Bypassing with multiple types of capacitors it not a good practice for me as well. It just doesn't sound good enough. In the worst case I just prefer a simple electrolytic! Often I can see that the chioce of bypass capacitor value is made with the wrong assumptions (i.e. not considering the actual contribution of the device but just the value of the catode resistor). However this is not the only choice. One can use in most cases 250V Solen/SCR MPK axial capacitor, for istance. They go up to 220 uF! A bit expensive but cost is subjective so I don't tend to include it in pros and cons. It's a passion, a hobby at the end of the day....
I should have been more clear - I tried all kinds of caps both singly and in combinations. I didn't find multiple caps offered any solution - I didn't like any version of a cathode bypass cap.
Yes - you're right, dynamics and soundstage are low priorities for me. Timbre is probably the most important thing. I just have all the sounds of acoustic instruments imprinted in my head after years of hearing them live right next to me on gigs.
But however much energy I put into building tube amps, my brother has a better sound system with Apogee Caliper Signature full-range ribbon speakers. And that's even with Krell amps (Lord preserve us....), but he does have a very good Schiit Gungnir DAC. If I had the room I'd use Apogees. There's nothing as good as panel speakers. I'm just doing the best I can with conventional cone units because of the small footprint of thin column enclosures.
I am not sure about this. The smaller the speaker can be for a target performance the better! The "shoes box" speaker as a general idea is the best invention in home audio of the last 50 years, IMHO. The rear emission of the speakers must be controlled to not interfere with the direct radiation as much as possible. One can't just use it equalizing to a target curve. It doesn't work because it creates masking and thus reduces dynamics, information despite an eventually higher SPL....
I tend to agree with Andy re DHT sound despite I took a slightly different route of experimentation. I found the mu-follower gyrator load in a DHT to sound really nice and provides a fantastic bass. It's very cheap in comparison to a very high quality IT. What is best, it can be used with DC-coupling to avoid a coupling cap and provide sufficient current to drive whatever output stage.
My latest implementation involves a crazy idea of using the 46 in filament bias. Why wasting such amount of power in filament bias? Simply because I love the sound of the 46 as much as 45. By using a step-up transformer and with the gyrator load which sets the anode voltage, helps me not to suffer from the miller capacitance and I can get the 46 to be the only stage to drive a 814 which is as demanding or more like the 300B. I simply love the sound and can't perceive any loss from the input transformer (LL7903). In fact this driver performs much better than many of the ones tried before. I can get less than THD 0.08% at 200Vpp with a good quality 46 valve:
http://www.bartola.co.uk/valves/2013/11/23/814-se-a2-amplifier/
I personally like the sound of 4P1L PSE. Given its low grid voltage swing requirements a nice quartet could be used, and still much cheaper than a 300B.
Having said that, I'm looking to build a 300B system for a friend with a 4P1L as driver in filament bias, input step up transformer as LL7903 and IT with a slight step up (LL2746/30mA) which provides additional 1:2. I will probably test other driver configurations but will be keen to listen to this configuration as well.
Back to the original topic, 4P1L is a great sounding valve and if implemented correctly a fantastic one for a line stage.
Ale
I think the discussion of total topology of the amplification is useful - the "line stage" can be moved into the amplifier in many cases where there isn't a long interconnect, and this opens up all sorts of possibilities. I'm keeping all my 10Y and 46 tubes just for the sort of projects Ale is talking about. They're both among the best sounding tubes ever made, to my ears. What's the mu of the 814, Ale? I remember we discussed this tube quite a bit. Is your amp ready yet?
Hi Andy,
The 814 is a superb valve. Paul was right, I run it in triode-mode and the screen voltage limitations are not affecting the valve in this mode. I changed it from 4-65a to 814 back in November and am really happy with the sound of it. You should come over to listen to it at some point.
The mu is about 8.3. You can read about the amp and look at the pictures in the previous link I posted. Here are the curves:
I rarely play it in A2 mode, but it does have an impressive dynamic response though. Bass is tight as ever had before in a SE amp.
