I've been thinking about some basic vacuum tube physics - specifically what the most important determinants of mu are. I would imagine the strength of the vacuum would be one, although I don't know whether it would vary enough between individual tubes of a type made at the same plant in the same batch. I would also imagine that stupid things like the cross-section of the filament and the size/shape of the anode/cathode/grids might matter as well and not-so-stupid things like the amount of material deposited for the getter. Everything I have listed except the size/shape of the anode/cathode/grids would be shared by both sides of a double triode/tetrode/pentode etc - my question is, by how much will these parameters vary within a double tube, and, more generally, how much will mu vary between the two sides of a double tube? Enough to matter if I were using one for, say, a long-tailed pair or for two sides of a push-pull circuit? Would I just be using these to save space?
Current=Purveyance*(VGrid+VPlate/Mu)^1.5
Current appears to be controlled at some virtual point of balance
between the influences of grid and plate. In the proportion Mu.
Vacuum definately makes a tube work better, but has nothing to
do with Mu. Flame Triodes have been demonstrated that require
no vacuum at all.
A real triode is many triodelets in parallel. That can be modeled
more truthfully by adding many (oh, say 9) copies of the above
equation, with Purveyance distributed across a 2:1 spread of Mu.
Not every path through a real triode presents equal Mu...
Current appears to be controlled at some virtual point of balance
between the influences of grid and plate. In the proportion Mu.
Vacuum definately makes a tube work better, but has nothing to
do with Mu. Flame Triodes have been demonstrated that require
no vacuum at all.
A real triode is many triodelets in parallel. That can be modeled
more truthfully by adding many (oh, say 9) copies of the above
equation, with Purveyance distributed across a 2:1 spread of Mu.
Not every path through a real triode presents equal Mu...
Last edited:
within 1tube should be minimal
http://www.diyaudio.com/forums/tube...alves-grid-voltage-vary-much.html#post3535886
Adding air changes permeability away from that of the vacuum, which will have an effect on the number of electrons boiled off by thermionic emission, which will sure as hell have an effect on the gain. Suppose the permeability in the tube were zero (as it would be if the tube were filled with an infinitely dense gas). Then the electrostatic potential would everywhere be infinite, and no electrons would ever leave the filament. We would thus have zero gain.
u= gm X rp and represents the max gain with an infinite plate load.
What determines this is basically the closer the control grid is to the cathode, the higher the gm. For rp, it varies in direct proportion to the control grid mesh. The measured u-Factor is actually an average taken over the entire plate surface. There will be variations over small sections, just as the hfe of a BJT is an average as well. In both cases, this can cause hot spots. One small part of the plate may start to show color as you approach red plate conditions. That's where the local u-Factor is greater for whatever reason: cathode oxide conditions, uneven electrostatic fields due to sub optimum designs (quite common: rectangular plates) variations in control grid pitch. In the same way, a BJT might burn out in one tiny spot on the collector even before the rest of the collector has a chance to get very hot.
You can also have cathode burn-outs caused by I2R losses in the oxide coating, though this isn't a problem until very high plate voltages are used.
As for the quality of the vacuum fill, VTs are quite forgiving in that regard, otherwise, the earliest VTs wouldn't work at all. As it is, the best vacuum ever attained in a lab still had 100E6 atoms and molecules per cm^3, and that's a lot of particles. It's been estimated that there are 1.0E9 ionization events per second taking place in a (not so) vacuum tube.
As for variations across types, it's no worse than for solid state devices. If you're doing an LTP, the balance between phases is best if you use an active tail load. In that case, AC balance is pretty much assured for all practical purposes. Differences in u-Factor between sections of a dual triode will show up as DC offset. That won't matter if you capacitor couple to the next stage. If you don't, then there are ways to null out DC offset.
For PP stages, the usual solution is either fixed bias with individual bias adjust pots so the DC currents of both sides can be set equal, or some other type of DC balance control, either manual or servo based.
A non-zero purveyance is required. Permeability is to do with inductors...
Neither changes the ratio of grid to plate field influence upon the cathode.
dense grid
-> hi gain
-> hi internal resistance
-> plate current less sensitive to plate U change (pentode like)
short distance between elements
-> low internal resistance
-> secondary emission risk
-> (higher steepness-more related to framegrid)
big heater/filament area
-> internal resistance
-> current capability
heater construction
->W-Th longevity, stability
->BaO SrO... peak current capability, ++eficiency
plate construction
-> common nickel
-> Tantal sheet, graphite block -runs at hi temp,TiO coating degass
grid construction
-> common Mo
-> Au plated -avoid grid emision
-> hi gain
-> hi internal resistance
-> plate current less sensitive to plate U change (pentode like)
short distance between elements
-> low internal resistance
-> secondary emission risk
-> (higher steepness-more related to framegrid)
big heater/filament area
-> internal resistance
-> current capability
heater construction
->W-Th longevity, stability
->BaO SrO... peak current capability, ++eficiency
plate construction
-> common nickel
-> Tantal sheet, graphite block -runs at hi temp,TiO coating degass
grid construction
-> common Mo
-> Au plated -avoid grid emision
Mu does vary somewhat with operating point. Because there are many paths of
different Mu through the same Triode. High Mu paths cut off first, leaving only
the lower Mu paths. Scrunched to the left at the bottom of the typical graph...
Flat triodes, like 300B, will express less variation in paths, less spread of Mu.
The resulting curves are nearly parallel.
Round triodes if they might be off-center, oval or other oddly shaped plates,
posts blocking some paths, these things present many paths of different Mu.
Resulting curves scrunch to the left at the bottom. For these, Mu does change
somewhat with operating point.
if we used nothing but the simplest triode equation with no other modifiers,
2:1 spread of Mu is required internally to produce realistic curves for 12AT7.
This is not to say the behavior experienced externally shows such a spread,
only expressed as the typical scrunch toward low Mu at the bottom...
different Mu through the same Triode. High Mu paths cut off first, leaving only
the lower Mu paths. Scrunched to the left at the bottom of the typical graph...
Flat triodes, like 300B, will express less variation in paths, less spread of Mu.
The resulting curves are nearly parallel.
Round triodes if they might be off-center, oval or other oddly shaped plates,
posts blocking some paths, these things present many paths of different Mu.
Resulting curves scrunch to the left at the bottom. For these, Mu does change
somewhat with operating point.
if we used nothing but the simplest triode equation with no other modifiers,
2:1 spread of Mu is required internally to produce realistic curves for 12AT7.
This is not to say the behavior experienced externally shows such a spread,
only expressed as the typical scrunch toward low Mu at the bottom...
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.