15G PQFP

First of all let me say that we are currently placing other 240 pin QFP'S with no problem. But none of them have the MASS of this beast. The part is a XILINX XC5215-6HQ240C which has a large metal heat sink covering most of the bottom of the part. The part is 32 mm sq., 3.65 mm high, and 15G in weight. It is 240 pins, pitch=.020", leads=.010". The pads on the board are also .010" and the stencil thickness is .006" with a 10% reduction on the apertures. The solder is on the pads and the part is aligned to the pads with in 10% before going into the IR oven. Our problem is when the board comes out of the oven we have a strange skewing problem that I have never seen before. The part is not twisted, that is the leads on all four sides misplaced the same amount, but rather about fifty percent of the parts are okay and the others are off the pads by 25% or more. This occurs on only one, two, or three sides at random. In general it seems to get progressively worse as we move from the first pin on a side(which is on the pad) to the last pin(which is not on the pad). What can the problem be? --Parts with a large mass do not self align? --Pad size to small? --Oven profile must be changed? --All of the above?

reply »

(snip)...The solder is on the pads and the part is aligned to the pads with in 10% before going into the IR oven. Our problem is when the board comes out of the oven we have a strange skewing problem that I have never seen before. The part is not twisted,...(snip)....In general it seems to get progressively worse as we move from the first pin on a side (which is on the pad) to the last pin (which is not on the pad). Hi there Jim! I've had something similar to this, and in fact, it maybe exactly the same problem I've seen before... The problem is a design issue...and I know there's gonna be a few people that are intimately involved with this assembly that probably don't want to hear that, but I'll almost betcha' when they really look into things, they'll be saying; "Oh crap, Jim's right!" What makes me think your problem is just like the one I've seen before is your statement; " In general, it seems to get progressively worse as we move from the first pin on a side (which is on the pad) to the last one (which is not on a pad)" I've had that same problem before with QFP's as well as TSOP's...all .020" pitch too. What the problem turned out to be was that the original PCB CAD file was in millimeters and got imported into a PADS system to update and change a few things on the board. The exact problem was that the designer had his default CAD environment set in inches with all the coordinates defaulting to 2,4 (two digits leading the decimal place, and then four digits trailing).

But to convert millimeters to inches accurately, you need to go out five digits after the decimal point (MM into Inches: multiply by .03937) So PADS either ignores the fifth digit after the decimal place or rounds it off somehow. (I'm not a designer, so I'm not absolutely sure precisely what goes on), but what you wind-up with is exactly what you've described; the first pin on a side will be dead nuts on, and as you go down the row it gets further and further off the pad. Like I said, there probably won't be very many people too pleased with what you tell them, and they'll probably try as hard as they can to try and prove it ain't so, but I'd be willing to bet that's what your problem is... Sorry I didn't have better news... -Steve Gregory-

reply »

I have to agree with Steve. If the problem is progessively worse the problem is the pitch conversion that was used, and I have seen it before. It may be masked by the solder paste which could be on pitch but after reflow the paste wickes to the pad carrying the part with it. Try placing on a clear double sided tape, carefully checking alignment. | First of all let me say that we are currently placing other 240 pin QFP'S with no problem. But none of them have the MASS of this beast. | The part is a XILINX XC5215-6HQ240C which has a large metal heat sink covering most of the bottom of the part. The part is 32 mm sq., 3.65 mm high, and 15G in weight. It is 240 pins, pitch=.020", leads=.010". The pads on the board are also .010" and the stencil thickness is .006" with a 10% reduction on the apertures. The solder is on the pads and the part is aligned to the pads with in 10% before going into the IR oven. | Our problem is when the board comes out of the oven we have a strange skewing problem that I have never seen before. The part is not twisted, that is the leads on all four sides misplaced the same amount, but rather about fifty percent of the parts are okay and the others are off the pads by 25% or more. This occurs on only one, two, or three sides at random. In general it seems to get progressively worse as we move from the first pin on a side(which is on the pad) to the last pin(which is not on the pad). | What can the problem be? | --Parts with a large mass do not self align? | --Pad size to small? | --Oven profile must be changed? | --All of the above?

reply »

This occurs on only one, two, or three sides at random. But I'll bet that the "good" sides have the same problem ,it's just that the center pins of the array are centered and it gets progressively worse towards each end, just not bad enough to be rejected. Metric conversion raises its ugly head again.

reply »

"...the part is aligned to the pads with in 10% before going into the IR oven." Sounds like the pad pitch might be okay (?). If that's not the problem, then what? I wish I had the answer, because I've struggled with heavy QFPs for some time now (2 different parts -- one 208pin, one 240pin, both having heat sinks molded into the package). My problem is usually obvious before reflow -- some (but not all) of the placements look poor coming out of the pick and place machine. I've eliminated many of the obvious ideas -- pad pitch, nozzle slip, fiducial recognition, board warpage, and many other factors. Calibrating our placer helps, but it doesn't eliminate the problem. Is our paste inadequate for holding a heavy component? My next step is a more rigorous, designed experiment. Has anyone studied the relationship between component mass and solder paste tack? Perhaps I should talk to my paste supplier...

reply »

| First of all let me say that we are currently placing other 240 pin QFP'S with no problem. But none of them have the MASS of this beast. | The part is a XILINX XC5215-6HQ240C which has a large metal heat sink covering most of the bottom of the part. The part is 32 mm sq., 3.65 mm high, and 15G in weight. It is 240 pins, pitch=.020", leads=.010". The pads on the board are also .010" and the stencil thickness is .006" with a 10% reduction on the apertures. The solder is on the pads and the part is aligned to the pads with in 10% before going into the IR oven. | Our problem is when the board comes out of the oven we have a strange skewing problem that I have never seen before. The part is not twisted, that is the leads on all four sides misplaced the same amount, but rather about fifty percent of the parts are okay and the others are off the pads by 25% or more. This occurs on only one, two, or three sides at random. In general it seems to get progressively worse as we move from the first pin on a side(which is on the pad) to the last pin(which is not on the pad). | What can the problem be? | --Parts with a large mass do not self align? | --Pad size to small? | --Oven profile must be changed? | --All of the above? In the good old days (last year or the one before), closely populated devices having considerable thermal mass were drawn togehter in IR type ovens. You did mention IR oven, didn't you - though I'm sure this is not of the direct type questions do arise. Anyway, though you cannot populate QFP's as closely as PLCC's, BGA's, and others, there might ought to be more thought given to profiling and/or design placement proximity. Earl Moon

reply »