Energy per step, per day, per season

A step can give up at most the walker's weight times how far the plate sinks. Springs take a share to lift the plate back; gears, generators and electronics lose more. What reaches the 24 V line is below.

Where one step's energy goes

Results

Electricity per step vs plate travel

Checks

Visitor figure: up to 60,000 on peak days (amNY, GrowNYC). Electricity price: Con Edison 2026 all-in residential average, about $0.29/kWh. Solar comparison is approximate. Steps per visitor are estimates to be checked with hand counts.

Bench tile: full build guide

A plywood plate on four springs pushes a rack. The rack turns a pinion, a belt makes a stepper motor spin almost four times faster, and the motor works as a generator. Two diode bridges turn its AC into DC, which heats a known resistor; an Arduino measures the voltage across that resistor and adds up each step's energy. Allow two afternoons: one for the frame, one for the board.

1 · How the circuit works

2 · Schematic

Arrows show the direction each diode conducts; the bar end is the stripe on the real part. Crossing lines without a dot are not connected.

3 · Parts (approximate prices)

Part	Qty	Why	Approx.
12 mm plywood, 60 × 60 cm	2	Top plate and base	$20
Compression springs, about 4 N/mm, ~25 mm long	4	Hold the plate up and return it	$12
Rack and pinion, module 1, 8 mm bore pinion	1	Turns plate travel into rotation	$12
KFL08 bearings and 8 mm × 100 mm shaft	2 + 1	Carry the pinion and big pulley	$10
GT2 pulleys 60T (8 mm bore) and 16T (5 mm bore), 200 mm belt	1 set	Spins the motor 3.75× faster	$12
NEMA 17 stepper motor	1	The generator	$14
1N5819 Schottky diodes	8	Two bridge rectifiers	$2
10 Ω, 10 W ceramic resistor (R1)	1	Known load	$3
47 kΩ (R2) and 10 kΩ (R3) resistors, ¼ W	1 each	Voltage divider	$1
1N4733A 5.1 V Zener (D9), 10 nF ceramic capacitor (C1)	1 each	Protect and filter the A0 input	$2
Arduino Nano or compatible, USB cable	1	Samples voltage, sums energy	$12
Perfboard 5 × 7 cm, 4-way screw terminal, standoffs	1 each	The circuit board	$4
22 AWG solid and stranded wire, heat-shrink, jumper leads	—	Rails, links, leads	$8
8 mm dowel, wood blocks, screws, washers for shims	—	Guides, mounts, end stops	$15
Total			≈ $127

4 · Tools and safety

5 · Build the frame

1. Cut two 60 × 60 cm plywood squares. Drill a 10 mm hole 3 cm in from each corner of the base; glue 8 mm dowels into matching spots on the top plate so they slide in the holes as plungers.
2. Slip a spring over each dowel between the plates. Unloaded, the top plate should sit about 10 mm above the end stops.
3. Screw the rack upright to the underside of the top plate, near one edge, teeth facing inward.
4. Screw two KFL08 bearings to a wood block on the base. Push the 8 mm shaft through them with the pinion meshing on the rack and the 60-tooth pulley on the other end. Tighten both set screws on the shaft's flat.
5. Mount the stepper on a second block with the 16-tooth pulley on its 5 mm shaft. Fit the belt, slide the motor back until the belt is snug, and screw it down. Without this 3.75× step-up the motor turns too slowly to get past the diodes.
6. Glue four end-stop blocks to the base under the corners. Stack washers on them until the plate stops after 6 mm of travel; you'll change this to 4 and 8 mm in testing.
7. There is no one-way clutch on the bench tile. On the way back up the springs turn the motor backwards; the bridges rectify that too, so you also measure the energy the springs give back.

6 · Solder the board, step by step

Click through the steps. The board view highlights what each step adds; faded parts come later. The view is from the component side.

