The sound - yes, certainly. I should start by saying what I listen for. My most important consideration, which is way ahead of everything, is the tone and timbre of acoustic instruments. I'm a classically trained musician and I spent years and years in orchestras and small jazz combos. As a consequence, the sound of acoustic instruments is indelibly imprinted on my brain. I hear a drumkit, a double bass, a grand piano, a vocalist, all the orchestral wind instruments and miscellaneous percussion so clearly in my head that all my amps are an attempt to re-create this sound. I pay attention to the woodyness of clarinets, the overtones and key action of the concert grand, the emotion and nuance in the words of the vocalists, the clang of the harp, the snap of the snare drum and so on.
Well, my whole journey in amp building has been an attempt to capture all this. When I discovered DHTs I used nothing else in all the amplification stages from that moment on - I knew I'd found the tone I was seeking. My crusade led me to several different interstages for coupling - that was the cleanest way I found of joining up stages. My preamp is 4P1L in filament bias just like the first stage here, but with a Lundahl LL1660/18mA. I like the combination of the Lundahl and the Hammond in these positions. Output stage could be any good OPT gapped for 70mA or more.
The big difference with the 300b SETs was in realism. I've known for a year that filament bias is exceptionally clean and clear, and it works great with the 4P1L in triode like it does with the 26. When I used 300b in cathode bias the cathode bypass cap - even a good polypropylene - smeared the sound unacceptably to my ears. I hate all cathode bypass caps, and all caps in general in the signal path. Yes, the 300b is a better tube and the strings in particular were more silky and smooth. If I could run the 300b in filament bias, believe me I would and the same goes for 2a3 or 6B4G. One day! It's a bit of an ask, though.
So for now I'm happy with the 4P1L PSE setup. There is a little treble graininess which is probably due to having three stages of 4P1L and nothing else - the odd harmonics probably multiply. And I haven't matched the 4P1Ls on the output for gain, just for filament draw, so that could be further optimised. But I can't live without the realism of the sound. The grand piano is just incredible - I listen to a lot of piano. Orchestral instruments are very good in brass and wind, a bit less smooth in upper strings but acceptable. Vocals are very real. If I could eliminate the hint of grain in the upper treble I would be very satisfied indeed.
I've had a few people listen to this setup with Alpair 10 fullrange speakers and we all agree that the sound is more real and immediate than with the 300b output stage. In the last shootout we had, the real contender was a single stage SE solid state amp with the Sony 2SK82. But that's another story.
Well, my whole journey in amp building has been an attempt to capture all this. When I discovered DHTs I used nothing else in all the amplification stages from that moment on - I knew I'd found the tone I was seeking. My crusade led me to several different interstages for coupling - that was the cleanest way I found of joining up stages. My preamp is 4P1L in filament bias just like the first stage here, but with a Lundahl LL1660/18mA. I like the combination of the Lundahl and the Hammond in these positions. Output stage could be any good OPT gapped for 70mA or more.
The big difference with the 300b SETs was in realism. I've known for a year that filament bias is exceptionally clean and clear, and it works great with the 4P1L in triode like it does with the 26. When I used 300b in cathode bias the cathode bypass cap - even a good polypropylene - smeared the sound unacceptably to my ears. I hate all cathode bypass caps, and all caps in general in the signal path. Yes, the 300b is a better tube and the strings in particular were more silky and smooth. If I could run the 300b in filament bias, believe me I would and the same goes for 2a3 or 6B4G. One day! It's a bit of an ask, though.
So for now I'm happy with the 4P1L PSE setup. There is a little treble graininess which is probably due to having three stages of 4P1L and nothing else - the odd harmonics probably multiply. And I haven't matched the 4P1Ls on the output for gain, just for filament draw, so that could be further optimised. But I can't live without the realism of the sound. The grand piano is just incredible - I listen to a lot of piano. Orchestral instruments are very good in brass and wind, a bit less smooth in upper strings but acceptable. Vocals are very real. If I could eliminate the hint of grain in the upper treble I would be very satisfied indeed.
I've had a few people listen to this setup with Alpair 10 fullrange speakers and we all agree that the sound is more real and immediate than with the 300b output stage. In the last shootout we had, the real contender was a single stage SE solid state amp with the Sony 2SK82. But that's another story.
I'm not quite sure what "smearing" means to be frank. I used the term for cathode bypasses because the sound goes slightly "foggy" to my ears - loss of detail and timbre and directness. Like if you smeared oil over a piece of glass. Maybe with PSE it refers to something different - one tube interfering with another. That might simply cause more distortion which might be a different subjective sound, like more edgy?
There is a slight increase, but minimal when matched mu pairs are used: http://www.bartola.co.uk/valves/2012/11/04/4p1l-pse-tests/
Ale
Right. The actual word and mechanism why this is so it's where the debate is. I think the consensus is that paralleling two tubes only sometimes sound better than one. I have to say I haven't played with this at all. Got a whole bunch of 4P1L for this but... so little time. I'm also curious of a 4P1L PP.
Right now I am developing PP with 4P1L. C3G (ltp) in front. B+ around 240V. OT 5K:8. I also plan to build PPP version.
