Konnichiwa,
A simple look at the triode wired 6550 Anode Curves Vs 300B/2A3 suffices, it is so blindingly obvious as to be well beyond any argument or discussion.
Of course, many people like to listen to a lot of distortion, so having a more linear valve does not per se imply better percieved sound.
Sayonara
SY said:
A simple look at the triode wired 6550 Anode Curves Vs 300B/2A3 suffices, it is so blindingly obvious as to be well beyond any argument or discussion.
Of course, many people like to listen to a lot of distortion, so having a more linear valve does not per se imply better percieved sound.
Sayonara
Konnichiwa,
Hmmm. If you connect a pair of identical real triodes in Class A1 PSE or Class A1 PP you will find the following:
1) The maximum output power for the same audibility of distortion (generally dominated in this case by 3rd Harmonics - so say < 1% 3rd HD) will be around the same.
2) At lower power levels (around 1/2 of Po-nom) the PSE circuit will have less 3rd harmonics and thus less audible distortion, absolute values of THD being higher (due top 2nd harmonics) nonwithstanding.
Thusly to classify SE Operation as "highest distortion" is highly misleading and to me at least indicates a fundamental failing in relating measurements to audibility.
Sayonara
SY said:
Hmmm. If you connect a pair of identical real triodes in Class A1 PSE or Class A1 PP you will find the following:
1) The maximum output power for the same audibility of distortion (generally dominated in this case by 3rd Harmonics - so say < 1% 3rd HD) will be around the same.
2) At lower power levels (around 1/2 of Po-nom) the PSE circuit will have less 3rd harmonics and thus less audible distortion, absolute values of THD being higher (due top 2nd harmonics) nonwithstanding.
Thusly to classify SE Operation as "highest distortion" is highly misleading and to me at least indicates a fundamental failing in relating measurements to audibility.
Sayonara
A pair of triode strapped EL34 in pp, self-bias, unbypassed cathode resistor, will give 14 watts out with 0.4% THD, dominated by third. And at a fraction of the cost of a single 300 and without the complications inherent in running that filamentary cathode. That's rather better than the 300 can do in SE- the best I could find was about 0.5% third and let's not even talk about the second, at only 9 watts out.
Konnichiwa,
In other words, 0.4% 3rd Harmonics.
But you are not making a fair comparison.
First, if I consider that you likely run the EL34 near dissipation limit, I need to consider a pair of 300B's running at 25W Anode dissipation. Nearest common operating I notice is 350V/60mA, as it so happens one of my favourites.
If we operate this in PSE with a 2K Load we get nominally 14W in PSE with 0.63% 3rd harmonics. In reality we would have to make sure the numbers are truely comparable.
BUT, the much more interesting happenings are in the "first watt" at any extent.
I have build quite a few Amplifiers SE, PSE, 300B, 2A3, 45, 211, 845, EL34, 6550, KT88.
If you only consider quantity (watt @ X% THD) per Dollar you are better of with traditional Pentode or UL Tetrodes/Pentodes in class AB Push-Pull. If quality is a greater concern than quantity things become more complex. In these the lesser uniformity of the gridline spacing and the greater curvature of them conspire to bringing the IDH Valves to a significant disadvantage.
A better way in some ways of looking at this is to plot the constance of the Mu (gain) for a given anode load line against a constant Vg/Va ratio. DHT's have much straighter lines (one reference would be Reiner Zur Linde "Roehrenverstaerker) than IDHT's.
Sayonara
SY said:
In other words, 0.4% 3rd Harmonics.
SY said:
But you are not making a fair comparison.
First, if I consider that you likely run the EL34 near dissipation limit, I need to consider a pair of 300B's running at 25W Anode dissipation. Nearest common operating I notice is 350V/60mA, as it so happens one of my favourites.
If we operate this in PSE with a 2K Load we get nominally 14W in PSE with 0.63% 3rd harmonics. In reality we would have to make sure the numbers are truely comparable.
BUT, the much more interesting happenings are in the "first watt" at any extent.
I have build quite a few Amplifiers SE, PSE, 300B, 2A3, 45, 211, 845, EL34, 6550, KT88.
If you only consider quantity (watt @ X% THD) per Dollar you are better of with traditional Pentode or UL Tetrodes/Pentodes in class AB Push-Pull. If quality is a greater concern than quantity things become more complex. In these the lesser uniformity of the gridline spacing and the greater curvature of them conspire to bringing the IDH Valves to a significant disadvantage.
A better way in some ways of looking at this is to plot the constance of the Mu (gain) for a given anode load line against a constant Vg/Va ratio. DHT's have much straighter lines (one reference would be Reiner Zur Linde "Roehrenverstaerker) than IDHT's.
Sayonara
http://www.triodeel.com/we86a.jpg
You get the prize lollypop, Bas. Push-pull indeed.
D'ya think maybe the WE engineers who designed that tube might have a good clue as to the best way to use it?
