Not yet settled, and not up to me. I'm recommending hard anodize.
Like Tim's project, the initial goals were somewhat unrealistic.
I'm not sure yet we met all the goals, but I am sure we came very close to the very best physically possible. Getting dissipation down to 5W meant n > 95%, input-to-output, through two SMPS. A large part lost is in the current sense resistors!
Cheers, James Arthur
Grin it's OK I screwed it up the first time too...
And then it sat in the middle of the upstream graph.
Oh.. ideally the time constant of something (a cube) goes as the Area, length squared. (R*C) (Hmm I don't think that's true once the cube gets bigger than a few cm.) The heat capacity of "everything" is about 3 J/(K * cm^3) (at room temperature)
George H.
I was thinking you'd want to make sides pointing at the sun nice and shinny... Nickle? (5W of full sun is ~50 cm^2) And the side facing down and to the north.. black. But I know nothing of things in the sun. White might be good. Some paint that reflects the sun (visible) but emits in the IR.
I like to drop as much voltage as possible across sense R's. (more signal) But I like class A too. :^)
George H.
100mV drop across 0.010 ohms @ 10A is 1W, which is 1/5th of my total dissipation, for just one sense resistor! And the design has 4--two current-mode switchers, plus input and output overload sensing.
If I had to take efficiency higher my next step would be to chop all the sense-resistor values, and regenerate the original sense voltage scales with op amps. I think that could save 1W, possibly a little more. That doesn't sound like much--just 1% of output--but that would cut the heat generation by 20%.
Cheers, James Arthur
There's a psychological advantage to the bare light bulb. It's something that everyone can relate to. Nearly everybody has burned themselves at least once on a light bulb. And it's hard for them to say, "just make it lower power" as they watch the light dim. And the client can do the experiment himself, five minutes after the phone call. Making the customer quickly understand the unintended consequences of his desire to have it all and then some is helpful in constraining the objective.
Right, but has the problem that it's not fair. A light bulb doesn't have comparable surface area to an enclosure.
Even the crudest Al enclosure gives Tim a starting point for how far expectations have to be adjusted.
(Apologies to Lasse, BTW, who suggested this first & I missed it.)
My customer offered to mill fins in the case, so I wrote a spreadsheet to ball-park the performance of 'n' fins of dimensions 'x' * 'z', etc., to recommend fin numbers, spacing, and height.
Cheers, James Arthur
The shape issue can often be reduced to a simple surface area calculation. The CDE data dealt with cylinders, the magnetics data with toroids (inside diameter detail ignored if fill
(Don't make the fins too close together in the race for more surface area, at least 1/4 inch or something like that... copy other heat sinks!)
George H.
Considering that the trials are performed with the load in a sealed container, you'd be better off using a fixed resistor bolted to a familiar shape - to represent the physical target. This requires less adjustment to external power sources and less frequent monitoring of the voltage and current that generate the controlled power burden.
There is also less confusion about the method of heat transfer to the internal ambient air and container walls. Though irrelevent to the calculation and results, it adds confusion that can dilute credibility of the physical trial method.
RL
"we as engineers" can do a quick in-mind guess of how hot something will get at a specific dissipation. We have all tried to put 1 or 5 watts into a resistor (rated at that wattage or not), have had to cool a classic series regulator that dissipates 20 watts or so, etc. We have the gut feeling that a small device dissipating 20 watts is going to get quite warm, and a calculation confirms it.
But indeed, relating it to a light bulb is good as well. People know how hot a lightbulb gets, especially those of 40 watts and above. However this experience will fade away now that classic light bulbs are not widely used anymore...
Heat spreaders aren't the only use for hot spots; if you can arrange a vertical airflow channel (a chimney) and put a hot spot in the middle, it'll start an updraft. Then, you can think about heatsinking more sensitive components at the bottom of the chimney, where cool air comes in.
As long as you keep it vertical!
Does putting a tube around a hot object cool it any better than leaving it in free air? I guess it could.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Decades ago, the original Macintosh and 512k "Fat Mac" were built without fans, to keep them quiet. They depended on convection cooling. They (specially the Fat Mac) were notorious for overheating, and cooking their power supplies.
A popular and cheap workaround was to build a chimney out of cardboard, and tape it around the convection vent slots on the top of the case. The resulting thermal updraft effect significantly improved airflow through the case, and cooled down the innards by several degrees.
Putting a tube around a hot object, and sealing the bottom against the entry of cool air, would certainly *hurt* cooling... lots.
Having a tube which starts some distance above the hot object, so that the tube channels the hot air upwards and doesn't constrain the arrival of cool air from the sides, can help quite a bit. Since air can't flow into the chimney tube from the sides, the chimney develops an upwards draft which pulls air in from the bottom.
That's one of the disadvantages of household fireplaces. The chimney effect is very effective, and the resulting draft up the chimney pulls up a lot more air than is actually going through the fire. Ordinary living-room fireplaces are a *negative* heat source for the house as a whole... the draft creates a partial vacuum in the house, outside air is pulled in, and those rooms away from the fireplace tend to get colder due to the influx of cold outdoor air.
Not the way I recall it; there was a common problem with solder joints at the yoke and/or power connector failing, possibly for cold-joint reasons. The joints would overheat after cracking, scorching connectors Sometimes, the nonpolar 3.9 uF electrolytic capacitor would fail (but it wasn't a frequent issue). Film capacitors used as replacements rarely failed.
After the Fat Mac, with the Mac Plus (it had a hard disk interface) folk started adding internal hard drives and auxiliary power for the hard drives; THOSE all needed a fan.
I think it was a combination of all of the above. I know the Fat Mac did run hot in my office, and I recall other people saying likewise. I think the heat buildup problems may have contributed to the failure of those NP 'lytics. I remember the film caps being part of the repair-and-upgrade kits which were commonly available.
Yeah. I recall warnings that the original supply in the Fat Mac didn't have enough power to run a hard drive in addition to everything else.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.