Anish Laha/Boeing 737 Yoke and Autopilot PanelSHEET 03 / 12 ›
DWGAL-03

Boeing 737 Yoke and Autopilot Panel

A two-axis flight-sim yoke and an autopilot panel that I designed, built and wired myself. Push and pull for pitch, rotate for roll, each axis sensed through its own mechanism.

Role
Designer and builder, solo
Context
IGCSE Design Technology (Systems and Control) coursework, Bangkok
Constraint
8 weeks, about 24 hours of workshop time, school tools only
Tools
Fusion 360Laser cuttingFDM printingArduino MegaSolderingC++ (Arduino)
Status
Built and tested: every function spec fully met
2 axesPitch and roll, each on a potentiometer
4 + 10 + 2Rotary encoders, push buttons, toggle switches
37-segment displays (MAX7219)
2,847 THBTotal cost of materials and components
The finished module on the bench
The finished module on the bench
Flying a 737 in the simulator with it
Flying a 737 in the simulator with it

What it is

Most people fly a simulator with a keyboard and mouse. I wanted a physical yoke that moves the way the real one does, plus the autopilot panel from the glareshield, so heading, speed, altitude and vertical speed are set by turning real knobs.

TRY ITYour pointer is the yoke. Left and right rolls, up and down pitches.

The mechanism

Exploded view of the internal mechanism
Exploded view of the internal mechanism

Pitch: linear motion to rotation

The handle sits on a PVC control rod. The rod is carried by a platform that slides on two steel rods through linear bearings, so push and pull is smooth and doesn't bind. A rack is fixed to the base, and a 72-tooth pinion on the platform rolls along it and turns a potentiometer.

The pinion size was the design decision. A potentiometer only turns through about 300°, so the gear had to be big enough that the yoke's full stroke used most of that range without running past the end stop.

CAD: the pinion on the rack, with the potentiometer behind it
CAD: the pinion on the rack, with the potentiometer behind it
As built: printed pinion, rack, and the platform on its rods
As built: printed pinion, rack, and the platform on its rods

Roll: rotation, sensed directly

The control rod turns inside its holder. A second potentiometer sits on the end of the rod in a laser-cut mount cut to the shape of the potentiometer shaft, so the shaft is driven by a friction fit and a slot.

CAD: yoke, control rod and mechanism
CAD: yoke, control rod and mechanism
As built, from above
As built, from above

Centring and feel

Rubber bands pull the yoke back to centre. The user sets the stiffness by adding or removing bands. It works, but it's the weakest part of the design, and it's the first thing I redesigned (see below).

Electronics

Final circuit diagram
Final circuit diagram
PartQtyJob
Arduino Mega1Reads every input and talks to the flight sim
B10k potentiometer2Pitch and roll position, on analogue inputs
Rotary encoder4Heading, airspeed, altitude, vertical speed
Push button + LED10 + 10Autopilot modes, each with a status light
Toggle switch2Auto-throttle and similar on/off functions
MAX7219 7-segment display3Shows the autopilot values
Stripboard15 V and ground distribution to every component

I started on an Arduino Uno and it couldn't hold a stable connection with the sim once the code grew, so I moved to a Mega for the extra memory and pins. Power is distributed on a soldered stripboard, with jumper wires running to the Arduino.

The three displays under test
The three displays under test
Soldered power distribution board
Soldered power distribution board
Wiring inside the case
Wiring inside the case

Case and build

Printed yoke handle before finishing
Printed yoke handle before finishing
Fitting the autopilot panel
Fitting the autopilot panel
Printed pinion on its laser-cut hub
Printed pinion on its laser-cut hub

What didn't work, and what I changed

ProblemCauseFix
The first potentiometer mount caught fire in the laser cutterCut lines were too close togetherSimpler mount shape with wider spacing
Potentiometer wouldn't fix to 3 mm MDFThe threaded section was too short for the nutMounted it on 1 mm stripboard
Unstable connection to the simNot enough memory on the UnoMoved to an Arduino Mega
Displays flicker and lagNot enough power, and too many displays for the board to drive directlyv2: multiplexer board and a separate supply
Yoke feel is staticRubber bands can't change force with airspeedv2: stepper-motor force feedback

Version 2 (designed, not built)

v2: stepper motors on printed brackets replace the rubber bands
v2: stepper motors on printed brackets replace the rubber bands

Result

Every function specification was fully met and checked by user testing: two-axis control mapped accurately to the sim, a working autopilot panel with a gear lever, adjustable tension, and a case that opens for repair.

The finished module
The finished module
And from the other side
And from the other side
X 0000 Y 0000 HDG 000 ALT 35000 FT