You get the prize lollypop, Bas. Push-pull indeed.
D'ya think maybe the WE engineers who designed that tube might have a good clue as to the best way to use it?
Konnichiwa,
First, THD is definitly NOT a figure of any merit.
Secondly, the original enquiry seemed to be more concerend about quality.
You can find both SE Class A1 and PP Class A1 and PP Class AB1.
You can also find in the same time Amplifiers which where in effect boosters and used the line output from one of the smaller SE Poweramps as input to drive a big'n'bad pair of PP 212's if things got real big.
Which was chosen depended on the power needed for the cinema size and speaker system.
At the same time you also find Germany SE, DH Tetrode Amplifers as Studio Monitor Amplifiers driving Eckmiller Coaxial Studio Monitors, in stereo for the first High Fidelity, stereo Music recordings in 1943.
Any more questions on History?
Sayonara
SY said:
First, THD is definitly NOT a figure of any merit.
Secondly, the original enquiry seemed to be more concerend about quality.
SY said:
You can find both SE Class A1 and PP Class A1 and PP Class AB1.
You can also find in the same time Amplifiers which where in effect boosters and used the line output from one of the smaller SE Poweramps as input to drive a big'n'bad pair of PP 212's if things got real big.
Which was chosen depended on the power needed for the cinema size and speaker system.
At the same time you also find Germany SE, DH Tetrode Amplifers as Studio Monitor Amplifiers driving Eckmiller Coaxial Studio Monitors, in stereo for the first High Fidelity, stereo Music recordings in 1943.
Any more questions on History?
Sayonara
Konnichiwa,
Then one should perhaps also compare like with like....
When comparing SE & PP this is not the case.
BTW, I still reject THD as measure of linearity, as it does give no information about the nature of the nonlinearity and for many applications (not just Audio) the nature of the non-linearity is often more crucial than the absolute value.
Sayonara
SY said:
Then one should perhaps also compare like with like....
When comparing SE & PP this is not the case.
BTW, I still reject THD as measure of linearity, as it does give no information about the nature of the nonlinearity and for many applications (not just Audio) the nature of the non-linearity is often more crucial than the absolute value.
Sayonara
You brought that up.
As a measure of linearity, reject it all you want, but it is still correct. What I think you mean is relating distribution of harmonics to audibility thresholds, and there I would agree 100%. If a box of gain has a THD of 0.5%, all second and third, it will not be audible in a bypass test. If it's got the same THD, all 7th (say), then it very well might be audible, and rather unpleasantly so. But that's a different question, and a theoretical one since all the tubes under discussion are dominated by 2nd and 3rd in their spectra.
> If you want low distortion get a tube that was designed before WW2
6L6 predates WWII, and is deliberately non-linear. It is optimized for push-pull, where the curvature (mostly) cancels and the active side of the curve gives better power sensitivity (cheaper drivers).
There seems to be a natural non-linearity for commercial tubes. It isn't Childs Law or anything simple; has a lot to do with how accurately they can wind grids. But from there they bend to what the market wants.
6F6 is kinda middle of the road, which is why we don't see much of it.
6L6 is deliberately bent. This gives good push-pull results. But it also became VERY popular, and thus "cheap", which made it (and the 6V6 cut-down) a go-to tube for mass-market single-ended work. Dang the distortion, give the suckers what they pay for.
At the other end, 2A3 and especially 300B fight the natural curve for better linearity, which makes them lower-distortion in single-ended work.
What you perhaps mean is: tubes designed after Good Sound became possible (around 1928) but before NFB became the common way to tame a nasty amp (around 1935). Or after buyers began to (think they could) understand Specifications. If the amp has a big power number and a low THD number, it has to be good, right? Before the mid-30s, people had to listen to amps; after that they could just read the specs.
I'm really starting to doubt my decades of looking at THD numbers. I'm not sure that 5% or even 10% 2nd harmonic is at all audible. 6L6 single-ended is rated 15% where 2nd becomes overwhelmed by 3rd, and my last SE 6L6 was not dirty, while some P-P amps with 1/10th the THD number seem to grate.
> If you want more power for your money get 6550's...
And all simple distortion drops as you fall back from maximum power. If output-iron was not so costly, I'd say that several 6550s capable of 20W but working at 5W might be sweeter than a 2A3 rated 5W and working to its limit. Four 6550 cost about like one good 2A3. Of course when you factor the iron, the total cost may favor 2A3, and 2A3 is a handsome bottle (though four 6550 have more attitude).
> how many watts a two a side 2A3 or 300B could generate in ultralinear?
There is triode and there is pentode. Pentodes can also be worked as triodes. Power and damping are very different. Ultralinear just splits the difference between triode and pentode mode. It is only possible on pentodes. (Even if they are called tetrodes, any power tube needs something like G3 to force secondary electrons back to the plate, even if not implemented as a grid.)
