I used the ECC40 quite a bit for a while, a few years ago. That was before I went over entirely to DHTs. I can't remember the operating point, but that didn't matter to me. Although the ECC40 saw off any of its close competition, I just found the sound so inferior in absolute terms to the best DHTs that there was no point in using it.
I realise that many people plan their amps on a rational basis - they need a stage with this amount of gain, that plate resistance etc. I don't. I listen to everything. Small signal DHTs were literally a jaw dropping experience for me. When I first fired up a 26 I almost jumped in the air for joy, and I may have even shouted out loud "this is what I've spent the last 30 years looking for!!!". I just hear enough of a difference for me not to want to use indirectly heated tubes, and that has determined how I've designed everything ever since - filament bias and transformer coupling.
So I'm not the person to ask about indirectly heated tubes - it's almost years now since I've used any.
I realise that many people plan their amps on a rational basis - they need a stage with this amount of gain, that plate resistance etc. I don't. I listen to everything. Small signal DHTs were literally a jaw dropping experience for me. When I first fired up a 26 I almost jumped in the air for joy, and I may have even shouted out loud "this is what I've spent the last 30 years looking for!!!". I just hear enough of a difference for me not to want to use indirectly heated tubes, and that has determined how I've designed everything ever since - filament bias and transformer coupling.
So I'm not the person to ask about indirectly heated tubes - it's almost years now since I've used any.
No, I am sorry I have not tried this. You can use this if you don't have high input signal. However I have noticed that grid current starts at about -0.5 V (average). For some tubes it might even be -0.7V. For such application I would select the valves for grid current occuring as close as possible to 0V.
I actually have some ECC40's with short glass that are branded AEG Telefunken (who never made this valve) and others branded Siemens from 1978 and 1972, respectively. They are in original military boxes. However I never used short glass ECC40.
Those I used are all long glass and branded Philips Miniwatt (but actually they all have the same factory code which should mean made in France at the Radiotechnique factory) or Siemens & Halske. I should try....
In principle there can be a difference but this doesn't mean one is bad and the other is good. You might need to adjust the op. conditions to get the desired result.
I have seen this especially with PCL82's. Tungsram valves (made in Eastern Europe) are quite different from West European types.
FYI & FWIW a few of data points from the real world.
My 4P1L is now running thus:
I don't know if that helps anyone, but there you go.
For the most part and most music, I run with 6dB or more attenuation at the inputs.
My 4P1L is now running thus:
- 6SN7GT on the input, LED bias. One section for each channel. AC heaters
- 4P1L triode mode, battery bias @ -21V, cap coupled. DC heaters 4.1V
- Plate at 252V, 38mA draw.
- Small vintage 5K:8 OPT
- Clipping at 1.3W output
- Clipping at -1dB of max soundcard output at 6SN7 grid.
- 14.2V RMS at 4P1L grid for onset of clipping.
- Distortion @ 1V RMS out 1Khz
- H2, -53dB, H3, -71db, H4 -85dB
I don't know if that helps anyone, but there you go.
For the most part and most music, I run with 6dB or more attenuation at the inputs.
This is encouraging me to try Ale's setup, eat my words and use a 2C22. If I input 2v using the ODAC that Ale uses or Pano's DAC with a slight step-up, there's a theoretical chance that I could get away with 2 stages with my PSE 4P1L outputs. I'd certainly like to hear it!
Pano - how would you get 2v out of your DAC?
Pano - how would you get 2v out of your DAC?
I have no real measuring equipment except for my multimeter which I use constantly, so I can't comment on linearity. I only go by what I hear. I used the ECC40 and 2C22 in a few different projects and I liked the 2C22 a lot at the time - it was my favourite small tube before the 26. I didn't like the sound of the ECC40 quite so much in the projects I used it in. I tried a few different operating points for both. I know that many here will say tubes don't have any distinctive sound, but all my years of experience of working with tubes leads me to believe that tubes have quite distinctive sounds. Or maybe the better ones do, since it's easier to distinguish fine details. I build almost entirely by the sound of what I hear. Obviously I spend a lot of time reading up on the theory, but given that, it all goes by what I hear. I listen very carefully and try to be as methodical as I can when I do A-B comparisons, always looking out for the nul hypothesis - that there is no discernable difference - and grading differences very cautiously.
Listening is the most important thing but I think one should check at least basic objective things. Maybe the lower linearity of the 2C22 was canceling some H2 distortion of the output valve. Who knows....
I don't believe in A-B comparisons because it has been proven several times that they are tricky and often give false results. Even a simple switch can be the cause! I have learned to listen one thing per time optimized as well as possible because one can tell if is right or not anyway. The importance of listening is the only point we have in common in this case.
I don't believe in A-B comparisons because it has been proven several times that they are tricky and often give false results. Even a simple switch can be the cause! I have learned to listen one thing per time optimized as well as possible because one can tell if is right or not anyway. The importance of listening is the only point we have in common in this case.
IKO - I'm using speakers. My custom Altec VOTT A5. I don't know if I'll go with the parallel pair. It does not gain me much, but even a little could help.
Andy: Depends on which chip I use for the DAC. Some have 2V out, some are 1.5V. If needed, I can use a 1:2 step-up.
What I will probably do is dedicate this to the midrange 1005 horns, and maybe the tweeters. It sounds soooo good up there. I'll put more power in the bottom end.
