The indirectly heated power valves are not inferior in most applications: instrument amps, PA systems etc, as well as noncritical domestic audio.
In these applications, the cost of the equipment, and the amount of output power, are far more important than distortion. This kind of equipment will be used with the worst kind of speakers, and the DHT advantages would be wholly wasted.
For instance, the MARSHALL guitar amps of the 1960s get 60W from a pair of EL34s, and they get vibrated to extremes in use (then thrown in a van).
this kind of abuse would kill 300Bs in one evening, and you'd only get 15W or so from them. And yet the equipment price would have to be higher, to account for the expensive DHTs, and separate filament supplies.
DHTs have no advantages in this application - which accounted (and still does) for more tube usage than hi Fi.
In these applications, the cost of the equipment, and the amount of output power, are far more important than distortion. This kind of equipment will be used with the worst kind of speakers, and the DHT advantages would be wholly wasted.
For instance, the MARSHALL guitar amps of the 1960s get 60W from a pair of EL34s, and they get vibrated to extremes in use (then thrown in a van).
this kind of abuse would kill 300Bs in one evening, and you'd only get 15W or so from them. And yet the equipment price would have to be higher, to account for the expensive DHTs, and separate filament supplies.
DHTs have no advantages in this application - which accounted (and still does) for more tube usage than hi Fi.
The Keith Thrower explanation is about grid current caused by small grid-cathode spacing i.e. high gain. Not direct heating. I think the general story is that the higher the gain of a valve the harder it is to make it linear and well-behaved in other ways.
Gold-plating(?) was later used to reduce grid emission. Frame-grids (e.g. ECC88/6DJ8, some ECC83) put a very fine grid very close to the cathode to make a planar structure and can achieve high gain and high linearity. With even more modern manufacturing techniques it ought to be possible to do even better, but this is unlikely to happen.
Gold-plating(?) was later used to reduce grid emission. Frame-grids (e.g. ECC88/6DJ8, some ECC83) put a very fine grid very close to the cathode to make a planar structure and can achieve high gain and high linearity. With even more modern manufacturing techniques it ought to be possible to do even better, but this is unlikely to happen.
Now you are twisting what he is saying - and at the same time ignoring the practical impact of the two cathode designs.
For instance, I am looking now at a 6L6. The height of the cathode is 35mm. The grid to cathode spacing is very small.
Now, I pick up a JJ 300B. The filament has four peaks, (8x the height of the anode, which is 55mm): total run-length of the filament: > 430mm.
The grid-to-filament spacing is very obviously greater than the 6L6's.
Now, compare the performance.
The 300B needs 6W of filament power, and achieves 4mA/V approximately.
the 6L6 needs 6W of heater power, and gets about 5,5mA/V.
Therefore, you have only a slight increase in gm, but the concentration of heater power on the grid is hugely increased. The grid is much closer, and the same heater power is concentrated on a smaller grid area.
This can only mean one thing - much higher grid temperature for the 6L6.
All else being equal, the LONG filament of the 300B will give more gm than a short filament/cathode. Therefore, the DHT can relax the grid-filament gap, and get comparable gm, and lower temperature on the grid
He said Direct Heating, and I believe he meant it!
For instance, I am looking now at a 6L6. The height of the cathode is 35mm. The grid to cathode spacing is very small.
Now, I pick up a JJ 300B. The filament has four peaks, (8x the height of the anode, which is 55mm): total run-length of the filament: > 430mm.
The grid-to-filament spacing is very obviously greater than the 6L6's.
Now, compare the performance.
The 300B needs 6W of filament power, and achieves 4mA/V approximately.
the 6L6 needs 6W of heater power, and gets about 5,5mA/V.
Therefore, you have only a slight increase in gm, but the concentration of heater power on the grid is hugely increased. The grid is much closer, and the same heater power is concentrated on a smaller grid area.
This can only mean one thing - much higher grid temperature for the 6L6.
All else being equal, the LONG filament of the 300B will give more gm than a short filament/cathode. Therefore, the DHT can relax the grid-filament gap, and get comparable gm, and lower temperature on the grid
He said Direct Heating, and I believe he meant it!
