Printed Circuit Board Assembly & PCB Design SMT Electronics Assembly Manufacturing Forum

Printed Circuit Board Assembly & PCB Design Forum

SMT electronics assembly manufacturing forum.


Another in a long line of Philips Topaz questions....

Views: 3683

#52031

Another in a long line of Philips Topaz questions.... | 11 October, 2007

We purchased a used machine and can't get the block fiducial function to work properly (PCB fids work fine). I can teach them and the machine will use them but the correction is worthless. We first tried it on a 2x2 panel where the top two boards were turned 180� from the other two and it was not pretty. We haven't tried it since. I suppose it's possible that it can't handle the rotation, but why wouldn't it?

The workaround is to nail down the first board, then just use PCB fiducials and tweak the block repeat info until they all get placed accurately. It works, but it's a waste of time and I can't believe that's the way it's intended to work. Are we doing something wrong or is there something wrong with the machine, a setting, or what?

reply »

#52034

Another in a long line of Philips Topaz questions.... | 11 October, 2007

Have you created block offsets for each board in the panel? We have rotated baords in panels but have not had to use block fiducials, we just use board fids. We place 0402s and 16mil fine pitch on multiple boards in a panel with boards rotated 180 from each other. Also is you system a "DOS" based or windows based system? If you like I can send you a working file of one of our boards you can use as an example. Let me know.

reply »

#52035

Another in a long line of Philips Topaz questions.... | 11 October, 2007

Yeah, James, we've got block repeats set up so we're getting the boards built with just pcb fids, and very accurately.

I want to be able to build with block fids, though, in case we have any issues with the panelization. It's happened in the past and I suspect it'll happen again.

The machine just doesn't correct accurately with block fids and so we end up with 1 board placed dead on and 3 with major misalignments. I find it hard to believe that this is normal or by design (what else would block fids be used for?).

It's a DOS system, by the way.

reply »

#52036

Another in a long line of Philips Topaz questions.... | 11 October, 2007

In Block repeat information did you put the rotation in before optomizing? This is where the machine gets the rotation of these individual boards. The Block fiducials are good to use, but they aren't for defining rotation. If you optomized using note data you can go back to a single image. teach your block fids on that first image and then make sure all your placements are dead on, then reoptomize (with the rotations typed in to the "R" column in block repeat information, as long as you taught them to the same board origin when you optomize this should have all the fids in correct location. (maybe that is what you meant by "tweak the block repeat info", not sure), but this is how I have always done it, and I have ran one of these machines for going on 8 years now.

reply »

#52037

Another in a long line of Philips Topaz questions.... | 11 October, 2007

Yep, we did it just the way you described, I think. What we did NOT do was a block conversion. We just built it a board at a time, to see if things were accurate when we first programmed it. They weren't, so we did away with the block fids, adjusted the block repeat locations to get all boards working the same, THEN did the block conversion.

I should clarify that the placements were close, but not acceptable, as in the corrections are somewhere in the neighborhood of .5 mm or less off. It's not some wild random scattering of components, just not usable.

reply »

#52049

Another in a long line of Philips Topaz questions.... | 12 October, 2007

We always do block conversion and have never had a problem. Is there a problem with the accuracy of your PCBs from the vendor? If you can get a panel placed correctly the next panels should also place correctly. I just created a 2x8 panel with half of the boards rotated 180. I create 1 board and the board offsets with rotation and then distribute the parts and badmarks across the PCB. It works perfectly every time unless I have given it bad offsets.

reply »

#52063

Another in a long line of Philips Topaz questions.... | 12 October, 2007

Hope that your PCB is supported evenly, I mean really plane to the conveyer rail. If not this may warp the PCB upwards, causing fiducials to be identified not accurately. Also I suggest you to check Mark Info of the fiducials and have a look at the tolerance (normally 30% works fine). Just in case it won't work I am more than available to come over and solve your problems.

reply »

#52085

Another in a long line of Philips Topaz questions.... | 16 October, 2007

Are you converting your block fids to local fids or just not using them?

reply »

SMT fluid dispensing

See Your 2024 IPC Certification Training Schedule for Eptac