Double-throw (case) latch

On this page

1. Parts
2. Print files
3. How it works
4. Mechanical advantage
5. What limits the force
6. Analyst's calculation
7. Adversarial review
8. Calibration adjustments
9. Preload retention outdoors
10. Materials
11. Hardware notes from the research
12. Design guidelines
13. Failure modes
14. Existing designs and commercial references
15. Evidence gathered in the sweep
16. Sources

Parts

Per latch: 1 part to print, 4 items to buy, 3 features printed into the box or lid and the TPU 85A seal. Parts marked † are not drawn in the 3D model. In the 3D gallery you can click a part to see where it sits. The last column links the print file, or an example of the part in a shop (see the shopping list).

Print files

Download all print files for this latch (ZIP, 56 kB)

Print in ASA (TPU 85A for the seal), 0.2 mm layers, at least 4 walls; settings and the clearances used are in print/README.md. The parts are already turned the right way up for printing. The box and lid coupons are short test sections with this latch's box-side and lid-side features, so you can try the latch before building it into your own enclosure.

All printed parts in their assembled positions, for CAD: double-throw-butterfly.step.

How it works

The Penn Elcom page calls this a recessed butterfly latch. The usual flight-case construction is a recessed dish with a spring-loaded folding or twist butterfly handle that engages a catch plate or keeper. The 'double-throw' label is the researchers', and neither the page nor my fetch confirms it. I analyse the printable principle as a quarter-turn cam draw, where a ramped cam or hook on the butterfly pulls the keeper down about 1.5 mm. A folding-lever over-centre reading has similar force limits. Commercial versions are 1.2 mm stamped steel with a steel spring, built for lid weight and shock, not gasket compression.

Mechanical advantage

Butterfly wing arm 18 mm / cam radius 8 mm = 2.25:1. The cam rises 1.5 mm per quarter turn (6 mm per revolution, lead angle 6.8 deg). With mu 0.3 the wedge factor is (1 - mu tan lam)/(tan lam + mu) = 2.3, so F_axial is about 5.2 x the finger force (155 N from 30 N). It is self-locking, since the 6.8 deg lead angle is below the 16.7 deg friction angle.

What limits the force

First hand torque (~0.5 N.m of finger effort gives ~155 N), then tension creep in the thin printed hook arm (a 10x5 mm arm at 3-5 MPa and Kt 1.5 holds ~100-170 N). The latch has a fixed geometry with little or no adjustment and no hard stop, so the gasket squeeze depends on tolerances. The commercial 90 kg rating (~883 N) is a lid-holding working load for a 1.2 mm steel latch and not a gasket draw force, and a printed copy will not match it.

Analyst's calculation

Analyst's sustained estimate: 70–200 N, typical 120 N; peak 400 N.

Commercial anchor (verified by re-fetch): Penn Elcom medium MOL recessed butterfly latch L905/928PBZ, max working load 90 kg (~883 N), 1.2 mm steel, zinc finish, 105x102 mm, 13 mm dish. The page does not say twist or double-throw. Printed version, own calc. Finger force 30 N at 18 mm arm = 0.54 N.m. Cam radius 8 mm, 6 mm lead per revolution (lead angle 6.8 deg), mu = 0.3 for printed-on-steel. F_axial = F_hand x (18/8) x 2.3 = 5.2 F_hand = 155 N (103 N at 20 N finger force, 207 N at 40 N). Hook arm tension: 10x5 mm = 50 mm^2, sustained 3 / 4 / 5 MPa, Kt = 1.5 gives 100 / 133 / 167 N. A thinner 36 mm^2 arm at 4 MPa gives 96 N. With the lead's conservative bearing and hot-sun margin I set low 70 N, typical 120 N, high 200 N. Peak: 80 N finger force x 5.2 = 414 N, and wing bending at 80 N x 18 mm over a 12x5 mm section is ~29 MPa, near the ASA limit, so peak ~400 N. On the reference lid at 100 mm pitch, 120 N is 1.2 N/mm, below the 2-4 N/mm window, and 250 mm pitch would be needed to hit ~0.5 N/mm. Series stiffness: hook arm EA/L = 1800 x 50/25 = 3,600 N/mm against ~625 N/mm gasket stiffness per latch at 100 mm pitch (own calc, 6.25 N/mm^2 x 100 mm). Loss = 0.85 x set, i.e. 25-49 % force loss for a 30-58 % compression set. Evidence note: the WebSearch budget was exhausted in this run, so no printed-butterfly test data was found.

