|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.13.65.111
My understanding of the way tubes wear out is that the few billion molecules of gas left in tubes get positively ionized by the bombardment of electrons passing from cathode to plate. A few of these massive (relative to an electron, anyway) objects make it past the protective cloud of electrons protecting the cathode and strike the cathode. These impacts physically abrade off the oxides that emit electrons. After a while (determined by evacuation of the tube, quality of the electron cloud, voltage the plate and grid are operated at, and probably other factors) enough of the oxide has been removed so that the emission of the tube is meaningfully decreased. Residual gas is the bad guy.
I've never seen an analysis of the gases left in tubes. My hypothesis is that since as a good approximation hydrogen leaks through anything, that at least a disproportionate amount of the gas in tubes is hydrogen that has leaked in through seals and even straight through the metal parts such as the leads. Helium is right behind hydrogen in leakiness, so there cold be a little helium in there too.
So the idea, at least for signal tubes, is to surround the glass envelope with a second glass envelope leaving a small gap between the two. A nipple would be fitted to the outer envelope and the signal tubes in a device would be connected to a a common continuously running vacuum pump. This should keep down leakage into the tube of light gasses by limiting the amount of gas in contact with the inner envelope. The vacuum pump wouldn't have to be very good to have a significant effect on tube life, if my conjecture is correct. Quietness of the pump would be important, though.
Follow Ups:
Some primary reasons for premature tube failure is over-zealous design (running the tube near its maximum dissipation limits) and cathode stripping.Many designers, especialy nowadays, run tubes way beyond the CCS rating for dissipation of the respective components, be it plate, screen or grid dissipation. For instance, the absolute maximum plate voltage for a 12AX7 is around 330V, but it is very common to see voltages around 370v or even higher. These higher voltages are used to increase gain yet also increase signal to noise ratios as well, with the compromise of higher distortion. Usually NFB is employed to compensate for these distortions and then creating a problem of increased dissipation on grids as well. Higher plate voltage is also used to improve plate impedence matching with stock transformers currently available (of which most have fixed values). For the best distortion figures and long term reliability a plate voltage of just under 300v is best for this tube, 285v being ideal.
It is very common to see this "hot-rodding" of tubes, especially in Class A high gain designs. NFB schemes can, if not implimented well, can exacerbate the problem. While these designs work well and sound warm (primarily due to higher & pleasant sounding THD), they tend to make tubes fail faster. It's kind of like running around town in your car in first gear all the time because it sounds cool. It may sound cool but your engine isn't gonna last long.
Cathode Stripping comes from applying plate voltage before the emission cloud has a chance to fully form by heating. Applying plate voltage before the tube reaches its quiescent temperature causes the negatively charged material used to offer up electrons on the cathode to directly strip off the surface of the cathode and fly to the high level positive charged plate. Eventually this stripping of material is substancial, leaving not nearly enough free electrons on the cathode to allow for full emission. This is the purpose of a standby switch, something you see less and less of in modern tube gear, and a fatal design flaw in my opinion. Plate voltage should only be applied to a tube after it has reached its quiescent operating temperature. This is responsible for about 80% of tubes failing.
Keeping a tube cool is not always going to extend its life. In fact if cooled too much it may actually lead to the same cathode stripping by lowering the quiescent temperature below that which causes emission cloud formation. Only in adverse conditions where it is recommended to cool a tube by external means is this advisable. Cooling all tubes in all cases for longevity is a classic audiophile myth.
There are some circumstances where this is good, like in tight cabinet designs where many tubes are closely spaced and ventilation is poor, or in cases where individual tubes are dissipating at very high levels, like in the hundred of watts each. You see these conditions more in communications gear and not very often in audio designs. Most modern designs the tubes are in the open. Tubes that operate at those higher levels (813, 833, 572A, 811A, etc) generally have described in their specs as to the proper method for cooling. But in most audio designs you do not see this and most regular convection cooling, on most normal commonly used tubes in the open, is usually all that is necessary, and in fact recommended.
Incidentially, Cathode Stripping due to plate voltage being present is also why it's bad to use a variac to "bring up to voltage" an old vintage piece of gear and NOT remove the tubes first. This creates massive cathode stripping in all the tubes. The plate voltages comes up to reasonably high levels and the heaters do not, or at least high enough to form a cloud from cathode heating. Also running tubes at lower filament voltage causes the filament to draw more current than it is design to handle and the heater may fail as well. Remove the tubes first, then bring the gear up on the variac. The capacitors in the supply will charge regardless of whether the tubes are present.
Any leakge of gas, of any kind, as you propose would radically alter the overall perfomance of the tube and would be evident by loss of the getter material within the tube. This issue you mention does exist (gassy tube) but always renders the tube useless long before the cathode would break down as you theorize.
Edits: 11/25/12 11/25/12 11/25/12 11/25/12 11/25/12
Tubes 'wear out' mostly due to decreasing
cathode emissions. Hydrogen does leak in, but
the cathode loss is the greater 'loss'.
Some cathode loss can be reduce by using a reverse
voltage plate to cathode, or with the control grid.
This rejuvenation process is temporary, the cathode
will become weak and have very low emissions.
Tube life can be prolonged by keeping the tube
cool. An added glass cover would cause too much heat
to build up in the tube elements.
Actually a lot of the internal contamination comes from contaminated parts: glass as well as the metal work. Sulfur comes from the lubricants used for drawing wires and rods, The metal and glass itself have dissolved molecular gas imbedded in them, typically the lighter gasses.
In the old days, precision military types all went through a pure hydrogen bake out in order to reduce the oxides, in particular. The oxides and hydrogen are responsible for the cathode "poisoning". The nitrides don't help any either. The hydrogen bake(pure hydrogen at 1200 degrees) out insures that all heavier gasses are displaced by hydrogen, which is then easier to pump out.
Glass is slightly hydroscopic, and thus an acid wash helps eliminate some of that contaminated surface.
Hydrogen bake outs are rarely utilized today. The NOS types are thus much lower in vacuum after use than most modern manufacture, with a few exceptions.
Of course YMMV
Stu
With a vacuum in the outer envelop? Instead, pressurize it with a heavy inert gas like argon. That'll keep the air out and the vacuum in: seems much easier than employing a pump.Also the getter flash is installed within a tube to "pump" out the residual gas in a vacuum tube. One of the many issues is that the materials used in fabrication often have molecular gas dissolved into the glass and metals. A pure hydrogen bake out helps alleviate this molecular gas. This involves a bake out in a pure hydrogen atmosphere at about 1K degrees. The lighter hydrogen tends to pull out the heavier gasses like oxygen and nitrogen, the principle constituents of the atmosphere, and then the lighter hydrogen is easier to evacuate when the vacuum pumps are hooked up.
Obviously this can be rather dangerous, but is still employed in developing mu metal, and it was used on critical tube elements in the US and Britain during the cold war era. I doubt if the procedure is employed in the Communist manufactured tubes, which is why they tend to become very gassy fairly quickly. Older American and Western tubes have significantly better vacuums and treated subassemblies.Glass often develops a layer of oxidation on the surface: not visible, but present. An acid wash just prior to assembly helps eliminate this oxidation layer. These procedures can be found on Pete Millet's excellent website, BTW, where he has scanned a large number of documents related to tube manufacture.
Nothing you an really do about osmosis through the leads though. That primarily due to the glass not adhering to the metal. The oxidation on the tube pins, curiously, is a necessary step in order to insure the glass adheres to the metal
For tubes which develop residual gas after use, the easiest way to get a better vacuum is to reactivate the getter. Theoretically the getter flash is activated whenever the tube is heated (as in use). However, I find that the residual gases often ionize and then cling to the cathode or anode, never reaching the getter flash. Activating the getter flash by placing the tube in a simple oven activates the flash without ionizing the internal gas contaminants. This uses the existing barium getter to work on the developed contaminants.
I've been using this procedure for years now. Tubes which start glowing blue after usage, have the blue glow disappear, after a prolonged bake out. In addition the initial top extension returns.
Stu
Edits: 09/07/11 09/09/11
They should use semiconductor grade silicon glass/ Si02 for the envelope.
Hydrgen meeting oxygen at 1000c will instantly form steam.
Semiconductor oxidation is most rapidly done in steam. The steam is made by flowing oxygen and hydrogen into a special 'torch' chamber where it is flame on. Keep the mix 'oxygen rich' to avoid possible left over hydrgen, which is a no-no.
Hydrogen is WAY reactive and if I were doing the process would flow a TCA mix into the glass tube to do a chemical getter. Works in semiconductor industry.
Too much is never enough
a 2nd envelop is not a great idea. You can't force the vacuum 'in' with external pressure.
As for an acid wash of the glass prior to pumpdown / seal, maybe. I don't know what is used for envelops of tubes, but in the silicon industry, a Silicon wafer will develop 20u of 'native oxide' nearly instantly after immersion to hydrophobic in an HF solution. Severe time limits are imposed from clean to the next process. Usually an hour or so.
Now, vacuum is a funny condition. Mechanical pumps like piston or the more common rotary vane and in some cases a roots 'blower' can only go down so far in vacuum. 10e -5 militorr is common for these. Most systems will cross over to what is called 'hi-vac' at about 25 to 50 mt where the NEXT stage of vacuum is used.
The real hi-vacuum pumps are of several different types.
1. Oil diffusion. Don't make me explain this one. Very clean when used with a LN2 'cold trap' to capture backstreaming oil, the will go to 10e-7 vacuum range easily.
2. Turbo-molecular. A VERY hi-speed driven turbine. Any atmosphere molecules which come in contact with the spinning blades are driven down, into the pump and a higher pressure region .
3. Cryo Pump. Using an external compressor of helium, the coldest part of these pumps is about 10degrees kelvin. Other parts of the array are warmed, but still everything freezes out except hydrogen and helium.
4. Gettering is used as the last stage.
Vacuum doesn't 'pump' in the water or pressurized air sense. Once below a certain pressure molecules don't 'flow'. They will bump into walls more often than one another. So, the flow regimes of gas in a vacuum system start with a viscous flow....sort of like water.
As higher and higher vacuums are reached, flow is called molecular. This is pretty random and pumps rely on entrapment or entrainment...see above.
TUBES? Well, that is a good question. It would be very difficult to reduce the atmosphere in one of them to very low levels. The small hole and odd shape work against it. The getter is that last best chance. And as Unc points out, the seal between glass and metal pins is problematic. Helium will diffuse thru glass. Our leak standards were a helium bottle with a valve and inset piece of thin glass. Leak rates were on the order of 10e-6cc per second. The bottles, smaller than a coffee thermos last for decades. I never knew one to run out....in over 30 years in the industry.
Too much is never enough
Stu,
How do you do this bake-out in a regular oven? If so, at what temp and for how long?
Thetubeguy1954 (Tom Scata)
SETriodes Group --- Central Florida Audio Society
Space Coast Audio Society --- Fullrange Drivers --- Front & Back Loaded Horns
======================================================================
"The man that hath no music in himself nor is not moved with concord of
sweet sounds is fit for treasons, stratagems and spoils." - William Shakespeare
I've written about this before.I use an old toaster over and stuff it full of tubes, If you have inquisitive household members, cover the tube with some fiberglass cloth to avoid thermal shock from curious eyes.
Once loaded I turn the oven on at the lowest setting and let the load sit for 3o minutes, I raise the temperature every 3o mites another 5o degrees till I reach about 300 degrees, and then I just let the tubes bake at that temperature for as long as I feel comfortable ( I've let the tubes bake over night). Once the baking period is over, simply unplug the oven and let it cool slowly to ambient room temperature, about3 or 4 hours.
I often do this to even new tubes as some residual gas seems to leach out after sitting for long periods ( NOS types). The result is a more dynamic sounding tube with a better top end.
Incidentally polishing tube pins also gives a much improved top end ( at least for 7 and 9 pin tubes where the pins enter directly through the glass). You also can effect big sonic changes in octal type tubes by changing the solder in the pins!
StuPS I should add that you could go higher in the bake out temperatures, but I do not recommend going past 350 degrees, primarily because the printing will yellow. Makes it difficult to resell. Ive gone up to 450 degrees with no issues to the tube however.
Edits: 09/19/11
You know that vacuum is a heat insulator, of course?
The tube envelop will get hot and the pins will conduct heat to the interior structures, but inefficiently.
I don't have a clear idea how hot you could get the innerds of a tube by your method.
Nothing is going to 'leach out' of a tube.....ever...while it is under vacuum.
Too much is never enough
Can I use a fan-forced oven?
Note that a post in response is preferred.
Warmest
Timothy Bailey
The Skyptical Mensurer and Audio Scrounger
And gladly would he learn and gladly teach - Chaucer. ;-)!
'Still not saluting.'
Yeah, you know us metrically deprived Americans. Convection ovens are fine, the key is to heat and cool the tubes slowly, or at least slow enough to insure that the differential between expansion rates of the metal and glass do not cause any issues.
Stu
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: