Since the unforced air doesn't move really fast between the fins I don't think there are any benefits to mount the devices at the top or bottom. I'd say the best is to mount them in a center, which provides more even heat distribution in the sink.
i agree a little bit with all of you.
heat will want to move to paths of lowest resistance, like most everything in nature and physics. think of it as cooling or heating eachother. ie - the transistor heats the heatsink, which heats the air. OR, the air cools the heatsink which is cooling the transistor. its really just a shift in thermal energy.
so, if the transistors are at the bottom, heat will naturally move up, but, heat will spread naturally too. but, when the heat reaches the bottom, it has nowhere to go, thus accumulating. you break the effect... now, the transistors are heating the heatsink, which is heating the transistors, because the heat is getting blocked or bottleknecked.
if transistors are on top (which is probably WORST place for them), all heat accumulates on top of case. heat travels down, but with less resistance than up.
a good place would be slightly down, offcentre. this would utilize the fact that heat travels up, while utilizing the entire heatsink.
the IDEAL would be an inverted tent on the top. like this \/ (just not such extreme angles). the heat would travel up and out. the transistors would be situated at the crux of the V (middle, or intersection). the heat wiould naturally travel across the vain of heatsink up and away from the middle, where the heat source is located.
heat will want to move to paths of lowest resistance, like most everything in nature and physics. think of it as cooling or heating eachother. ie - the transistor heats the heatsink, which heats the air. OR, the air cools the heatsink which is cooling the transistor. its really just a shift in thermal energy.
so, if the transistors are at the bottom, heat will naturally move up, but, heat will spread naturally too. but, when the heat reaches the bottom, it has nowhere to go, thus accumulating. you break the effect... now, the transistors are heating the heatsink, which is heating the transistors, because the heat is getting blocked or bottleknecked.
if transistors are on top (which is probably WORST place for them), all heat accumulates on top of case. heat travels down, but with less resistance than up.
a good place would be slightly down, offcentre. this would utilize the fact that heat travels up, while utilizing the entire heatsink.
the IDEAL would be an inverted tent on the top. like this \/ (just not such extreme angles). the heat would travel up and out. the transistors would be situated at the crux of the V (middle, or intersection). the heat wiould naturally travel across the vain of heatsink up and away from the middle, where the heat source is located.
They say that heat is naturally moving up because of air. Warm air expands and since its density is less than cold air it weigh less and moves up. When heat is transfered in a heat sink it has nothing to do with air so I don't quite understand why it would move up easier. It is probable though that with tall sink the air moving up between the fins is warmer at atop than air at the bottom and doesn't absorb as much heat, making the top part of a sink warmer. That is only my speculation and of course I might be wrong.😉
you are right on about densities. heat only rises because of density of air. however, keep in mind on the fins of the sink, the air surrounding the fins will be warmer, because, well, thats how sinks work!
what i wrote isnt "technically" correct, but the concept is. heat will not as much rise up the heatsink, but the heatsink will essentially be less effective due to the air surrounding the fins becoming saturated.
what i wrote isnt "technically" correct, but the concept is. heat will not as much rise up the heatsink, but the heatsink will essentially be less effective due to the air surrounding the fins becoming saturated.
definately.
i think putting the heatsinks on top would be ideal, but the top surface area would have to be far too large for this to really work out right...
i still havent decided what i will do once i begin.
i think putting the heatsinks on top would be ideal, but the top surface area would have to be far too large for this to really work out right...
i still havent decided what i will do once i begin.
great... now you throw another wrinkle
hmm, so I thought we were in agreement that slightly lower than center would be ideal (heat spread within the sink will be maximized) I was originally thinking more along these lines than all the way at the bottom, and I was confused about why you would put it at the top. But now you're suggesting that a heatsink on top would be best? I assume that you mean horizontal, with the fins pointing up? It seems like that would reduce the flow of cool air to the sinks, since there isn't any sort of chimney to get the air moving.
hmm, so I thought we were in agreement that slightly lower than center would be ideal (heat spread within the sink will be maximized) I was originally thinking more along these lines than all the way at the bottom, and I was confused about why you would put it at the top. But now you're suggesting that a heatsink on top would be best? I assume that you mean horizontal, with the fins pointing up? It seems like that would reduce the flow of cool air to the sinks, since there isn't any sort of chimney to get the air moving.
