How hot is hot?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Ok, my dillema: i need to heatsink the output transistors on my JLH. Now, i found some great .6°C/W heatsinks (not to large ,but bulky) at $25 each; with two of those (one per channel, holding the two output transistors), they would work at about 50°C, as each transistor dissipates 44W, give or take. Now, would that be enough? Or getting another pair and give each transistor it's own heatsink (25°C each) is worth the extra $50? Comments appreciated.
there can NEVER be enough heat-sinking, by definition! That being said, I would have to see the H.S. to get a feeling. Anyways, good heatsinks are hard to find and $25 seems like a fairly decent price. If you like them, buy' em. If the amp is over-cooled maybe that will motivate you into building a bigger and better amp later on.
Anyways that's my 2 cents.

With one heatsink for a pair of transistors, the heatsink will be at a temperature of 53degC above ambient, i.e. around 75 to 80degC. If you are using 2N3055s, the junction temperature will be at least 90degC above this (165-170degC). Though this is within the maximum limit of the 2N3055, it is too high for component reliability.

Use one heatsink for each transistor. The heatsink temperature will be reduced to around 50degC and the junction temperature to 140degC. The junction temperature is still higher than ideal so it would be worthwhile considering the use of either paralleled output devices or transistors with a better junction to case thermal resistance. email me if you would like more details.

I would agree that 0.6c/w to dissipate 88w is not enough. I am building an a40 and am using one 0.67c/w heatsink for each transistor that dissipates 25w. The predicted heatsink rise in this situation is ~17c +25c ambient = ~42c. Junction temp at this level will be about 77c. Nelson recommends trying to keep the heatsink itself in the 45-50c range, meaning that you can touch it briefly, but you wouldn't want to keep your hand there all day long...

You might want to have a look at the <a href="">large heat sink</a> here. It looks like an R-Theta 61570 which is spec'd at 0.8c/w per 3". This one is 17" long, so should be 0.14c/w for about $45

[Edited by Eric on 07-25-2001 at 03:33 PM]
while the math involved in figuring the approximate size of the heatsink is trivial, the point I was trying to make is how do you know that the heatsinks are really 0.6 C/W. In my experience, the figures get somehow reduced as the heatsink goes from the manifacturer, through the middle man on down to the shop. If you are talking surplus then everything is fair game. Case in point the link that Eric gave...a while ago they (S. S. of Ne) had this heatsink for sale with a power dissipation quoted at 260w...yeah, right, maybe at the sun core temperature!
Bottom line, be very careful.

[Edited by grataku on 07-25-2001 at 12:46 PM]
Well, this is the story... heatsinks are kinda hard to get here (Argentina) at decent prices, and this local shop had this old massive sinks sitting arround (they're not getting they any longer). The salesman had the catalog datasheet for them and they were rated at "0.5-0.6°C/W typical", and, at $25 each, it sounded quite ok to me! These are not very large sinks, but blocky ones, and very heavy too.

Now, i'm kinda short on cash right now :) But i think i'll end up spending the extra $50 and getting another pair; it's just i can't wait any longer to work on this proyect... been posponing it a lot.

Oh, BTW, for those who care; i tried that "Lavardin-style" headphone amp published in the web as a preamp (basically the same design with a tad more gain) and.. well, it sounded just like a preamp, nothing spectacular really. Not to say it works bad, it's just it wasn't worth the hassle i beleive. On the other hand, on headphones the sound is VERY VERY good; clean, sharp, defined, with no nastinesses at all. I'm considering it for a headphone out on my future preamp!
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.