6L6 predates WWII, and is deliberately non-linear. It is optimized for push-pull, where the curvature (mostly) cancels and the active side of the curve gives better power sensitivity (cheaper drivers).
There seems to be a natural non-linearity for commercial tubes. It isn't Childs Law or anything simple; has a lot to do with how accurately they can wind grids. But from there they bend to what the market wants.
6F6 is kinda middle of the road, which is why we don't see much of it.
6L6 is deliberately bent. This gives good push-pull results. But it also became VERY popular, and thus "cheap", which made it (and the 6V6 cut-down) a go-to tube for mass-market single-ended work. Dang the distortion, give the suckers what they pay for.
At the other end, 2A3 and especially 300B fight the natural curve for better linearity, which makes them lower-distortion in single-ended work.
What you perhaps mean is: tubes designed after Good Sound became possible (around 1928) but before NFB became the common way to tame a nasty amp (around 1935). Or after buyers began to (think they could) understand Specifications. If the amp has a big power number and a low THD number, it has to be good, right? Before the mid-30s, people had to listen to amps; after that they could just read the specs.
I'm really starting to doubt my decades of looking at THD numbers. I'm not sure that 5% or even 10% 2nd harmonic is at all audible. 6L6 single-ended is rated 15% where 2nd becomes overwhelmed by 3rd, and my last SE 6L6 was not dirty, while some P-P amps with 1/10th the THD number seem to grate.
> If you want more power for your money get 6550's...
And all simple distortion drops as you fall back from maximum power. If output-iron was not so costly, I'd say that several 6550s capable of 20W but working at 5W might be sweeter than a 2A3 rated 5W and working to its limit. Four 6550 cost about like one good 2A3. Of course when you factor the iron, the total cost may favor 2A3, and 2A3 is a handsome bottle (though four 6550 have more attitude).
> how many watts a two a side 2A3 or 300B could generate in ultralinear?
There is triode and there is pentode. Pentodes can also be worked as triodes. Power and damping are very different. Ultralinear just splits the difference between triode and pentode mode. It is only possible on pentodes. (Even if they are called tetrodes, any power tube needs something like G3 to force secondary electrons back to the plate, even if not implemented as a grid.)
That oddly echoes the results from SE Amp CAD (not set up to test this stuff with real electrons yet), which models 2nd and 3rd in the following manner. For a given B+ a SE 'real' triode's percentage distortion of 2nd and 3rd both rise with decreasing bias. For a triode connected pentode, starting at 100% rated anode dissipation 2nd rises but 3rd drops with decreasing bias, reaching a minimum somewhere in the 55%-75% anode dissipation range depending tube and B+. The triode models included with the software are almost universally lower in distortion under equivalent optimization than the triode/pentodes.
Here's the interesting part. The 3rd harmonic of the triode/pentode can be nulled really, really, low at the expense of high 2nd (5% and up), in some cases reaching below -90 dB relative to Pout and staying low across a very wide range of output power. Anyone know how accurately this models reality? If correct it implies triodes are generally better for SE and lightly biased triode connected pentodes doubled up can make a real kick-hiney PP.
Here's the interesting part. The 3rd harmonic of the triode/pentode can be nulled really, really, low at the expense of high 2nd (5% and up), in some cases reaching below -90 dB relative to Pout and staying low across a very wide range of output power. Anyone know how accurately this models reality? If correct it implies triodes are generally better for SE and lightly biased triode connected pentodes doubled up can make a real kick-hiney PP.
Konnichiwa,
Not terribly close, but yoes, triode connected pentodes have gobs of 2nd HD and comparably lower 3rd than a triode for the same 2nd HD. I mentioned this in my SE Amplifer Primer on my (old) website ( http:\\thunderstoneaudio.nav.to ) and suggested that this alloed for a good methode of cancelling an output triodes 2nd HD without unduely raising 3rd HD....
The conclusion that triode wired tetrodes/pentodes seem more suited to PP than to SE is IMHO not alltogether incorrect.
Sayonara
rdf said:
Not terribly close, but yoes, triode connected pentodes have gobs of 2nd HD and comparably lower 3rd than a triode for the same 2nd HD. I mentioned this in my SE Amplifer Primer on my (old) website ( http:\\thunderstoneaudio.nav.to ) and suggested that this alloed for a good methode of cancelling an output triodes 2nd HD without unduely raising 3rd HD....
The conclusion that triode wired tetrodes/pentodes seem more suited to PP than to SE is IMHO not alltogether incorrect.
Sayonara
jeff mai said:
Compare apples to apples, what about EL34 vs 300B when operated in the same way?
Why would one do that? Bad engineering practice.
That I can believe. The one thing about high second order (and to some extent, third) is that the error tends to make the sound richer and fatter. That can be quite attractive unless you're one of those annoying types like me who insist that the amplifier's output replicate the input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.