I hope to get back to my new version of the 4P1L line stage shortly so will post results then
Ale
The 814 is a superb valve. Paul was right, I run it in triode-mode and the screen voltage limitations are not affecting the valve in this mode. I changed it from 4-65a to 814 back in November and am really happy with the sound of it. You should come over to listen to it at some point.
The mu is about 8.3. You can read about the amp and look at the pictures in the previous link I posted. Here are the curves:
I rarely play it in A2 mode, but it does have an impressive dynamic response though. Bass is tight as ever had before in a SE amp.
I hope to get back to my new version of the 4P1L line stage shortly so will post results then
Ale
In my opinion the best IT for driving the 300B is the Tango NC20. Now it is no more available but that should be the target. With the NC20 or similar if one wants 2x gain from the driver the 6H30 does it nicely. For Class A1 operation it can deliver the full voltage swing with low distortion like a DHT....
I know, but couldn't help myself and went and run the tests:
All test details here: 4P1L – LL2746 driver test | Bartola Valves
Not bad at all, but not as good as other drivers like triode-strapped pentodes. However, this is a DHT in filament bias so I'd be tempted to use it as I love the sound of them. I will have to do final build and let my ears to judge it!
Ale
You can try I Ale but that Lundahl doesn't look good as other models. I cannot see an advantage in comparison to a 6H30 with 1:1 interstage. Your Zout is also 4x higher...so no chance to get even near the onset of grid current. Maybe you will not need 200 V p-p.
I would have done it differently as don't like Zobel networks at the driver stage. In a low signal area is a different story. An alternative to the NC20 could be the Sowter 9525 which is ideal for source impedances between 1K and 3K. You get minimum 75KHz bandwidth. It has 50H for 50 mA primary DC and I guess it could be specified for 20-25 mA for even higher inductance.
I would have done it differently as don't like Zobel networks at the driver stage. In a low signal area is a different story. An alternative to the NC20 could be the Sowter 9525 which is ideal for source impedances between 1K and 3K. You get minimum 75KHz bandwidth. It has 50H for 50 mA primary DC and I guess it could be specified for 20-25 mA for even higher inductance.
Hi there -
It's a DHT thing. Ale and I belong to that band of people who prefer using DHTs to anything else.
I've used 6N30 and a great deal of indirectly heated tubes in the past, but no more. I can hear a difference with DHTs and enough of a difference to know that's what I want to use. Made that decision about 3 years ago now. Ale's pretty much the same. It's purely a question of sound. You design around the fact you use DHTs. But it does give you the charms of filament bias!
It's a DHT thing. Ale and I belong to that band of people who prefer using DHTs to anything else.
I've used 6N30 and a great deal of indirectly heated tubes in the past, but no more. I can hear a difference with DHTs and enough of a difference to know that's what I want to use. Made that decision about 3 years ago now. Ale's pretty much the same. It's purely a question of sound. You design around the fact you use DHTs. But it does give you the charms of filament bias!
I don't belong to any sect. My point however was a bit different. You just can't do a 300B amp in the same manner as your 4P1L PSE. If you prefer, the 801A with the Sowter 9525 is far superior. You need 3 stages.
I don't like it as a general thing. Filament bias is convenient where there is not too much waste of power. It is not better than fixed bias that is just more complicated but at least doesn't waste that much. One of the advantages of the 4P1L is its low filament consumption and with filament bias it's just wasted. You can only do it without caring too much because you have a 3W amplifier......
Yes - of course my system is happy with around 5 watts output. But I can add more 4P1L - next amp will have 3 in parallel to continue the experiment. Bear in mind I came from a 3-stage amp with 300b outputs and I prefer 2 stages of 4P1L. I'm quite happy with the extra supply needs of filament bias - it's easy to implement and also sounds better than all the fixed bias experiments I tried. No doubt I could have continued to optimise fixed bias, but since I already have a solution with no coupling caps, no cathode bypasses and 2 stages, what's the point?
