Author Topic: Gerbers different from PCB  (Read 898 times)

0 Members and 1 Guest are viewing this topic.

Online Ice-Tea

  • Frequent Contributor
  • **
  • Posts: 911
  • Country: be
    • Freelance Hardware Engineer
Gerbers different from PCB
« on: September 29, 2017, 07:03:10 pm »
Oy!

So, the Gerber files are just a little bit too different from the actual PCB (small 5x5mm thingie). I can easily solve the not-solid planes (it are tracks next to each other, adding another one to bridge the gap works) but I'd like to know why there's such a big difference..

But my bigger problem is that the arcs aren't exaclty arcs anymore. Any idea how to improve? Already tacked "software arcs", improved a bit but not enough...
« Last Edit: September 29, 2017, 07:06:19 pm by Ice-Tea »
An engineer never has a problem. He just needs more time.

FS: TTi TSX1820P, TDS754A (w SPC fail), Agilent Infinium 54815A, 54825A, R&S CMU 200 (multiple units, various options) UPL, SMIQ03,Tektronix CSA8000B, HP 8594E, 8595E, Marconi 6201B (8GHz), IFR 2390A (22GHz), 2383 (4GHz)
 

Online wraper

  • Supporter
  • ****
  • Posts: 7500
  • Country: lv
Re: Gerbers different from PCB
« Reply #1 on: September 29, 2017, 07:11:38 pm »
But my bigger problem is that the arcs aren't exaclty arcs anymore.
What makes you think there will be any difference on actual PCB? On such scale it's very unlikely that etching process will be so precise that there would be any sharp edges.
 

Online Ice-Tea

  • Frequent Contributor
  • **
  • Posts: 911
  • Country: be
    • Freelance Hardware Engineer
Re: Gerbers different from PCB
« Reply #2 on: September 29, 2017, 07:15:54 pm »
Mh, good point. But why are the solid surfaces not solid anymore?

Annoys me...
An engineer never has a problem. He just needs more time.

FS: TTi TSX1820P, TDS754A (w SPC fail), Agilent Infinium 54815A, 54825A, R&S CMU 200 (multiple units, various options) UPL, SMIQ03,Tektronix CSA8000B, HP 8594E, 8595E, Marconi 6201B (8GHz), IFR 2390A (22GHz), 2383 (4GHz)
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 10580
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Gerbers different from PCB
« Reply #3 on: October 02, 2017, 10:50:28 pm »
I've never noticed holes in polygons before.  Bug report?

Tim
Seven Transistor Labs, LLC
Electronic Design, from Concept to Layout.
Need engineering assistance? Drop me a message!
 

Online Ice-Tea

  • Frequent Contributor
  • **
  • Posts: 911
  • Country: be
    • Freelance Hardware Engineer
Re: Gerbers different from PCB
« Reply #4 on: October 03, 2017, 12:58:04 am »
Not polygons, though. Tracks side by side. Still should not happen.
An engineer never has a problem. He just needs more time.

FS: TTi TSX1820P, TDS754A (w SPC fail), Agilent Infinium 54815A, 54825A, R&S CMU 200 (multiple units, various options) UPL, SMIQ03,Tektronix CSA8000B, HP 8594E, 8595E, Marconi 6201B (8GHz), IFR 2390A (22GHz), 2383 (4GHz)
 

Offline dfnr2

  • Regular Contributor
  • *
  • Posts: 222
  • Country: us
Re: Gerbers different from PCB
« Reply #5 on: October 13, 2017, 07:16:46 am »
1) What gerber viewer are you using?  It may not be rendered 100% accurately.  For example, gerbv has 3 modes: "fast", "normal", "high-quality".  If you choose "fast" or sometimes "normal", you can see artifacts like this that disappear in "high quality" modes.

2) If you created the regions by juxtaposing tracks, try overlapping them a bit more.  It will make no difference in the PCB, but will cut down on the artifacts due to inaccurate rendering.

3) From the triangular void in your picture, it looks like the tracks are not all entirely parallel.  Perhaps some are 1 or 2 pixels off from top to bottom.  You can edit each track and check the X and Y coordinates of the endpoints, and fix.  Again, no real impact on the PCB, but may contribute to this artifact.

4) Why not just draw the outline of the region using primitives, then convert to a polygon. Then you can control the resolution of the arcs as well.   Also, it will result in a smaller, quicker rendering Gerber.
 

Online Ice-Tea

  • Frequent Contributor
  • **
  • Posts: 911
  • Country: be
    • Freelance Hardware Engineer
Re: Gerbers different from PCB
« Reply #6 on: October 13, 2017, 08:06:48 pm »
1) What gerber viewer are you using?  It may not be rendered 100% accurately.  For example, gerbv has 3 modes: "fast", "normal", "high-quality".  If you choose "fast" or sometimes "normal", you can see artifacts like this that disappear in "high quality" modes.

The Altium viewer.

Quote
2) If you created the regions by juxtaposing tracks, try overlapping them a bit more.  It will make no difference in the PCB, but will cut down on the artifacts due to inaccurate rendering.

Which is what I did in the end. Still sucks though ;)

Quote
3) From the triangular void in your picture, it looks like the tracks are not all entirely parallel.  Perhaps some are 1 or 2 pixels off from top to bottom.  You can edit each track and check the X and Y coordinates of the endpoints, and fix.  Again, no real impact on the PCB, but may contribute to this artifact.

This is the part that baffles. They *were* parallel.

Quote
4) Why not just draw the outline of the region using primitives, then convert to a polygon. Then you can control the resolution of the arcs as well.   Also, it will result in a smaller, quicker rendering Gerber.

Because using tracks immediatly gave the curves I needed. Otherwise, I'd have to lay a square *and* arcs. This was just a *lot* faster. I like fast. :D
An engineer never has a problem. He just needs more time.

FS: TTi TSX1820P, TDS754A (w SPC fail), Agilent Infinium 54815A, 54825A, R&S CMU 200 (multiple units, various options) UPL, SMIQ03,Tektronix CSA8000B, HP 8594E, 8595E, Marconi 6201B (8GHz), IFR 2390A (22GHz), 2383 (4GHz)
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf