Wafer Analysis¶

Now we will run a die loss cutback analysis on the die data we uploaded earlier.

As before, make sure you have the following environment variables set or added to a .env file:

GDSFACTORY_HUB_API_URL="https://{org}.gdsfactoryhub.com"
GDSFACTORY_HUB_QUERY_URL="https://query.{org}.gdsfactoryhub.com"
GDSFACTORY_HUB_KEY="<your-gdsfactoryplus-api-key>"

import getpass

from tqdm.notebook import tqdm

import gdsfactoryhub as gfh

project_id = f"spirals-{getpass.getuser()}"
client = gfh.create_client_from_env(project_id=project_id)
api = client.api()
query = client.query()
utils = client.utils()

Lets Define the Upper and Lower Spec limits for Known Good Die (KGD).

For example:

waveguide width (nm)	Lower Spec Limit (dB/cm)	Upper Spec limit (dB/cm)
300	0	3.13
500	0	2.31
800	0	1.09

As for waveguide loss you can define no minimum loss (0 dB/cm) and you only define the maximum accepted loss (Upper Spec Limit)

Lets find a wafer pkey for this project, so that we can trigger the wafer analysis on it.

wafer_data = query.wafers().execute().data
wafer_pkeys = [wafer["pk"] for wafer in wafer_data]

from gdsfactoryhub.functions.wafer import aggregate_die_analyses

aggregate_die_analyses.run(
    wafer_pkey=wafer_pkeys[0],
    die_function_id="propagation-loss",
    output_key="propagation_loss",
    min_output=0.065,
    max_output=0.10,
)

{'output': {'mean_propagation_loss': np.float64(0.08296279661949407)},
 'summary_plot': <Figure size 1000x450 with 4 Axes>,
 'wafer_pkey': '89edb89b-520d-44f5-ac44-c0210f364ac9'}

png

with gfh.suppress_api_error():
    result = api.upload_function(
        function_id="aggregate_die_analyses",
        target_model="wafer",
        file=gfh.get_module_path(aggregate_die_analyses),
        test_target_model_pk=wafer_pkeys[0],
        test_kwargs={
            "die_function_id": "propagation-loss",
            "output_key": "propagation_loss",
            "min_output": 0.065,
            "max_output": 0.10,
        },
    )

Duplicate function

results = []
for wafer_pk in (pb := tqdm(wafer_pkeys)):
    pb.set_postfix(wafer_pk=wafer_pk)
    result = api.start_analysis(  # start_analysis triggers the analysis task, but does not wait for it to finish.
        analysis_id=f"wafer_propagation_loss_{wafer_pk}",
        function_id="aggregate_die_analyses",
        target_model="wafer",
        target_model_pk=wafer_pk,
        kwargs={
            "wafer_pkey": wafer_pk,
            "die_function_id": "propagation-loss",
            "output_key": "propagation_loss",
            "min_output": 0.065,
            "max_output": 0.10,
        },
    )
    results.append(result)

  0%|          | 0/1 [00:00<?, ?it/s]

Let's have a look at the last analysis:

analysis_pks = [r["pk"] for r in results]
utils.analyses().wait_for_completion(pks=analysis_pks)
analyses = query.analyses().in_("pk", analysis_pks).execute().data
succesful_analyses = [a for a in analyses if a["status"] == "COMPLETED"]
analysis = succesful_analyses[-1]
img = api.download_plot(analysis["summary_plot"]["path"])
img.resize((530, 400))

Waiting for analyses:   0%|          | 0/1 [00:00<?, ?it/s]

png

Die Analysis Generate Layout

Wafer Analysis

Wafer Analysis¶

On This Page