Agree with Kevin. Starving filament will decrease gm and increase rp. Not probably what you would want for a driver stage.
Have a look at the attached 4P1L traces. I've paralleled the filaments and did two tests with nearly nominal filament current (e.g. 670mA) and starved. You can clearly see that at your bias point you need to run the valve at 20mA to get the best out of it:
My Photos
With regards to the sound, I have to say that I like it a lot as a preamp stage. Great bass and clear tone, one of my favorite DHT despite its microphony.
Hope this helps?
Cheers,
ALe
Have a look at the attached 4P1L traces. I've paralleled the filaments and did two tests with nearly nominal filament current (e.g. 670mA) and starved. You can clearly see that at your bias point you need to run the valve at 20mA to get the best out of it:
My Photos
With regards to the sound, I have to say that I like it a lot as a preamp stage. Great bass and clear tone, one of my favorite DHT despite its microphony.
Hope this helps?
Cheers,
ALe
Hi there,
26 is superb but only with the right anode load (e.g. OT). Other DHTs are more suitable for gyrator or choke loads as have lower anode resistance (e.g. 71, 4P1L).
A difficult question to answer, but from all my tests I have to say 26, 71a, 4P1L and 46 are my preferred ones. It all depends how much gain and the end to end configuration of your system. Also power supplies are critical and as important as the stage. Using Rod Coleman's regulators takes the DHT to a next level....
Cheers,
Ale
Hi!
Don't just dial up the current. getting the op point away from the knee of the curves is one thing to do. Set the op point such that you have maximum available swing in positive and negative direction.
Thomas
Don't just dial up the current. getting the op point away from the knee of the curves is one thing to do. Set the op point such that you have maximum available swing in positive and negative direction.
Thomas
Now have the 4P1L running at 14mA, with 140v on the anodes. Also the filaments were previously more starved than I thought - I've restored them to 3.8v in series. This has all made the sound a lot better - more pleasant tone. It's not where I want it yet. I think I need more voltage on the anodes for a start. Then there's the whole question of the anode chokes. The Hammond 124b may be a bit small, though rated at 40mA. I have some LL1635 and some LL1621 which I can try. The LL1621 have a rather unusual implementation - it shows 0.1uF across the transformer between anode of V1 and grid of V2. Not sure why this is the case - treble response no doubt.
andy
andy
Hi mogliaa.
Have read a lot about your experiment with the 4p1l tube , as a line stage.
I'm a total newbie, but can get some help from a builder here in Denmark.
It your schematic "the siberian" the best you have tried so far, or have you done
any further progress with it ??
Any recommandation for a PSU, which voltage on the secondary ??
I saw ron coleman has a link on some 150VA which he recommend...
/Thanks Michael...
Have read a lot about your experiment with the 4p1l tube , as a line stage.
I'm a total newbie, but can get some help from a builder here in Denmark.
It your schematic "the siberian" the best you have tried so far, or have you done
any further progress with it ??
Any recommandation for a PSU, which voltage on the secondary ??
I saw ron coleman has a link on some 150VA which he recommend...
/Thanks Michael...
Hi Michael, 4P1L likes 20mA of anode current. Both gyrator load or transformer are very good options. I rather use transformer if not gain is needed in your system. A step down transformer like LL1660/20mA in 4.5:1 is ideal for good driving and low output impedance. You can get a very good bass with 4P1L. Filaments need starving at 550mA and in parallel to reduce microphonic noise.
Good luck in your project
Ale
Good luck in your project
Ale
I use the 4P1L as the input stage in my 4P1L PSE amp, fed by a 9023 DAC outputting 2v. With 2v input microphonics are less of an issue than with say 0.5v in where you have to be creative. I use a LL1660 in 4:4.5 because I want to keep the gain for a 2-stage amp. Works fine. Two things you can do for microphonics:
1. Wrap plumbers teflon tape around the bulb. Helps a little.
2. Use a big and heavy chassis
Here are a couple of photos.
Attachments
Hi Ale.
You referer to the LL1660/20mA , but when I check supply out on JACmusic
page I only see these models.
10,14,18,25 mA - http://jacmusic.com/html/order/jacmusic-pricelist.pdf.
Which one would be the best for the given job ??
Have you done any improvement on "the siberian" on your page ??
Looking at your schematic from your page : Siberian DHT pre-amp | Bartola Valves
PS. Sorry I'am newbie so please have a little patience.
/Thank Michael.
Hi Michael,
Ask Big Bear Audio or VinylSavor for quotes on LL1660. You can get it gapped to the current you need. If not, 25mA would be great.
Unfortunately haven't had time to document all the variations I tried so far. I'm actually rebuilding mine again with an LL1660, filament bias and Rod's filament boards. Concepts and parts are similar to the Siberian and 26 preamps.
