I hope you don't work in any professional studio. I would fire you immediately with no hesitation.
Headroom is the difference between clarity and mud.
Some people love rolling in mud (the DHT people for instance).
They love all that crappy 2nd harmonic rubbish.
If you believe in progress and clarity of purpose you won't go there.
🙄
That Thomas guy in Germany is a nice guy, but he muddles up lots of stuff,
he even admits it
and through his endless testing often ends up putting the prices of some useful and relatively nice NOS valves through the roof, thanks to his efforts to publicise stuff which should be kept under wraps for anti-hero people who don't want the Chinese to buy up all the stocks of x y or z for the sheer sake of it and mostly for speculation...or take the P...ss out of us stupid westerners then claim ignorance.. .
Last edited:
This is going a bit off topic..., but thats OK...
Personally, I like high transparency and correct tone.
I tried a lot of PP Amps and got to really great results, but found that the SE can give you a better, very natural presentations of tone, expecially for voices and wood instruments...the SE-amp as well does not have to sound muddy, just use fixed bias and avoid poor RC-loading...and use lots of current on the drivers to have good slew rates and stuff...and I try to compare at same frame parameters: So same type and number of tubes (one pair of 300B in PP vs. two in PSE, both runnig at similar bias, both IT-coupled, all powered by the same PSU etc)
BUT:
I can see that some people call PP Sound machines as they erase H2, even though when H2 belongs to the recording....Unfortunately, my current SE-impression is, that SE can do the same, it changes the original recording by adding much more H2 than in the recording.
So, my current path of research will go towards SE and try to find a combination of stuff (using a spectrum analyser) that avoids adding unnecessary H2 while preserving that H2 which was in the recording...which is what I cant teach PP to give it back...
If that will be of success..we will see...and I love the transparency and speed of PP, so I am for sure not a die hard SE-guy at all.
Personally, I like high transparency and correct tone.
I tried a lot of PP Amps and got to really great results, but found that the SE can give you a better, very natural presentations of tone, expecially for voices and wood instruments...the SE-amp as well does not have to sound muddy, just use fixed bias and avoid poor RC-loading...and use lots of current on the drivers to have good slew rates and stuff...and I try to compare at same frame parameters: So same type and number of tubes (one pair of 300B in PP vs. two in PSE, both runnig at similar bias, both IT-coupled, all powered by the same PSU etc)
BUT:
I can see that some people call PP Sound machines as they erase H2, even though when H2 belongs to the recording....Unfortunately, my current SE-impression is, that SE can do the same, it changes the original recording by adding much more H2 than in the recording.
So, my current path of research will go towards SE and try to find a combination of stuff (using a spectrum analyser) that avoids adding unnecessary H2 while preserving that H2 which was in the recording...which is what I cant teach PP to give it back...
If that will be of success..we will see...and I love the transparency and speed of PP, so I am for sure not a die hard SE-guy at all.
Last edited:
Sorry I have heard this hypothesis 100s of times.
It's utter nonsense.
PP amps cannot remove information from the original recording, they only remove the distortion inherent in SE amps (loads of dirty old H2).
A push pull amp IS a SE amp,(esp if it's operating in the class A region) it's just there's 2 of them acting in counter phase.
It's in the transformer that the even order DISTORTION gets cancelled out, nothing whatsoever to do with the original signal which it restores pretty well.
The paraphrase phase splitter does a similar trick, as well as removing microphony and hum yet, we don't hear rumours about it removing bits of signal.
(One demo I did was to prove valves don't need to be microphonic as one hi end freak was claiming, if the 2 halves of the dual triode are in anti phase...)
If you don't believe me, just do some IMD distortion tests on a good PPP pentode design and watch what comes out the other end.
The IMD signature on the SE will be almost invariably a whole mile worse, no matter what some people claim.
It must also be said a TOTEM POLE SRPP when operating correctly into a correct load is ALSO a P-P amp.
That's exactly its mode of operation dumping AC current into a load, yet we don't here zillions of people coming out with the same nonsense about the SRPP.
