AI Sol on the fritz

I'm already wringing my hands in anticipation. It looks like they are using a fairly fast switching reg for the driver (inductor is obvious, no big caps, must have internal switch since there isn't one present).

Those proprietary MCPCBs are interesting.

I am surprised there are no provisions for ventilation on the LED side of the heatsink. These have a big fan on the top, right?
 
Ooooh ooooh, these are my favorite moments, when you get to rip something like this apart and learn it's secrets :D
 
Unfortunately most of the "real" secrets are going to be code on the processor(s) up near the power jacks, and there's likely no way get it off or reverse engineer it. I'm more interested to see what chip they're using for the LED drivers.
 
I have a board as well, just got my replacement in today, free to whoever comes and gets it or can bring to the swap
 
Yeah I'll bring it to the swap, the one data port solder fell off which is why they sent me a new one, power can still be connected to it, I just noticed you were in buffalo, I'm downtown if you want to grab it earlier. Pm Me if that sounds good
 
Ha, I think that was the same thing Nate said about my board, something about the data port solders being sloppy.
 
Yup. It definitely looks like the solder joints on the data jacks are all cold. Clearly the rest of the board, being all small SMD parts, was reflow soldered, then the jacks were hand-soldered afterwards. AI needs to get their assembly house to fire the guy doing the hand soldering.

Cully, if you're interested, I may try to re-solder the jacks and we can plug it in to one of your units and see if it works. Then you'll have a spare.

FWIW the LED drivers are LM3404 ICs. This is in the same family as the LM3409 most of us DIY'ers are using now, except the mosfet is internal and it's not quite as flexible (1A limit vs 5A on the LM3409).

Once I get some close-up photos and a chance to analyze the rest of the components on the board, I'm planning on doing an "anatomy of a commercial LED driver" post over in the DIY driver thread. I'll post a link here when I get to it.
 
Yup. It definitely looks like the solder joints on the data jacks are all cold. Clearly the rest of the board, being all small SMD parts, was reflow soldered, then the jacks were hand-soldered afterwards. AI needs to get their assembly house to fire the guy doing the hand soldering.

Cully, if you're interested, I may try to re-solder the jacks and we can plug it in to one of your units and see if it works. Then you'll have a spare.

FWIW the LED drivers are LM3404 ICs. This is in the same family as the LM3409 most of us DIY'ers are using now, except the mosfet is internal and it's not quite as flexible (1A limit vs 5A on the LM3409).

Once I get some close-up photos and a chance to analyze the rest of the components on the board, I'm planning on doing an "anatomy of a commercial LED driver" post over in the DIY driver thread. I'll post a link here when I get to it.

Very interested in that post
 
I'm more interested in a cheaper way to replicate them for supplementation toward the back of the tank like we talked about. When I'm ready I'll have to consult with you about what pieces parts I should buy.

Having a spare would be nice though :thumbsup:
 
I'm more interested in a cheaper way to replicate them for supplementation toward the back of the tank like we talked about. When I'm ready I'll have to consult with you about what pieces parts I should buy.

Having a spare would be nice though :thumbsup:
 
Back
Top