snipped-for-privacy@highlandsniptechnology.com wrote in news: snipped-for-privacy@4ax.com:
I even saw some where they slice up each fin into little fingers and the sink was "bristling" with them. Probably pretty good, but again, the air movement over and through them is required. Air flow means nothing if it does not intertwine with the heated elements to get hot and then be carried off as more air is added to be heated and moved. It's a bucket brigade.
Anthony William Sloman snipped-for-privacy@ieee.org wrote in news: snipped-for-privacy@googlegroups.com:
My decades old idea about making steam deep down in the heart of a slat dome (VERY hot) won't work either. However, pipes from the sirface taken down there to a heat exchanger that then sends the heated water back up would be easier to bring to a steam head than cooler water at the surface. So it could be used to preheat steam boiler water.
Jan Panteltje snipped-for-privacy@yahoo.com wrote in news:t3hijn$cus$ snipped-for-privacy@dont-email.me:
That is a very poor heat sink. Essentialy it conduction cools the device it is mounted to, but not much gets radiated away. More like a homogenizer.
There is a reason that motorcycle air cooled cylinder jugs and heads are NOT painted.
I know I'm wasting my time trying to explain anything to you, but I'm willing to give you the benefit of the doubt.
Very well compared to the interface between the plate and the air. That's the point. Even a thin plate has lots of conductivity compared to the metal-air interface unless that interface area is large.
But you seem to have missed another aspect. The important thermal conductivity in the metal, is the *aggregate* thermal conductivity of the many plates. That would be similar to thicker fins spaced further apart.
Zero evidence of that. The thick fins are spaced widely and much of the air flowing between them does little to cool the fins because it is not turbulent enough to come into contact with the fins.
If you need a case, that's great, but that's not whats being discussed here.
That's always true in engineering.
"Blasting air" is only useful if it contacts the heat sink enough for the heat to transfer. Simply blowing between fins doesn't necessarily transfer the heat.
All aspects of the design need to be balanced.
Yes, but nothing relevant. I don't see any images of skived heatsinks.
Convection cooled, right? For convection you need minimal obstruction of the air flow since the pressure is very, very low. It has to be *very* large relative to other designs to provide the same heat flow.
Radiation cooling goes as the 4th power of the temp difference between the sink and the universe. Since semiconductors can't be allowed to get very hot above ambient, rad cooling is usually minor. Inside a metal box, a lot of the radiation is trapped and reflects back.
The emission from a black body goes like T**4, so the thermal coupling between two surfaces goes like the difference, i.e. T1**4 - T2**4. For small temperature differences, that's basically 4 * T**3 * delta T.
Cheers
Phil Hobbs
snipped-for-privacy@highlandsniptechnology.com wrote in news: snipped-for-privacy@4ax.com:
All of the MIDS JTRS tactical transcievers in the US and ally military are all populated with canned, conduction cooled cards. Whether it is being slid into an F-18, or an F-35 or an F-22 they all go on the same tray Which has the PS unit and the MIDS unit and the RF amplifier unit. It is the most advanced SDR there is. ALL conduction cooled 'cards'. All temps get homogenized throughout the init. The RF amp is the only part that has air passages ported through it.
That is a micro sized VAX but way faster than a normal VAX. It manages AEGIS targetting and all of the battlefield waveforms. Mainly the still in use, still not crypto broken Link 16 waveform.
Don't forget to use absolute temperature. I remember a discussion here about inflate-gate where it took four iterations on the calculations and a large number of posts before anyone got the math right, because both the temperature and the pressure had to be absolute, not relative. Basically, there are two independent mistakes that could be made and all three erroneous calculations were done before the correct calculation was stumbled upon.
I wonder when it will support full self driving?
I wrote a little PowerBasic program to do the calc.
It's still pretty small for semiconductors on heat sinks.
For two black bodies at 300K, the thermal impedance of radiation is about the same as 6 mm of air.
Cheers
Phil Hobbs
That's a fun number. That's a lot of thermal resistance, and the insides of a metal box won't be a black body so things are worse in real life.
Until convection gets going, right. That's also the reason that foam and fibre-batt insulation isn't just immobilized air, as is often said--you need to scatter the IR radiation every few millimetres at most.
Photon budgets are a lot easier to do when you have a bag full of fun-fact numbers like that: spheres scatter light into 4 pi steradians; if your feedback resistor drops more than 50 mV, you're in the shot noise limit; and so on and so forth.
Cheers
Phil Hobbs
I experimented with covers over heated blocks, namely with various types of insulation. Given a pretty close fitting cover, the best insulation is none.
I don't think so, those fins are almost paper thin.
How does that even work? The bit at the end of each one where it magically bends up to the vertical.
Oooooh fewer eh?
Have something to add? Share your thoughts — no account required.
Ask the community — no account required