Prototyping a 100W CFB amp...

[OliverD]

Here's another project I'm currently working on... It's a preliminary PCB stuffed with everything absolutely necessary to get the beast running. Frequency compensation is not yet satisfactory, I'm working on that. You can see my mockup-heatsink (which has seen many many amplifiers on it in the past years - I sand-blasted it for the photo, it was even more ugly (uglier?!) before). Two complementary pairs are still missing, but there are lots of cheapish capacitors, even tantalum caps in the DC servo, all of which will be replaced. I messed up one tapped hole, so one of the driver transistors will have to stay away a bit from the board. The design is very close to a commercial unit, lets see who finds out first

[OliverD]

Input section details. Input filter and output short circuit protection are missing for now. The 2 x 4 small signal transistors will thermally be coupled by using thermal grease and shrink tube.

[OliverD]

If the amp works as good as expected, I will design and later post a universal PA amp driver board to which you can connect a PSU and an output stage. Power output will be scalable from 50W to 400W per channel into 8 Ohms. Let's see, no promises made yet.

[OliverD]

This way of mouting the transistors allows for an easier single-sided PCB layout (paralleling devices without using bridges on component side). It is also easier to repair the amp, as you don't need access to the solder side of the PCB to change power devices. BD139/BD140 will have to be replaced by 2SA1507/2SC3902 as their voltage rating is right at the limit.

[PRR]

Very pretty. Love the heatsink.

> I sand-blasted it for the photo

Rough sand-blast surfaces may give poor thermal conduction. Grease is not as good as smooth metal-to-metal contact. I can't see the actual surface condition, but if you are stressing the transistors you might need to smooth the surface for best heat transfer. If you have been trained with a file, a smooth file can make a very smooth surface (or a very bad one if you are a clumsy worker). An electric (or hand) sander with Fine and Extra-Fine grit sandpaper might be better than coarse sand-blasting. A stainless-steel spoon, rubbed HARD over the transistor contact area, would peen-down the high-points and be better than a sand-blast surface.

[Upupa Epops]

I see, that you was inspirated by machines made by Accuphase with all their advantages and also absences .

[OliverD]

Rough sand-blast surfaces may give poor thermal conduction. Grease is not as good as smooth metal-to-metal contact.

Don't worry, for now I used tons of thermal grease and the transistors stay cool. It's only a mock-up after all. For the final product, I will sandblast the fins and simply mill the back afterwards.

[OliverD]

I see, that you was inspirated by machines made by Accuphase

How did you know?

[JojoD818]

AMT-freak,

Horray for single sided pcb design! I will watch the thread with great interest, thank you sir!

JojoD

[Upupa Epops]

I had many times open this machines when I was write tests for magazines .