tsl8.app / examples / cartridge-laundry
The Cartridge Laundry System
A flytrap pod for a single shirt. One at a time, quietly, around the house.
I know how this sounds. But every piece of it exists in some other industry — peristaltic pumps are how every dialysis machine works, biaxial braided tubes are standard in stents, heat pump dryers are the current efficiency leaders. The thing that doesn't exist yet is the combination and the per-item deployment model. — Tom Ellsworth · sent to a soft-robotics PI · Feb 2026
This is a TSL8 — a structured envelope for ideas you want examined, not applauded. It was sent by Tom Ellsworth, verified at the top of the page via linkedin.com/in/tomellsworth. The cargo below is Tom's v0.4 design brief for an ambient per-item laundry system, along with eighteen soft spots he has already named and a specific ask to pressure-test the silicone-braid interface.
Trust model: open pre-print, not confidential. The envoy at the bottom of the page will answer your questions before you commit to writing back. Tom has given the envoy permission to read his public LinkedIn.
The rotating-drum washing machine is a local optimum we got stuck at a century ago. This proposes a dozen bowl-shaped silicone pods distributed around the house. You drop one dirty shirt into one. It recognizes the shirt (RFID tag in the care label), twists itself closed into a sealed pod, washes by tightening a helically-wrapped ball-chain around its outside, wrings, vacuum-dries, opens back up, and presents the shirt clean on a hanger.
A utility core in the basement handles water, heat, and chemistry. The pods are dumb terminals. The system's throughput per pod is slow; total household throughput is high because the pods run in parallel and are always available.
- Per-item is the real workflow. Nobody produces a full load of dirty laundry at once; people produce one dirty shirt at a time. The drum washer was a batching machine for a non-batch input. Every piece of friction in household laundry — sorting, waiting, pile accumulation, shared-machine scheduling — comes from batching.
- Every subsystem exists elsewhere. Peristaltic pumps (dialysis). Biaxial braided compression (stent deployment). Heat-pump drying (Samsung, LG, Miele flagships). RFID textile tagging (Decathlon, Inditex inventory). Silicone bladder chambers (food-grade sous vide, infant feeders). This is an integration problem, not a materials problem.
- Parallelism beats batch throughput. Eight pods × 38 minutes × 16 hours/day = 200 garments of daily capacity. A family of four produces ~30. The system is throughput-oversized by design so that waiting is never the user's experience.
- Water and heat are amortized. The utility core is where the expensive subsystems live — heat pump, filtration, dosing, greywater reclaim. The pods move water in and out through quick-connect hoses. Per-pod hardware stays cheap.
- Care is per-garment. RFID in the care label means the wash parameters travel with the garment. Delicates, denim, wool, performance fabrics each get their own cycle. No more damage from a mixed load.
The pod is a silicone hemisphere, inverted, with a hinged rigid lid. When a garment is dropped in, the pod closes and a helical ball-chain braid wrapped around its outer surface tightens — the same geometry a biaxial braided stent uses, but in reverse: compression rather than expansion.
The braid tightening squeezes the pod's contents against themselves and against the inner silicone wall. Water and detergent flow in through the base; agitation is produced by the braid pulsing on and off at 0.3–2 Hz (variable by cycle). Wring is the braid held at maximum tension while a vacuum is pulled. Dry is a warm low-humidity airflow through the still-sealed pod, pulled by the utility core's heat-pump loop.
The whole thing is ambient. No drum, no spinning mass, no vibration, no imbalance, no anchor to the floor. A pod can sit on a shelf in the bedroom.
Problems I already know about. Full list available on request; these are the ones I most want pushback on.
Silicone fatigue under this strain regime has no precedent.
I can't find any production silicone component cycled through this deformation profile at the frequencies I'm proposing. Medical-grade silicone bladders see fewer cycles, at lower strains, with less thermal loading. This is the one I'd most want a materials scientist to push on.
Candidate mitigation: hybrid construction — silicone inner wall, braided textile outer restraint, replaceable pod bladder as a consumable.
Vacuum seal reliability through a flexible aperture.
The hinged rigid lid must form a vacuum seal against the silicone opening after a garment, which is not flat, has been placed across it. I have no confidence this is solved.
Candidate: the garment sits in an inner mesh sleeve during loading; the mesh's edges form the sealing surface, not the garment itself.
RFID adoption in consumer apparel is zero.
The system is substantially worse without per-garment care instructions. Without RFID, the user must confirm the cycle manually on every load — not a dealbreaker, but it collapses one of the big UX advantages.
v1 ships with a default-safe cycle and an app-based override; inline manual RFID tagging with a printer accessory ($40) for power users.
Single-item throughput is inherently slow.
38 minutes per garment is slower than the per-garment wall-clock of a drum washer at equilibrium. The pitch depends on parallelism being enough to overcome this. If a family of four hits a peak-load event (camping trip, gym class, travel day) the system is stretched.
Utility core handles a mild overflow via a larger "communal" pod (10-garment capacity). Ships with v1.
Per-pod cost at scale is $240, not $24.
Eight pods plus the utility core is a $2,500–$3,500 system. That's the price of a high-end washer-dryer pair, which is reassuring, but the distribution is weird: a lot of small units rather than one big unit. The sales conversation will be harder.
Modular rollout — start with three pods and grow. Rentable pods for college dorms. Shared-pod housing retrofits.
This might be a very expensive solution to not folding.
The real pain is not washing; it's the pile on the chair. If the pods solve wash-and-dry but still produce a cloud of clean garments that need to be folded, the UX win is smaller than I want it to be. The "present on a hanger" detail matters more than it sounds.
v1 commits to the hanger output. Mesh sleeve designed to drop-fold around a garment hanger as the pod opens. Not fully validated.
Talk to the envoy before you reply.
The envoy has read the full brief, the eleven industry precedents, the BOM estimate, and Tom's public LinkedIn. It will answer specific questions (thermal, mechanical, regulatory) and tell you when it doesn't know. Your conversation is yours; Tom sees only what you choose to send.