Lifeline / Quickly-Draw! — A Sub-S$10 Open-Source Blood Sampler

The constraint

Currently we draw blood from a vein. It hurts, the vein is not always easy to find, and it can rupture. If you are elderly you also book in advance, travel to the hospital, and wait your turn.

Now move that same problem into a field clinic with no mains power, or a rural maternal-health post, or the first 72 hours after an earthquake. The hard part stops being technique and starts being equipment. A vacutainer system is a supply chain. When the supply chain breaks, the cost of drawing 4 mL of blood becomes infinite.

Lifeline — the repo calls it Quickly-Draw! — is one answer. It prints on an Ender 3 plus a budget SLA printer, it runs on the strength of a human hand, and it empties into the same VacuTube the lab already uses.

No electricity. No trained operator. Sub-S$10 per unit. 4 mL — forty times the volume of a competing microneedle patch.

How it works

Prime. Squeeze the TPU 95A bellows. The accordion folds compress and store −80 to −96 kPa. Squeeze force is 15–25 N, which most adults including elderly users can produce.

Apply. Press it to the skin. A dual-spring mechanism fires a 5×5 solid microneedle array at high velocity and retracts it inside 50 ms.

Draw. Release. The bellows springs back, the inlet check valve opens at 5–8 kPa, and capillary blood moves through five radial microfluidic channels into the vial.

Why it does not hurt much

The stratum corneum has no pain receptors. Below about 1,500 µm you start reaching nerve endings in the papillary dermis, so our needles stop at 600–800 µm with a sub-micron tip radius and a 3.0 mm pitch. Combine that with a sub-50 ms dwell time and the literature puts pain reduction at roughly 95% against conventional venipuncture.

Solid needles rather than hollow ones, because they are stronger, they do not clog, and SLA can actually make the tip sharp enough.

Decisions I would defend

Bellows, not a bulb. A bulb was cheaper. It was also unreliable across operators with smaller hand spans, and that failure mode lands hardest on exactly the community health workers this is for. The bellows geometry follows Tawk et al. (2019) — 45° fold angle, 1.2–1.5 mm wall at the root, printed vertically so the layer lines run perpendicular to the fold. About 21,500 cycles before fatigue failure.

Two check valves, not one duckbill. A single valve fouled by a clot loses the whole sample. Two opposing one-way flexures cost an extra hour of print time per unit and remove the failure mode.

Earlier revisions were wrong and that is on the record. The first design I sketched used a 7×7 array and about 25 kPa from a silicone bellows squeezed between PETG plates. The current one is 5×5 at −80 to −96 kPa in printed TPU. I am leaving that trail visible because the reason for the change — needle-force distribution against achievable vacuum — is the actual engineering content.

How it was made

Written in OpenSCAD with a lot of iterative prompting, sketched on paper first, then rendered and exported to STL. I am not going to pretend the parametric CAD wrote itself; I prompted my way through it and learned the geometry on the way. The whole thing is open-sourced under CC-BY-4.0 — CAD source, print profiles, renders, and a technical report — because a hospital or an NGO should be able to adapt this without negotiating with me. Attribution is the only ask.

If you want to print one, or fix one of my valves, the files are in hardware_projects/lifeline.

Next

A clinical-grade variant with a moulded silicone gasket replacing the printed flexure valve. Field testing with a partner clinic in central Vietnam, scoped through SAIL. Built for the Lifeline Challenge; still open, still unfinished, still printable tonight.