It turns out to be rather popular.
The cascode amp even works very much like a pentode, with the top grid being the screen grid, and the bottom one the control grid, and is so good it's commonly a circuit of choice for RF amps.
+
There were 1000s of very good amplifiers made with EF86 as the input stage, which horror of horrors runs a pentode as a class A amplifier for the i/p stage and part of the phase splitter.
Again we don't find stacks of horror stories recited about the Radford ST25 which is very well designed and people tend to think sounds rather good.
I really don't get this SET/DHT stuff, they just sound dirty and muddy.
How come the vast majority of studio mixers like Neve would never dream of using anything but solid state OPAMPS? cos they need to monitor stuff cleanly?
Hey presto, being as they are fully balanced the vast majority of studio Opamps are push pull by definition because it cancels out all that undesirable parasitic stuff like hum and noise.
Try one compared with a top grade of headphones some day or even dare I say in a studio against a powered Genelec.
You know one of those things with sand in it, and input via a SPDIF digital signal 😉
Last edited:
People tend to think PP beam tetrode amps sound harsh because of the higher levels of odd harmonic distortion (making them sound more like those things made with sand).
This is not a fatality and there's lots of ways to get around this, one of which is using modern valves with those really FAT high current cathodes/lower impedances.
Funnily enough, people thought early colour TVs have horrible quality valves, despite the fact a vertical scan design is basically a big class A amp, and has to be dead reliable under nice big overloads. (never mind the Horizontal stuff).
The ability to withstand giant overloads is part and parcel of the whole headroom argument, which is why it's possible to fry a 100V line transistor amp to smithereens with some nasty overloads, ...
where our dear old friends with nothing inside will carry on eating up on some numpty's bad circuitry for ever and a day without exploding.
You can get big anodes to glow red for hours on end without really complaining, they will even swallow a whopping EMP from a nuclear weapon or a CME from the sun, without it killing anything except the mains cable leading into the building, and the grid outside supplying the PSU.
One thing for sure, the next "carrington event" that comes along (we missed one giant one on July 23, 2012,), my good old valve amps will be still running fine, (on emergency backup power) while the entire GPS and mobile phone system and electricity grid goes dead.
This is not a fatality and there's lots of ways to get around this, one of which is using modern valves with those really FAT high current cathodes/lower impedances.
Funnily enough, people thought early colour TVs have horrible quality valves, despite the fact a vertical scan design is basically a big class A amp, and has to be dead reliable under nice big overloads. (never mind the Horizontal stuff).
The ability to withstand giant overloads is part and parcel of the whole headroom argument, which is why it's possible to fry a 100V line transistor amp to smithereens with some nasty overloads, ...
where our dear old friends with nothing inside will carry on eating up on some numpty's bad circuitry for ever and a day without exploding.
You can get big anodes to glow red for hours on end without really complaining, they will even swallow a whopping EMP from a nuclear weapon or a CME from the sun, without it killing anything except the mains cable leading into the building, and the grid outside supplying the PSU.
One thing for sure, the next "carrington event" that comes along (we missed one giant one on July 23, 2012,), my good old valve amps will be still running fine, (on emergency backup power) while the entire GPS and mobile phone system and electricity grid goes dead.
Last edited:
Ok, I still was wondering why Thomas would design a driver which can output the double voltage (given you have digital 0db as input from your DAC)...
Especially since I started to do some spice simulation which shows that if the power stage needs 100Vpp for full power (read we have a bias of -50V and after that the Power tube goes into A2) and we drive it with 200Vpp (without special circuit for class A2): We have a heavily distorted output.
Which would be the opposite of what the whole discussion is about.
Now, I just build a GM-70 Amplifier which is running with -88V and 950V. I was really a bit concerned as my driver stage can present 240Vpp when my DAC is playing a 0DB sinus-testsignal from flac. but the music I got acutally did not sound really distorted at all. Why is that ?
