At visible light frequencis, glass reflects only a small proportion of the incident energy. Otherwise all our windows would look like mirrors! The proportion reflected rises for glancing angle incidence, which is why a valve looks more shiny on the edge. Most of the rest of the light is transmitted through.
At infra-red glass is not so transparent. A significant proportion is absorbed and heats the glass. This is why valve envelopes get hot. When two hot valves are next to each other, some of the heat radiation will heat the other envelope and some will go through and heat the other anode. Some of the extra heat in the envelope will then be radiated inwards, so indirectly heating the anode.
Can we estimate the effect of other valves being too close? We can surround a valve with 6 other valves, all touching to form a hexagon. Then the middle one has virtually no cooling. We could say that each extra valve reduces cooling by 1/6 =16.7%. We don't put them that close, but we can assume something like 5-10% reduction in cooling for each close valve i.e. we would need to reduce anode dissipation by 5-10% per near neighbour. A stereo amp would have 4 output valves in a line, so two of them have two neighbours: 10-20% reduction needed. The data sheet max dissipation would be for one valve under ideal conditions, and ideally we want to be a bit below that anyway to get good life (unless we are rich or foolish).
At infra-red glass is not so transparent. A significant proportion is absorbed and heats the glass. This is why valve envelopes get hot. When two hot valves are next to each other, some of the heat radiation will heat the other envelope and some will go through and heat the other anode. Some of the extra heat in the envelope will then be radiated inwards, so indirectly heating the anode.
Can we estimate the effect of other valves being too close? We can surround a valve with 6 other valves, all touching to form a hexagon. Then the middle one has virtually no cooling. We could say that each extra valve reduces cooling by 1/6 =16.7%. We don't put them that close, but we can assume something like 5-10% reduction in cooling for each close valve i.e. we would need to reduce anode dissipation by 5-10% per near neighbour. A stereo amp would have 4 output valves in a line, so two of them have two neighbours: 10-20% reduction needed. The data sheet max dissipation would be for one valve under ideal conditions, and ideally we want to be a bit below that anyway to get good life (unless we are rich or foolish).
Edit - the below with reference to your earlier of the two posts - the later I larger agree with!
with respect DF, you are wrong.
Warm isn't hot, its warm. Moreover, neither of us have defined either in terms of a specific figure...
Lets face it, the heater is orders of magnitude hotter in respect of the other internal components of the tube than ANY realistic radiated heat that may enter the envelope. The ONLY condition I am aware of that has caused excessive internal heating of the tube is over-current conditions - I have yet to see anyone describe a failure based on external heating of the tube (in sane applications) - if you have an example let me know.
You cannot convect in a vacuum - its an oxymoron. Thats why the shadowed areas of outer space are incredibly cold while those areas in direct sunlight are incredibly hot...
Therefore the glass is NOT the main way of getting heat out (in a glass envelope tube) except for the fact that it allows radiation. Conduction through terminations has a far greater effect on cooling than convection.
While heat is obviously life-threatening to a valve, it doesn't follow that radiated heat from other sources is a serious problem. The heat that kills valves is internally generated - its from too much current.
TV valves - wild operating parameters and cheap production values. Go figure.
Ask George.
Again, the datasheets tell the story by omission - there is generally NO information on this aspect of use. Odd, given the amount of data on other factors.
Having said that, if wider spacing is available and makes you happy - go for it. Its not going to hurt.
with respect DF, you are wrong.
Warm isn't hot, its warm. Moreover, neither of us have defined either in terms of a specific figure...
Lets face it, the heater is orders of magnitude hotter in respect of the other internal components of the tube than ANY realistic radiated heat that may enter the envelope. The ONLY condition I am aware of that has caused excessive internal heating of the tube is over-current conditions - I have yet to see anyone describe a failure based on external heating of the tube (in sane applications) - if you have an example let me know.
You cannot convect in a vacuum - its an oxymoron. Thats why the shadowed areas of outer space are incredibly cold while those areas in direct sunlight are incredibly hot...
Therefore the glass is NOT the main way of getting heat out (in a glass envelope tube) except for the fact that it allows radiation. Conduction through terminations has a far greater effect on cooling than convection.
While heat is obviously life-threatening to a valve, it doesn't follow that radiated heat from other sources is a serious problem. The heat that kills valves is internally generated - its from too much current.
TV valves - wild operating parameters and cheap production values. Go figure.
Ask George.
Again, the datasheets tell the story by omission - there is generally NO information on this aspect of use. Odd, given the amount of data on other factors.
Having said that, if wider spacing is available and makes you happy - go for it. Its not going to hurt.
Last edited:
Conduction through terminations is not the main way of getting heat out of a glass valve. Almost all the heat starts as radiation from the anode. Some comes through the glass and heats up nearby objects. The rest is absorbed by the glass and heats it; in turn it loses heat partly by radiation and partly by convection (outside!!). Half of the radiated heat from the glass goes back inside - some of this will hit the anode, some will miss and hit the glass elsewhere. So there is a complex set of heat flows going on.
External heat on its own will not damage a valve. What it does do is reduce the allowable dissipation, by perhaps 10-20%. There is not a sharp cutoff on dissipation: fine below, instant failure above. All that happens is that hotter valves don't last as long.
