SMD pads and tracks misaligned in Eagle on mm grid

Archive BatchPCB
1

I have recently attempted to use Eagle to lay out a simple breakout board for an LCD module that uses a 30 pin 0.5mm SMD flat cable connector. When I printed the layout I noticed a really peculiar problem: the tracks to the connector’s pads are not exactly aligned to the pads - they are offset by a small amount. After some checking and rechecking, I still don’t know what the problems is. Here are some things I am pretty sure I got right:

  • - the connector's pads are aligned on a 0.5mm grid; the pads are exactly 0.25mm wide
  • - the tracks on the board are 0.25mm wide. Initially I used an inch grid for the board and only aligned to the pads by starting the tracks from them, but since I have also tried a mm grid on the board with no difference to the outcome
  • - if I redo the part footprint on an inch grid (0.508mm spacing between pads, 0.254mm pad width), and use 0.254mm tracks, the problem disappears (but those numbers are outside of the tolerances specified in the part datasheet).
    • All in all, I doubt that this would be a problem for the board, but it's just annoying.

    Any ideas are much appreciated.

    Regards,

    Tomislav

    2

    I’ve seen something similar from autorouting. what happens if your ripup the trace coming out of the pad and redo it starting from the pad?

    If the part was from an eagle lib, you will need to take a closer look. I don’t trust the eagle libraries until I’ve looked through them. Make sure the device origin is aligned to the same grid as the pads. (use the lib editor and edit the device).

    I’ve seen weird misalignments occur due to switching between english and metric.

    Also, make sure that when you drop the part, the grid is .5 mm

    3

    Another thing to watch for is that you’ve actually aligned the part to the grid when you placed it - hold down Ctrl when you click the part with the Move command and it’ll snap the mouse to the parts origin; otherwise you may be moving the part by grid-increments but offset from the actual grid.

    I may have seen something similar to this actually, assuming it is a problem, however I didn’t pay much attention to it as it was about a 0.01mm offset on 0.2mm traces/pads. I figured it was close 'nuff. :slight_smile:

    4

    Hm. I tried posting a reply but got ‘forbidden word’ error. No idea which one, though - everything looks ordinary enough. I’ll try it in pieces…

    5

    The traces were manually routed, but I have tried autorouting as well and it produces the same wrong results.I have ripped out a misaligned track too many times to count and rerouted it in many different ways. No improvement.

    Now, the original connector part is something I have done myself, but I have started from an existing Eagle part (XF2H-3015-1LW from con-omron; mine is slightly different: XF2M-3015-1A). I think it’s correctly aligned to the same grid as the pads.

    6

    However, it probably doesn’t matter any more since I have later simplified the whole setup by creating simpler footprint (for an imaginary part) and used that to test out a few things, all with the same consistent results: if the pad spacing and size is a whole number in millimetres, the tracks are offset by a small amount (regardless of the board’s grid). If the pad spacing and sizes are in inches, all is fine.

    7

    I have also tried using Ctrl key to align the part to current grid, but still no difference.

    As for how much of an offset it is, it’s difficult (was:hard - produces a warning) to tell. It is definitely small. I have since made the breakout board in question and it looks like it will be usable (though offsets are visible). For prototyping like this it doesn’t matter, but once I send the board (to be) made properly, it may look unprofessional. ← without the paretheses around ‘to be’ I get the warning. What is happening here? Anyone?

    Text of the warning:

    Usage of forbidden word

    We’re very sorry, your post request hold for moderation.

    A site administrator will be notified and will approve your post.

    8

    The traces were manually routed, but I have tried autorouting as well and it produces the same wrong results.I have ripped out a misaligned track too many times to count and rerouted it in many different ways. No improvement.

    Now, the original connector part is something I have done myself, but I have started from an existing Eagle part (XF2H-3015-1LW from con-omron; mine is slightly different: XF2M-3015-1A). I think it’s correctly aligned to the same grid as the pads.

    However, it probably doesn’t matter any more since I have later simplified the whole setup by creating simpler footprint (for an imaginary part) and used that to test out a few things, all with the same consistent results: if the pad spacing and size is a whole number in millimetres, the tracks are offset by a small amount (regardless of the board’s grid). If the pad spacing and sizes are in inches, all is fine.

    I have also tried using Ctrl key to align the part to current grid, but still no difference.

    As for how much of an offset it is, it’s hard to tell. It is definitely small. I have since made the breakout board in question and it looks like it will be usable (though offsets are visible). For prototyping like this it doesn’t matter, but once I send the board to me made properly, it may look unprofessional.

    9

    Generated gerber file is perfect - no misaligned tracks. Something in Eagle’s printing (conversion to postscript?) routines is at fault maybe…

    Also, running the layout2 cam job to generate an EPS file first and then printing that has no misalignment problems. Only printing directly from Eagle. This is now good enough for me.

    Thanks a lot to people who came up with suggestions.