Re: great... now you throw another wrinkle
well to answer the questions in order...
first, off centre would be good, but there would be no real way to tell just HOW off center it should be. plus, it really wouldnt make a HUGE difference. so, center would be fine, and if they had to go a little lower, it wouldnt hurt at all.
and as far as mounting them horizontal, this is just a thought. plus, to counter your arguement, air doesnt inherently "move" by itself. so, if you mounted them vertical, it doesnt necesarrily induce any sort of flow. with them horizontal, air is warmed in between fins, and becomes warmer than the air above it, and will rise. rinse and repeat. the cooling effect is the same for vertical. however, it doesnt work when the heatsink is vertical and the fins are horizontal, because the heat gets trapped in the fins.
if mounted on the tip with fins pointing up, heat has a clear path to escape, more heat is kept away fromt he PCB, and the heatink is heated/cooled more evenly.
this is just an idea though, and probably not the best way to impliment heatsinks either, but, hey im just throwing it out.
well to answer the questions in order...
first, off centre would be good, but there would be no real way to tell just HOW off center it should be. plus, it really wouldnt make a HUGE difference. so, center would be fine, and if they had to go a little lower, it wouldnt hurt at all.
and as far as mounting them horizontal, this is just a thought. plus, to counter your arguement, air doesnt inherently "move" by itself. so, if you mounted them vertical, it doesnt necesarrily induce any sort of flow. with them horizontal, air is warmed in between fins, and becomes warmer than the air above it, and will rise. rinse and repeat. the cooling effect is the same for vertical. however, it doesnt work when the heatsink is vertical and the fins are horizontal, because the heat gets trapped in the fins.
if mounted on the tip with fins pointing up, heat has a clear path to escape, more heat is kept away fromt he PCB, and the heatink is heated/cooled more evenly.
this is just an idea though, and probably not the best way to impliment heatsinks either, but, hey im just throwing it out.
Evaas said:
Oh for crying out loud, enough about the heatsinks. Somebody pay attention to me and my Dillema!!🙂
Anthony
Anthony
Regarding vertical and horizontal heat sink placement:
I asked one member of the Forum (PedroPo) why he placed heat sinks in his SOZ in horizontal position. He replied that it made only 2 deg C difference so he chose more aesthetic look.
I asked one member of the Forum (PedroPo) why he placed heat sinks in his SOZ in horizontal position. He replied that it made only 2 deg C difference so he chose more aesthetic look.
I took my thermodynamics book...... (i hope you can undrstand my bad english)
-if you mount an heatsink horizontally the efficiency is extremly low (the flux of air at the center of the heatsink is extremely low so is the cooling)
-the thermal energy move in every direction IN the heatsink (the conduction of heat don't know where up is...)
-the air move between the fins because the specific volume raise with the temp, so the hot air is much lighter than cool air (peter daniel is correct)
-the best place for placing mosfet i think is slightly higher than center, you have to create a continuous increasing in the temp of the heatsink from bottom to top, so the air is continuously accelerated (if you put the mosfet at the bottom of the heatsink the top is cooler than the bottom, and the air is not accelerated). The best place for placing mosfet is in function on the specs of the heatsink (fins dimension and thickness of the wall, thicker the wall less important the placing of the mosfets because the temp is more stable).
-spacing of fins is in function of the equilibrium temp and the phisical dimension of the heatsink (for an heatsink 250mm high, with fins 75mm long and an quilibrium temp of 55-60 Celsius the optimum spacing of fins, if i remember, is between 7 and 9 mm)
i hope you can understand what i wrote (and i hope is all correct of course)
F.
-if you mount an heatsink horizontally the efficiency is extremly low (the flux of air at the center of the heatsink is extremely low so is the cooling)
-the thermal energy move in every direction IN the heatsink (the conduction of heat don't know where up is...)
