Easier method for entering parts to Personal Libraries

[amphibious]

Mod. edit: Moved to a new thread to avoid hijacking.

Hi everyone,

I am hijacking this thread as I think Cynthias issue has been adequtely answered and I have a an issue associated with the title

I’m using KiCad professionally and I have to integrate new footprints into projects on a regular basis. I spend a lot of time searching for footprints that aren’t in KiCads libraries. It’s mostly footprints and 3D models I am talking about. I might get them from the manufacturers page, Snapeda or the like. Then I need to integrate them into my libraries.

1. Generate a symbol library, if that category doesn’t exist
2. enter information into symbol (name, datasheet, etc.)
3. Generate a footprint library, if that category doesn’t exist
4. enter information into footprint (name, datasheet, etc.)
5. associate symbol with footprint
6. Generate a 3D library, if that category doesn’t exist
7. rename 3D-part
8. associate footprint with 3d-part
9. correct 3D-part placement

There is a lot of back an forth between these and I’d love to have some kind of interface that will just let me take the three files I got (symbol, footprint, 3D), enter the metadata, choose the destination libraries and then moves everything to where it should be, editig the paths accordingly.

Does that make sense?
Has anybody done this yet?

Cheers
ff

[paulvdh]

Looks like normal library management to me.

@amphibious

I’d love to have some kind of interface that will just let me take the three files I got (symbol, footprint, 3D), enter the metadata, choose the destination libraries and then moves everything to where it should be, editig the paths accordingly.

How would you solve the Create new … if it does not exist part sin steps 1, 3 and 5?

There is a gitlab issue for:

Wishlist: one file per symbol in eeschema libraries. (lp:#1842206) (#2490) ·...

And similarly, there is also another issue for packing multiple footprints into a single footprint library. That would make it symmetrical again, as both approaches have their on sets of (dis) advantages.

One file per symbol makes library management more flexible and it makes it easy to use scripts to move files around. A way to pack and unpack those libraries is then also going to be a part of the interface. KiCad-cli is then probably going to be extended (too) to do this function.

@amphibious

Has anybody done this yet?

The database driven library system is getting more popular, especially for people working in companies and with bigger libraries. It takes some work because you have to set up the database first, but after that working on individual parts is easier.

[amphibious]

@paulvdh

How would you solve the Create new … if it does not exist part sin steps 1, 3 and 5?

I’d have a list of my existing libraries to choose from as in the library editors now.
One entry would be named “Create New” which would prompt me to give a name and a location when I save the part.
The paths for the list(s) could be edited in a prefereces window as is done in Kicad already.

[amphibious]

@amphibious

I’d have a list of my existing libraries to choose from as in the library editors now.
One entry would be named “Create New” which would prompt me to give a name and a location when I save the part.
The paths for the list(s) could be edited in a prefereces window as is done in Kicad already.

My hope was that some convenient tool to enter new part into the databases would spill into the community

[paulvdh]

Not much chance for that. KiCad only provides an interface to a database. You have to setup the whole database itself. This way it’s both the simplest for KiCad (there are few developer resources and there are many new features that people wish to be implemented), and it provides the most flexibility. It is common that companies already have a pre-existing database from some other EDA suite, and they can create an interface to integrate their existing database into KiCad.

I am just a hobbyist and only use the “old fashioned” / simple libraries, but as far as I know, when you use the database driven libraries, you do all part creation (symbols, footprints, 3D models) from within the database system and it also handles where those parts are stored, and how they are combined.

[Piotr]

@amphibious

• Generate a symbol library, if that category doesn’t exist
  …
• Generate a footprint library, if that category doesn’t exist
  …
• Generate a 3D library, if that category doesn’t exist

To decide and generate my libraries structure was my first task when planning to use KiCad.
I have done it in 2017 and since than I have done only one change - this year I separated the terminal block footprints from the connector library.
So listing these 3 points as ‘to do’ when adding each part seems like an exaggeration to me.
My library structure:

Directory Organization

Agree. I use only global libraries and I use only symbols and footprints and 3D models from my libraries. All symbols I define myself. Footprints I copy from KiCad libraries and modify a little. Before even trying to design my first KiCad PCB I have checked what are the possibilities to prepare files for documentation I need. And I found being it a problem with KiCad footprints. So I decided to have them defined a little differently (I have Reference and Value at CrtYd layer). I also don’t li…

[joojala]

I dont think its a good idea to hijack this thread. Should be moved in another thread.

@amphibious

There is a lot of back an forth between these and I’d love to have some kind of interface that will just let me take the three files I got (symbol, footprint, 3D), enter the metadata, choose the destination libraries and then moves everything to where it should be, editig the paths accordingly.

Now in general i disagree with library comments, but this case is actually true. That’s the import should be easier. My suggestion is that any single footprint library opening should trigger a wizard of some kind for import.

But realistically somebody should make a workable suggestion of how such a thing should work.

But also i think people are exaggerating on how hard a database library is to make. But again make a suggestion on how it should work then maybe somebody can get an idea to build that.

[jmk]

@joojala

Should be moved in another thread.

Done

[amphibious]

I already started to draft up a concept for a python tool some time a go. Hopefully I’ll find the time to get it together. Whoever wants to join is welcome to PM me so we can meet up and do it together!