Otto Lock

Otto LLC, 2016 · Product development co-op: production parts, test fixtures, vendor bring-up

A smart deadbolt from a Bay Area startup, which I helped take from CAD to a production line during a nine-month co-op. I designed production parts, built the fixtures that tested them, and then taught two vendors to run those tests on their own lines. Two patents were earned and can be seen in the patent section.

Exploded render of the Otto lock: interior and exterior units, deadbolt and strike on a door — Fig. 1. The Otto lock and its bolt assembly.

Table 1. Summary
My part	Production components and packaging, assembly and test fixtures, vendor training and on-site FAI in China
Tools	Siemens NX, GD&T drawings, first article inspection, optical measurement, in-house and outside 3D printing, machining
Test hardware	Mark-10 gauges and desktop torque tester, FUTEK torque sensor, toggle clamps, linear rail, voltmeter
Patent	US 10,683,677, intelligent door lock system for use with a door assembly

TL;DR. Otto made a smart lock. I designed a few of the parts inside, then built the fixtures that check each lock before it ships. Every station is simple enough for anyone to run and easy to hand over to a machine later, as we scaled. Vendors were taught how to do assembly verification and quality control with these fixtures.

1. Production parts

I modeled several production components using Siemens NX, then took them through the full engineering steps. When the first article inspections came back I measured them myself against those drawings, partly on an optical measuring machine where I had designed automatic protocols.

Render of a cork box with its lid lifted, holding the lock's parts in fitted pockets — Fig. 2. Packaging for the lock.

2. The bolt test

A predecessor to my later work. The bolt ships as its own assembly and the customer puts it in the door, so it has to pass on its own before it leaves. My fixture checks seven things about the bolt in four moves by the operator. A Mark-10 reads every force, and a voltmeter confirms the whole circuit.

Every operator step is either flipping a toggle clamp or taking a reading. That was on purpose. Otto planned to automate this once volume went up, so swapping the clamps for electric actuators and letting a computer take the readings is all it takes.

Bolt test fixture on an aluminum plate with toggle clamps, a linear slide and a Mark-10 gauge beside it — Fig. 3. The bolt test fixture.

3. Bezel torque

The bezels on the lock have to turn with the same feel every time, so I adapted the bolt fixture to measure their torque. A toggle clamp pushes a FUTEK torque sensor into the bezel. The trick is a spring on a linear rail behind it, which keeps that contact force identical from part to part no matter who's pulling the clamp.

For the operator it's simple and very difficult to do incorrectly. Close the clamp, pick inner or outer bezel in the program, hit record, and turn the crank. The program logs two full turns, just to be sure. Put a motor on the crank and an actuator on the clamp, and the whole thing runs itself off one button.

Bezel torque fixture: a red FUTEK torque sensor inline on a rail with a toggle clamp and crank, a PC showing the torque program behind it — Fig. 4. The bezel torque fixture and its test program.

4. Eject bezel

The eject bezel only turns 70° clockwise and then springs back, so there's one number to check. It has to be stiff enough that nobody trips it by accident, and light enough to turn without a tool. A plate was designed that fits into a Mark-10 desktop torque tester. You set the bezel on it, turn it the full 70°, and read the peak torque.

Mark-10 desktop torque tester with a custom orange mounting plate holding the eject bezel — Fig. 5. Eject bezel fixture on the Mark-10.

5. Pogo pins

The bolt talks to the lock through six pogo pins, one for power and five for data. I wanted bad pins caught before the bolt was ever assembled. This board clamps five bolt backs at once, the operator puts a voltmeter on the power rail, and then touches each data pin in turn. A reading means the pin is good. It's the simplest fixture of the bunch.

Orange fixture board with toggle clamps holding bolt backs while an operator probes the pogo pins with a multimeter — Fig. 6. Continuity check on the pogo pins.

6. Teaching the vendors

Once the fixtures worked in house, I traveled to two first-tier vendors in China. These lines had moving parts, wiring and Otto's own test software all in one station. One vendor was great at making parts but had never done assembly, and our schedule had just been pulled in. So I took on teaching their technicians to build mirror copies of my fixtures and bring the lines up.

Vendor technicians assembling a copy of the bolt test fixture on a workbench — Fig. 7. Vendor technicians building a mirror of the bolt fixture.

7. How we proved it

A test station is only worth something if the vendor runs it the way we meant. Step by step:

Run every fixture in house first, against my own operating procedure.
Hand that procedure to the vendor and have them write their own.
Stand at their line and check their procedure, and their operators, against ours.
Run first article inspection on the vendor's critical parts on site, and work through design changes, tooling and fit checks until they built to spec.

An operator at the vendor's line running the bezel fixture while their procedure is reviewed — Fig. 8. Checking the vendor's procedure on their copy of the bezel fixture.