Directly-heated cathodes are more efficient as electron emitters so 6W of heater power gets you more emission for them than 6W for an indirectly-heated cathode. This means the DH cathode is effectively bigger, so the grid can be further away and still give the same gm. Compare mu. I expect you will find that the DHT has lower mu. Mu measures geometry. gm is more a measure of size. For a triode mu matters more than gm, even for an output stage.
Yes, that's another way of confirming what I said.
The fact remains - the grid will be very much hotter in the (roughly) equivalent indirectly heated device.
The fact the gold grids are used in some 300Bs suggests the grid emission problem remains a risk, even in the better DHT device.
Entirely plausible, then, that for indirectly heated valves, the problem is a serious source of degradation.
The fact remains - the grid will be very much hotter in the (roughly) equivalent indirectly heated device.
The fact the gold grids are used in some 300Bs suggests the grid emission problem remains a risk, even in the better DHT device.
Entirely plausible, then, that for indirectly heated valves, the problem is a serious source of degradation.
What about the 26? Or the 71A (is it a 'power' DHT?), or the 31. To my eyes it seems that there's a notion going around that DHT preamp tubes make the overall sound more transparent than DHT power tubes, if you should pick only one. I haven't my mind up on this, and anyway prefer all DHT.
I find this talk of tube geometries fascinating. I noticed some time ago on my own that all planar tubes sound better than other kinds. Also all round with a spiral filament.
What about the cathode cooling effect? When an IDH cathode emits some electrons, it cools a bit, and takes some finite amount of time to regain full emission strenght. Filament supposedly doesn't do this. This of course should not be an issue in constant current situations.
Grid current due to temperature is a driver issue, not an output linearity issue. At most it would be a minor source of degradation. As, for example, an ECC83 can drive an EL34 quite happily it cannot be a big problem.Rod Coleman said:
Cathode cooling due to emission will be small and, as SY said, probably worse for DHTs. Their smaller thermal mass means that any such effect, if present, might affect LF. Unlikely though.
I am happy with the explanation, even though I largely had to arrive at it myself. When the facts change (or emerge) I change my mind. Some DHT power valves are linear because they approximate a planar structure. Some DHT smaller valves may be linear because they approximate a concentric structure.
Note that this does not mean that an SE DHT amp will necessarily be more 'transparent' than a more conventional circuit with properly-applied GNFB. That is a separate argument!
Depends on what the level of grid emission actually is - which you have no idea. We have no measurements. But unless it is low numbers of microamps, it is a degradation we can do without. It is likely to be noisy and highly variable with grid voltage.
Without definite information, the view of a professional designer is worth far more than an armchair designer - and Thrower has called it a serious effect.
Have you actually built a Directly Heated Triode amplifier, properly and carefully designed and implemented, with low-distortion driver, and no-compromise power (HT and filament)?
Those of us that have made this undertaking are in no doubt as to the value of the DHT, and listening to a properly implemented example renders the idea of ECC83 driving EL34 laughable in comparison.
Sad to say for lovers of guitar tubes in audiophile amps, I'd have to say that this isn't an exaggeration. I used to have a few of those amps, and built several similar ones. I loved them dearly at the time, of course, but that's many steps back in the journey.
I agree completely, an ECC83 driving an EL34 is a joke compared to a good 300B system.
It still depends on what kind of the speaker. If it requires more power I would rather use PP on pentodes with nested feedbacks than an electric stove of the industrial size.
That's actually discussed a bit in VA3 and VA4- for tubes that many of us use (e.g., EL84, 6L6GC), it's well under a microamp. You can get a relative sense of the number by looking at the specified maximum grid resistance value- interestingly, that doesn't seem to be lower with DHT power tubes.
For the second bit, it depends on what you mean by "no compromise." I didn't buy the DHT heater supplies you sell (you probably didn't exist then), but we were using some pretty tight and solid regulators for heaters, B+, and bias. Our preference ended up being AC for the heaters to avoid the systematic effect of uneven emission from different parts of the filament. We made perhaps a half a dozen different amps, then went on to better things. Our experience paralleled Pete Millett's description of his own no-compromise 300B amp (AudioXpress 2009).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.