening though.
He has a little mono RCA console that runs 45's in push-pull. It runs all day and can play any kind of music the radio can throw at it. He often gets offers for that console.
He has a little mono RCA console that runs 45's in push-pull. It runs all day and can play any kind of music the radio can throw at it. He often gets offers for that console.
Member
Joined 2009
Paid Member
Iintetrsting observation and given the price of AN gear maybe he should sell it rather than leave it on a shelf? By the way, I too prefer mono for casual listening.
Last edited:
He of course would sell it, but it seems these days that buyers for such things are not so common anymore.
The AN circuit is also not that remarkable. Perhaps the SRPP gain stage with 12AY7 sweetens up the sound with some 2nd order harmonics? I am not sure about the cathode follower stage using both triodes of a 5687 either - today a single mosfet would likely do a better job. I can only speculate due to having only heard it once, with no reference and no measurements.
In any case, all those (likely) dissimilar paralleled output triodes are not my thing.
The AN circuit is also not that remarkable. Perhaps the SRPP gain stage with 12AY7 sweetens up the sound with some 2nd order harmonics? I am not sure about the cathode follower stage using both triodes of a 5687 either - today a single mosfet would likely do a better job. I can only speculate due to having only heard it once, with no reference and no measurements.
In any case, all those (likely) dissimilar paralleled output triodes are not my thing.
Last edited:
It's not a SRPP. The output is taken from the plate of the bottom tube.
The Neiro is a 8W amplifier where 2A3's are only lightly driven into positive grid. The cathode follower is more than enough and also addresses other things. Of course working conditions were chosen by listening when Mr. Kondo designed his amps. That has always been his rule.
I have tried both driving a 2A3 PSE with individual half 6SN7-5687-6350 triodes + individual bias networks and then just like the Neiro. In most case there was no difference in terms of THD, IMD and sound. When there was a noticeable difference I found the latter was better!
No one would want a mosfet follower on such expensive gear!😀
For me is nice amp (the REAL Japanese one) with the right speakers (which is true for any valve amp).
Member
Joined 2009
Paid Member
What were the things Kondo was listening for when he designed his amps, if he was old how could he hear the hf ?
Old? I think he could hear well enough 40 or more years ago.
Actually, the kind of circuit employed in the Neiro and other Japan made AN amps certainly dates back to seventies and was rather well known in Japan by most. It was the era when the rest of the world had no idea how a "modern" valve amp could perform and sound, yet. Without the Japanese "revolution" and tube revival you would have never known valve amplifiers today....
Actually, the kind of circuit employed in the Neiro and other Japan made AN amps certainly dates back to seventies and was rather well known in Japan by most. It was the era when the rest of the world had no idea how a "modern" valve amp could perform and sound, yet. Without the Japanese "revolution" and tube revival you would have never known valve amplifiers today....
Ah yes - you are correct. But of course a FET CCS cascode will still put it to shame.
I think its fair to suggest that Kondo San might create different amplifiers if he had access to todays parts.
Ok, post your measurements to back up your claim, but it might be that the dual triodes in your 2a3's already create significant THD.
I will still prefer a single triode. If your name is anything to go by, then I think you already know what I mean. 😉
I prefer it to the tube follower, with a few special exceptions. If I had just the Kondo Iron to work with, I would certainly consider it. This is only hypothetical... I have used mosfet followers on Tango and Tamura iron to great effect. Nobody ever sees original Japanese Kondo transformers for sale....
Thus, nobody would seriously modify an original Japanese Kondo amplifier. That would be like modifying a Ferrari of old.
And a Ferrari of old can't be compared to a Ferrari of today. So let's leave it at that.
Last edited:
The 2A3 is certainly not the champion of linearity among DHT's however, by general standard, is still very good. You can get 8W with THD within 5% from a PSE easily! Carefully selecting tubes it can be 3% and get a bit more power. I don't have the measurements at hand now but it was my first build with DHT's. Hence so many tests....
One single 2A3 will get half power and overall the PSE wins because even if you introduce some distortion by paralleling triodes the PSE will demand half power to each triode to get the same level. The PSE will distort less at same Pout. Then when the SE is clipping the PSE will have another 3 dB's headroom left (if made in the same way). With standard efficiency speakers at 87-88 dB/1W it makes a clearly perceptible difference.
Regarding the cathode follower made with paralleled triodes, it's very unlikely you will see a difference respect to a single one. It's distortion is negligible in comparison.
Nowadays the "Neiro" by AN Japan has been renamed Souga and is built in monoblocks. In this case what would you do with the other half of the 5687?😀
P.S.
You can actually see distortion figures of 3x 4P1L's in parallel by Hiro here (post 756):
http://www.diyaudio.com/forums/tubes-valves/183329-one-more-4p1l-se-76.html#post3901747
and here the final result:
http://www.diyaudio.com/forums/tubes-valves/183329-one-more-4p1l-se-91.html#post4071585
From one single 4P1L you won't get better THD figures even at 0.1W....
Last edited:
5687 is a twin triode and the AN Neiro parallels both triodes to drive the output stage, if you built them as mono amplifiers nothing would change except the common power supply.
Also, Souga is a stereo amplifier and uses a considerably different driver tube to the 5687 in the earlier amplifier.
Part of the attraction of either circuit to me is the use of a split rail supply for the driver, something you still don't see very often, DC couple to the output stage with output filament referenced direct to 0V.. theres a lot to like about it, also easier to wind a quality ~1k25 OPT for PSE than ~2k5 for SE.
L.H/S
- Status
- Not open for further replies.