Update on diagonal cross-hatching of polygons? Feature request!

[scott]

This topic was discussed in a forum post from almost a year ago.  is this on the future features list?  One of the comments was that 45-degree hatching it wasn't functionally different from 90-degree, but that's not accurate.  For controlled-impedance situations, where the hatched pour is the reference plane, it matters a lot.  Imagine that the traces are running left-right, i.e. 0 degrees; if the hatching is 0 and 90 degrees, you'd have some traces with a hatch strip under them, and others with mostly no copper under them, leading to very different impedances.  The 45-degree hatching is therefore commonly used in this situation.  Why not just use a solid pour in that case?  If the design is a flex circuit, a solid pour is often too stiff for the intended bending of the flex.

So, this is a plea for that feature to be added!

Thanks,

Scott

[jorge_garcia]

Hi @scott,

I'll put in the feature request for this. With that said, I think that value of the 45 degree hatching is being overestimated here. Going with the example you've stated of a trace moving from left to right, a 45 degree hatch would never follow the path of the trace correctly, the return path would be a zig-zag whereas the signal path would be a straight line.

In this scenario you would be better off using a localized solid polys to define the return path for critical traces and then a hatched poly for the rest of the board. In this way you avoid making the flex circuit too stiff but still have a better impedance profile. That's just my opinion though.

In any case, I'm always for having the option so like I mentioned I'll put in the request for it.

BTW. You can workaround this currently by rotating your entire board by 45 degrees. That will give you a diagonal hatched poly. I know it's extreme but it can solve the problem in the meantime.

Best Regards,

[scott]

Hi Jorge,

Thank you for entering the feature request. My understanding is that the behavior depends on the wavelengths of the signal vs. the physical dimensions of the copper features. For what I'm talking about, it's not a return-path situation, but rather a wave traveling in a transmission line, so it won't zigzag but rather it sees the effective RF impedance, which can be thought of as being made up distributed inductance and capacitance.  The 45-degree hatching ensures that all trace in, say, a high-speed bus, see a similar impedance.  The approach you're proposing can work in some situations, but not for a bus with a large number of signals spaced closely together.  Overall, it's nice to have more options, so hopefully the 45-degree feature gets added.  If nothing else, it's a feature which competitors such as Altium have offered for many years!

Thanks again,
Scott