That's essentially what I did, I've done plenty of soldering with delicate components, but none of them were surface mount. I suspect my 15/30W radio shack soldering iron gets way too hot and overheated the chip while I was trying to solder the ground plate onto the PCB, It's the same iron I've used for years, and have never burned anything out before.
Could be the iron, but if you've soldered lots of sensitive things I bet you'd see a trend by now. The CAT4101 is pretty durable from a soldering perspective. Read below...
Faulty as in, didn't work, and didn't pass the continuity test. Two of the drivers I ordered were DOA, the LEDs never lit using them, and the only component that tested faulty, was the CAT4101 chip. I checked the chips by checking for continuity from the ground pin to the other four pins, from what I understand, if there is continuity, the chip is fried. If there isn't, it's good.
If you're testing the chip in this manner
in circuit then you're not proving the chip itself is faulty, just that there's a short
somewhere on the circuit. I point this out because I fried a bunch of these due to a bad trace on a PCB. I suspected it was the chip so I replaced it, only to kill the new chip, too. Finally I tested the PCB with no components and cut leads until I isolated the short. Surely enough there was a bad trace that was shorting, it was never the chip(s).
Right out of the package the chips had no continuity, I checked them after tinning the leads and the ground plate, all but one were good, then I soldered them in place and checked them again, all but one tested bad. the one that tested good, tested bad after I soldered the leads to the PCB and slid it across the table. This boggles my mind because I wouldn't think I would get enough static from sliding PCB across a laminate top table while wearing a ground strap that was properly attached to a ground point. And I spent a less than half a second on each lead of the chip, so I know I didn't overheat it while performing that action.
Take the chips off and check for shorts on the bare pcb. It's possible that your theory is correct and you killed the chip(s) but it really sounds like you've got a bad pcb.
i know in this thread is has been mentioned but what PSU would be recommended ? Was it potrans?
I'd like to not see a failure since I'd be completely loading the transformer.
There are a bunch of different power supplies in use. Most of us use the open-enclosure adjustable style, like these:
http://www.mpja.com/24V-65A-150W-Power-Supply/productinfo/16032+PS/
Search eBay, they're pretty easy to get. If you're using CAT4101 drivers, the adjustability is a key feature. It's almost never mentioned in the specs explicitly but you can almost always see it in the product photos - look for a little trimpot next to the screw terminals.
I'm almost done soldering my driver. I do have a question though. Soldering the surface mount things was a real PITA. Is there some reason why those were chosen over "regular" capacitors and resistors that have legs that you solder through the PCB?
Thanks for all the work you guys have put into this. Letting others reap the benefits of your time and effort is a great thing to do. I know lots of us appreciate it.
Surface mount can be a huge pain compared to through-hole components if you're used to through-hole. But once you get your head around surface mount, IMHO it's easier and much faster. You don't have to flip the board, trim leads, or screw around. The LM3409 is only available in a surface mount package, so at that point, it just makes sense to use surface mount for other components on the board.
The key is modifying your approach and having a few handy tools ready. I use a pair of super-fine tweezers and sometimes a tiny blob of bluetack (the stuff you use to hang posters without thumbtacks). Depending on your skill with the tweezers, the bluetack can make things much easier. You put a TINY little dot of it on the end of the tweezers, or even on a toothpick, and then you can just touch it to a resistor or cap or other small component to pick it up.
Generally I solder the passives (caps and resistors) first unless the board is really tight around a key IC, in which case I'll do the IC first. Big stuff like inductors, transistors, terminals, etc is almost always last.
To start out, I'll wipe the board with rubbing alcohol.
The passive components all generally have two pads to solder. Before I solder anything, I'll march around the board and tin one pad on the pcb for each passive component. Then, I'll line up the bags of components in order and pull them out one by one. For each component, grab it with the tweezers by one end, and hold it in position on it's pads. Then, touch the soldering iron to the edge of the pad you tinned and it'll solder that half of the component down. I march around the board tacking components on by a single pad like this until all of the passive components are down. Then, I'll get the solder and march back around the board, soldering the other pad for each component.
ICs can be a little more challenging depending on their pitch (the distance between pins). Bigger ICs can be soldered by tacking one or two pins down, usually on opposite corners, then going back and soldering individual pins as you would the pads on the passive components.
ICs with smaller pins require very different approaches IME. The LM3409 is a great example of a chip with "smaller" pins. I have a hard time seeing the individual pins without my glasses on.
For this chip, I do the "swipe" method. I hold the part in place with tweezers and tin the tip of my iron with a (very small) blob of solder. Then, in one swift motion, I wipe the blob of solder across all the pins on a given side of the chip. This will almost always get each pin soldered down, but it will also almost always create gigantic solder bridges from pin to pin. So, afterwards, you go back over the chip with a piece of solder wick (copper braid made for removing solder). You lay the wick over the pins and heat it from the top with the iron and it wicks the extra solder off the pins, removing the bridges. Sometimes this requires multiple attempts.
The LM3409 has an exposed ground pad UNDER the chip that needs to be soldered down. In a production setting this is easily accomplished using solder paste and a reflow oven. Since we are all likely to be doing this by hand, we put vias (holes) under the part on the PCB. This allows you to get a big blob of solder to that ground pad from under the board. I do this last after everything else is in place.
There are some really good videos on youtube showing these techniques if pictures work better than words. Or, getting a friend who's done this stuff before can be helpful. This is part of what I wanted to show you Rochester guys at our LED meet but that seems to have fallen through.
