Roland AC-33 low output

[drewl]

Supposedly 10 - 15W to each speaker, at most I can push maybe 5w to each.

It looks like the MAX9744 flat pack 40 some pin class D amp is bad. Nice clean sine wave going into it power amp from the headphone board.
Output to the speakers looks like only the top of each sine wave is coming out.

Anyone work on these yet or have the same problem?

[drewl]

Dang, the MAX9744 power amp is only around five bucks, but it has 44 pins, 11 per side in a .2" space.

it's between the silver surface mount caps in this fuzzy photo.

[Chuck H]

Dayam! I saw a tutorial on soldering such things once. Made it look easy. I'm still intimidated by it.

https://www.youtube.com/watch?v=tlSY1uaw0GA

https://www.youtube.com/watch?v=9Sfggumc-Tc

[Enzo]

I tried that first method once, and promptly tore up a couple traces. I know it can be done well, but I screwed it up. Not happy

The second one works as well as it seems, assuming you do it well.

[Mick Bailey]

We've discussed this elsewhere, but Chip Quik is the easiest way to get them off. The lead version if you can get it - much better characteristics.

I find that SMD rework flux + very fine tin/lead/silver solder is the answer to re-soldering. Much cleaner job and no solder bridges.

[drewl]

I have the equipment for micro stuff like this at work.
Getting the old chip off is easy, resoldering the new one is definitely going to be hairy, even with a microscope.
At least this one is easy to get to.

Some of the boards on the Tubemeister 36 in this week look like they are impossible to get to and service.
They probably toss it and replace the board.

[guitician]

We used to use a solder paste, solder dust/flux, and screen print it only were needed. Then lay the part onto the pasted landings and heat with vapor until it melted. I guess you could use a heat gun too depending on the parts tolerances.

[drewl]

Well like I said getting the old chip off was easy, installing the new one a royal pain.
It's the type of chip with no protruding leads, so you need to use a ton of flux, spend a while lining it up under a microscope, then you just run the iron tip along the pads while feeding small amounts of solder.
Then you clean, check under the microscope for pads that didn't get soldered, and repeat this process over and over until all leads are soldered.
A half hour later, all leads soldered making good contact.

Aaaannnndddd....the problem is still there.

Need to read up on class D amps to see why.
Must be something in the output wave shaping components.
About a dozen little surface mount caps that take the output and convert it to a clean some wave.