Hi,
First I was so glad to see that I ordered 3 pcb and received 6. I was telling myself that my round board let fill the panel faster and everybody on the same panel use to have their PCBs faster…
Than, a friend engineer told me that would probably mean that they were some errors on 3 tries and Batch send them to me anyway… sounds fair but after a closer look only 2 on th3 I’ve order seems to don’t have shorts on them… and I ain’t no finish my inspection…
So, not very glad of my move at Batch. I know, someone will say it’s only a prototype… but I was expecting at least 3 on 3, not 2 on 6…
Somebody else on panel 764 had the same problems ???
Seb
… heu… Finally… 5 on 6 with shorts… :evil:
What was your trace-to-X spacing (X is trace, plane, via, pad, etc)? What about X-to-board edge? Yes, it must have met minimums to pass the DRC, but if they are AT minimum, that could cause the problems.
Can you post pictures of your boards and gerbers? For years I’ve been ordering from BatchPCB and Gold Phoenix (the board house BatchPCB uses) and have never had a problem with a board that wasn’t my fault.
-Bill
Hi,
Batch requiring at least 8 mil of clearance and I set 10 in my design rules to be sure… I’ve use the same rules for the polygon planeIf they were only one spot where I get a short it would be fine… but it’s at opposite place on the board… I can undertand a “general offset” of all a layer but in my case, that was the clearance between a std 2x3 pins header vs the ground plane…
The other places are with 2 different vias (one isolated thermal) and the ground plane…
Anyway, I’ll try to send some pics for fun in the next days…
Seb
I get more boards than ordered pretty routinely. I generally assemble all of them (why waste?) and I’ve only found one major problem in one so far with a broken trace.
Then again, mine aren’t very complicated.
tjncooke:
What was your trace-to-X spacing (X is trace, plane, via, pad, etc)? What about X-to-board edge? Yes, it must have met minimums to pass the DRC, but if they are AT minimum, that could cause the problems.
Would that mean that a grounded pour polygon should not be run up to the board edge-- that the pour must be 0.008 or more away from the board edge?
analogon:
Would that mean that a grounded pour polygon should not be run up to the board edge-- that the pour must be 0.008 or more away from the board edge?
Yes! Just keep in mind that the isolation between an entity and the edge of a board is usually larger than the minimum trace to trace spacing elsewhere on the board.
-Bill
analogon:
Would that mean that a grounded pour polygon should not be run up to the board edge-- that the pour must be 0.008 or more away from the board edge?
At least in Eagle, if you have your DRC set up properly, the pour will be automatically kept away from the board edge, even if you draw it right to the edge. I like to draw my ground polygons 25 mil away from the edge, but the SparkFun DRC has a copper to dimension of 10 mil.
One reason for keeping planes away from the board edge is that you want the edge to be insulating. Copper at the edge may not be covered by the stop mask and thus could short to an enclosure, components lying on your workbench, etc. Also, if the board house used a shear to cut boards apart (BatchPCB does not!), copper at the edge will create whiskers. If a top and bottom feature exist at the same edge location, they could easily short together!
Sorry for the hijack…
I appreciate the need to keep a copper-dimension clearance, but in my case I also want to maximize the usable board area while still having a precise outline to fit in a small handheld enclosure. It appears that in the dimension layer, the dimension line has a thickness and the board will be cut to the outside of this line. So the edge clearance becomes slightly larger, and so does the overall board size, and I end up filing the board to fit the case 
Are you saying this as fact or conjecture? I don’t have any BatchPCB boards with me right now to check, but I don’t remember seeing the copper outline on the actual boards. This would imply that they were routed to the inside of of the dimension line.
Mind you, I haven’t remeasured my boards, as finished size is not a concern for me atm.
Sorry, the last part was conjecture. I have put my design through the DRCbot but have not ordered a PCB yet. The DRCbot appears to take the outside of the dimension line, I am not sure that is definitive as I know there are also manual steps later.
I really should have phrased the question “How is the final board cut with respect to the dimension line?”
So, now I am home and have access to my boards… I used a dimension line of 0mil, and see no copper around the board. The board was 2.5" wide in Eagle, and measures exactly 2.5", maybe a smidge less. Hope this helps.