2P29L + 6J5 (choke loaded) == enormous gain (over 40dB).
Possible hiss, hum and 6J5 distortion profile. Is this worth?
Even 6J5/6SN7 has enough gain to drive 2A3 .. and you have "ordinary" 2A3 SE amp ... with ordinary tone.
Yes exactly - 6J5 isn't really a driver tube, though it's a very clean sound. As I was saying, it would need a voltage divider in front of it. I tried that and it did work, but it's not a very elegant solution. I would expect more of a 2P29L in the driver position. With resistor load the gain is around 7.5. If a 2P29L can drive a 2P29L and sound good then that's one option. I'll rig it up and try it. Might try a 4P1L as well.
6J5 driving 2a3 is exactly what I have been using. If you choose a good version of 6J5 like CV6 the tone is a bit more than "ordinary" I would say. And this is my "default" 2a3 amp with the virtues of a 2 stage amp - clean and detailed. But put a DHT in the input and the tone improves quite noticeably, which just leaves you with the choice of a driver tube, which is what I've been exploring.
Andy - you may try 2-stage with 6J5 type at hotter operating point. Choke loaded, Ua=300 V, Ik=12 mA. This will exceed 6J5 plate dissipation rating, but you can use 1/2 6CG7 (Pmax=4 W). You will get 4-5 times less driver THD than 150 V 6 mA resistor-loaded OP.
With 2P29L, you need a very low mu driver, like 45, for optimal gain structure. 10Y and 46 are a little too much.
With 2P29L, you need a very low mu driver, like 45, for optimal gain structure. 10Y and 46 are a little too much.
Last edited:
Yes, I could do that but I prefer CV6 to 6CG7. It depends, though, on whether the tone is better with the choke or the resistor. A choke has to be good to better a simple resistor in my experience. For me it's all about tone - transparency and tone.
And yes again, a low mu driver like mu=4 would be ideal. I don't have 45. I may be able to scare up a pair of 71A, unfortunately sold my stash of them. I'll have a think - there are some low mu DHTs if you look for them. A 6C4C would work of course and I have plenty, but not in filament bias. But I'll start with 2 x 2P29L in filament bias. Resistor loads reduce the gain a little. Direct coupling the first 2 stages might help also.
And yes, I put a grid resistor in the top cap. Usually 390R but 100R or 220R might be better.
1G4 is a nice tube. It's not as good as 2P29L and microphonic as I remember. I had boxes full of various DHTs but sold them all. I sold off a lot of tubes like 30, 30SP, 31, 71A, 1G4, 3a5 etc etc. I kept 01A, 112A, 26, 2P29L, 4P1L, 10Y, 46 and 47 which sounded the best of the bunch to me. I did audition a whole stack of DHTs before thinning down my collection.
The 31 would be useful as a driver in fact, but they're all gone. I might buy some more - it looks perfect as a driver and it did sound pretty good.
... and you will give about 36dB gain (R loaded 2P29L direct coupled to choke loaded 2P29L).
I'm really going round the houses on this one. What I've found out so far is that an indirectly heated driver tube following a DHT in a 3 stage amp doesn't provide any significant benefit over a good indirectly heated input tube in a 2 stage amp. With 3 stages you lose the benefits of the DHT input tube. So that's a dead end, leaving me with some surplus 1626 tubes, offers welcome.
I then went back to 2P29L on its own driving the 2a3 directly as my starting point. This gave me everything I want in terms of purity and tone. So back to how to add gain. Having rejected an IDHT driver it has to be a DHT driver tube. Type 31 looks good but I don't have any and they're pretty rare now, so back to the plentiful and cheap 2P29L, which I have yet to build as a driver stage. I don't think I'd go to direct coupled - too complex and would require a choke and cathode bypass cap. The 2P29L can be used as a driver tube in filament bias with a resister load.
For the moment I thought I'd stick a step-up transformer in front of the 2P29L. A LL1540 was disappointing, but a pair of Hammond 124B turned out to be quite good. This is a 49% nickel transformer, nominally 10K into 90K. In the UK they are £36 each, so a friendly price. This is better than I expected and will stay for a while to do some listening. Initial impressions are that you do get almost all the tone of the 2P29L, but with just a slight touch of coarseness overlaid on top of it. Not very bothering on jazz and rock - it showed up most on strings and orchestral music. This leaves the door open for a better step-up transformer. This is quite a good stand-in setup until something better emerges that keeps all the purity and tone of the DHT sound. It's easy to build and not expensive. The gain is around the same as using a 6J5 in a 2 stage amp. Tone is better, purity not quite as clean as just a 6J5. Swings and roundabouts.
