they are two to three times thicker than real 0.2C/W Kapton isolators.rhysh said:
I don't believe Bergquist.
Look for 0.05mm thickness or less for the Kapton and 0.05mm or less for the adhesive.
Maybe <=0.08mm total thickness for adhesive coated Kapton.
1 thou (mil) for each of Kapton and adhesive ~=0.05mm total thickness. This when thermal paste coated on the non adhesive side may achieve 0.2C/W in To247 size, but I doubt it.
Some active device manufacturers quote 0.2C/W for direct thermal paste connection to the heatsink.
Andrew,
Where did you get that 40% from the bottom figure? It would seem that by the rational you used, the source should be placed in the center.
I can't locate the info I found last week, but it was info from an engineering paper or technical blurb provided at a heat sink supplier. It was mentioned that contrary to the popular notion that heat sources are better placed closer to the bottom (due to irrelevant notion of heat rising), the contrary is true. The popular notion that heat rises is not a factor in heat conductivity THROUGH the heat sink.
I don't remember the rational presented off hand, but I suspect the one I mentioned in my previous post; i.e., that you don't want the air warmed so soon in its rise, because then it could take less heat away rising through the rest of the heat sink. However, since I was also reading a lot about air flow and pressure drop through the fins, it could have played a part. I'm not that good in fluid dynamics, but I wonder if pulling the air through the fins via a heat source at the top is better, from a flow bypass issue, than heat at the bottom pushing air to the top of the fins.
Where did you get that 40% from the bottom figure? It would seem that by the rational you used, the source should be placed in the center.
I can't locate the info I found last week, but it was info from an engineering paper or technical blurb provided at a heat sink supplier. It was mentioned that contrary to the popular notion that heat sources are better placed closer to the bottom (due to irrelevant notion of heat rising), the contrary is true. The popular notion that heat rises is not a factor in heat conductivity THROUGH the heat sink.
I don't remember the rational presented off hand, but I suspect the one I mentioned in my previous post; i.e., that you don't want the air warmed so soon in its rise, because then it could take less heat away rising through the rest of the heat sink. However, since I was also reading a lot about air flow and pressure drop through the fins, it could have played a part. I'm not that good in fluid dynamics, but I wonder if pulling the air through the fins via a heat source at the top is better, from a flow bypass issue, than heat at the bottom pushing air to the top of the fins.
go to Avvid and/or Wakefield.
They show arrangements of devices to obtain optimum dissipation.
~40% is not halfway.
If I remember correctly, it's slightly less than 40%, but if the hole is drilled at tapped at 40% up and the device oriented such that the main heat source is below the hole, then one has achieved very near optimum location for a string of devices that run horizontally across the vertically aligned heatsink.
They show arrangements of devices to obtain optimum dissipation.
~40% is not halfway.
If I remember correctly, it's slightly less than 40%, but if the hole is drilled at tapped at 40% up and the device oriented such that the main heat source is below the hole, then one has achieved very near optimum location for a string of devices that run horizontally across the vertically aligned heatsink.
pooge said:
You mean XLR instead of RCAs? I'm not that fussy about connectors and I use cheapest Cardas gold plated RCAs.
The type of connectors is mainly dependant on interconnect and if it's RCA terminated I would need ot use adaptor anyway if amp had XLRs. OTOH, I'm not tempted to modify interconnects I already have.
rhysh said:All the thermal insulation i can seem to find from Farnell that are the right size are all made by Bergquist.
I use the Bergquist on F4 and F5 with the Conrad heatsinks, no grease, and they seem to work just fine: Digikey BER120. This particular one is TO-247 size (25.40mm x 19.05mm x 0.15mm thick).
Andrew is right about the heatsink placement.
That hot air isnt staying stationary, its moving up and quickly. The heat doesnt spread DOWN the heatsink very far so mid sink placement is a waste of sink. Also all these heatsinks with the fins facing anything but straight up are being used inefficiently, but they still will work if they are big enough to make up for the fact that the heat isnt getting out as quickly as it would like.
A good thing to consider is a nice 1/4" thick piece of copper as a heat spreader and your heatsink will be used even more efficiently.
Uriah
That hot air isnt staying stationary, its moving up and quickly. The heat doesnt spread DOWN the heatsink very far so mid sink placement is a waste of sink. Also all these heatsinks with the fins facing anything but straight up are being used inefficiently, but they still will work if they are big enough to make up for the fact that the heat isnt getting out as quickly as it would like.
A good thing to consider is a nice 1/4" thick piece of copper as a heat spreader and your heatsink will be used even more efficiently.
Uriah
You could use thermal paste to thermally connect the spreader and the sink. Connect the device to the spreader the same as you would to the heatsink. The copper will diffuse the heat quicker and more evenly than the aluminum and then the aluminum will take it from there.
So if you could get a hold of a piece of copper, shine it up really nice so it makes great contact with both other surfaces, or dont, if its really large probably not having perfect contact will still be fine. Since a spreader isnt necessary but is an advantage you can go with any size and I think that something about 2" by 3" would be fine. Ebay probably has metal vendors that will offer to cut it for you. You could go larger, but going larger gets into problems when the copper or aluminum is a bit warped and may not be making contact over the whole surface but you wont probably see that as it will be between the copper and aluminum. You will have to drill both the copper and the aluminum and bolt/screw it into place. Use lock washers so it never backs out.
Uriah
So if you could get a hold of a piece of copper, shine it up really nice so it makes great contact with both other surfaces, or dont, if its really large probably not having perfect contact will still be fine. Since a spreader isnt necessary but is an advantage you can go with any size and I think that something about 2" by 3" would be fine. Ebay probably has metal vendors that will offer to cut it for you. You could go larger, but going larger gets into problems when the copper or aluminum is a bit warped and may not be making contact over the whole surface but you wont probably see that as it will be between the copper and aluminum. You will have to drill both the copper and the aluminum and bolt/screw it into place. Use lock washers so it never backs out.
Uriah
Peter Daniel said:
Peter,
I took a peek at this and the parts count is almost like an SET amplifier (lower than most actually). Can you comment on the use of the Vishay/Caddock and which positions you used them in? I think it might be interesting to do the entire thing in nude Vishays as the parts count is so low.
Chris
Aren't copper and alumimum together a no-no as far as corrosion.
Regarding my earlier comments about transistor placement on a heat sink, I found a blurb on ESP site. Don't know if it's the info I recall.
ESP: Heat sinks
See section 14.
Not saying that this is the last word. No rational is given. I'd like to find a good explanation.
Regarding my earlier comments about transistor placement on a heat sink, I found a blurb on ESP site. Don't know if it's the info I recall.
ESP: Heat sinks
See section 14.
Not saying that this is the last word. No rational is given. I'd like to find a good explanation.
udailey said:
That must be because the temperature differential between the air and heat sink are is greater below. It would be nice to see a good work on this issue, like they show for fin density and pressure drop, etc. I couldn't find anything. Looking back, Elliot's point about not having them mounted on the bottom is not really inconsistant with Andrew's. He may just be saying don't mount them on the VERY bottom.
I would think that if you do not allow air inbetween them they would not corrode. I didnt know that btw but I think corrosion needs oxygen.
I would say Rod Elliot is the end all be all answer on most questions, however Andrew's suggestion of 40% from the bottom is not nearly the first time I have heard this.
While it may be true that heat doesnt care which direction it spreads in a solid it will care if it is trapped under that solid once it hits air. It will also make a difference if those fins are not facing directly up since once the heat hits the air convection is now in play and if the only surface facing straight up is 2" by 5" rather than lets say the face of the sink at about 12" by 5" by a hugely larger area since the fins have surface area as well, yes it will matter a lot. The heat will get to the top of the sink that is standing on end and not be able to get OUT of the sink whereas it will be able to dissipate much quicker with a few hundred square inches facing upward like when the fins are pointing up.
Uriah
I would say Rod Elliot is the end all be all answer on most questions, however Andrew's suggestion of 40% from the bottom is not nearly the first time I have heard this.
While it may be true that heat doesnt care which direction it spreads in a solid it will care if it is trapped under that solid once it hits air. It will also make a difference if those fins are not facing directly up since once the heat hits the air convection is now in play and if the only surface facing straight up is 2" by 5" rather than lets say the face of the sink at about 12" by 5" by a hugely larger area since the fins have surface area as well, yes it will matter a lot. The heat will get to the top of the sink that is standing on end and not be able to get OUT of the sink whereas it will be able to dissipate much quicker with a few hundred square inches facing upward like when the fins are pointing up.
Uriah
Heatsinks are usually anodized, at least for external use, so copper/aluminum corrosion shouldn't be a problem.
I used copper spreader in F3 with good results:
In this amp, I used what I already had in stock, but my choice was also biased by previous experience with other amps.
So I used Vishay S102 for R9, R13 and R14.
Caddock MK132 for R10 R5-R8.
Riken for R1-R4 but only because I had that particular value only with Rikens.
Mills for R11/R12.
I used copper spreader in F3 with good results:
An externally hosted image should be here but it was not working when we last tested it.
chrismercurio said:
In this amp, I used what I already had in stock, but my choice was also biased by previous experience with other amps.
So I used Vishay S102 for R9, R13 and R14.
Caddock MK132 for R10 R5-R8.
Riken for R1-R4 but only because I had that particular value only with Rikens.
Mills for R11/R12.