The questions people actually search for at midnight, answered with the limits stated rather than hidden. We write them because we build a design tool, and we would rather be judged on whether we understand the problem than on adjectives. Every note says where the model stops being true.
hydraulics
Why your parallel channels don't split flow evenly
Four identical branches, one shared manifold, a 1.9× spread in outlet flow. Where it comes from, why depth is the dimension that bites, and three fixes with their trade-offs.
microfluidics
What T-junction droplet correlations can and cannot predict
The squeezing law is one of the most useful results in the field and one of the most misquoted. Its regime, its ±14% coefficient band, and the five things it was never about.
manufacture
GDSII for microfluidics, in plain words
What the format carries, what it does not — it has no field for channel depth — and the checklist that stops a mask run going wrong.
Pressure control or flow control: which one your chip wantsIn preparation
Syringe pumps and pressure controllers do not produce the same experiment, and the difference shows up in your droplet size distribution.
How much does a 3 µm depth error actually cost you?In preparation
A tolerance study on the dimension nobody controls, and what robust design means when depth is cubed.
This is what the notes are about
Fluidraft is microfluidic CAD with a validated 0D/1D flow solver, DFM checks and fab-ready export. It runs entirely offline — nothing is uploaded, ever.
Run the complete product in your browser →The complete product runs client-side. Nothing to install, nothing uploaded. Subscriptions from $79/month; academic labs pay half.