Request capabilities

[fixrouter4400]

Hi Doug,

First of all thank you for writing this wafer_map app. It was great and nice to use.
Tool is working fine for me and this is not an issue but a nice to have capabilities request.

At the moment I'm working mainly in Chip RETICLE(Mask) preparation and planning.
Most of our designs has a multiple combination of die i.e. 4 die in 1 RETICLE shot(MASK) with different die arrays X=6, Y=6 and sub arrays of die i.e die1 X=3, Y=4; die2 X=3, Y=4; die3 X=6, Y=1 and die4 X=6, Y=1

1. Will you be able to add capability with an option to input mutiple die i.e Die 1....n
2. Input die X and Y of each die i.e. 4 die we can input die1 X:Y and so on
3. Input die array and sub arrays in 1 Reticle shot
4. Then calculate number of total die per wafer

I hope that my request is clear again thank you for making this wonderful wafer map app.

Thanks!
Fixrouter

[dougthor42]

Thanks for the suggestion! Can you explain a little more what you want?

I'm not understanding how a single subdie coord could be two different die. "die1 X=3, Y=4; die2 X=3, Y=4;" Do you mean you have two different reticle shots like so?

# reduced to a 2x2 reticle shot for simplicity
# numbers denote die type
# Shot A      # Shot B
+---+---+     +---+---+
| 1 | 2 |     | 1 | 3 |
+---+---+     +---+---+
| 3 | 4 |     | 5 | 6 |
+---+---+     +---+---+

This tool is not intended to be a gross-die-per-wafer calculator, so I don't think it makes sense to support reticles.

Instead, this is supposed to be for viewing die data to see any wafer-location-dependent patterns that may show up. For example, when plotting oxide thickness, a hot spot in the oxidation chamber could cause thicker oxide in the lower-left part of the wafer.

When I first wrote this, I did think about reticle support but I couldn't come up with a use case. Perhaps you can persuade me.

Also, note that I have very little time these days so I may not get to adding the feature 😢.

[fixrouter4400]

Hi Doug,

Thank you for your reply and YES I'm mainly working on a RETICLE internal design.
These are the wafer info constants:
Wafer Dia = 300mm
Exclusion Edge =3mm
Flat or Notch Height =7.75mm
Street Line =0.08 -0.12mm

It is only a single reticle shot but with combination of up to 6 different die and different array and sub arrays.

Reticle shot =1 (Reticle X, Reticle Y)

Combination of 4 die with 4 x 4 array

Die Sub Array - Note All die X dimension is equal, Die 1 & 2 Y is equal and Die 3 & 4 Y is equal

#1 = 2 x 2
#2 = 2 x 2
#3 = 1 x 4
#4 =1 x 4

Please see attached sample reticle with an array of 4 x 4.
Thanks!
Fixrouter

[dougthor42]

Ah, I see what you mean. Sadly I didn't design the program to handle multiple die sizes 😞.

If you don't care about the difference in Y sizes (maybe it's a difference of only a few percent and won't be noticeable?), you can still use this to plot data - it might just look a little funny.

What you want to do is figure out function that tells you what the die type is. Based on your image, something like this might work:

def die_type(x, y):
    """`x` and `y` are the integer die coordinates"""
    # Top two rows
    if y % 4 == 0 or y % 4 == 1:
        if x % 4 == 0 or x % 4 == 1:
            _type = 1
        elif x % 4 == 2 or x % 4 == 3:
            _type = 2
        else:
            raise ValueError("This shouldn't be possible.")

    # Bottom two rows
    elif y % 4 == 2:
        _type = 3
    elif y % 4 == 3:
        _type = 4
    else:
        raise ValueError("How did you even get here???")

    return _type

You can then use that function to define the discrete bins. In example.py, you'd replace line

wafer_map/wafer_map/example.py

Line 142 in 5e2ea25

discrete_xyd = [(_x, _y, random.choice(bins))

with:

discrete_xyd = [(_x, _y, die_type(_x, _y))
                for _x, _y, _
                in xyd]

You'll end up with something like this:

That's about all you'll be able to do with the program in it's current state. You're welcome to try and make changes though! But I warn you - I think this code is pretty damn bad. My past self wasn't as good of programmer as I am today.

[fixrouter4400]

Hi Doug,

Thank you for help and reply. Will try your suggestion.

Cheers,
Fixrouter4400