Thermal paste gets dried out and it is not just because of the heat, but it can speed it up. My fried didn't understand why games became slow. Turned out CPU was throttling frequency and voltage to keep under 100 degrees. It was used with the factory grease for 4 years. He sad it was like dry clay. Replaced it with new and cpu didn't go over 60 degrees. Amplifiers can also easily go over 60.
There are no long term grease tests. You get nearly the same result from tooth paster for about a week, but then they dry out lot faster. Some even tested with jelly
A good grease should not dry out for many many years. Not sure if there are types that do not dry out.
Here is one very efficient method to improve thermal conductivity of IC to heat sink
Those metal parts are used in one very unique Macola's High-End AMP design build with 4 x bridged/parallel LM3886 per channel (or 6x, I forget, was long time ago) … There is no schematic for that AMP currently …
Those metal parts are used in one very unique Macola's High-End AMP design build with 4 x bridged/parallel LM3886 per channel (or 6x, I forget, was long time ago) … There is no schematic for that AMP currently …
Attachments
Don't be discouraged so easily.
That appears to be copper, or a copper alloy. Anyone knows that copper is second to none when it comes to absorbing heat. The second advantage you have there is mass. That sink can absorb a lot of heat in a short period of time.
However, you then need to move the heat out of the block. The best conductor of heat is aluminum, but the trick is to have as much surface area as possible; fins.
The ideal situation would be a fairly thick layer of copper, bonded to a thinner layer of aluminum with lots and lots of fins to increase surface area.
Browse photos of CPU heat sinks and you will see this repeated in many styles. This is because it works.
" here is picture, no explanation, worship me" doesn't cut it I am afraid. I asked for a simple clarification and got a hissy fit.
True, it won't do much, but it'll conduct heat better than air - especially if thermal compound is used conservatively.
Still, you are quite right that all of that is certainly unnecessary if the tab side has a well designed and implemented sink.
I just had a funny thought... does anyone water-cool their amps?
I seriously looked at water cooling when I started the planning of my build as you can run the water block uninsulated from the tab then insulate the mounts for the block. But as every enthusiasm greater than time or fabrication skills. Also you realistically its about bragging rights for most power out the chip rather than a need. If you really have to run that close to the limit build a bigger amp!
In which case you might as well buy the non-isolated LM3886T version.
Using some sort of clamping bar across the LM3886 package can assist in the distribution of the clamping pressure, which may lower the thermal resistance between the device and the heat sink ever so slightly. In the case of the clamping blocks shown a few posts back, one has to be careful to get the clearances right. Too much slop in the clamping block results in a worse thermal interface. Too little, and you risk cracking the IC package during mounting.
Haters or not, physics apply to everybody.
For more information than you ever wanted to know about semiconductor mounting, see ON Semi's AN-1040.
~Tom
You know, some of those fat metal blocks have holes (larger than for mounting screws) in them, and I wonder if ...
LOL! What a coincidence!
Yeah, exactly. The main reason water cooling is popular with PCs is you just can't get a bigger/faster microprocessor, you can only try to do a better job at cooling and overclocking the ones available.
Imagine if a chip amp was Da Bomb (absolutely no distortion) but couldn't be paralleled...
I agree that the notion of water-cooling an audio amp is ever-so-slightly silly, however the benefits of water-cooling in the PC world are not limited to just being able to rev the cycles up higher. The primary benefit for some (including me) is the reduction in noise. As soon as one moves from passive cooling to active cooling, the noise reduction benefit of water/liquid becomes a big factor. It truly can mean the difference between a very loud PC and a barely audible one.
The fact that I can run my triple core 3.2Ghz CPU at 4.1Ghz and maintain core temps only 17C above ambient is the icing on the cake.
This dramatic reduction in noise (afforded by water/liquid cooling) would also be a benefit with amplifier designs that rely on active cooling, or designs that may have been dismissed in the past due to the impracticality (fan noise) of actively cooling an audio amplifier.
Last edited:
they are yelous becouse they spent too much time actualy studying maths or similar when others were having fun chasing girls and actualy building stuff.
no offence i am joking .
i dont understand one thing ,if you dont want v- on heatsink what else you can do ?buy 3886t and put fat mica insulator and two layers of heat paste? is that realy better than tf version with little bit of paste?
in datasheet says
so isnt that the real limit of chip? junction to case resistance?
yu3ma solution puts less case to heat sink resistance but does it lessen junction to case resistance , i hope that with your expert maths you actualy can calculate this things in more than 2 dimensions of chips back.
Considering that low-noise CPU fans are available that generate less than 10 dB(A) of noise, I think water cooling an amp is a tad overkill...
In a PC, the need for cooling is there from the moment it powers up. In an amplifier, the need for cooling is there when the user really cranks it. The rest of the time, it muddles along at barely above idle dissipation.
~Tom
In a PC, the need for cooling is there from the moment it powers up. In an amplifier, the need for cooling is there when the user really cranks it. The rest of the time, it muddles along at barely above idle dissipation.
~Tom
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.