ESP8266-12E: Wrong pinout in schematics

[tobiassoltermann]

Great library, thanks for the amazing work. There's just one thing I've discovered. The PIN number order from 9 to 22 in the schematics does not correctly match to the one in the footprints.

The datasheet tells that pin ordering is the form of an U. It starts on top left on 1 then goes down the left edge (that part is okay). Then PIN 9 to 14 continue on the bottom edge and then go up again on the right edge from 15 to 22.

Here's the schematics (label ordering is correct, pin numbers are not)

Here's the footprint (numbering is correct, labels are not)

Pinout as it should be (unfortunately rotated by 180°):

Thanks for looking into it.

[jdunmire]

Hi Tobias, Thank you for the complements and feedback. While it is true that the pin numbering differs from the data sheet, it is internally consistent within this library. So if you use the schematic symbol here it will work properly with the footprint here.

The pin numbering was chosen specifically to allow existing designs to upgrade from a ESP-12 to and ESP-12E/F without redrawing the schematic.

This was discussed last March when tadas-s submitted #38. After the discussion I rejected the pull-request for two reasons: First and foremost, there are existing designs using the ESP8266-12E part from this library which the pull-request would break, and second the useful feature of being able to easily upgrade old designs.

I am still open to adding a component with the vendor pin numbering if we can come up with clear naming to make sure that the schematic symbols and footprints don't get mixed up.

I will leave this open for a bit so that you can respond if you like.

Thanks again, ...jerry

[phschoen]

hi Jerry, this is an odd approach. Then the compatible part should have separate part. The normal esp-12 and esp12e should be like they are in real. And i mean if you would break existing designs this is partial true since you can always use git sub-module with the exact version of the lib you where doing your design. Otherwise this flaw can never be corrected and even new one would have it.

WBR Philipp

[jdunmire]

Hi Philipp, I understand your position on this approach. It was very useful at the time (which actually pre-dates this thread) and it remains internally consistent.

An approach to clarifying the problem might be to create a 'vendor numbering' branch going forward, or at the risk of accidental updates to old designs (yes, I know sub-modules could be used to deal with the problem), maybe a 'compatibility' or 'historic' branch.

What ever approach is pursed, I no longer have the time to support this library. I sincerely hope it has been useful to others and that someone has the time and interest to take it up.

Regards, ...jerry