tsl8.app / examples / cardboard-appliance
The Cardboard-to-Packaging Desktop Appliance
Your weekly Amazon boxes, reshaped on the counter, shipped back out by Saturday afternoon.
Desktop appliance converts household waste cardboard to molded-pulp protective packaging on demand, using commodity mechanical hardware and non-toxic natural binders. v1 ships two forming stations (vacuum former, brick press) with three accessory stations planned.
I got tired of the one-way economics. The Amazon boxes come in, the bubble wrap goes out, and nothing about that is supposed to make sense forever. This is a desktop appliance that closes the loop — using proven molded-pulp processes, commodity hardware, and natural binders. I'm looking for someone who wants to push on the engineering before I commit to a prototype build. — Tom Ellsworth · November 2025
This is a TSL8 — a format for ideas you want examined. It was sent by Tom Ellsworth, whose identity is verified at the top of this page via linkedin.com/in/tomellsworth. The cargo below is Tom's v0.3 design brief for a household packaging appliance, along with honestly-labeled soft spots and open questions.
The trust model here is email/website, not NDA. Treat this as a proposal shared openly with someone Tom trusts to pressure-test it — not confidential IP. Respond by writing back, pushing on the engineering, or relaying it to someone in your network who'd have better intuition.
A desktop appliance that turns a household's waste cardboard into on-demand protective packaging, compressed bricks, and other pulp-based products. The user loads flattened boxes; the machine shreds, wets, pulps, removes contaminants (tape, staples, plastic), dewaters, and deposits the slurry at a swappable forming station.
The flagship output is custom-fit molded-pulp cradles generated from a 3D-printable mesh mold — the biodegradable equivalent of foam-in-place packaging. Secondary outputs from the same core include compressed bricks for planters, loose fill, 3D-print paste, and high-performance aligned-fiber aerogel foam. The system closes a packaging loop that is currently one-way plastic waste, using commodity mechanical engineering and non-toxic natural binders.
- Real consumer need. Household packaging waste is ubiquitous; return packaging for small online sellers (Etsy, eBay, Reverb, Shopify makers) is an unsolved recurring cost. Bubble wrap and air pillows are cheap but universally thrown away.
- Proven underlying technologies. Molded pulp packaging (Apple, Huhtamaki), cellulose foam (Stora Enso Papira, VTT, Nature-Pack Fibrease), expanding honeycomb paper (Ranpak Geami), and paper briquette presses all exist at industrial or commercial scale. The appliance is a miniaturization-and-integration play, not a materials breakthrough.
- Commodity mechanical stack. Every subsystem has a consumer-price analog: cardboard shredders, texture hopper guns, clay pug mills, hydraulic sausage stuffers, peristaltic pumps, 3D printers. No subsystem requires inventing new hardware.
- Non-toxic by default. Native lignin in kraft cardboard acts as a thermosetting binder above 140 °C — no added chemistry needed for baseline operation.
- Packaging is the killer app, planters are the demo. Molded cradles and pour-in-place void fill address a weekly household need; bricks and planters give the marketable hero shot.
Five stages shared across outputs, plus a swappable forming station.
Infeed and shred
Hopper accepts flattened boxes up to ~30×30 cm. Dual-shaft shredder targeting 25–50 mm chunks. Current-sensing auto-reverse for jams. Tape, staples, and plastic windows pass through without pre-sorting.
Wet pulping
Heated stainless tank (~20 L, 40–60 °C). High-shear rotor produces slurry in 10–20 min. Dwell time and blade geometry controlled via a coarse/medium/fine dial.
Coarse: long fibers, structural outputs, better tear strength. Fine: short fibrillated fibers, surface fidelity, extrusion, printing.
Contaminant separation
Mill-style cascade, scaled down: magnetic junk trap (staples, plastic windows, tape wads), 0.5 mm vibratory slot screen (fibers pass, sticky fragments don't), small hydrocyclone (heavies), optional thermal disperser (v2) for residual stickies at ~100 °C.
Dewatering and binder dosing
Perforated screw press brings solids to target consistency (1–3% for vacuum form, 20–30% for paste extrusion, 30–40% for bricks). Metering pump injects binder by dry-fiber mass.
Forming station (swappable)
Standard wet-slurry port + 24 V / USB-PD control connector. v1 ships with a vacuum former (custom cradles) and a brick press. Accessory stations for paste extrusion, long-fiber pour-in-place, foam cushioning, and freeze-casting follow.
Ten outputs share the same pulping core — what changes is the forming station, the slurry grade, and the dewatering target.
| Output | Forming station | Slurry & binder | Cycle |
|---|---|---|---|
| Custom molded cradle | Vacuum former + 3D-printed mesh mold | Fine, 1–3% solids · none | ~1–2 h |
| Pour-in-place void fill | Spout + vented liner in shipping box | Fine/medium, 3–10% · none | 30–60 min drying |
| Compressed brick | Brick press | Coarse, 30–40% · native lignin or +5% lignosulfonate | 5 min press + 24 h cure |
| Loose fill (peanut equiv.) | Paste extruder + chopper | Fine, 20–30% · none | 30 min air-dry |
| Corner/edge protector | Paste extruder, linear die | Fine/medium, 25–30% · none | 30–60 min dry |
| Foam cushioning (EPS equiv.) | Foam head (whipped) — v2 | Fine, 2–5% + surfactant · none | hours |
| Expanding honeycomb sheet | Sheet table + die-cut | Fine, 3–5% · none | 30 min + die-cut |
| 3D-printed paste part | Paste extruder on 3D printer | Fine beaten, ~25% · 3% methylcellulose | Print + 2–4 h dry |
| Aligned-fiber aerogel foam | Freeze-caster — v2 | Dilute, 1–2% · none | Overnight |
| Brick-built planter / structure | Brick press + mortar | Coarse · optional lignosulfonate | Stack + cure |
Pumping & transport
Fiber length drives pump choice. Fine fiber accepts any pump; long fiber requires positive-displacement equipment with large passages.
| Fiber grade | Primary option | Price | Notes |
|---|---|---|---|
| Fine, thin | Hopper gun / texture sprayer | $30–75 | Needs 2 HP compressor. Carriage House rebadges as "pulp sprayer." |
| Fine, thick | Peristaltic / diaphragm pump | $50–200 | DIY builds reliable. Compatible with 3D-print paste extrusion. |
| Long fiber, thick | Clay pug mill | $300–500 | TECHTONGDA, EQCOTWEA on Amazon. 60 mm bore, handles fibers natively. Best consumer-grade option. |
| Long fiber, batch | Hydraulic sausage stuffer | $200–500 | Food-grade stainless. Good for pour-in-place. |
| Continuous long fiber | Progressive cavity grout pump | $500–2000+ | Industrial but consumer-adjacent. v2 upgrade. |
Binder strategy
Heat-activated natural preferred; UV reserved for v2.
- Native lignin activation (v1 default). >140 °C activates kraft's ~15–25% native lignin as a thermoset. No added chemistry. Mechanism behind MDF and binderless coconut boards.
- Lignosulfonate. 5% by dry fiber mass. Non-toxic, cheap, heat-set. Improves outdoor durability for bricks/planters.
- Starch (cassava/wheat). Food-grade, low-temp gel. Indoor only.
- Soy flour + citric acid. Cures 180–200 °C. Better water resistance than starch.
- UV bio-resins deferred to v2 — good for thin prints, unsuitable for thick bricks (poor UV penetration).
Core appliance (shipped): chest-freezer footprint, single 15 A / 1500 W circuit. Dual-shaft shredder, 30×30 cm hopper. 20 L heated pulper at 60 °C, coarse/fine rotor dial. Magnetic junk trap + 0.5 mm slot screen + hydrocyclone. Perforated screw-press dewaterer. Lid interlocks, two-hand press actuation, thermal fuses.
Included forming stations: vacuum former (diaphragm pump at −60 kPa, 10 L slurry bath, heated drying rack to 90 °C) and brick press (2-cavity steel mold, heated platens to 150 °C, hydraulic or lever ram, 5-min hold).
Software: STL-to-mesh-mold generator (open source), recipe library, shipment helper (object dims → recommended output type).