If you don't care about valve life, put them right next to each other and run them right up to the data sheet max dissipation. Otherwise, keep them about 1 diameter apart or derate by at least 10% (20% is better). Beyond 1 diameter there is little extra benefit from higher spacing.
External heat on its own will not damage a valve. What it does do is reduce the allowable dissipation, by perhaps 10-20%. There is not a sharp cutoff on dissipation: fine below, instant failure above. All that happens is that hotter valves don't last as long.
If you don't care about valve life, put them right next to each other and run them right up to the data sheet max dissipation. Otherwise, keep them about 1 diameter apart or derate by at least 10% (20% is better). Beyond 1 diameter there is little extra benefit from higher spacing.
Yeah amps are the opposite of women... maybe that's the reason they don't share our enthusiasm for this hobby.
As for what MJ recommends... maybe he should have called his book Tubes for Dummies. You can't generalize, can't compare a 6C33 with an el84, etc. Here's an amp I build long time ago. Those tubes lived a long and healthy life. Again common sense is your friend, keep them below maximum dissipation and don't hide your amp in a closed space.
.
Attachments
MJ talks more sense than some of the people on here. He does follow fashion - his own fashions, rather than other peoples'. We all do that. Recognise and filter out his bias, and what you have left is an excellent book.
Once again: here is my answer for heat problems. Tubes as far as reasonable from each other; anodes turned such a way they minimally radiate on each other; mirrors behind the tubes; perforated chassis.
It's Pyramid-VII-M amp prototype, it will go soon to cobra2 in Norway, his friend I hope will be happy to use it, as soon as I finish testing and some cosmetics.
It's Pyramid-VII-M amp prototype, it will go soon to cobra2 in Norway, his friend I hope will be happy to use it, as soon as I finish testing and some cosmetics.
Attachments
The mirror will make things worse, as each valve will see an image as well as the real thing, but you have the valves far enough apart that this should not be a problem.
If valves are close together better to make the back wall matt black so it absorbs heat, then gets rid of it by convection.
If valves are close together better to make the back wall matt black so it absorbs heat, then gets rid of it by convection.
Hi,
All dark😱.
Must be a power station trip due to excessive current at start up. If you use an inrush suppressor the lights don't dip.
Very nice!
Regards
M. Gregg
All dark😱.
Must be a power station trip due to excessive current at start up. If you use an inrush suppressor the lights don't dip.
Very nice!
Regards
M. Gregg
Absorbs heat and heats transformers? No, thank you! To absorb heat I am going to use heatsinks around 6С19П in another project, but you if I remember right you yourself said that it is bad as well. 😉
As I said, anodes oriented such a way they radiate minimally to each other and perpendicularly to mirrors, so rays are reflected under angles between tubes.
Pretty good book but far from excellent. I could point out its weak points but this thread its not about MJ.
I agree on that.
I have fried tubes using external heat, but the applications weren't sane. I was melting them intentionally to watch the glass get vacuum formed around the guts!
Tubes can absorb quite a lot of heat from the outside world. Just set a few outside in the noon time Florida sun during the summer. You can't pick them up with your hand. Considering how hot the tubes get in normal operation it's still a small increase in added heat.
Tubes shed most of their internally generated heat through IR radiation. There was a rather heated discussion on this forum a few years ago about OPT's that get hot. I stated that they get hot by absorbing the heat radiated by the surrounding tubes since the dissipation inside an OPT should be very low. After much disbelief I offered this test. Put a shiny piece of foil between the tubes and the OPT and see if they still get hot. They don't.
Actually the internal heat is from excessive dissipation. It can be caused by too much current, too much voltage, or both. You can melt a tube with the plate current and plate voltage both within spec, if the total plate dissipation is too high.
I went to a technical high school that had a 3 year long electronics program. Our "textbooks" were Army training manuals from the late 50's and Philco manuals from the same era. We were taught the "one diameter spacing" rule, and it seems to work OK for me. If you are going to push a tube to it's max, I would give it all the space you can. The 6V6GT works at 8 to 10 watts of dissipation. The spacing on the Magnavox amp shown looks adequate to me unless you are going to stuff in bigger transformers and crank things up.
Different brands of tubes have different diameter.
It also works at 10 to 12 watts of dissipation. In guitar amps up to 15 watts.
This is a DIY forum; some people are conservative, others like to push the envelope. Freedom for DIYers please. Tubelab sells his designs, if I did I would do things in a different way. What I'm trying to say is that the more space the better (up to a point). Not only with tubes but also with transformers, caps, etc. Keep them away from the heat!
Someone spiked my drink today - I can't stop posting.😀
Hi,
All dark😱.
Must be a power station trip due to excessive current at start up. If you use an inrush suppressor the lights don't dip.
Very nice!
Thank you!
I dimmed the light to take a picture of glowing toobs. No excessive current at startup; it has a wirewound resistor in series with primary and relay powered from filament rectifier that shorts the resistor as soon as voltages are up. Nice feature, highly recommend!
Yes, and somewhere in one of my old posts there is a picture of one running at about 40 watts of dissipation! SE amps do burn the tubes hotter than P-P ones, and of course there are plenty of guitar amps out there, including some I have built, that run the output tubes well above the dissipation limits.
The thread starter showed a pictures of an old Magnavox P-P amp. They were not known for running the tubes hard. The amp uses a aingle 6X4 rectifier. That's where I based my 8 to 10 watts from.
- Status
- Not open for further replies.