advice on chassis as heatsink

I am not trying to start another flame war on this topic, just looking for advice if this is a reasonable thing to try.

I am building a BrianGT 3875 kit into a mostly wood box. While the sides are wood, the bottom is a 0.09" thick 12"x12" piece of T-5052 aluminum. (To give an idea of the thickness, it is slightly thicker than the plate on a TO-220 heatsink where a chip is connected to it. -- about 2.29mm) Would it be a reasonable thing to just bolt the chips to the bottom of the chassis, or will I need something more substantial to dissipate heat? The transformer has dual 18V secondaries if that is important.

There's plenty of room (the inside of the case is about 12"x12"x3") I just don't happen to have sinks around, and want to get this thing running without placing yet another order for parts.

I do have some fairly large TO-220 sinks, but they are too narrow to attach the chips to. I thought about attaching them to the chassis right next to the chips, but I am not sure this would actually do any good. (as I reread your post, maybe this is what you are suggesting. Perhaos I'll give it a try.)

Placing the chips horisontally on the bottom plate will work fine in this case, only thing is that hot air goes upwards in this case around the chip, dissipation would be better if you placed the chips vertically, on a heatzink or a piece of alu/cubber.
LM3875 doesnt, really need much heatzink.
But still - it will work fine....maybe you could drill some holes in the bottomplate to get some airflow, but i don't really think that it is necessary.
Nuuk said:

If he joins the chip to that flimsy piece of aluminium base, it is going to pick up all the vibration. Adding something more substantial will help minimise microphony and 'take away' the heat.

The aluminum is pretty solid, and is attached to the case pretty rigidly in six places. I am not sure I understand why there would be more microphonics from the bottom of the case than from heatsinks attached to the bottom of the case.

The case itself does have rubber half-sphere feet, so hopefully that will reduce transmission of vibration from the shelf to the chassis.

Additionally, I have a sheet of copper. I was thinking I might attach the chips to the copper with HS compound, then attach the copper to the aluminum with the same compound. I can bend a bit of the copper up to provide a big fin while also effectively increading the thickness (and heat transfer rate) where the chips are attached. Perhaps the biggest benefit of this is that the big fin should provide a RF shield between the chips and the PS.

That can be a bit of a hotly debated subject in "audiophile" circles... Some will say that they definately do. These are likely to be the same folks who will tell you what a tremendous difference the new feet on their amplifiers made, or the power cord etc. There is apparently a LOT of money to be made on such upgrades and a LOT of people who can (or think they can) hear the effect of, say, a different *style* of brass spike-foot for their speakers.

Anyhow, MY speakers (or ears) apparently are lacking but I have a different opinion. If you hear microphonics, try to get rid of them. I don't (yet, perhaps).
Since this a DIY forum, may I suggest the obvious answer to the original post -- try it and see what happens. Ditto the microphony question. If it works, bravo! If not, try something else.

When you get opinions as varied as you have here, the only way to resolve it, is haul your behind away from the PC and go build something.

More to the point, whether the plate will be adequate heatsinking depends in part on how much current power you will run through the chips, which in turn depends on the power supply and your listening habits.
I'm running a P3A amp from a 18-0-18 V transformer and the heatsinks barely get above room temperature. The plate by itself should be fine. Give it a shot. The chips should shut down if they get over temperature.
"SPiKe protection means that these parts are completely safeguarded at the output against overvoltage, undervoltage, overloads, caused by shorts to the supplies, thermal runaway, and instantaneous temperature peaks."
For minimal heatsinking, look here..


More pics available here

This is the amp I use for my workbench. It's classic BrianGT LM3875 stuff. Power supply is a 2 x 16V toroid. The speakers I use are labelled as being 4 ohms. I never really cramp the amp up to full power as it's in a small room, and I'm usually pretty close to the speakers. Even when I set the volume to "shower mode" (the bathroom's in the next room) the bottom plate of the amp rarely feels more than warm to the touch. I'm sure I could activate the thermal protection if I really tried, but I'd probably need a 4 ohm dummy load to do it, otherwise it would be just TOO loud.

In summary, I think it depends on the size of your room, your speakers impedance and sensitivity, and your preferred listening level. Looks like you've got more heatsink than me though!

But it's not the aluminium making the music!
Why should the aluminium disturbe the sound then ?
Actually 2,25 mm alu isn't that flimsy with those measures.
Additionally, I have a sheet of copper. I was thinking I might attach the chips to the copper with HS compound, then attach the copper to the aluminum with the same compound. I can bend a bit of the copper up to provide a big fin
This will make it better, copper leads the heat from the source 2 times faster than alu paint it black and dissipation will be even better, but still i don,t think that it is necessary.