Adversarial review

Independent recalculation: lead angle = atan(6 / (2 pi 8)) = 6.81 deg; wedge (1 - mu tan l)/(tan l + mu) = 2.299 at mu 0.3; efficiency tan l x wedge = 0.274; gain 2.25 x 2.299 = 5.17. F = 103 / 155 / 207 N at 20 / 30 / 40 N finger force, 414 N at 80 N. Including 4 mm pin friction (mu 0.3) the closing force at 30 N is 132 N. Hook arm 10x5 mm: factor 1 + 6e/t = 1 / 2.2 / 3.4 at e = 0 / 1 / 2 mm. Capacity at 4 MPa with Kt 1.5 is 133 / 61 / 39 N. Series fraction 3600/(3600+625) = 0.852, so loss is 0.85 x 30-58 % = 25-49 %. Sustained is therefore about 66-116 N from the closing force and about 60-90 N from the arm at e of about 1 mm. Typical 85 N is 29 % below the analyst's 120 N. I used 'corrected' because the arm capacity ignores eccentricity and the model is not the cited product. Fetched and confirmed: Penn Elcom (90 kg working load, 1.2 mm steel, zinc, 105x102 mm, 13 mm dish, no 'twist' or 'double-throw' wording); Lee Spring (music wire strength drops above about 121 C, rusts when wet; stainless superior outdoors); CNC Kitchen hook test; CNC Kitchen PC Blend review not re-fetched by me. The Albany fastener guide PDF text could not be extracted by my fetch, so the salt-spray hours (12-72 h, 96-120 h) are unverified by me. The WebSearch budget was exhausted, so I could not look for the real meaning of 'double-throw' or for printed butterfly or cam latch tests.

Issues raised:

• The mechanism identity is unverified. The Penn Elcom page (re-fetched) calls it a 'Recessed Butterfly Latch' and gives 1.2 mm steel, zinc finish, 105x102 mm, 13 mm dish, 231 g, 10 holes, and 'Maximum working load capacity: 90kg'. It does not say twist, cam or double-throw. The 'double-throw' label and the quarter-turn cam are the analyst's own model of a hypothetical printed latch. The 90 kg (883 N) figure is a lid-holding rating for a steel latch, correctly not used as a draw force.
• The wedge arithmetic reproduces: lead 6 mm/rev on r = 8 mm gives a 6.8 deg lead angle, wedge factor 2.30, gain 18/8 x 2.30 = 5.17, so 155 N from a 30 N finger. It ignores pivot friction. With a 4 mm pin at mu 0.3 the closing force is about 132 N. A pure cam is only about 27 % efficient.
• Hook-arm capacity is inconsistent with the snap-fit analysis. The 100 / 133 / 167 N figure for the 10x5 arm at 3 / 4 / 5 MPa and Kt 1.5 assumes zero eccentricity (e = 0). With the hook line offset e = 1 mm the arm holds 45 / 61 / 76 N (68 / 91 / 114 N without Kt). With e = 2 mm it holds 29-49 N. The arm therefore limits sustained force to about 60-90 N, below the 120 N typical.
• The latch is force-limited by the user's twist, not stroke-limited. At 132-155 N through the series stiffness (hook 3600 N/mm in series with 625 N/mm of gasket per 100 mm pitch) the gasket compresses only about 0.25-0.29 mm. The 1.5 mm cam throw is mostly consumed by FDM clearance. Clamp force therefore varies about +/-35 % with 20-40 N finger effort (103-207 N), and there is no stop or adjustment.
• Sustained = closing force x (1 - set loss). With 25-49 % loss (ks/(kg+ks) = 0.85) a 132-155 N closing force falls to 67-116 N. At 100 mm pitch this is 0.5-1.0 N/mm, below the 2 N/mm floor. Typical lowered from 120 to 85 N.
• Peak. 80 N finger x 18 mm x 2.3 gives 414 N. With pivot friction it is 350 N, and wing bending (12x5 mm) is 28.8 MPa, consistent. Peak set to 350 N.
• Outdoor durability. A zinc-plated stamped steel latch only has 12-72 h salt-spray class (yellow zinc 96-120 h) per the analyst's Albany source, which I could not read. A printed cam under 100 N at 60 C will creep and wear, with back-out in vibration unless a detent holds it. The recess collects water. Durability 5 reduced to 4. Usability 8 reduced to 7 because the twist effort sets the preload.

Calibration adjustments

• sustained_N.typical: 85 (reviewer; range 45-150) → 65 (range 40-120). The reviewer deducted gasket set loss from a hand-limited closer, which the cap, bayonet and multi-point reviews did not do, so that deduction was inconsistent. The structural limit is the 10x5 mm hook arm with 1 mm of eccentricity (45-76 N at 3-5 MPa with Kt 1.5), the same hook-arm limit as the cam-hook (55 N) and bail (50 N). Typical set to the arm limit.

Preload retention outdoors

Poor. The printed hook arm is stiff (3,600 N/mm vs ~625 N/mm gasket per latch), so ~85 % of the gasket compression set at 70 C (30-58 %) appears as force loss, 25-49 %. The geometry is fixed, so there is no way to retension unless the keeper is a screw. The cam is self-locking at rest, but vibration and thermal cycling plus creep of the contact surfaces can back it off.

