Xilinx Symbol Creation using kipart: unbundled option creating too big "Units"


#1

While the -b option is provided to limit the size of the units being put on the schematic, I may want to do some thing creative with the layout and want all the individual pins called out in the schematic. So here is what I did:

  1. Edited the csv package file and named the column named “Bank” corresponding to VCCINT, VCCAUX, VCCAUX_IO_G0, VCCBRAM, VCCPINT, VCCPAUX and VCCPLL as NAA instead of NA.
  2. Renamed the csv file as xc7z030ffg676an32pkg.csv
  3. Executed the following command line: kipart -r xilinx7 -s name -o xc7z030ffg676an32pkg.lib -w xc7z030ffg676an32pkg.csv

However the resulting .lib file is empty and is 0 bytes in size!!! Can someone explain what maybe going on?

I also tried to change the Bank column annotation from NAA to an arbitrary numerical number: 59 and I am still haveing the same problem.

I also moved the file to a different folder and renamed it to xc7z030ffg676pkg.csv and ran the following command line: kipart -r xilinx7 -s name -o xc7z030ffg676pkg.lib -w xc7z030ffg676pkg.csv and I still had the same problem. I also had changed the package name in the very first cell to reflect the package I am trying to create a lib of and I am still having the same issue.

Infact I also changed the file name to the original name and the original format and I am still having the same issue of a .lib of 0 kB size with an empty file.

Can someone point out what I may be doing wrong here?

-GigaVolt


#2

I did the same operations as you and I got a non-zero .lib file. I used kipart 0.1.24, Windows 7, python 2.7.12 and 3.5.2.


#3

I used kipart 0.1.24 on Mac OS 10.12.1, python 2.7.12
What worked for me:
kipart -r xilinx7 -s name -o xilinx7.lib -w k7all.zip a7all.zip v7all.zip 7zSeriesALL.zip
What did not work for me:
kipart -r xilinx7 -s name -o xc7z030ffg676an32pkg.lib -w xc7z030ffg676an32pkg.cs
kipart -r xilinx7 -s name -o xc7z030ffg676pkg.lib -w xc7z030ffg676pkg.csv

Does it matter that if I modify the csv using MS Excel?


#4

I edit the CSV file with Libre Calc, but it shouldn’t make any difference if it’s just a CSV file.

Email your file to devb@xess.com. I’ll run it through my stuff and see if it causes a problem.


#5

The interesting thing that I find is that once I save the file using excel even with no modifications and run the kipart command, the .lib file is 0 Bytes!!! I think it has something to do with the MS Excel saving the file in the Mac format. The original file from the uncompressed .zip file creates a perfectly fine .lib (non-zero size).

Regards,
Gigavolt.


#6

I sent you the file and a few variations.

-gigavolt


#7

Seems I have the same problem of gaining zero byte lib files. Mac OS X as well.

Downloaded the Xilinx Artix7 pinout files from Xilinx themselves.

Unfortunately, these are just plain whitespace delimited *.TXT files. Did not work with KiPart. So I ran them through Regular Expressions in Notepad++ to replace whitespaces with (a consistent number of) tabs. Still no good in KiPart. I then imported these modified tab delimited *.CSV files in Libreoffice Calc (worked flawlessly) and saved them as comma separated *.CSV files instead. Still no go. Neither did removing header and footers do any good.

So I am stuck. Short of parsing the Python code itself, is there any chance of having the header of a working Xilinx 7-series pinout file dumped here? Just so I can deduce what I need to adhere to?

( I can of course convert this Xilinx file to a generic KiPart pinlist. But this process is error prone. )


#8

Ah, I finally managed to locate the culprit. It is indeed Xilinx themselves. They offer three different versions of the same pinout files.

• The first version is the plain space delimited text version without much in terms of header information, that is directly downloadable from their web page.

• The second and third versions are embedded in the Zip archive you get by pressing the “Download ZIP” button at the bottom of the same web page. (Tag all the BGAs you are interested in first.)

The latter ZIP archive will give you both the same whitespace delimited TXT version – but with more extensive header information – and the usable comma separated CSV version, also with extensive header information ( and with a consistent number of commas on every line, including the commented lines: )