board stretch and IPC -D-300G

Subscribe to Thread

We've got a problem with a fine pitch board that has about .... - Nov 12, 2001 by Steve Thomas

Yes!!! How can we assemble boards with 20 pitch when we all... - Nov 12, 2001 by davef

Thanks, Dave. Three courtesy cheers for specifications ... - Nov 12, 2001 by Steve Thomas

Maybe. While it doesn't help with your current problem, t... - Nov 13, 2001 by davef

Steve Thomas

#18150

board stretch and IPC -D-300G | 12 November, 2001

We've got a problem with a fine pitch board that has about .005" registration error at each end (towards the center of the board) when it's printed. Apparently we've got a case of board stretch generated at the fab shop. I still have to do some measuring to verify this.

Anyway, IPC-D-300G says that the Class C (tightest) conductive pattern location tolerance is .008" over 12", which would allow .004" at each end when centered. Still seems like a lot to have to deal with on a fine pitch board. Is this what the industry (you guys) typically sees, or is this extreme?

The other question is, what do you do? I know I can order a custom stencil to match the boards, but is there a chance that the next batch will be undersized? Is the chance of their artwork shrinking just as good as the chance of it expanding? I have no idea what kind of environmental controls they have on that process, but obviously it deserves some investigation. Thanks.

reply »

davef

#18155

board stretch and IPC -D-300G | 12 November, 2001

Yes!!! How can we assemble boards with 20 pitch when we allow 0.008� error to the master art? Excellent question!!!

IPC-D-300G has nothing to do with assembler requirements. It is written by and for fabricators, so that they can be comfortable.

Being at the 0.008� limit is extreme. Your supplier is too comfortable. Sure they can say they met spec, but that sure doesn�t meet your needs. You�re paying the price of your buyer being a hero [and probably getting a BIG bonus] by reducing the bought-in price on that board. Check some boards from your other suppliers. They probably are not as bad as these.

If your supplier is hanging-out the upper spec limit, there�s no telling what they�ll do in their next shipment.

I want to say Mikie was frothing about this within the past two months on SMTnet. Search the fine Archives.

In this stretch effect, the larger the board, the more severe the growth (or shrinkage) it will have. Stretch also depends on the board house capabilities and controls and board thickness. The best way to minimize stretch effect is to group components that require tight tolerance into the central location of the PCB.

Other contributing factors are: * Stencil stretch: 0.001" per inch for chem etched stencils was the best we could get a few years back. * PCB warpage: Contributes to the problem. * PCB solderability protection: The thermal gradients seen in the HASL process contribute to both warpage and stretch. If using a HASL finish, consider OSP or ENIG finish. * Screen print: Vision system on the printers? Optical alignment of stencil pattern (round half-etched fiducials on the bottom side of the stencil seems to work well) to board pattern is crucial. Of course foil thickness, aperture design, and squeegee material all play a role.

reply »

Steve Thomas

#18157

board stretch and IPC -D-300G | 12 November, 2001

Thanks, Dave.

Three courtesy cheers for specifications written to reflect the capabilities of the part suppliers, even if they fail to meet the purchaser's (not unreasonable) process req'ts.

In the meantime, I'm sure my stencil guy would be more than happy to make 3 ea. for every board over 12" long, cover the potential error from our pcb fab. guy. Maybe they're cousins?

reply »

davef

#18163

board stretch and IPC -D-300G | 13 November, 2001

Maybe.

While it doesn't help with your current problem, there is nothing that says that you cannot impose requirements in your purchase order that are stricter than D-300.

reply »