BGA shifting

Dear all,

We've got a curious problem. We have had a major problem with BGA's occasionally being placed one to several rows or columns off the pads. We are using Universal's GSM2 platform for our general purpose placements. Our Universal rep here didn't have a ready answer for us and even after we followed their advice, we were still not getting any improvements. We've checked our component database settings, our lighting levels, our BGA's, the pickup tolerance, and nothing helped. Then the problem went away for about a half year. we never did find the answer to it. But now it's back again. We've had two boards with BGA's shifting on us in the past week.

Has anyone any idea what may be the cause of this problem? I'm open to any suggestions because we are baffled!

Thanks, Hon

reply »

| Dear all, | | We've got a curious problem. We have had a major problem with BGA's occasionally being placed one to several rows or columns off the pads. We are using Universal's GSM2 platform for our general purpose placements. Our Universal rep here didn't have a ready answer for us and even after we followed their advice, we were still not getting any improvements. We've checked our component database settings, our lighting levels, our BGA's, the pickup tolerance, and nothing helped. Then the problem went away for about a half year. we never did find the answer to it. But now it's back again. We've had two boards with BGA's shifting on us in the past week. | | Has anyone any idea what may be the cause of this problem? I'm open to any suggestions because we are baffled! | | Thanks, | Hon | Amazing coincidence. I just had such a conversation with our pnp tech. She had the same problems repeating recently both with plastic types and our ceramic BGA's and CCGA's.

The first problem was oxidized solder balls. This, as you already know, has the affect on cameras and lighting of not allowing proper recognition, alignment, and placement - of a row or two sometimes.

For the ceramic types, some operators were not aligning to edges properly and at times even picking up adhesive "creep" oozed out from under metal heat sinks on top. This too caused serious problems - though neither may apply to your concerns.

Earl Moon

reply »

| Dear all, | | We've got a curious problem. We have had a major problem with BGA's occasionally being placed one to several rows or columns off the pads. We are using Universal's GSM2 platform for our general purpose placements. Our Universal rep here didn't have a ready answer for us and even after we followed their advice, we were still not getting any improvements. We've checked our component database settings, our lighting levels, our BGA's, the pickup tolerance, and nothing helped. Then the problem went away for about a half year. we never did find the answer to it. But now it's back again. We've had two boards with BGA's shifting on us in the past week. | | Has anyone any idea what may be the cause of this problem? I'm open to any suggestions because we are baffled! | | Thanks, | Hon | I have seen so many things that could happen during automatic assembly and at the first sight no logical reason. Easiest are those malfunctions that happen all the time, I hate that stuff that occurs only now and than unpredictable. So we had sudden loss of accuracy only help was to reboot the system . Defect was the ram bank sometimes loosing one dataline. Low vacuum causing angle misalignment, strong enough to hold the part but not to counteract the rotationforce. Vision-missreadings because of lightchanges, very hard to detect, occured only if certain lamps where on or visitors shadowed the normal light. So in each case you have to consider all influences of your specific surroundings, take a close look at the error if there�s any hint on what maybe could have happend to the part ( is there a smear of paste? ). List all possible causes and check one at a time, the easiest first. You are the person with the best suppositions to find the cause. If you have the sparetime take some sample components, a sample-board with translucent tape and a small programm and ran it over and over.

Good luck and enough coffee ( or do you prefer tea?)

Wolfgang

reply »

| Dear all, | | We've got a curious problem. We have had a major problem with BGA's occasionally being placed one to several rows or columns off the pads. We are using Universal's GSM2 platform for our general purpose placements. Our Universal rep here didn't have a ready answer for us and even after we followed their advice, we were still not getting any improvements. We've checked our component database settings, our lighting levels, our BGA's, the pickup tolerance, and nothing helped. Then the problem went away for about a half year. we never did find the answer to it. But now it's back again. We've had two boards with BGA's shifting on us in the past week. | | Has anyone any idea what may be the cause of this problem? I'm open to any suggestions because we are baffled! | | Thanks, | Hon | Hon, There is a few things you might want to try. First you need to find out what algorithms your machine is using. You can use Bump process A and set the minimum inspection to 100% but you need to realize this does not actually check for 100%. You are probably going to have to change to Bump process E. When doing this, depending on what type of camera you are using (i.e. 4-mil, 3-mil, ect. ) try to side light the part using a low light level( <1 ). With side light you will get rid of the light that is reflecting from the sides of the part. I don't believe it will work on a 3-mil or 1-mil. Another thing you need to check is whether or not you are using a PTF. If you are is the parts that are being shifted larger than the others. If this is the case you might want to check if your PTF has been retro-fitted with the Theta motor. If it has, and you have pre-rotate checked in the PD, you might need to get a bigger PTF nozzle. The parts being picked off center is normally the reason why they are placed off center. You might want to slow down the pallet transfer speed and the belt transfer speed. If none of this helps you can always call it a chip! ( just kidding )

reply »