Are u lacking of gain in your system? what is the input stage of your amplifier? are you driving long cables? You need to consider these questions when choosing the best configuration for your system
cheers
Ale
Ask Big Bear Audio or VinylSavor for quotes on LL1660. You can get it gapped to the current you need. If not, 25mA would be great.
Unfortunately haven't had time to document all the variations I tried so far. I'm actually rebuilding mine again with an LL1660, filament bias and Rod's filament boards. Concepts and parts are similar to the Siberian and 26 preamps.
Are u lacking of gain in your system? what is the input stage of your amplifier? are you driving long cables? You need to consider these questions when choosing the best configuration for your system
cheers
Ale
Hi Ale.
Too be honest , we are four die hard HIFI alcoholic located in Denmark, who just has a new life on out 2a3 DC amp, which we would like to extend r'further.
As you properly know life is now the same anymore, when you touch the sweetness of a DHT tube amp.
I can't tell the difference, but something has changed.
Therefore I would like to listening to the "4P1L" because I feel that it's the right preamp for me.
Can It be run In direct coupled mode, without any caps in the signal path ????
/Best Michael.
Too be honest , we are four die hard HIFI alcoholic located in Denmark, who just has a new life on out 2a3 DC amp, which we would like to extend r'further.
As you properly know life is now the same anymore, when you touch the sweetness of a DHT tube amp.
I can't tell the difference, but something has changed.
Therefore I would like to listening to the "4P1L" because I feel that it's the right preamp for me.
Can It be run In direct coupled mode, without any caps in the signal path ????
/Best Michael.
Hi Michael,
Glad to hear you are loving the DHT amplifier.
If you want to minimise the use of capacitors in the signal path, you can look at the popular design I replicated in the 26 preamp. This is what I also used lately with the 4P1L and what I'm planning to rebuild shortly.
26 Pre-amplifier Gen3 | Bartola Valves
As you can see, most of the effort is put into the power supplies. Is key to get an absolutely clean DC filament supply to feed the Rod Coleman regulators if you are looking to implementing filament bias. That is where most of your budget will be burnt! Think about split bobbins EI transformers and good quality chokes for choke input supplies.
For the HT, I'd go for oil caps and LCLC feeding either an SSHV2 shunt regulator or a CCS+VR or Morgan Jones' statistical shunt regulator.
Wiring of the preamp and special care to ground wiring and loops is a must to get the preamp dead quiet. I can stress enough how important this is. You can spend a lot of money and screw up your build just because how you solder your components to ground!
good luck!
Ale
Glad to hear you are loving the DHT amplifier.
If you want to minimise the use of capacitors in the signal path, you can look at the popular design I replicated in the 26 preamp. This is what I also used lately with the 4P1L and what I'm planning to rebuild shortly.
26 Pre-amplifier Gen3 | Bartola Valves
As you can see, most of the effort is put into the power supplies. Is key to get an absolutely clean DC filament supply to feed the Rod Coleman regulators if you are looking to implementing filament bias. That is where most of your budget will be burnt! Think about split bobbins EI transformers and good quality chokes for choke input supplies.
For the HT, I'd go for oil caps and LCLC feeding either an SSHV2 shunt regulator or a CCS+VR or Morgan Jones' statistical shunt regulator.
Wiring of the preamp and special care to ground wiring and loops is a must to get the preamp dead quiet. I can stress enough how important this is. You can spend a lot of money and screw up your build just because how you solder your components to ground!
good luck!
Ale
Yes - I'm saying the same on the "One more SE 4P1L" thread. You need to think of filament bias as another world - you're not just heating the tube up. The signal goes through the filament supply, so you need to think of it just like any other signal PSU - choke input if possible, and Rod's regulator or the equivalent. This is what we're also saying for the signal PSU - choke input and some kind of regulator.
But consider your options for the whole system. I started off with conventional thinking - you use a 2a3 or a 300b output because they are what you use for best sound. OK. Now I need a DHT to drive it because I don't like the sound of indirectly heated tubes any more. So what's the most gain I can get? I could use a 3a5 with both halves together for a gain of 15, but 4P1L sounds better for a gain of around 10. OK. So now I need a preamp. I'll build a 26 preamp. That's what I did, in fact I built about 3 amps like this.
Then one day I built an amp with parallel 4P1L as output - why not, I have 150. Hmmmm.... this sounds damn good in filament bias. I'm still using a 4P1L driver and a 26 preamp. Then I get a ES9023 DAC which puts out 2v, and I connect it straight to the 4P1L driver. WOW - sound is cleaner. This is GOOD - 2 stages at last, no coupling caps, no cathode bypass caps, all stages in filament bias. I haven't been able to improve on this for a year and a half, and I don't see how I can right now except change the transformers. I have LL1660 in 4:4.5 and LL1620/80mA. I'm thinking Tango, Tamura, Hashimoto.... whatever. Bear in mind that with PSE 4P1L outputs you are adding enough gain to your system to make a 2-stage all-DHT amp a possibility. There's not many other ways you can do this, at least not all in filament bias so you're back to cathode bypass caps or fixed bias or various other problems you have to find solutions for, like a 211 output (mu = 11) but that's high voltage, PX25 but that's rare and costly, or 10Y (mu 8) which is a very good idea (also in PSE) but a bit more work and a lot more money. The 10Y could just about be used in filament bias - if you want to throw money at an amp I'd consider it, but PSE 4P1L is much simpler and cheaper and a little more gain. All this assumes 2v input and reasonably sensitive speakers. Yes, the 2a3 and 300b are great tubes but by the time you add cathode bypass caps and an extra stage because the gain isn't high enough and you're not using filament bias, then you lose all the advantages you would otherwise have. Which is actually very little, because PSE 4P1L is in the same ballpark for sound quality.
That's an idea to think about!
Last edited:
I have reached different conlcusions over the years about 2 or 3 things.
First it is not true that less stages are better in general. It always depends on the real gain of the single devices, the amp design and all the other boundary conditions. An amplifier with two stages is not better than one with three by default. For some devices it is not possible to get decent results with two stages and three is always the best solution. If another amp can do it with two then it is good but superiority is not a given. Another problem with many 2 stage amps is that they are power amps, not integrated amps. It might not be enough for everyone. As an extreme example, I have recently seen an amp just made with one device, the EML 20B, however this must use a step-up input transformer which can work but in my experience is difficult to drive and rarely sounds good. That's just shifting problems upstream the chain and making them worse, IMHO.
From what I am seeing, the 4P1L -although is a pretty linear device in triode mode - it is not any better than 2A3 and 300B as it has been claimed many times in practical working conditions as an output device. The 300B in particular is superior. Of course if one buys Chinese un-selected 300B's anything can happen....
The only real advantages are gain, low filament consumption and tight tolerances. One has to put 4 in parallel to compete with a 300B and distortion at low power has to be seen if it can be as good as a single 300B or a 2A3 PSE which can sport 0.2-0.3% @ 1W quite easily. The 2A3 need to be selected but some Chinese types and the 6C4C are good enough. Often one gets even less than 0.2% with the 300B....
Second, cathode bias with 2A3 and 300B is good for cheap solutions otherwise fixed bias is superior. Better if there is a DC coupled CF or SF. One thing I really don't like about filament bias for output devices is that they cannot be efficiently driven into positive grid. So even four 4P1L's in parallel could only manage just over 6W before clipping hard. A 300B SE or 2A3 PSE with the same total plate dissipation (around 30-32W) can manage 9-10W easily and clipping is much softer with power drive. The difference, when listening, is not really in terms of SPL but overall presentation with highly dynamic music.
Third, above 92-93 dB true efficiency speakers have too many tradeoffs for my taste, especially if they are less than 3 ways.
First it is not true that less stages are better in general. It always depends on the real gain of the single devices, the amp design and all the other boundary conditions. An amplifier with two stages is not better than one with three by default. For some devices it is not possible to get decent results with two stages and three is always the best solution. If another amp can do it with two then it is good but superiority is not a given. Another problem with many 2 stage amps is that they are power amps, not integrated amps. It might not be enough for everyone. As an extreme example, I have recently seen an amp just made with one device, the EML 20B, however this must use a step-up input transformer which can work but in my experience is difficult to drive and rarely sounds good. That's just shifting problems upstream the chain and making them worse, IMHO.
From what I am seeing, the 4P1L -although is a pretty linear device in triode mode - it is not any better than 2A3 and 300B as it has been claimed many times in practical working conditions as an output device. The 300B in particular is superior. Of course if one buys Chinese un-selected 300B's anything can happen....
The only real advantages are gain, low filament consumption and tight tolerances. One has to put 4 in parallel to compete with a 300B and distortion at low power has to be seen if it can be as good as a single 300B or a 2A3 PSE which can sport 0.2-0.3% @ 1W quite easily. The 2A3 need to be selected but some Chinese types and the 6C4C are good enough. Often one gets even less than 0.2% with the 300B....
Second, cathode bias with 2A3 and 300B is good for cheap solutions otherwise fixed bias is superior. Better if there is a DC coupled CF or SF. One thing I really don't like about filament bias for output devices is that they cannot be efficiently driven into positive grid. So even four 4P1L's in parallel could only manage just over 6W before clipping hard. A 300B SE or 2A3 PSE with the same total plate dissipation (around 30-32W) can manage 9-10W easily and clipping is much softer with power drive. The difference, when listening, is not really in terms of SPL but overall presentation with highly dynamic music.
Third, above 92-93 dB true efficiency speakers have too many tradeoffs for my taste, especially if they are less than 3 ways.
Last edited:
As posted here:
http://www.diyaudio.com/forums/tubes-valves/248829-rh300b-12w-1-thd-2.html
a simple 2 stages 300B amp with RC coupled ECC83 (sections in parallel) and cathode bias can have very low distortion and pretty good sound. This is less than 0.2% at 1W and less than 2% at 9W. Without 2nd harmonic cancellation the figures were 0.3% and about 3.5%, respectively. I don't think a 4P1L PSE can compare. The limits of that 300B amp only come out when used at its full power in comparison to another 10W low distortion amp....
http://www.diyaudio.com/forums/tubes-valves/248829-rh300b-12w-1-thd-2.html
a simple 2 stages 300B amp with RC coupled ECC83 (sections in parallel) and cathode bias can have very low distortion and pretty good sound. This is less than 0.2% at 1W and less than 2% at 9W. Without 2nd harmonic cancellation the figures were 0.3% and about 3.5%, respectively. I don't think a 4P1L PSE can compare. The limits of that 300B amp only come out when used at its full power in comparison to another 10W low distortion amp....
Hi there. I can see that you have come to these conclusions by careful calculation, all of which is very interesting.
This is where our priorities start to differ - the actual sound of the system. I used indirectly heated tubes for most of my life until recent years, like most people. When I first heard a 26 my jaw dropped (literally) and I believe I actually said out loud "this is the sound I've been waiting for all my life". I was stunned at the transparency, dynamics and the timbre of acoustic instruments. I then spent about a year comparing the very large number of indirectly and directly heated tubes I acquired - which included the absolute best 6SN7 and all the rest of them. At the end of the year I could find NO indirectly heated tube that had exactly the same transparency as a DHT. The 2C22 came about the closest but not quite. So at that point I simply stopped using indirectly heated tubes and I've never regretted it or gone back. At some other point I spent quite a lot of time comparing cathode bypass capacitors. I made a test rig and tried out just about every combination I could of electrolytics, polypropylenes, polystyrenes, and even banks of teflon caps, like 20 soldered together. I didn't like any of the combinations (some were downright awful) and at that point I decided to design in such a way as to eliminate cathode bypass caps. My experiments led me to conclude that two stages of filament bias with DHTs were better sounding even than a directly coupled stage where the second tube had a cathode bypass capacitor. I also concluded that a good interstage sounded better than cap coupling.
I spent a very large number of hours building stuff, breadboarding, comparing and so on, and that led me to the point I'm at. I listen to a lot of opera and orchestral scores (I'm an ex-pro musician who used to play in orchestras) and the two things I've always demanded of my system are transparency so I can hear all the instruments and parts, and faithful timbre and tone so each acoustic instrument sounds to my ears like the real thing. Naturally I spend time calculating operating points and all the rest of it, but in the end I need the transparency and timbre I get with DHTs. So that has pretty much decided my priorities for me.
Last edited:
It was no a critic to you Andy but rather a general approach. If you are happy like that no problem. It's the generalized comparison to 300B and 2A3's which I disagree with. It's not just calculation, it's about real amps which I listen to. I just don't promote an amp that doesn't sound good to me. The point is that there are too many options and issues. Getting to a unique best solution is impossible, IMHO. I am not that extremist. I prefere a DHT as power output device but that's it.
One thing where my approach is very different is that you care a lot about the sound of instruments whereas I care a lot more about instruments playing which means that dynamics, physical sensation with a real holographic 3d presentation just like when you are in a concert hall come first (not bizzare audiophile things like a stable soundstage with point-like or undefined contours sources etc...). This is not a constant and dependes hugely on listening room. If the listening room is not big enough you can just forget it. However the timbre and tone of an instrument are not costants as well, they are variables even in the real world direct event. They depend on the actual instrument, on ageing of the instrument (if acoustic) as well and the musician himself! Actually it is true even for an electric guitar.
One thing where my approach is very different is that you care a lot about the sound of instruments whereas I care a lot more about instruments playing which means that dynamics, physical sensation with a real holographic 3d presentation just like when you are in a concert hall come first (not bizzare audiophile things like a stable soundstage with point-like or undefined contours sources etc...). This is not a constant and dependes hugely on listening room. If the listening room is not big enough you can just forget it. However the timbre and tone of an instrument are not costants as well, they are variables even in the real world direct event. They depend on the actual instrument, on ageing of the instrument (if acoustic) as well and the musician himself! Actually it is true even for an electric guitar.