Accessories (v1 sold separately): paste extruder for loose fill, corner protectors, 3D-print-compatible; hydraulic sausage stuffer for long-fiber pour-in-place.
v2 roadmap: foam head for whipped cellulose cushioning, freeze-caster for aligned-fiber aerogel foam, thermal disperser for residual tape, UV-cure accessory for bio-resin 3D printing, rotational nozzle for programmed chiral fiber orientation.
These are the parts of the pitch I'd expect a skeptical engineer or investor to push on. Honest assessments. Where I have mitigations I've named them; where I don't, I haven't pretended.
Drying dominates cycle time and energy
Drying is the rate-limiting step for every wet-formed output. A small molded cradle needs 30–60 min at 60–90 °C; a brick needs multiple hours; foam-formed cushioning is worst because internal pore structure traps moisture. At 1500 W household power, an end-to-end packaging session is ~1–2 h. For structural outputs it's closer to half a day. This is probably the single biggest UX problem the product has to solve.
Mitigation: dedicated drying cabinet, waste-heat recovery from the pulper, overnight batch modes. Without solving this, the "on-demand packaging" pitch is a stretch.
Feedstock is abundant; storage and pre-processing is the real friction
Supply isn't the problem — free flattened cardboard is everywhere, from recycle piles behind any retail store to your own porch every week. The real friction is that cardboard is bulky, awkward to store, and has to be shredded before it's useful. A household that commits to this appliance has committed to a staging area: a corner of the garage where a week's boxes live until the next run.
Mitigation: the appliance probably wants a built-in high-compression cardboard "hopper storage" stage — shred-on-intake, buffer dry fiber in a bin, pulp on demand. That changes the UX from "load the boxes when you need packaging" to "feed the machine whenever you break a box down." Worth prototyping early.
The corner in your garage has to be worth more than a pallet of bubble wrap
This is really the same argument as Soft Spot 02, told from the customer's side. Bubble wrap is priced insanely — a pallet is nothing; $25 at Staples covers you when you have no other choice. The question the buyer is really asking isn't “what costs less per shipment?” It's “do I store bubble wrap in that corner, or do I store this machine and a week's cardboard?”
We do not win on per-unit cost and we should stop pretending we might. We win when the corner itself is the reward: the appliance produces better packaging than bubble wrap (custom-fit cradles, no plastic waste, the planter bricks and loose fill as side benefits), on inputs the household already generates, and replaces a standing supply order with a standing habit. The pitch is environmental values + custom fit + “I made this myself” — and we're only worth the corner for people who already feel the plastic-waste friction.
Food-contact regulatory exposure
Many Etsy/maker sellers ship food (candles, cookies, soaps). FDA 21 CFR 176 governs paper/paperboard for food-contact use. Mixed recycled cardboard with printing inks, receipts, and tape residues almost certainly fails this standard without additional processing. The product cannot honestly be marketed for food-contact packaging in v1.
Mitigation: explicitly position as non-food-contact; optionally sell clean-pulp feedstock packets for food users.
Tape and contamination in real-world cardboard
Laboratory slot screening tolerates ~1% contamination. Actual consumer cardboard — especially moving boxes and e-commerce boxes with kraft tape and fiberglass strapping — can push 5–15% contamination by volume. Screens will clog faster than a lab test suggests; shredder blades will wear; the hydrocyclone will need regular cleaning.
User prep instructions (remove large tape strips) help but are onboarding friction. Expect more maintenance than a typical appliance.
Moisture sensitivity of finished parts
Cellulose-fiber bonds are water-reversible. A molded cradle or cardboard brick loses 30–50% of its strength when re-wetted. Lignosulfonate and starch acetate help, but commercial molded pulp uses wax sizing, AKD, or fluorochemicals — none of which fit the non-toxic story.
Honest framing: indoor/covered use is fine; outdoor use needs a breathable seal (linseed oil, beeswax) applied by the user.
Mold wear and replacement cost
3D-printed vacuum-form molds are cheap to make but have a limited life. PLA softens at 60 °C (the drying temperature); resin molds have release issues after 20–50 cycles; PETG/ABS last longer but cost more. Effective mold cost is probably $5–15 per unique shape over ~30 uses.
A library of pre-validated STLs should ship with v1.