I don't believe it is able to do 5W with reasonable distortion even if you run them hot. If you run them hot I don't believe they can sound good. A 4P1L is not a 845.....
You can only get that power in class A2 which is not possible with filament bias and 1:1 IT coupling with another 4P1L unless you like a lot of distortion. Otherwise I can tell I can get 6W from a single 2A3 without breaking its 15W plate dissipation limit?
With the primary load you are using (more than 6K per valve) the grid current increases rapidly and the driver will not be able to cope with it even if this is a 4P1L running at 20-25 mA.
I don't care! The discussion was about a driver for the 300B. Using a 1:2 IT for driving the 300B might well result in a better performance of the 4P1L PSE. But this doesn't mean that the latter is better in general. Can you bear this in mind?
A 300B performing at its best is not something for everyone. It is not easy and this is the main reason why I don't like the 300B for a SE amplifier. I have made some of them, from simple ones to complicated ones, and in my opinion it is too much effort for what you get anyway. The 300B becomes convenient in PSE or PP when the target is in the region of 20-30W. At that point its convenience is obvious if one looks at the other options for that sort of power.
Again. A 2 stage amp like that is not possible with the 300B. Or it might be possible but one has to accept limitations. Probably this is the main reason for some myths about 300B and its "tuby" sound......
Frankly I don't believe you are in the position to state in general what sounds better. So discussions should only be limited to technical arguments.
I repeat, you can only use filament bias because you have a small amp and the extra power and heat might not be relevant. It is your choice.
Regarding the 300B, the power wasted for using filament bias for the 4P1L driver could light up a 300B filament! That is quite relevant to me.
The 4P1L can be a convenient driver for the 300B because of its low filament consuption. If one has to waste it because of filament bias then there are better options. This is the point.
In the context of set amps 'wasted power' i think is a spurious argument!
Filiment bias is easy to implement. Its just a higher secondary voltage tx and a resister. Nothing to keep you up at night!
I think we all accept sound quality to be subjective and dependant on lots of factors other than the amp.
Your milage may vary
Filiment bias is easy to implement. Its just a higher secondary voltage tx and a resister. Nothing to keep you up at night!
I think we all accept sound quality to be subjective and dependant on lots of factors other than the amp.
Your milage may vary
I agree with the previous poster that the "wasted power" argument for filament bias is not something a lot of us worry about so I'll put that aside. The thread in general is also not about driving a 300b - look at the first post - but about using the 4P1L as a "line stage". The thread is also uniquely about using a DHT so while it's always interesting to make comparisons e.g. with indirectly heated tubes, Ale started with the desire (shared by others on this thread) to use DHTs.
I've NEVER (have to use large caps here!) understood why so many people accept that using a DHT like 300b or 45 or 2a3 or 845 in the output is better on principle and in terms of sound, but can't understand that it is EQUALLY or MORE important to use a DHT for the input section for the same reasons of sound quality. Hence the whole point of this thread. So many people drive a 300b with a 12AX7 (for instance) and think they are proud owners of a "DHT amp". It isn't - it's a 12AX7 amp since that's where most of the gain comes from.
I wouldn't build an amp if "subjectively" I didn't like the sound - you can use the word "subjective" though I'd prefer a more nuanced analysis. So technical considerations have to be at the service of the overall sound, not visa versa. I don't think we need to continue this argument - we've all been round the houses on that one.
The reason I'm talking about a 2 stage amp on this thread, is that if you have two stages, the line stage IS the input stage of the amp. It's both the line stage and the driver, whether you situate the stage in the amp or out of it. If 2 stages can be made to work as well as 3 stages I see this as a worthwhile goal. That goal demands an output tube with a gain of around 10, partly depending on the sensitivity of your speakers.