I then went back to 2P29L on its own driving the 2a3 directly as my starting point. This gave me everything I want in terms of purity and tone. So back to how to add gain. Having rejected an IDHT driver it has to be a DHT driver tube. Type 31 looks good but I don't have any and they're pretty rare now, so back to the plentiful and cheap 2P29L, which I have yet to build as a driver stage. I don't think I'd go to direct coupled - too complex and would require a choke and cathode bypass cap. The 2P29L can be used as a driver tube in filament bias with a resister load.
For the moment I thought I'd stick a step-up transformer in front of the 2P29L. A LL1540 was disappointing, but a pair of Hammond 124B turned out to be quite good. This is a 49% nickel transformer, nominally 10K into 90K. In the UK they are £36 each, so a friendly price. This is better than I expected and will stay for a while to do some listening. Initial impressions are that you do get almost all the tone of the 2P29L, but with just a slight touch of coarseness overlaid on top of it. Not very bothering on jazz and rock - it showed up most on strings and orchestral music. This leaves the door open for a better step-up transformer. This is quite a good stand-in setup until something better emerges that keeps all the purity and tone of the DHT sound. It's easy to build and not expensive. The gain is around the same as using a 6J5 in a 2 stage amp. Tone is better, purity not quite as clean as just a 6J5. Swings and roundabouts.
2A3 estimated grid bias voltage is -45V at A1 (datasheet 250V, 60mA, -45V).
If you reach this with 2VRMS input signal (2.83Vpeek), you need 24dB gain.
With some headroom this requirement would be 26-27dB.
2P29L (filament biased, R loaded) stage gain about 17dB.
Requiring more 9-10dB gain.
1:4 SUT has 12dB, so it's enough.
BTW 2P29L stage output impedance isn't low, about 3k.
If you reach this with 2VRMS input signal (2.83Vpeek), you need 24dB gain.
With some headroom this requirement would be 26-27dB.
2P29L (filament biased, R loaded) stage gain about 17dB.
Requiring more 9-10dB gain.
1:4 SUT has 12dB, so it's enough.
BTW 2P29L stage output impedance isn't low, about 3k.
Andy - your experience is priceless.
As to paralleled 2P29Ls as driver, they simply don't have the umph. A good driver should use as high plate voltage as possible, preferably close to that of the output stage. 2P29L falls short.
45 is not as outrageously expensive as 10Y.
For detail, and purity of tone, you cannot go wrong with old 80% Ni input transformers, such as UTC A-20, or, even better, HA-108. They are not much more expensive than current production offerings.
With 2P29L, it is better to put 1:2 step-up after it than before. This way the tube has 1/2 output swing, resulting in less distortion with limited plate voltage.
I recon that in your designs you put more emphasis on parts quality than on operating conditions for minimum distortion. You should not neglect the latter.
As to paralleled 2P29Ls as driver, they simply don't have the umph. A good driver should use as high plate voltage as possible, preferably close to that of the output stage. 2P29L falls short.
45 is not as outrageously expensive as 10Y.
For detail, and purity of tone, you cannot go wrong with old 80% Ni input transformers, such as UTC A-20, or, even better, HA-108. They are not much more expensive than current production offerings.
With 2P29L, it is better to put 1:2 step-up after it than before. This way the tube has 1/2 output swing, resulting in less distortion with limited plate voltage.
I recon that in your designs you put more emphasis on parts quality than on operating conditions for minimum distortion. You should not neglect the latter.
Like you say CV6 is better than 6j5 or 6CG7. It is on the assumption of 150 V 6 mA operating point, which is very far from optimal, technically speaking, for a 2A3 driver. Who knows, if you use the best OP, maybe your preference will change?
In my experience, operating point is more important for sound quality than tube brand or variety.
In my experience, operating point is more important for sound quality than tube brand or variety.