I went with a recording analysis program into my flac files and analyzed the maximum levels which are in the loudest recordings I got. What I found:
- The Sinus-Testsignal of 0db has 0 db 🙂
- The loudest music recordings have their absolute peaks max at -9db, typically -12DB for the loudest in "normal loud" recordings
- Normal loud music plays around -24db
Well, now I believe stuff comes together:
By coincidence, my Gm-70 amplifier has already a bit of Thomas' design principle build in: A far too loud theoretical voltage swing for the 0DB-Lab-Sinus. 240Vpp should create severe distortion at an output stage with -88V.
But it does not. Because in reality I never present more than -85Vpp (-9db) input signal to the Power stage as there are simply no recordings (and I chose the loudest from hundreds of CDs) going louder than 9db. not even transients as I did a spekrum analysis.
Which in return means that I could (for my 240Vpp@0DB-Input/driver-stage) even design a power stage with only -45V bias and it would still be clean and not distorted...
OR...
to my surprise it looks like for real life recordings: I have with my -88V bias even 6db (factor 2) headroom build in already. So, instead of feeling worried and try to get the negative bias for the GM-70 high to -125V, Ishould feel pretty relaxed.
Any flaws in that thinking ?
(and therefore as well Thomas design philosophy makes perfect sense as he simply knows that nobody has digital 0DB recordings available, so you would waste power...which is not a good thing with a 300B SE amp I guess)
Last edited:
It all makes sense here. I just like the idea never to be short of headroom. I also think that a beefier driver is able to react faster on these transients and therefore gives more microdynamics. This will be difficult to measure as it is subjective. Frequency range and FF plots can give an idea but somehow all this subjective perception is something else all together...
Yes, this is a different dimension I believe we have when you say beefier...the old vt52.com had a nice formular spreadsheet which calculated for correct transient reproduction a driver should be run with 10mA at least for a 300B as an example I believe...a function of input capacitances and mu. But typically 10 to 20mA is what a driver should have...I am running 25mA with a 801A.
From my practical experience, I am building nowadays my driver with IT as I have found a very good winder (50Ae) and the IT acts - additionally to the current I send through the driver tube - as a kind of energy storage for transients...I found a 6sn7 always a good tube, but a bit boring when driving a El34, but when you give it an IT it is a different animal..much better dynamics.
From my practical experience, I am building nowadays my driver with IT as I have found a very good winder (50Ae) and the IT acts - additionally to the current I send through the driver tube - as a kind of energy storage for transients...I found a 6sn7 always a good tube, but a bit boring when driving a El34, but when you give it an IT it is a different animal..much better dynamics.
I use cathode followers to drive my power tubes (any). But it is a modified follower as the standard followers are too slow for me (again subjective).
Attachments
Last edited:
The driver that can deliver 200V pp has roughly half distortions at 100 V pp , besides headroom .
Who would design things that barely meet the specs ? This is a very simple concept , for tubes , transistors , mechanics , everything .
Last edited:
@ Lampie519: So you mean an Allen Wright style like a tube /ccs above and below the actual cathode follower tube ?
I think we have to be clear what we mean by "...can deliver..." can deliver under which circumstances / input level signal for example. That B+ needs to be high enough or negative Bias voltage low enough etc...that is all a given. But that the gain actually can deliver double the Vpp when using digital 0db as the baseline input level...that is quiet a difference and quiet some consequences as shown above.
That is why I gave a practical example and real world numbers to make the point clear.
Usually i do not share any of my schematics, so please send me a PM.
The power supply i can share to give an idea...
My schematics all are incomplete as things that are obvious are sometimes not drawn or omited. So connections are missing on paper but not in the amp.
The power supply i can share to give an idea...
My schematics all are incomplete as things that are obvious are sometimes not drawn or omited. So connections are missing on paper but not in the amp.
Attachments
![IMG_9840[1].jpg](/community/data/attachments/855/855030-549b1933c864db4350e93a010b4497b7.jpg?hash=VJsZM8hk20)
Last edited: