Thread split from: http://www.diyaudio.com/forums/tubes-valves/184558-just-why-dht-tubes-so-expensive-2.html#post2496446
____________________________________________________________________________
On further reflection, a dedicated triode version of the EL34 and KT88 might suffer from the "fall between two stools" problem in marketing. Not as good as a "real" direct-heated triode, and too weird for the guitarists. On other words, no love from either camp, even if it was technically a better solution than a triode-connected pentode. It would be a fairly expensive experiment to try - not just a few prototypes, but seeding them with opinion-makers in both camps.
Still, it would be worth trying, just to answer the question of the sonics of IDHT versus DHT triodes, as well as having a genuine single-grid triode that was rugged enough for professional duty. In terms of ruggedness, it least it wouldn't suffer from the problem of the screen grid overheating because the screen is close to the same potential as the plate, which is the usual problem of a triode-connected pentode with the screen connected to the plate through a 100-ohm current-limiting resistor, or even worse, a direct connection to the plate.
Same problem for the so-called "ultra-linear" connection - even though there is a UL tap on the primary of the output transformer, the screen is still at plate potential, which does it no favors in terms of dissipation. Both connections abuse the screen and can exceed ratings under dynamic conditions. No screen, no worries about screen overheating or arc-over, and no distortion of electron flow between grid and plate. My guess it would take more abuse than conventional EL34's and KT88's, at least in triode and ultralinear circuits. Plugged into a pure pentode circuit, the pentode bias would simply be ignored, automatically converting the circuit to triode.
If the EL34 Super and KT88 Super were to take off, it would make sense to omit the suppressor and screen pins on the base of the tube, as a subtle designator of the tube type. It would be interesting to audition them - a triode version of the KT88 would have about the same power-handling as a 300B, but would probably sound (and measure) different than either a triode-connected KT88 or a 300B. It might even be better than either! Heresy!
____________________________________________________________________________
On further reflection, a dedicated triode version of the EL34 and KT88 might suffer from the "fall between two stools" problem in marketing. Not as good as a "real" direct-heated triode, and too weird for the guitarists. On other words, no love from either camp, even if it was technically a better solution than a triode-connected pentode. It would be a fairly expensive experiment to try - not just a few prototypes, but seeding them with opinion-makers in both camps.
Still, it would be worth trying, just to answer the question of the sonics of IDHT versus DHT triodes, as well as having a genuine single-grid triode that was rugged enough for professional duty. In terms of ruggedness, it least it wouldn't suffer from the problem of the screen grid overheating because the screen is close to the same potential as the plate, which is the usual problem of a triode-connected pentode with the screen connected to the plate through a 100-ohm current-limiting resistor, or even worse, a direct connection to the plate.
Same problem for the so-called "ultra-linear" connection - even though there is a UL tap on the primary of the output transformer, the screen is still at plate potential, which does it no favors in terms of dissipation. Both connections abuse the screen and can exceed ratings under dynamic conditions. No screen, no worries about screen overheating or arc-over, and no distortion of electron flow between grid and plate. My guess it would take more abuse than conventional EL34's and KT88's, at least in triode and ultralinear circuits. Plugged into a pure pentode circuit, the pentode bias would simply be ignored, automatically converting the circuit to triode.
If the EL34 Super and KT88 Super were to take off, it would make sense to omit the suppressor and screen pins on the base of the tube, as a subtle designator of the tube type. It would be interesting to audition them - a triode version of the KT88 would have about the same power-handling as a 300B, but would probably sound (and measure) different than either a triode-connected KT88 or a 300B. It might even be better than either! Heresy!
Last edited by a moderator:
Lynn, minor correction, there are boutique guitar amps using DHTs!!
Not many, but they are out there...
As far as eliminating the screen and suppressor grids from and EL34 or a 6L6, it don't woik quite like dat. You'd have to redesign the grid and the spacing, afaik. But still it's not a bad idea. Sounds a bit like - ummmm, brain failure alert - the verison of the 6A3 that is indirectly heated. At least this is what the echos and shadows in my mind are telling me. Could be illusion or delusion.
_-_-bear
Anatoliy, the 4P1L is a DHT pentode?? Got curves for it strapped for triode? (too bad it's a pentode...)
Not many, but they are out there...
As far as eliminating the screen and suppressor grids from and EL34 or a 6L6, it don't woik quite like dat. You'd have to redesign the grid and the spacing, afaik. But still it's not a bad idea. Sounds a bit like - ummmm, brain failure alert - the verison of the 6A3 that is indirectly heated. At least this is what the echos and shadows in my mind are telling me. Could be illusion or delusion.