Thanks Bela - very useful calculations.
sser2 - point taken about operating points. What has been restricting my choice is a B+ of around 300V and a preference for resistor loads. With a plate choke the choice of operating point would be much wider, and I'll try that at some point. As I always say, my most important factor is the tone of acoustic instruments. I'm a conservatoire trained musician and was pro for many years, so the tone of acoustic instruments is indelibly imprinted on my brain. So yes, in my case I prioritise the variety of tube in terms of sound quality and tone and that means DHT tubes wherever possible. Parts are important too - I always use teflon coupling caps and DC Link caps for larger sizes. But I will pay more attention to operating points as you say.
As for higher quality input transformers, I have no idea where to start - there are so many available new and used. So all suggestions welcome. I would need 1:3 or 1:4. I'm in the UK so anything European works, as well as US types. The amp is driven by a typical DAC with AK4490 chip and op-amp output giving 2v.
sser2 - point taken about operating points. What has been restricting my choice is a B+ of around 300V and a preference for resistor loads. With a plate choke the choice of operating point would be much wider, and I'll try that at some point. As I always say, my most important factor is the tone of acoustic instruments. I'm a conservatoire trained musician and was pro for many years, so the tone of acoustic instruments is indelibly imprinted on my brain. So yes, in my case I prioritise the variety of tube in terms of sound quality and tone and that means DHT tubes wherever possible. Parts are important too - I always use teflon coupling caps and DC Link caps for larger sizes. But I will pay more attention to operating points as you say.
As for higher quality input transformers, I have no idea where to start - there are so many available new and used. So all suggestions welcome. I would need 1:3 or 1:4. I'm in the UK so anything European works, as well as US types. The amp is driven by a typical DAC with AK4490 chip and op-amp output giving 2v.
Last edited:
I absolutely agree on directly heated tubes and quality parts. And instrument tone is as important to me.
You would agree that all pianos have different tone, and big part of listening to piano music is savoring the tone. For great artists use best of the best instruments, and performances are closely miked. With a resolving system, you get all of it. You hear how instrument opens up from piano to forte, wonderful decays, delicate pianossimos and violent fortissimos. You hear all extraneous sounds - pianist breathing, a faint squeak of piano bench, tap of keys - which are all parts of live performance. You also hear the shortcomings, like poor job of piano techs on Yundi Li' s recordings, or tampering by recording engineers. Just piano alone deserves the best sound system, not even speaking of vocal and orchestral music.
You would agree that all pianos have different tone, and big part of listening to piano music is savoring the tone. For great artists use best of the best instruments, and performances are closely miked. With a resolving system, you get all of it. You hear how instrument opens up from piano to forte, wonderful decays, delicate pianossimos and violent fortissimos. You hear all extraneous sounds - pianist breathing, a faint squeak of piano bench, tap of keys - which are all parts of live performance. You also hear the shortcomings, like poor job of piano techs on Yundi Li' s recordings, or tampering by recording engineers. Just piano alone deserves the best sound system, not even speaking of vocal and orchestral music.
Why not ask Ale to lend it for a trial his LL7903 transformers?
This transformer (wired 1:4) as good as my Sowter 9063 ones.
BTW if you use 1:4 SUT, must to increase filament bias resistor value up to -about- 68R.
See sample:
Last edited:
Thanks for the circuit, Bela. Very useful.
I've been enjoying the Hammond 124B into 2P29L for several hours today, and it's a good start on the road to a SUT input. It's so nice to have an all-DHT amp. Makes all the difference to the tone. Not only Ale is using the LL7903, Thomas Mayer has used it into his 10Y > 45 amp. But there are plenty of choices besides the LL7903, which isn't cheap.
I've been enjoying the Hammond 124B into 2P29L for several hours today, and it's a good start on the road to a SUT input. It's so nice to have an all-DHT amp. Makes all the difference to the tone. Not only Ale is using the LL7903, Thomas Mayer has used it into his 10Y > 45 amp. But there are plenty of choices besides the LL7903, which isn't cheap.
LL7903 is a good transformer, but it is huge overkill for use as 1:4 input SUT. There is no need for 1:8 capability and 27 dBu at the line level.
I modified the input circuit to give me 12V on the cathode of the 2P29L. This sounds comfortable with the SUT. The filament supply is choke input with a 158T Hammond choke. This sounds like it works well. The whole input stage sounds pretty good to me. And it's very economical to build - no expensive parts. I'm liking this.
Last edited: