cool BGA replacement thing

Pretty big, a pain when you have a gazillion R and C around the periphery terminating and bypassing various and sundry.

Only thing better would be epoxy backfill to doubly hold it down and seal it, which of course is difficult to "sort-of-self-aligning". Or, if you want to get really out there, direct wirebond, but, yeah...

Tim

-- Deep Friar: a very philosophical monk. Website:

We always leave a little whitespace around a BGA so we can get our little inspection prism thing in there. We can see 3 or 4 balls deep, and spot any crud deeper down between rows. We don't usually use many bypass caps.

Sounds cumbersome. When replacing a BGA I clean the pads, add flux, position the chip, add a little more flux and heat the PCB from the other side.

are you using IR plates? I go a small kit of different sizes, I find it is a trial an error to get it right. I normally perform a time test with close examination of the solder melting on the board before actually applying the bga package. This seems to work well because I've found different board layer count gives different results to time required to heat the underside for proper ball hand soldering.

Btw, I've removed some bga's in a unique way, well maybe not unique to others here but is to me. I've taken a small steel wire and connected it to the supply to heat it up enough to melt solder. I pull this wire under the package while it is under current keeping the heat on. It plows the solder and disconnects the chip rather nicely most of the time.

I've also applied that trick to SOL types too!..

Jamie

I'd send them out to a company who specializes in such things. IMO, there's no way users can do enough of these to perfect and keep the necessary skills.

I recognise the SGS Thomson logo, but what is Compaq's name doing there?

That's neat! like hot wire cutting polystyrene foam wings for model airplanes...

Nope. A heatgun. First I preheat part of the board using the low setting and then I use the highest setting to quickly bump the temperature to the solder melting point.

Yes, I do that also, but I put a chimney around the area to keep the hot air enclosed. I also have a suction cup holder that I attached to my vacuum port on the station where I can either hold or pull the chip.

I guess it all depends what you have at the time of the job.

I've seen but not done, some one using a micro torch on the bottom side to reflow bad solder connections. This person has very good luck doing this.

Jamie

We place and solder, and rework, BGAs in-house. Our yield is way over

The way we rework them is to use an IR+hot air thing to pull the old chip. Then we stencil-paste the new chip (not the board) and place on the PCB pads, using a split-image optical alignment/placement gadget to lower the part precisely onto the pads. Then run through the reflow oven. My production folks like this new stickon thing, because they think the paste-and-place part will be easier. We'll see.

I'm thinking this stickie thing would be a good way to do small BGA runs and protos, especially for people who don't have the stencil and alignment/placement equipment. It could all be done by hand.

I'll let you know how it works.

The stick-on stencil demo came with that chip and the test board. It's a dummy/throwaway. There is a market for dummy solder-test chips, and I guess they buy up the industry's mistakes.

I wonder if a dummy 0603 resistor costs less than a real 0603 resistor.

I've never personally done it but at the last job, the rework percentage (above just another pass through the oven) was much closer to zero. Fortunately, rework of BGAs was a real rarity. The real problems were QFNs but that was mostly self-inflicted.

Please do. It's interesting.

Most of the PCB's that I have to work with have components on both sides. Under the BGA chips are usually a mess of bypass caps. Heating from the bottom will cause these caps to fall off. I protect the PCB with aluminum foil, use tons of flux, heat from the top of the BGA chip, monitor with an IR thermometer, and pray to my favorite diety that it will work.

For reballing, I've had limited luck with stencils, frames and fixtures. Instead, I just clean off the solder from both the chip and PCB, shield the area with aluminum foil, apply a drop of solder paste with a syringe to both chip and PCB, add flux everywhere, carefully position the BGA chip (that's the hard part), and apply heat with a hot air reflow wand.

I've also done some experiements with conductive silver two part epoxy, to just glue the BGA chip contacts to the PCB pads. I don't think it's going to work with high currents, but seems to with with low power small BGA chips such as Wi-Fi, audio, and LAN BGA chips. I haven't perfected the method quite yet, but it looks promising. The down-side is that the connection is mechanically fragile. While soldering has some strength, tolerates some board flexing, and can be reworked, epoxy is brittle, can tolerate little flexing, and is permanent. Ten minute work time is really not enough for a monster BGA array. While the surface tension of the solder tends to center the BGA chip in the middle of the pads (if the chip is unattached and floating), epoxy requires perfect positioning. Therefore, I suspect this method will be limited to small BGA chips and possibly repair work.

Just a thought... Are you using RoHS paste? That could be the difference in our experience. RoHS sucks.

We don't do that. Most people use way too many bypass caps. If we do put bypass caps on the bottom, we keep them a bit away from the ball array.

Sometimes we use ROHS. I hate the stuff. The 780-ball replacement will use regular solder.

What's on there now? Getting them off, without damaging the board or the part is harder than putting them back on. I agree, RoHS sucks. The first thing I do when I replace a component is get that crap off the pads.