_-_-bear
Anatoliy, the 4P1L is a DHT pentode?? Got curves for it strapped for triode? (too bad it's a pentode...)
Last edited:
Too good it is an extremely linear triode! 😉
And too good it costs much less than worse DHT tubes! 😀
But what is not so good, it is hard to find proper sockets for them.
http://www.diyaudio.com/forums/tubes-valves/183329-one-more-4p1l-se.html
Good to hear they exist - a specialty item to be sure, but they boost demand for DHT's, which has to be a good thing.
Going back to the concept of a "Super" EL34 and KT88, they would of course have the grid spacing adjusted so that mu and Rp would be the same as the triode-connected versions of standard EL34's and KT88. That way they would be plug-compatible with existing EL34 and KT88 "triode-connected" circuits with no modification.
I suspect distortion in a true triode would be lower than the triode-connected equivalents, since the screen is at plate potential in a triode-connected pentode, and that will attract a portion of the electrons towards it, away from the plate, and altering the flow of the rest of the electrons from the grid to the plate. A small effect but probably measurable.
Last edited:
I could be wrong, but I wonder given the large difference in surface area and the fact that the screen and plate are effectively tied together in triode connection whether it matters at all that a few electrons are collected by the screen as opposed to the plate, differences in electron transit time would be pretty insignificant, and all of the current is still flowing in the plate circuit. The plate curves of a triode strapped EL34 seem to say it makes a reasonably good triode.
I am not sure very significant changes in geometry would be required if you removed the screen grid from the existing EL34 design. Mu in triode mode is determined by grid winding pitch, and distance from cathode, I suspect the distance to the plate is perhaps somewhat less significant than the other two.
Bear, Sylvania made a replacement version of the 6B4G in the 1980s by strapping the screen grid internally to the plate of a 6AV5.. It wasn't particularly great sounding compared to the original 6B4 which is an octal version of the 6A3..
There are lots of IDHTs out there, like the 6CK4, and 6AH4, and doubtless countless others I have not encountered yet.. (Haven't looked for them actually.)
I think the actual geometry of the filament, grid and plate relative to each other plays a larger role than whether or not the cathode is directly or indirectly heated. It appears that DHTs in general were designed with geometries that resulted in better linearity than their IDHT cousins, but I am not sure it pays to over generalize the point.
I've ordered some 4P1L based on Wavebourne's comments, this is a direct heated pentode which when triode connected seems to exhibit some of the best looking plate curves I have ever seen even amongst the better dhts. I think other design issues than whether or not a screen and suppressor are present in the envelope may play a bigger role. (I bet you take one of these apart and you find things like ideal heater, grid, and plate geometries, as well grids that are carefully aligned to each other, etc.)
When this is correct, it seems to make more sense to make an indirect heated 300B and 2A3, instead of a pure triode El34 and KT88.
Many of current production DHT's have problems with the filaments (durability; break).
An indirect heated 300B / 2A3 would also solve the "problem" of the filament supply, as simple AC supplies could be used 🙂
Pieter
I am not sure on this, but the ID heater tends to be a round affair... whereas the geometry of the typical triode tends to rely on a large surface area cathode in a rectangular shape.
Having said this - I think we should split this OFF to another THREAD?
Seems like it is its own topic that will get lost in the midst of this thread - lost for future reference and won't grab the interest of others who may or may not have significant input to the topic. Whaddya say?
_-_-bear
Having said this - I think we should split this OFF to another THREAD?
Seems like it is its own topic that will get lost in the midst of this thread - lost for future reference and won't grab the interest of others who may or may not have significant input to the topic. Whaddya say?
_-_-bear
The one thing I do wonder about (just now in fact) is the relative area of emission of a large dht filament as compared to an idht. The filament wires are spread out over a relatively larger area (even if the actual area doing the emission is no greater) - I wonder whether emission over this area is more consistent than is the case with most idhts? It may possibly just point back at geometry issues.. Something we can ponder..
Bear: It might be worth asking Pano to split the thread into a new topic about dht vs idht linearity and geometry issues.. I'll ask him now.
Last edited:
Interesting comment because the 9 pin high transconductance idht and idhp I prefer sonically all have rectangular cathodes. (5842, D3A/7721, 7788, etc.) This is also true incidentally of the JJ 12AX7A and 12AT7A which as SY will attest are among the best sounding of their types. This is also true of a certain variant of the Mazda/Cifte (French made) 12AU7A which I find far superior linearity-wise and sonically to the often mediocre types sharing this designation..
Geometry of a large indirect heated cathode would certainly be the challenge.
Many small tubes have in fact a rectangular cathode structure; look for instance at a 12B4, or the larger 6AS7.
