Making an array of cells using python

Comments

• Matthias

  September 2018

  Hi Shoval,

  I think I don't understand precisely what's the issue. If you're able to place one cell in Python, just use this code in a loop and place the cell multiple times.

  Or am I missing something?

  Matthias
• Shovalzo

  September 2018

  Hi Matthias,

  I managed to solve the problem by using the CellInstArray to insert it multiple times using a transformation with dx, dy. Previously I inserted the die into the wafer without it, and it would place the (0,0) of the die on the (0,0) of the wafer.

  Shoval
• CreatedByCreator

  January 26

  Hi Matthias,

  I am currently using the following code, but I currently only see one instance (n = 8)

  i = 0
  while i < self.n/2:
    t_dev_x = startpos+i*pitch
    placement = pya.DTrans(t_dev_x,0)
    self.cell.insert(pya.CellInstArray(pcell_var, placement))
    i= i+1
  

  Am I missing something? I guess the other alternative is inserting an array instead of single cells, but any help is appreciated.

  Thanks!
• CreatedByCreator

  January 29

  So actually, I found that it is stepping and placing instances. However, for whatever reason, the scale seems to be off.

  pitch = self.pitch+2*self.length

  Where self.pitch and self.length are TypeDouble

  When using the debug points, these numbers look normal and equal 36.5. When I zoom in, I see that it did indeed place the number of instances I had asked for except that the step size is only 0.0365.

  When I "correct" the pitch by multiplying by 1000, it looks to be correct, but I would like to get to the bottom of why the scale is off by 1000 here, but not elsewhere in my code?

  pitch = (self.pitch+2self.length)1000

  If anyone can clarify why this is happening that would be much appreciated.
  Thanks in advance!
• Matthias

  January 29

  Hi @CreatedByCreator,

  You're using "DTrans" for the transformation, but put it into a "CellInstArray" object, which is by definition using integer DBU units. That is not consistent and the effect is that "CellInstArray" will convert the "DTrans" into integer units, but not applying the DBU.

  You have to use "DCellInstArray" instead of "CellInstArray" to consistently work with micrometer units.

  In short, KLayout can work with database units - that is integer multiples of the DBU value, which in your case seems to be the default of 1nm. Or it can work with micrometer units with some limitations. In that case you have to use different object types and some limitations apply - like boolean operations, which are only available in integer space. You mixed up these spaces, hence your micrometer values will be seens as integers and taken as DBU multiples. Hence instead of 36 micrometer, you get 36 nanometer.

  There is an explanation of that concept here: https://www.klayout.org/klayout-pypi/overview/layout/.

  Matthias
• CreatedByCreator

  January 31

  Thanks Matthias! I didn't realize there was the DCellInstArray, but that makes sense now.

  I also didn't realize there was a whole other python documentation here: https://www.klayout.org/klayout-pypi/

  Thanks for putting together all the documentation!