-the air move between the fins because the specific volume raise with the temp, so the hot air is much lighter than cool air (peter daniel is correct)
-the best place for placing mosfet i think is slightly higher than center, you have to create a continuous increasing in the temp of the heatsink from bottom to top, so the air is continuously accelerated (if you put the mosfet at the bottom of the heatsink the top is cooler than the bottom, and the air is not accelerated). The best place for placing mosfet is in function on the specs of the heatsink (fins dimension and thickness of the wall, thicker the wall less important the placing of the mosfets because the temp is more stable).
-spacing of fins is in function of the equilibrium temp and the phisical dimension of the heatsink (for an heatsink 250mm high, with fins 75mm long and an quilibrium temp of 55-60 Celsius the optimum spacing of fins, if i remember, is between 7 and 9 mm)
i hope you can understand what i wrote (and i hope is all correct of course)
F.
I like being horizontal
hehe... This reminds me about that other post where someone mentioned the incessant heatsink questions as leading to the death of this forum (ok well maybe it wasn't that apocolyptic). But I enjoy these side discussions. And now for some content...
The 2 degree differential is interesting, and is less than I would have imagined, so that free's me up a bit. I was considering placing the sinks on both sides and top but I didn't think it would be thermally matched, but this makes me think it would work. I'll just make some small changes to match the temps, and It should look pretty sweet.
Anthony, what was your dilemma? I looked through a few earlier posts and I couldn't find it.
hehe... This reminds me about that other post where someone mentioned the incessant heatsink questions as leading to the death of this forum (ok well maybe it wasn't that apocolyptic). But I enjoy these side discussions. And now for some content...
The 2 degree differential is interesting, and is less than I would have imagined, so that free's me up a bit. I was considering placing the sinks on both sides and top but I didn't think it would be thermally matched, but this makes me think it would work. I'll just make some small changes to match the temps, and It should look pretty sweet.
Anthony, what was your dilemma? I looked through a few earlier posts and I couldn't find it.
Heatsinks
Now that's alot of hot air for sure.......
http://www.amazon.com/exec/obidos/ASIN/0791800741/ref=pd_sxp_elt_l1/104-5272109-0907141
Now that's alot of hot air for sure.......
http://www.amazon.com/exec/obidos/ASIN/0791800741/ref=pd_sxp_elt_l1/104-5272109-0907141
Attachments
Peter Daniel
Be careful! A 2°C difference may turn into a 12°C difference very quickly when you raise the power dissipation requirement. Remember, this is a geometric coefficient we are dealing with here. I just thought that I would tell you guys that Nelson puts his mosfets just about at the top of his heatsinks on the X-250 so I guess someone in his production department thinks the thermal differential seems to be the better way...or maybe that was the only place left to stick them because of poor parts planning? Ha!
zero to sixty
hmm, the air acceleration idea is fascinating, and I'd love to see some data on a few heatsinks. For example, what is the air flow across the fins when the transistor is mounted in different locations, and when the sink is mounted in different directions, and what is the junction temp in all these scenarios.
As for practical use, I'll probably just mount near the center and get enough sinks that it doesn't matter.
hmm, the air acceleration idea is fascinating, and I'd love to see some data on a few heatsinks. For example, what is the air flow across the fins when the transistor is mounted in different locations, and when the sink is mounted in different directions, and what is the junction temp in all these scenarios.
As for practical use, I'll probably just mount near the center and get enough sinks that it doesn't matter.
My Chassis
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
sorry i forgot a couple of things:
- the best color is black (black painted is good too)
-if you use fans is all different..... (less space between fins=more performance)
F.
- the best color is black (black painted is good too)
-if you use fans is all different..... (less space between fins=more performance)
F.
a75 to aleph
well the aleph is more efficient than the a75, so if you build an aleph of the same power rating, then the chassis should be fine. but if you upgrade your transformer then Idunno, but your "chimneys" look pretty big. Of course this is all my own reasoning and has no basis in fact.. 🙂
btw, where do you get those cheap irfp244's?
well the aleph is more efficient than the a75, so if you build an aleph of the same power rating, then the chassis should be fine. but if you upgrade your transformer then Idunno, but your "chimneys" look pretty big. Of course this is all my own reasoning and has no basis in fact.. 🙂
btw, where do you get those cheap irfp244's?
- Status
- Not open for further replies.