Appliance footprint, noise, install location
The full stack (shredder + pulper + press + dryer) is chest-freezer sized, 1500 W, and loud during shredding. Realistic install locations are garages, basements, and workshops — not a kitchen counter. This narrows the audience significantly versus an "every household" positioning.
Mitigation: modular split (shredder in the garage, former indoors), acoustic insulation, smaller form factor in later versions.
Water and effluent
Pulping consumes 20–50 L per session; most is recaptured. Screened fines, evaporation losses, and wash cycles still consume fresh water. Effluent contains dissolved lignin and fiber fines — generally legal down a household drain, but not everywhere, and not if the user runs it daily.
Closed-loop water recycling is a desirable v1.5 feature.
Cold-start problem
Packaging is a "need it now" product. The appliance workflow is a "need it in two hours" product. The user must buy the appliance before a specific shipping need arises — a discretionary purchase, which is harder to sell than a utility.
Mitigation: users batch-produce generic inventory (corner protectors, sheet dunnage, loose fill) ahead of time, then produce custom cradles only for high-value shipments.
IP and prior-art landscape
Molded pulp packaging is a century-old industry with thousands of patents. Foam forming, vacuum-forming, binder chemistries, and drying methods are all covered in some form. Some specific techniques (thermal disperser, foam chemistries, proprietary mesh-mold designs) may require design-arounds.
FTO review needed. Favor public-domain mechanical approaches over novel chemistry to minimize exposure.
The "green" story is only partially true
A full packaging session at 1500 W for 90 min uses ~2.25 kWh, roughly 1 kg CO2-equivalent on a mixed US grid. A piece of bubble wrap embodies maybe 10 g CO2. On carbon alone, the appliance only wins if the alternative was landfill bubble wrap AND the electricity is largely renewable.
The stronger environmental story is material circularity (cardboard reused rather than recycled) and landfill diversion of plastic packaging, not direct CO2 savings. Marketing should reflect this.
End-of-life fiber degradation
Each recycle cycle shortens fibers ~15–25%. After 3–5 cycles through the appliance, average fiber length drops to levels unsuitable for structural outputs.
Honest framing: "3–5 more lives," not "infinite closed loop." Long-term users will need to top up with fresh cardboard; the appliance extends fiber life, it doesn't create it.
Where I'm not sure
- Drying is slow — a packaging session is really 1–2 hours. "On-demand" is a stretch until we solve the dryer.
- Household supply is spiky — some weeks you'll have no cardboard, some weeks you'll have a pile. The sweet-spot user both ships and receives weekly.
- Bubble wrap is really cheap — we don't win on cost. We win on environmental values and custom fit.
- Not food-safe — recycled cardboard has ink and tape residue. Can't ship cookies in it. Yet.
- Moisture re-wets the bond — indoor use is fine, outdoor planters need a linseed-oil seal.
- It's chest-freezer sized — this lives in a garage, not a kitchen counter.
- Carbon math is shakier than the circularity math — we should sell the loop, not the kilograms of CO2.
One Saturday afternoon, end-to-end.
- Load a week of Amazon + grocery boxes into the hopper.
- Produce four bricks → stack them with slurry mortar into a small planter base.
- Switch to the vacuum former → produce a molded cradle for a ceramic pot.
- Switch to the extruder → run off a cup of loose fill.
- Ship the pot (protected by the cradle + loose fill) while planting in the brick planter.
Inbound cardboard becomes outbound packaging becomes garden — the product's narrative in one afternoon.
I've been thinking about this one for a while. The version I'm sending you is honest about the places I'd expect you to push back — drying, food-safety, the appliance footprint — and I'd genuinely rather hear that the whole premise is wrong from you now than from a customer in a year. Tear it up.
Molded pulp & cellulose foam
Binders & fiber properties
Pumps & transport
Questions for the engineer behind this?
The envoy can take your hardest question and answer in Tom's voice — drying economics, binder tradeoffs, commercial wedge, whatever you'd push on if he were in the room.
Something to add?
You can write back to Tom directly — or, if this cargo wants to land in your network, relay it onward with your own opener. Relayed TSL8s show From Tom · Relayed by you at the top. Recipients see who trusted whom along the chain.