I only do mils to 2 places normally, but you can go 3 or 4 by setting an env variable (or user preference). Patrick Westfeldt, Jr. 720-406-0887 _____ From: icu-pcb-forum-bounce@xxxxxxxxxxxxx [mailto:icu-pcb-forum-bounce@xxxxxxxxxxxxx] On Behalf Of Ed Caldwell Sent: Monday, May 22, 2006 7:39 AM To: icu-pcb-forum@xxxxxxxxxxxxx Cc: icu-pcb-forum@xxxxxxxxxxxxx; icu-pcb-forum-bounce@xxxxxxxxxxxxx Subject: [PCB_FORUM] Re: mysterious via shifting Hi Patrick, You are probably on the trail... I placed the vias in the BGA footprint in mm (0.8mm). The shift is likely derived from the rounding to mils in the tool which may have occurred when I experimented with the "Fanout By Pick" command. I never ran fanout on the BGA, my experiment was on a bypass cap for another IC. I also tried to add another decimal point to the database but Allegro would only allow 2 for mils. Oh the woes of unit conversion! I was curious if anyone else had similar experience... I opted NOT to use the "Fanout By Pick" command in the master copy and the 0.8mm BGA is holding up so far. Thanks for the input. "westfeldt" <westfeldt_nbcd@xxxxxxxxx> Sent by: icu-pcb-forum-bounce@xxxxxxxxxxxxx 05/21/2006 08:21 AM Please respond to icu-pcb-forum@xxxxxxxxxxxxx To icu-pcb-forum@xxxxxxxxxxxxx cc Subject [PCB_FORUM] Re: mysterious via shifting I'm not sure when Allegro decides it has an unconnect, but if you fanned out, copying at 31.5 mils, 5 copies gets you a .02mil error. I don't really see how this is the mechanism for your problem, but you do have a tiny error in your 31.5mils spacing, should be 31.496. Patrick Westfeldt, Jr. 720-406-0887 _____ From: icu-pcb-forum-bounce@xxxxxxxxxxxxx [mailto:icu-pcb-forum-bounce@xxxxxxxxxxxxx] On Behalf Of Ed Caldwell Sent: Friday, May 19, 2006 6:11 AM To: icu-pcb-forum@xxxxxxxxxxxxx Subject: [PCB_FORUM] mysterious via shifting Anyone experience a mysterious shifting of vias in Allegro 15.2 s126 (v15-2-51EU) (12/14/2005) i86... ? The design is in mils to 2 decimal places. The vias are in a 0.8 mm BGA (spaced at 31.50 mils apart). Somehow the group of vias within the rectangle pictured below shifted by 0.10 mils sometime during my edits - cause unknown... It is important to note that I did not change the database resolution - it has remained (mils to two decimal places). The only thing I changed was the routing grid which I switched between 25 mils, 5 mils and 0.1 mils as I was working in different areas of the layout and I was not working in the area where the vias shifted. Puzzl_Ed Ed Caldwell Electronic Design Services, LLC Canton, GA - USA