Andy: Depends on which chip I use for the DAC. Some have 2V out, some are 1.5V. If needed, I can use a 1:2 step-up.
What I will probably do is dedicate this to the midrange 1005 horns, and maybe the tweeters. It sounds soooo good up there. I'll put more power in the bottom end.
No - I quite agree that A-B testing is tricky and can give false results. I'm a psychologist in my day job, so I've done quite a lot of testing of various things and I'm just as wary as you are of coming to the wrong conclusions. That's what I mean about being very careful and always looking for the nul hypothesis. All those hyperbolic "night and day" differences are as foreign to me as they would be to you. I'm expecting to hear small differences or indeed nothing some of the time. But for all its faults, I need A-B comparisons for how I build!
This is all rather OT, but maybe useful by way of explanation. Andy
7193 is not the most linear valve, agree. However is not that bad. I need to check my notes but distortion at full tilt was very good one samples were picked for such purpose.
I agree with Andy, the sound is really nice. However, I'm more inclined to try a 6e5p or 6e6p in triode mode and shunt cascode (Rod's circuit). You can get all the gain you need and distortion is outstanding at less than 0.2% for 200Vpp. Some strange push pull-ness is coming out when driving it at full tilt as the H3 component increases over H2. Either way is very promising due to gain and also results reported by Rod in terms of sound.
Ale
This is actually a common misconception. For example the hottest high-end headphone right now is an HE-6. It takes a serious 6Watts at 50 ohms. Thats more voltage swing than a 35 watt amplifier. Only a 1000V GM70 SET can deliver that sort of Voltage Swing properly. They are really a ss amp only type of headphone.
For senns they are efficient so you may be able to get away with a 5k:300 preamp OPT. Just watch the secondary DCR, a 300 ohm secodary can kill your damping factor and power transfer.
Even 32 ohm 97db/mw headphones are a problem, I had a pair of custom electraprint 5k:32. Secondary DCR is 12 ohmsSo I am losing 1/3 of my power down the drain, if I had known that I wouldn't have designed my SET around a we417.
I guess what I am trying to say is make sure you love your headphones cause you have to build the SET amp for the specific model and start with the OPT. Its very difficult to build a one size fits all.
With speakers sometime in the 60's all the manufacturers got together and said "we will make 8 ohms the standard" at the same time headphones were all standardized to a 120 ohm "tap", but the last decade that standard has been thrown out the window. So designing an amp for headphones is really tougher than speakers (if you want the best.)
Main thing is define your scope of work, the beauty of DIY is we can build to match. I definately think the 4P1l would be an ideal output tube for senns, just finding the right OPT will take some legwork, but it should be able to serve double duty as a preamp.
Hi Regal,
I followed the thread that you started SET headphone amp, and I completely understand your frustration. Have you looked at One Electron UBT-1. It is SE 1.6k to 4-8-16. Measured DCR on 16 ohm out and is under 1ohm. I would like to try a PSE 4P1L with this OT and 32-38 ohm load. Could be a decent compromise.
Best,
Radu
Just an update:
I'm running my SET 4P1L full range at the moment and it works very well. Sounds very nice. A single 6SN7GT as the input tube, LED biased, AC heater. The 4P1l outputs are running at about 240V, 35mA on battery bias, triode strapped. The batteries are bypassed with a film cap, if that makes any difference.
4P1L filaments are running at 4.2V DC, with the center point tied to ground via 1R. Supply is from the same 6.3V winding as the input tube, just rectified, filtered and regulated with a simple LT1083V adjustable regulator. Noise is low, lower than my bench supply. HT is vacuum tube rectified, then RCLC filtered.
OPTs are small, vintage 5K:8 units made by Roberts, IIRC. They have decent bandwidth and low distortion thru most of the range.
I'm very happy with it overall. It's a real spaghetti mess at the moment, but some day I'll get it in a proper chassis. Some day. ASAP, I'll switch back to an active crossover using this amp on the midrange horns - and maybe the tweeters. It should be well suited to that use, and will need about 10dB less output when I remove the passive attenuators on the horns.
The 4P1L is a nice tube, fairly easy to use and suites my needs perfectly.
I'm running my SET 4P1L full range at the moment and it works very well. Sounds very nice. A single 6SN7GT as the input tube, LED biased, AC heater. The 4P1l outputs are running at about 240V, 35mA on battery bias, triode strapped. The batteries are bypassed with a film cap, if that makes any difference.
4P1L filaments are running at 4.2V DC, with the center point tied to ground via 1R. Supply is from the same 6.3V winding as the input tube, just rectified, filtered and regulated with a simple LT1083V adjustable regulator. Noise is low, lower than my bench supply. HT is vacuum tube rectified, then RCLC filtered.
OPTs are small, vintage 5K:8 units made by Roberts, IIRC. They have decent bandwidth and low distortion thru most of the range.
I'm very happy with it overall. It's a real spaghetti mess at the moment, but some day I'll get it in a proper chassis. Some day. ASAP, I'll switch back to an active crossover using this amp on the midrange horns - and maybe the tweeters. It should be well suited to that use, and will need about 10dB less output when I remove the passive attenuators on the horns.
The 4P1L is a nice tube, fairly easy to use and suites my needs perfectly.