Ale I believe uses a Fostex unit which is sensitive enough for a 2-stage solution and has recently been experimenting with the 814, mu of 8.3. Other options are 813, mu of 8, or 211 with a very useful mu of 12. These options would give you the bass and power that would take the amp up a level. Once you design around a higher gain output tube (PX25 would be nice if it were not rare and $$$, 4P1L is common and cheap) you start to get some more interesting options for the preceding stage.
There are some interesting options around A2 operation, as you mention, but unfortunately I have no experience of this, and I've never built high voltage amps so I do admit that there are some more ingenious solutions out there (also some solid state hybrid ones) which your technical ability can contribute to. But for the purpose of this thread, I think the 4P1L should stay in the mix.
Last edited:
It's your opinion. I don't like unecessary stoves....
That's what I am trying to say. Discussions should only be about the technical side. A sentence saying this is better because it sounds better to me is not a good argument exactly because can be anything....
I don't worry about filament bias I just use it "cum grano salis", as always. Very simple concept, I think.
My comment was about Ale's 300B driver stage and I think that there is no problem with short off topics if people have a little bit of mental flexibilty and can look a little bit beyond their nose....it becomes boring and useless when quoting is not used properly.
Ok then. Looks like you are the guru and you tell me what is better and what is not. Can you accept the fact that you system is flawed for me?
If you came to Rome with me I can show you what is my reference. I can bet you will have a serious panic attack after you listen to that....
Sorry, it is you that bring in this argument all the time. Your theory of superiority of one thing in comparison to another is all based on the fact that it sounds better to you. Where is the techincal side? I can't see it.
It is not possible with the 300B by using a DHT with limited gain at less than 10 and IT transformer. Or it is possible but it is not ideal and here is where myths about the sound of the valves come from..... That's my opinion.
The 814 has nothing to do with the 300B. The 814 only needs about half of the voltage swing to produce 9-10W. So in theory there is not need for IT step-up. However it has to be class A2 and 1:1 IT coupling with any DHT would not work that good. In fact he isn't using this solution rather there is a mosfet in signal path! Can you explain why a mosfet is good in this case and IDHT cannot in another case?
A step-up transformer at the input is another matter although it is not my favourite.
It is the IT the issue not the 4P1L. Especially if that IT is not the best one can get. Stepping up to get more gain is not better than a well made 1:1 IT with a good IDHT like the 6H30, which has the target gain, just because there is no DHT. It is just speculation... You might like or not but that's what I have experienced and what I think. The 6H30 at low and medium levels is as good as any DHT and at 200V p-p output its about 2% THD (in "ideal" condition, not too difficult to get though if A2 is not required) can be used to get overall less distortion because it is mostly 2nd harmonic. 2nd harmonic is well audible when it is too much, in the form of IMD products for sure, and people confuse it with lack of damping factor....... Low distiortion is always the first requirement and when properly done the spectrum will be fine.
Harmonic cancellation works well if done right and will work forever. I have read funny comments about the fact that will change during the valve life. It's a legend. The amplifier will behave as any other when valves become worn it will give a certain THD at lower and lower power until it dies. Drivers and voltage amplifiers outlast power devices by a good amount usually and certainly this is the case of the 6H30.
I hope I have been clear enough because it is becoming boring now.
Last edited:
I think we're all trying to be as clear as possible, it's just that we start from different points, have different priorities and are considering different solutions and - inevitably - compromises.
I don't think any of this is boring - far from it. We're considering some unorthodox and ingenious solutions. Nobody ever used a 4P1L as a line stage because of microphonics - we've moved forward to starved filaments, filament bias, DC regs that work in filament bias, using a 4P1L as a driver stage, active loads, constructing chassis solutions to minimise microphonics and so on. Even 4P1Ls as output tubes. All this since the thread was started in 2011. That's some very good progress! Hopefully we still have places to go.
Did you mean on this forum.
In Russia, many gathered similar to this scheme. The results are good. Lamps in pairs can not pick up enough food separate glowing.
Sorry for not knowing the language.
Last edited: