Add Microchip_FCVG484_19x19mm_Layout22x22_P0.8mm and

[pthomas]

Microchip_FCG1152_35x35mm_Layout34x34_P1.0mm

These packages are defined for the PolarFire SoC's
in ug0902 here:
https://www.microsemi.com/document-portal/doc_download/1244577-ug0902-polarfire-soc-fpga-packaging-and-pin-descriptions-user-guide

---

I wasn't sure about the names, should the Microchip prefix be there? Should FCG & FCVG be there?
Here's the Microchip_FCVG484_19x19mm_Layout22x22_P0.8mm:

And the Microchip_FCG1152_35x35mm_Layout34x34_P1.0mm:

[codeclimate]

Code Climate has analyzed commit b73d747 and detected 0 issues on this pull request.

View more on Code Climate.

[myfreescalewebpage]

@pthomas comment about the script (I'm not 100% confident so if you have a reason why this is like that tell me):

Microchip_FCVG484_19x19mm_Layout22x22_P0.8mm

• pad diameter is 0.5mm +/- 0.05mm in the datasheet, not 0.4mm

Microchip_FCG1152_35x35mm_Layout34x34_P1.0mm

• overall_height should be 2.8 +/- 0.19mm
• pad_diameter: 0.64 +0.06/-0.14mm in the datasheet, not 0.51mm

@chschlue if you can have a look to :
mask_margin: 0.015 => this one is not clear for me and we find various values in the script for BGA packages. Is it computed from something else ?
paste_margin: 0.000001 => probably it should not be 0 ? Other BGA have the same value. Probably Ok even if I'm not sure what is the purpose of this field.

Joel

[chschlue]

0.000001 is a workaround addressing a limitation of the footprint format (0 in the file means "inherit from board setup" so if you actually want zero, you need to use a very small value instead.) Other than that, I believe @cpresser is more knowledgeable WRT the BGA generator.

[ghost]

First, it needs to be decided if this is going to be a IPC-compliant footprint, or should it be build according to vendor spec. O even mix both (which I think is horrible).
In this case, I don't see why it should be vendor specific. Its does not have any special properties, and the pitch is rather coarse.

Thus, I would suggest to

• specify ball_diameter instead of pad_diameter.
• keep all margins at 0 to use board defaults
• that will generate a NSMD footprint, as recommened by the manufacturer

To answer the original question. @chschlue is correct. The mask_margin will directly be applied to the footprint. It is not set on a per-pad-level.

[pthomas]

Thanks for looking at this everyone.

On Thu, Oct 1, 2020 at 6:45 AM cp-aquila ***@***.***> wrote: First, it needs to be decided if this is going to be a IPC-compliant footprint, or should it be build according to vendor spec. O even mix both (which I think is horrible). In this case, I don't see why it should be vendor specific. Its does not have any special properties, and the pitch is rather coarse. Thus, I would suggest to - specify ball_diameter instead of pad_diameter. - keep all margins at 0 to use board defaults - that will generate a NSMD footprint, as recommened by the manufacturer So if we do IPC-compliant and specify ball diameter what will the pad

diameter end up being? The tables lists the ball diameter as TBD, but I do see in the drawing it shows it, so it would end up looking like this for the 484: description: "FCVG484" ... body_size_x: 19.0 body_size_y: 19.0 overall_height: 2.92 pitch: 0.8 ball_diameter: 0.5 mask_margin: 0 paste_margin: 0 layout_x: 22 layout_y: 22

…

-Paul

[ghost]

You can omit the mask_margin and paste_margin parameters. Which is the same as having them at zero.