PP 4P1L is an obvious build, and PPP even more so. Are you thinking of using a CCS under the output stage, or a pentode sink? Allen Wright was using his kind of thing. It eliminates a cathode bypass cap which is always a good thing.
How do you propose to match up the 4P1Ls? I think Ale had a circuit to do this.
You could use one filament supply for each tube. I've thought of doing this with my output stage in filament bias. Easier to match up the operating conditions.
How do you propose to match up the 4P1Ls? I think Ale had a circuit to do this.
You could use one filament supply for each tube. I've thought of doing this with my output stage in filament bias. Easier to match up the operating conditions.
Does your driver have enough voltage to swing 4 of them in parrallel? Seems like a 1.25k OPT @ 10watts would be a nice solution.
Say like this?
http://www.edcorusa.com/p/465/cxse25-8-1_25k
Last edited:
Sure, trioded 6Ж9П can do that easily. You may try some modern low capacitance high voltage MOSFET, it may be even better.
I have some 30W 1.2K OPTs, but was shy to use them for 10W only. May be I will try later. Bass would be gorgeous, though. Very full range 10W SE amp. Too many projects already...
I need several apprentices may be.
Member
Joined 2009
Paid Member
How is it different? Well, look at the coupling cap.
When it is loaded on the non-linear resistance it creates dynamic distortions. So if you use SRPP (not a mu-follower, it has higher output resistance and introduces additional non-linear distortions driving the grid positively) with the cap between it's output and the control grid of 4P1L you would get dynamic distortions caused by rectification of the signal on the cap-grid shifting bias point.
It is the main reason why people believe that 300B amps sound bad: because of this dynamic distortions. Using the cap between anode of the driver and grid of the follower I eliminated this problem.
Now if you ask me to remove the coupling cap, I will ask you how to organize stable DC coupling. However, it is possible to have a 2-floor even 3-floor power supply, voltage stabilized, but it would be completely different project. I am working on something similar currently.
Member
Joined 2009
Paid Member
I see - it's the capacitor in the drive to the grid of output tube that is the beast to be avoided since positive grid current on clipping into A2 produces some charging of the capacitor.
So if I understand, with appropriate power supply there is an option of all-dc coupled (a) which means no capacitors. There is also mu-follower which puts the Cathode Follower in totem-pole and also can avoid capacitor before power tube grid but can't avoid capacitor in the totem pole (b)
unfortunately, I only get over to the West coast 4 times a year and it's always to Thousand Oaks - too far from your gaff
So if I understand, with appropriate power supply there is an option of all-dc coupled (a) which means no capacitors. There is also mu-follower which puts the Cathode Follower in totem-pole and also can avoid capacitor before power tube grid but can't avoid capacitor in the totem pole (b)
that could be funI need several apprentices may be.
unfortunately, I only get over to the West coast 4 times a year and it's always to Thousand Oaks - too far from your gaffAttachments
Last edited:
Member
Joined 2009
Paid Member
You could use a shunt regulator as a cathode bypass.
But I'm not so concerned about having zero capacitors, more interested in how the mu-follower differs from common cathode + cathode follower. Especially a hybrid, using a FET on top in a mu-follower might allow painless A2 of the output tube.
But I'm not so concerned about having zero capacitors, more interested in how the mu-follower differs from common cathode + cathode follower. Especially a hybrid, using a FET on top in a mu-follower might allow painless A2 of the output tube.
The main problem is, tubes amplify. It is easy to directly couple stages with resistive load: the worst thing that may happen is saturation. But when the tube has transformer load it pulls as much current as it can. And this current changes with change of the voltage between cathode and grid. Slight idle voltage change changes idle current such a way it can easy kill the tube. That's why cathode resistor is used, to keep the tube alive. But it decreases voltage amplification factor and increases output resistance. That's why it is usually shunted by capacitor.
Yes, we can solve the problem of bias by shunt regulator in cathode. But we can't solve the problem of it's stability: slight change of current through the driver will cause big current change on the output stage.
Yes, we can solve the problem of bias by shunt regulator in cathode. But we can't solve the problem of it's stability: slight change of current through the driver will cause big current change on the output stage.
hi guys
nice work on the 4p1l. i'm gathering parts to build my own version too.
looking at this curves traced by bartola valves, i'm curious about the operating point of 220vak, -18vgk, said to be the sweet point for this tube.
4?1?/4P1L triode curves | Bartola Valves
this point doesn't seem to be on the curves?
http://www.bartola.co.uk/valves/wp-content/uploads/2012/06/4P1L-triode-1024x832.jpg
unless the curves doesn't start at 0v?
appreciate any comments.
bg
nice work on the 4p1l. i'm gathering parts to build my own version too.
looking at this curves traced by bartola valves, i'm curious about the operating point of 220vak, -18vgk, said to be the sweet point for this tube.
4?1?/4P1L triode curves | Bartola Valves
this point doesn't seem to be on the curves?
http://www.bartola.co.uk/valves/wp-content/uploads/2012/06/4P1L-triode-1024x832.jpg
unless the curves doesn't start at 0v?
appreciate any comments.
bg