techmannih - @Abse

can you let me know bit about it

what we have an issue with polygon pads

After we added polygon pad support for kicad pads, you can see there is now there is difference

https://github.com/tscircuit/footprinter/pull/664

Check this PR

You can see some footprints changed

hey <@1067854999889137715> , me and <@757706909351411845> were talking and I need to go to bed a have an issue and we need your help

Yea I will check it soon, just now I wake up.

I'm going to bed in like 20 min let me know if you have any questions

(PR for gerber gen, needs to be tested after the footprinter update in core to make sure diodes have it marked) https://github.com/tscircuit/circuit-json-to-gerber/pull/116

<@398467566588264449> diode and polygon pads issue are different right? I was confused

<@1067854999889137715> we're looking to clarify the polarity of diodes by adding fabrication notes that makes it clear. This will make sure JLCPCB assembles properly

yea, I got the issue from the video, we have an issue with diode pin as compare to kicad, which should be flip, pin1 should be neg and pin2 should be pos.

yea but we're not going to do that yet

because there are issues with schematic symbols

the goal right now is just to make sure that the fabrication note text is present

Do we have an issue in core or footprinter related to fabrication note text for diodes, or should I check it myself?

Cc <@757706909351411845>

it is footprinter

you are just adding fabrication note text

Yea I am on it

yea that's good- it's weird that we don't load the fabrication note stuff from kicad on the right tho

we can do it

also

<@1067854999889137715> important video on how to make sure this is put together properly https://cleanshot.com/share/fqvNZfYQ

Abse and I are going to bed so try to solicit reviews from others

<@757706909351411845> Yes, I'll update all the diode footprints to match KiCad and also try to fix the Gerber issue. Right now, the fabrication note layer is visible there. We can tell JLCPCB that our diode footprints match KiCad. The remaining pin-related issue will need to be fixed in the core and schematic symbols.

<@1067854999889137715> DO NOT UPDATE TO MATCH KICAD - WATCH MY VIDEO CAREFULLY

Here is a video demonstrating how you can verify https://cleanshot.com/share/M5GsmxP7

DO NOT MATCH KICAD - WATCH THE VIDEO CAREFULLY - IT MUST POINT TOWARD THE CATHODE AND OUR CONVENTION DOES NOT MATCH KICAD- DO NOT MESS THIS UP!

WATCH THE VIDEO AGAIN MAKE SURE YOU UNDERSTAND! THIS IS CRITICAL - MESSING UP WOULD BE BAD!

make sure to watch this video carefully- do not mess up the note

do not trust ai, understand carefully what it means and confirm carefully, the board _will not work_ if the fabrication note text is reversed

in Abse's gameboy board- the diodes should all be correct when the footprinter strings are used- you should fork/modify to confirm you have gotten all the fabrication note text to be correct

this project has been delayed a lot, if there is one mistake we can't have this board at open sauce to demo

Sounds good. We need to add fabrication notes to all our diode footprints. For now, the important part is that the fabrication (symbol arrow) always points toward the cathode. I know that both our diode pin naming and KiCad's pin naming don't match correctly right now, but we don't need to fix that in the core yet. For now, we just need to ensure the fabrication (symbol arrow) clearly shows the correct diode direction by pointing toward the cathode.

<@757706909351411845> <@398467566588264449> https://github.com/tscircuit/footprinter/pull/679

this looks good but for small footprints like 0402 , 0201 and 0603 add a box like how kicad is doing not the symbol

also add the kicad fabrication to the snapshots if possible for the parity test

also the width and height should be symmetric some symbols look stretched

<@398467566588264449> yes, I check it, but we do not copy the kicad i think, seve said in the video we should make like that symbol and point towards the cathode

The kicad fabrication doesn't make sense if we are not matching kicad. I was showing that snapshot to seve, he said don't match with kicad

.

<@398467566588264449> if you are available, can you come in the voice

yes

<@1067854999889137715>

<@1067854999889137715>

<@398467566588264449> can you check now

<@398467566588264449> can you let me know what footprint you are using on your board,

there is ALOT of repeated code , just make one createManualDiodeFabricationNotes and reuse it

<@398467566588264449>, that's what I was doing before, but we couldn't modify every diode footprint according to its size and dynamically for all.

hmm did you try using transformation-matrix

not yet, i am trying to do

<@757706909351411845>

can you check it now https://github.com/tscircuit/footprinter/pull/679

i think there is also wrong parity?

<@757706909351411845>

even the jlcpcb diode point to the wrong pad

gasp

no that is almost certainly core

this feels like a bug planted by a spy

yeah like I mean while our pin one is anode all the fps will be wrong

to subtely break every board

yea

oof

I can flip the componet and connect it wrong ? like pin 1 as cathode and pin2 as anode to order ?

not sure if that is ok

just manually swap pins on the footprint maybe?

yeah

oh yeah

<@757706909351411845>

this should be right

looking good

I'm going to start the order process rn ?

do it!!!!

WOOO

lets goo

<@757706909351411845>

https://jlcpcb.com/partdetail/WinbondElec-W25Q16JVUXIQ/C2843335

this is the pico flash do you think this is enough for storage ?

yea we can make it work

might be limited but not hugely concerned

There are a few diode footprints left without fabrication notes. Can I add fabrication notes for them.?

yes