Many small tubes have in fact a rectangular cathode structure; look for instance at a 12B4, or the larger 6AS7.
Large triodes intended for use as a series regulator pass element like the 6C33 and the 6336 also have rectangular cathodes, and at least in the case of the 6C33 can sound rather good in the right circuit. My first brush with the suspicion that perhaps global NFB was not always required in a power amplifier design was with my ill-fated 6C33 circlotron design that sounded surprisingly good to my ears without any global feedback applied. The design was not otherwise successful since I had intended it to be used closed loop, and it had severe reliability issues, but it did open my eyes (and ears) to the possibility that those quaint little SE dht amps might possibly sound better than I suspected. (All ancient history now as I hardly ever build anything else.. 😀 )
I vaguely remember some years ago a (British Firm I.I.R.C.) doing just this, (or advertising it anyway) an internally strapped, G2 less EL34.
I think in this case we're actually contemplating a version where the screen and suppressor grids are not installed during assembly, so no internal strapping would even be required.
Perhaps someone like Mike Matthews will stumble across this thread and make a few out of curiosity. (Probably not..🙁) It would be so simple to just omit said parts with no other changes and see what happens parametrically speaking.
The material cost may be a little higher in a DHT. The cathode in a DHT is an alloy
ot tungsten and about 3% the radioactive metal thorium. Thorium besides being kind
of expensive makes the tungsten thorium alloy verry brittle. I expect it's pretty hard to work with because it's so brittle. These 2 things may account for some of the higher cost for DHT tubes.
ot tungsten and about 3% the radioactive metal thorium. Thorium besides being kind
of expensive makes the tungsten thorium alloy verry brittle. I expect it's pretty hard to work with because it's so brittle. These 2 things may account for some of the higher cost for DHT tubes.
True I think, and the filament design probably requires even more skilled assembly workers.
Some early "modern" derivative production 300B dhts had serious problems with both initial filament misalignments and sagging over relatively short periods of use.
Pano, thanks for moving this part of the thread to a new topic. 😀
Now if the forum software just automatically subscribed prior participants to the new thread as well that would be cool...
Now if the forum software just automatically subscribed prior participants to the new thread as well that would be cool...
My guess is that to make a 'triode EL34' you just have to put the anode where g2 currently is. It won't be quite that simple, but that is where I would start.
With regard to cathode shape, it doesn't matter whether it is cylindrical or rectangular. The important thing is that cathode, grid and anode are all the same shape, with the same ratios (e.g. width equals 5 times thickness). This means that every part of the valve has the same characteristics. A minor complication is that electric fields can 'fill in' an internal corner to make it look rounded, so the cathode and grid should be rounded too.
In addition, the grid pitch should be much smaller than the cathode-grid spacing, which means either lowish mu or frame grid construction. Grid support rods should be avoided or hidden away from the electron stream.
With regard to cathode shape, it doesn't matter whether it is cylindrical or rectangular. The important thing is that cathode, grid and anode are all the same shape, with the same ratios (e.g. width equals 5 times thickness). This means that every part of the valve has the same characteristics. A minor complication is that electric fields can 'fill in' an internal corner to make it look rounded, so the cathode and grid should be rounded too.
In addition, the grid pitch should be much smaller than the cathode-grid spacing, which means either lowish mu or frame grid construction. Grid support rods should be avoided or hidden away from the electron stream.
WRONG . BAM .
thorium is only used where your filaments are HOT . and where your fillaments are HOT , you are talking transmitters . 🙄
to my knowledge virtually all transmitter tubes use thorum filaments ( correct naming is thorium-doted filaments .
presuming you wheren't all degrading transmitter tubes to hifi use . you are wrong .
now what is used ? well virtually all* newly produced DHT's have barium oxidefillaments . this coating is sprayed or pasted on .
( i have seen a photo of a Chinese worker using his fingers to apply this to the heaters ) witch means you get **** tubes . because the coating is polluted whit human grease .
anyway these days only the barium coating is used . while its not the best . there are others witch perform better .
about the oxide . these are formed when carbonates are broken down so the barium coating is applied in the form of barium- carbonate
okay here it comes . i have read numerous articles about this . and if any one is interested i would happily share my knowledge about this topic .
* edit . all means im talking audio tubes . not transmitter . these are still made using thorium coating ,.
Last edited:
essentially you would need to remove G2 and replace it with the plate. In triode mode, G2 determines the characteristics and the plate handles the dissipation.
If you do a set of curves using G2 as plate you see essentially the same triode operation as when the plate is connected to G2.
dave
- Status
- Not open for further replies.