If that was the case it seems unlikely I'd ever have seen any temperature reductions, something I emphatically have, and repeatedly.
This is a many faceted problem, not one 'resolved' by over-simplified views.
It can't. A first, it has 2 sides. Second, area of anode is much smaller than area of heatsink. Third, unlike anode heatsink is exposed to an air convection.
Glass is cooled also, due to chimney effect that accelerates air flow.
Last edited:
The relative emissivity of specular and matte surfaces
If you are of a mind something of an article on this would be helpful to many I'm sure
If you are of a mind something of an article on this would be helpful to many I'm sure
Studying this many years ago I remember lots of surprises... But what I remember, is a fraction, mostly conclusion. Like, no matter how black some material looks, it's absorption and radiation of infrared rays is not always correlated with visual blackness. Glass have very low coefficients of absorption and radiation. Different metals added to glass vary coefficients of absorbtion and radiation of IR, while in visual band of spectrum it may be not so obvious.
I dream of that IR camera that is used in semiconductor industry, but I can't afford it now, unfortunately.
Yes, it has two sides. That is why I said half radiation could go back in. In practice convection and fins on the outside would mean that less than half would go back. Which radiates most into a valve: a hot black body ('cooler') just outside the glass, or a cooler grey body (chassis, transformer?) some distance way? The area of the anode is irrelevant, as much of the incoming radiation will be absorbed in the glass.
Glass is cooled by the chimney effect with a loose 'cooler' which has an air gap big enough not to restrict flow. Such a 'cooler' cannot use conduction. You choose: chimney or conduction?
OK, how about this: you pay the postage and I'll give you a couple of coolers?
I think some "Tubelab" type tests till destruction are going to be required to find out what cooling method works best for Wavebourn. Please have your camera ready so we can see the results! But it seems to me that a glowing plate is the ultimate limit, and no matter how one cools the glass, it is not going to make much difference there. Anti-photons? Cool rays? Cold dark matter circulation cooling? Sterile neutrinos? Parabolic mirror pointed out into the darkness of space?
Last edited:
Sure, I will take pictures! 6С19П looks like have nickel alloy anodes, at least they look exactly like ГУ-50 anodes, so I suspect them to be hard to kill.
I bought a box full of 6C19Pi from an Ebay seller in the Ukraine. So far I have killed only 1 and it was a real stupid move on my part. I was experimenting with cathode follower and augmented cathode follower output stages. There is simmilar US tube, the 7233 but they aren't cheap. Non compatible pinouts lead to the blown Russian tube.
Our department at work bought a Flir for measuring the thermal gradient across IC die, and evaluating IC packaging for thermal and RF performance. The object to be measured must be evaluated for emissitivity, or coated with a substance of known emissitivity. Glass and charcoal will be grossly different. With the emissivity known the camera can then be calibrated for direct temperature measurements. The Flir engineer spent a day here figuring out exactly how to do what we wanted and several experiments using known dissipations on the die were carried out for calibration. I am sure that there is more to it than I know, since I was present in the lab but busy with other work when all of this was done.
Last edited:
Right, and that's what I have figure out before I run up a rental bill of 20% of the $17,00.00 cost of the camera is question
Yes, and I don't want to be publishing my best efforts only to have them later shown off the mark because I overlooked or simply didn't know something important . . . Ah, the joys of self-employment ;-)
I still want to run 6С19П a bit above SOA.
Wavebourn,
Use a fan - they work. Ceramic bearing fans are now rated for 300,000 hrs
lifetime. You can run a low noise 80x80x25 12V DC fan on around 9~10V and
it will be virtually inaudable for most apps.
I have repaired/modified 100's of high powered Bass amps running up to
12 x 6550 OP tubes. The biggest source of tube reliability problems in these
beasts is heat. Many modern tubes go into grid emission and then bias
runaway. Installing 1 or 2 slow running fans makes a huge difference.
cheers
Terry
A good thread topic always endures, and having read through this one I suggest many comments are riddled with a misunderstanding of thermal physics and how heat from a power tube is mainly transferred. There was a good mix of diy solutions and discussion on the Pearl coolers, and the more typical forced air cooling path.
The main misconception I thought was related to the valve glass envelope and infrared radiation. There is no net radiated energy transfered between two blackbody radiators at the same temperature in a vacuum. We don't have blackbody plates or glass envelopes, but from what I can gather they behave pretty much as black bodies for infrared radiation at the tempertures involved in a power valve, ie. the glass envelope is best thought of as just being opaque to the energy radiated from the plate. The net transfer of energy from plate to glass wall is one way and depends on the difference in temperatures - in essence all the heat generated in the plate is transferred to the glass - if the glass is cooler then the plate will be cooler.
Removing more heat from some regions of the glass than others will stress the glass as glass is a very poor thermal conductor. For the Pearl option, the actual blend of heat transfer processes from the glass wall to ambient air in the vicinity of the vanes seems to result in temperature differences that are acceptable, but other techniques have caused problems.
If high temperature was a problem, and existing air cooling was not sufficient or you don't like the Pearl option then there may be a better path to take by using something like Bergquist's gap pad or q pad to provide lower thermal resistance transfer from the total glass surface to say a split copper tube that can then conduct the heat to fins - as it achieves primarily conductive transfer of heat from the glass, and would be the basis of a much lower thermal resistance method than the pearl option. There is still a need to transfer the heat to air, but a copper tube buffer would allow the use of simpler and better heatsinking to air. Obviously not an easy practical method for ST profile glass envelopes, but reasonable for straight sided profiles.
Ciao, Tim
The main misconception I thought was related to the valve glass envelope and infrared radiation. There is no net radiated energy transfered between two blackbody radiators at the same temperature in a vacuum. We don't have blackbody plates or glass envelopes, but from what I can gather they behave pretty much as black bodies for infrared radiation at the tempertures involved in a power valve, ie. the glass envelope is best thought of as just being opaque to the energy radiated from the plate. The net transfer of energy from plate to glass wall is one way and depends on the difference in temperatures - in essence all the heat generated in the plate is transferred to the glass - if the glass is cooler then the plate will be cooler.
Removing more heat from some regions of the glass than others will stress the glass as glass is a very poor thermal conductor. For the Pearl option, the actual blend of heat transfer processes from the glass wall to ambient air in the vicinity of the vanes seems to result in temperature differences that are acceptable, but other techniques have caused problems.
If high temperature was a problem, and existing air cooling was not sufficient or you don't like the Pearl option then there may be a better path to take by using something like Bergquist's gap pad or q pad to provide lower thermal resistance transfer from the total glass surface to say a split copper tube that can then conduct the heat to fins - as it achieves primarily conductive transfer of heat from the glass, and would be the basis of a much lower thermal resistance method than the pearl option. There is still a need to transfer the heat to air, but a copper tube buffer would allow the use of simpler and better heatsinking to air. Obviously not an easy practical method for ST profile glass envelopes, but reasonable for straight sided profiles.
Ciao, Tim
If the glass stopped all the IR radiation, then closing the glass doors on a fireplace would stop the IR and would leave the human absorbers on the sofa cold. (at least initially) I think the glass on most tubes is thin enough to allow some significant IR transmission. You can feel the heat from the filament before the glass gets hot. (Try putting ones hands around a light bulb that is just turned on and see how fast they get warm!) Silica envelopes (transmitting tubes) would likely transmit a majority of the IR.
Last edited:
Yes Mike, I am working on it, but still in my imagination.
Some real things need to be made to start some kind of money income, I work on them now. www.wavebourn.com • View topic - Gubernator - 8: parallel SE GU-50 amp
Member
Joined 2009
Paid Member
