In theory a bar across the device could apply more uniform pressure, but is the benefit significant? I've seen many products that use spring-metal clips. You screw the clip to the heat sink and it pushes the package against the heat sink. This also minimizes the precision needed in alignment. As Tom said, there are a lot of 'more perfect' options but is improving the performance from 90% of ideal to 91% of ideal (to use arbitrary numbers) of any real benefit? Six months' worth of accumulated dust on the heat sink probably negates that.
None of this thermal interface stuff is new. It's well-documented in Motorola (now ONsemi) application notes from over 40 years ago and other sources as well. Those have lots of plots of thermal performance vs interface material, screw torque, whatever.
None of this thermal interface stuff is new. It's well-documented in Motorola (now ONsemi) application notes from over 40 years ago and other sources as well. Those have lots of plots of thermal performance vs interface material, screw torque, whatever.
I wholeheartedly agree.
The Keratherm pads are now online: Keratherm Red (86/82) for LM3886, TO-247, TO-218, TO-3P
Tom
The Keratherm pads are now online: Keratherm Red (86/82) for LM3886, TO-247, TO-218, TO-3P
Tom
The Modulus-686 measures 8.25 x 2.3 inches. The assembled module stands 1.35" tall.
You can get 100 W into 8 Ω from the MOD686 by lowering the supply voltage to about ±24 – ±25 V.
There's really no compelling business reason for me to make a 100 W version of the MOD686 available. In theory, you should be able to get rid of two of the LM3886es and associated components and make the board a little smaller. So you'd save a few bucks in parts, but the assembly cost won't change much and neither will my module cost. Why not just spend $20 more and get a 240 W capable amp?
Tom
You can get 100 W into 8 Ω from the MOD686 by lowering the supply voltage to about ±24 – ±25 V.
There's really no compelling business reason for me to make a 100 W version of the MOD686 available. In theory, you should be able to get rid of two of the LM3886es and associated components and make the board a little smaller. So you'd save a few bucks in parts, but the assembly cost won't change much and neither will my module cost. Why not just spend $20 more and get a 240 W capable amp?
Tom
Or run the amp at clipping with a hot (>65 ºC) heat sink. In either case, you should probably take a close look at what you're doing.
Tom
Modulus-686: 380W (4Ω); 220W (8Ω) Balanced Composite Power Amp with extremely low THD
Different scaling comparison, however I do understand what you say. I don’t think it would make a difference to listening experience in 0.0001% either assuming nothing else changes.
The effort to squeeze power and distortion out of a design is the game.
Different scaling comparison, however I do understand what you say. I don’t think it would make a difference to listening experience in 0.0001% either assuming nothing else changes.
The effort to squeeze power and distortion out of a design is the game.
If that's the game, you're far better off putting your efforts towards obtaining larger heat sinks.
The main reason I went with the Keratherm pads is that I needed a large amount of pads for the assembled modules, thus could justify buying in bulk. Also, I figured the Keratherm pads would go over well with the thermal geeks here, hence result in some bragging rights and perceived value. They are nice pads. Nothing wrong with them.
Tom
The main reason I went with the Keratherm pads is that I needed a large amount of pads for the assembled modules, thus could justify buying in bulk. Also, I figured the Keratherm pads would go over well with the thermal geeks here, hence result in some bragging rights and perceived value. They are nice pads. Nothing wrong with them.
Tom
Quite agree about the pads. They are convenient with well published data. I would hate to have to ask about anodizing data for heatsinks.
As far as heat sink sizes, I would rather experiment from what I have around. But I would want to get my Mod86 running and listed to first. Today I just dug out a MyRef FE kit under a pile of stuff. Totally forgot about it.
As far as heat sink sizes, I would rather experiment from what I have around. But I would want to get my Mod86 running and listed to first. Today I just dug out a MyRef FE kit under a pile of stuff. Totally forgot about it.
If you add black dye to the anodizing soup, I believe it will improve the radiating efficiency of the heat sink. It's not the anodization that improves the thermal properties but rather the black dye, which makes the heat sink closer to a blackbody radiator.
Now, as far as I understand it, the heat sink just has to be "black" (absorptive) at the right wavelength (I'm guessing in the IR spectrum). Ironically, white snow is supposedly a blackbody radiator in the IR spectrum.
Tom
Now, as far as I understand it, the heat sink just has to be "black" (absorptive) at the right wavelength (I'm guessing in the IR spectrum). Ironically, white snow is supposedly a blackbody radiator in the IR spectrum.
Tom
Dude. That is seriously reckless. In particular in this forum. If any of the LM3886es short to the heat sink, you'll very likely destroy the board. If you're willing to blow $300+ that way, go right ahead. Apparently you have more money to burn than I do. However, recommending that others do so is unwise to say the least.
Geeking out is one thing, but what you are suggesting is taking it too far. If that's what you're into, I strongly suggest that you start a thermal geek-out thread. Those of us who would like to join you in the discussion about using a molecular layer of aluminum oxide as a thermal pad will do so there.
Like I said previously, you'd be better off getting heat sinks that are a little bit larger. Or by dusting your heat sinks every now and then and ensuring that the equipment has adequate airflow.
Tom
Last edited:
Modulus-686: 380W (4Ω); 220W (8Ω) Balanced Composite Power Amp with extremely low THD
I don’t think you will have LED heat sinks hidden, specially high powered ones. So the safety regulations all apply.
Packing power into as small package as possible with the best performance is never any easy task.
I don’t think you will have LED heat sinks hidden, specially high powered ones. So the safety regulations all apply.
Packing power into as small package as possible with the best performance is never any easy task.
Last edited:
Hard anodizing that is thick enough has been used as an insulator in heat sink applications if the voltage isn't too high. In our sort of application that would almost require drilling (and tapping?) the holes for screws, and then final surfacing and removal of any burrs before anodizing. Otherwise you risk poking through the anodizing and shorting. This is the sort of thing that is practical for specialized applications, but not DIY audio.
Also, this continues a pretty pointless discussion as the gain in thermal efficiency from such measures is insignificant and much too much cost and effort for negligible improvement. If one is trying to eek out more power output before tripping the chips' thermal shutdowns, then, as Tom points out, go for a bigger heat sink and/or more air circulation around the heat sink.
Also, this continues a pretty pointless discussion as the gain in thermal efficiency from such measures is insignificant and much too much cost and effort for negligible improvement. If one is trying to eek out more power output before tripping the chips' thermal shutdowns, then, as Tom points out, go for a bigger heat sink and/or more air circulation around the heat sink.