Materials

Buy the stamped latch in stainless for outdoor use. Zinc-plated steel only gives 12-72 h salt spray and yellow zinc 96-120 h (albanycountyfasteners, from the lead). For printed parts use ASA for the wings, cam and recess, with a stainless pin and keeper. PC (UV-stabilised) for more creep margin. Not PETG at 60-70 C.

Hardware notes from the research

Stamped stainless (A2/A4) butterfly latch and keeper (buy), or for a printed copy a 4 mm stainless pin, stainless keeper screw or bar and a stainless spring.

Design guidelines

• Use it as a secondary keeper or alignment aid beside a proper clamp. Do not use it as the primary clamp for a solid 85A TPU gasket.
• If printing a twist-cam version: cam radius ~8 mm, 6 mm per revolution lead (6-8 deg lead angle so it self-locks), quarter-turn stop, wing span 36-45 mm, hook arm >= 10x5 mm printed flat with the load in XY.
• Make the keeper an adjustable M4 stainless screw or bar so preload can be retensioned. Add a printed hard stop beside the latch at 0.75-0.8x the free bead height.
• Use a 4 mm stainless pin with bearing <= 8 MPa (4x8 mm = 32 mm^2 is good for ~250 N sustained). Pin axis along Z.
• Add a detent or spring so the butterfly cannot back out in vibration. Keep the recess drained.
• Fillets >= 1.5 mm on the hook and cam root. ASA has only ~3 kJ/m2 notched Charpy at -30 C.
• Place latches at the rim-bending pitch limit (~28 mm for a 3 mm flat lid edge, ~110 mm with a 20 mm skirt). Several latches of ~120 N are what the lid needs.

Failure modes

• Hook arm creep or fracture at 60-70 C
• Cam or wing wear and slip under vibration, so the latch backs out
• Wing breakage when over-torqued by hand (peak ~400 N)
• Gasket compression set loses 25-49 % of the force with no way to retension
• Zinc-plated steel corroding outdoors (96-120 h salt-spray class)
• Recess collecting water and debris
• FDM tolerance giving too little or too much squeeze

Existing designs and commercial references

• Penn Elcom medium MOL recessed butterfly latch L905/928PBZ (commercial steel reference, not printable) — Verified: 90 kg max working load, 1.2 mm steel, zinc, 105x102 mm, padlockable. Page does not say twist or double-throw.

Evidence gathered in the sweep

COMMERCIAL: Penn Elcom medium MOL recessed butterfly latch L905 series, 90 kg (~880 N) maximum working load, 1.2 mm steel, zinc finish, verified by me https://translations.penn-elcom.com/medium-mol-recessed-latch-with-padlock-bracket-in-deep-dish-l905-928pbz . No printed estimate; the sustained force of a printed copy was not estimated and is expected to be low. CONFIDENCE: low.

Sources

• Penn Elcom medium MOL recessed latch L905/928PBZ — Verified by re-fetch: 90 kg working load, 1.2 mm steel, zinc finish, 10.5x10.2 cm, 13 mm dish, 231 g, listed as a butterfly latch with no double-throw wording.
• Lee Spring: music wire versus stainless steel — Verified: music wire rusts and tops out ~121 C; 302/304 stainless to ~288-316 C. Use stainless springs outdoors.
• Albany County Fasteners: fastener materials and coatings guide — From the lead (zinc plate 12-72 h and yellow zinc 96-120 h salt spray); not re-fetched.
• CNC Kitchen: comparing PLA, PETG and ASA — Verified: standing vs lying hook strength (ASA 17 vs 57 kg); ASA softens 110 C.
• CNC Kitchen: Prusament PC Blend review — Verified: PC Blend hooks 74 and 77 kg lying vs 22-35 kg standing, softens ~110 C, fails 125 C; no creep data.
• filamentspecs: creep resistance of 3D printed parts — Verified: PETG joint lost ~11.5 % preload in 7 days vs ~3 % PLA.

Sources the reviewer flagged (dead, blocked or misdescribed):

• https://translations.penn-elcom.com/medium-mol-recessed-latch-with-padlock-bracket-in-deep-dish-l905-928pbz is quoted correctly (90 kg working load, 1.2 mm steel, zinc) but never says 'double-throw' or twist. The 'double-throw' naming is unsupported
• https://www.albanycountyfasteners.com/media/48/29/31/1764608481/fastener-materials-and-coatings-guide.pdf: the salt-spray hours (12-72 h zinc, 96-120 h yellow zinc) could not be read by my fetch and are unverified
• https://www.leespring.com/music-wire-versus-stainless-steel: the 121 C limit for music wire is right, but the 288-316 C figure is a general stainless range, not stated for 302/304, and '25-30 % weaker' is the analyst's derivation, not a page statement
• https://www.cnckitchen.com/blog/prusament-pc-blend-review was not re-fetched by me; the PC Blend hook numbers (74 / 77 kg lying, 22-35 kg standing) are unverified in this review