Hi everyone.
Does anyone know why the plates of these 12AX7 are so much different?
I have a third one around, which is closer to the one on the right (image attached).
According to analytical studies, the amplification factor of a triode increases as the distance between grid and plate increases. But the depth difference of the plates of these two models is just huge. I didn't find any specific information about this. Any comment is very much apprecited.
This is not exclusive of 12AX7s. Same thing happends with a pair of 12AU7.
Does anyone know why the plates of these 12AX7 are so much different?
I have a third one around, which is closer to the one on the right (image attached).
According to analytical studies, the amplification factor of a triode increases as the distance between grid and plate increases. But the depth difference of the plates of these two models is just huge. I didn't find any specific information about this. Any comment is very much apprecited.
This is not exclusive of 12AX7s. Same thing happends with a pair of 12AU7.
Attachments
Plate size depend of plate disipation, but as this tubes usually work well under their maximum ratings, perhaps decreasing plate size saves cost.
Compactron 6U11 has two units in a single 12 pin tube: a 12AX7 plus one unit from 12AU7, but wastes only 600mA of heater, so one unit is "free" saving 300mA.
Compactron 6U11 has two units in a single 12 pin tube: a 12AX7 plus one unit from 12AU7, but wastes only 600mA of heater, so one unit is "free" saving 300mA.
Attachments
Are they? I couldn't tell from the pictures. I do have many later manufacture 12ax7s that are rotated 90 degrees and don't have the welded plate seams in line with the grid support rods, so I thought this might be the same.
Little known fact…
Most of the mu of a valve depends on the inter-grid spiral spacing and the distance from the shell-of-the-cathode to each concentric grid. The anode can be near or 3× as far (on the scale of cathode-to-grid-(to-grid)*-to-anode measure), and won't change mu much at all.
However, anode-being-farther-away increases the maximum tension (i.e. anode-to-cathode voltage) that is allowable for a given design.
You can see this especially with high-tension TV flyback rectifier tubes.
The cathode is an itty-bitty thing, but the anode is large, and even has physically rounded edges, to suppress arcing … at 35+ kilovolts of blocking tension. If made smaller in radius, arcing would happen. The larger anode can also serves — probably obvious — to dissipate more waste-heat power.
Always good with flyback rectifiers.
So… so long as the mu and transconductance curves of the given 12AX7 (or any other family) design are adhere'd to, then whether the plates are larger, or smaller, doesn't much matter.
The exception would be if you're driving the AX7's at high tension, where the larger, further anodes would have a better chance of tolerating the tension without BLAM degeneration.
Just Saying,
GoatGuy ✓
Most of the mu of a valve depends on the inter-grid spiral spacing and the distance from the shell-of-the-cathode to each concentric grid. The anode can be near or 3× as far (on the scale of cathode-to-grid-(to-grid)*-to-anode measure), and won't change mu much at all.
However, anode-being-farther-away increases the maximum tension (i.e. anode-to-cathode voltage) that is allowable for a given design.
You can see this especially with high-tension TV flyback rectifier tubes.
The cathode is an itty-bitty thing, but the anode is large, and even has physically rounded edges, to suppress arcing … at 35+ kilovolts of blocking tension. If made smaller in radius, arcing would happen. The larger anode can also serves — probably obvious — to dissipate more waste-heat power.
Always good with flyback rectifiers.
So… so long as the mu and transconductance curves of the given 12AX7 (or any other family) design are adhere'd to, then whether the plates are larger, or smaller, doesn't much matter.
The exception would be if you're driving the AX7's at high tension, where the larger, further anodes would have a better chance of tolerating the tension without BLAM degeneration.
Just Saying,
GoatGuy ✓
It's an excellent question, and one that interests lots of people, but I don't think you'll get any useful answers on an Internet website, even this one. Valves were built from proprietary experience and largely ad hoc, and all before any ideas of computer modeling.
The idealized 3/2 power function comes from infinite parallel planes with equal charge everywhere, among other gross simplifications. If any big semiconductor manufacturer were to spend the Big Bux to adapt their proprietary design software to vacuum valves, amazing new valves could be designed. (not *built* - that's another Big Bux to spend just to duplicate ancient tech) But that can't possibly be paid for, so can't happen.
Our current knowledge of valve design isn't very different from archeology. The same will soon enough be true for our current semiconductor processes. You don't believe me now, but half a century will tell the tale.
All good fortune,
Chris
The idealized 3/2 power function comes from infinite parallel planes with equal charge everywhere, among other gross simplifications. If any big semiconductor manufacturer were to spend the Big Bux to adapt their proprietary design software to vacuum valves, amazing new valves could be designed. (not *built* - that's another Big Bux to spend just to duplicate ancient tech) But that can't possibly be paid for, so can't happen.
Our current knowledge of valve design isn't very different from archeology. The same will soon enough be true for our current semiconductor processes. You don't believe me now, but half a century will tell the tale.
All good fortune,
Chris
There are a few reference books about valve design, such as the well known RCA electron tube design compilation of articles: Vacuum Tube Manual: RCA 1962 Electron Tube Design : Free Download, Borrow, and Streaming : Internet Archive ; the paper at page 669, as example, gives many details about the internal structure of audio output power tubes (materials, spacings, shapes and so on). This is a lot more than archeology grade knowledge.
That would be because of their design as specs were already layed down for the specific type. Different factories arrived at different solutions, inter electrode capacitance would be the greatest obstacle. Telefunken novals were sold with a promissed life time, followed by Philips Special Quality and Mullard 10K. Same specs as ECC83 / 12AX7 but intended for high demanding circuits for scientific applications etc.
Did anybody notice there are lots of long plate 12AX7, and lots of short plate 12AX7?
I am pretty sure there are good sounding long plate models, and good sounding short plate models.
Put them in your amplifier.
Listen.
If they are used with lots of negative feedback around them, do not expect to hear much of a difference.
If you hear lots of difference, there is likely one of two causes:
No negative feedback, and one intrinsically has more distortion.
One is really bad.
I am pretty sure there are good sounding long plate models, and good sounding short plate models.
Put them in your amplifier.
Listen.
If they are used with lots of negative feedback around them, do not expect to hear much of a difference.
If you hear lots of difference, there is likely one of two causes:
No negative feedback, and one intrinsically has more distortion.
One is really bad.
Last edited:
Don't forget TESLA framegrid and JJ E83CC framegrid
All of the above has very short plates.
True. But I think that it is the distance between anode and grid what, among other factors, defines mu. It seems logical to think that long or short plates shouldn't compensate for narrow plates. And, as we all know, all 12AX7 have a quite high mu.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.