What a good joint looks like

7 · Multimeter checks before any power

Probe between	Setting	Expect	If not
+ pad (column 22, top) and GND pad	Ω	≈ 10 Ω	0 Ω: solder bridge between the rails. Open: an R1 lead is loose.
Each diode: red on the plain end, black on the stripe	Diode	0.15–0.35 V	Swap probes and it should read OL. OL both ways: open joint. Low both ways: shorted diode or bridge.
A0 pad and GND	Ω	≈ 8.2 kΩ	About 10 kΩ: R2 side open. About 47 kΩ: R3 side open. Much lower: D9 is in backwards.
Each J1 screw and its row-7 node	Continuity	beep	The jumper wire's joint is open.
Column 5 node and column 10 node	Continuity	no beep	Solder bridge between the two bridges.
Stepper wires, before connecting	Ω	1–5 Ω in a pair	Wires from different coils read open: that's how you find the two pairs.

8 · First power-up

1. Find the coil pairs with the ohmmeter. Put one pair in J1's A+ and A−, the other in B+ and B−. Which wire of a pair goes where doesn't matter; the bridge rectifies either way.
2. Meter on DC volts across R1. Spin the big pulley by hand: the reading should flick up to a few volts.
3. Move the meter to A0 and GND. Stomp on the plate as hard as you can: it must stay under 5 V.
4. Plug the Nano into a laptop by USB, connect the A0 and GND leads, upload the sketch below, and open Serial Monitor at 115200 baud.
5. Step on the tile. Each step prints one line: step number, joules, peak volts and the running average.

9 · Logger sketch and calibration

Before trusting the numbers, measure three things and type them into the top of the sketch: R1 with its leads (R_LOAD), R2 and R3 to work out (R2 + R3) ÷ R3 (DIVIDER), and the Nano's 5V pin while on USB (VREF; often 4.7–5.1 V). The output is comma-separated, so you can paste it straight into a spreadsheet.

// Harvest Steps bench logger v2: energy of each step into the load resistor
// Put your own measurements in the four values below.
const float R_LOAD  = 10.0;   // ohms: R1 plus its leads, measured
const float DIVIDER = 5.70;   // (R2 + R3) / R3, from measured values
const float VREF    = 5.00;   // volts on the Nano's 5V pin, measured on USB
const float THRESH  = 0.20;   // volts across R1 that count as "producing"
const unsigned long DT_US    = 500;  // sample every 0.5 ms (2 kHz)
const unsigned long QUIET_MS = 400;  // this long below THRESH ends a step

float joules = 0, peakV = 0, total = 0;
unsigned long quietSince = 0;
int steps = 0;

void setup() {
  Serial.begin(115200);
  Serial.println("step,joules,peak_volts,average_joules");
}

void loop() {
  static unsigned long next = micros();
  while ((long)(micros() - next) < 0) {}    // wait for the next sample
  next += DT_US;

  float v = analogRead(A0) * VREF / 1023.0 * DIVIDER;   // volts across R1
  if (v > THRESH) {
    joules += v * v / R_LOAD * DT_US * 1e-6;              // P = V^2/R, times dt
    if (v > peakV) peakV = v;
    quietSince = millis();
  } else if (joules > 0 && millis() - quietSince > QUIET_MS) {
    steps++;
    total += joules;
    Serial.print(steps);      Serial.print(',');
    Serial.print(joules, 4);  Serial.print(',');
    Serial.print(peakV, 2);   Serial.print(',');
    Serial.println(total / steps, 4);
    joules = 0;  peakV = 0;
  }
}

10 · Test plan

1. Weigh three walkers of different builds on a bathroom scale.
2. With the end stops at 4 mm, each walker takes 20 normal steps across the tile. Save the serial output.
3. Repeat at 6 mm and 8 mm. Average each walker's joules at each travel.
4. In the Energy physics tab, set the mass and travel, then adjust the efficiencies until the prediction matches your average. That tells you which stage loses the most.
5. If 6 mm gives under 1 J, raise the step-up ratio (a 20T pinion or a bigger pulley) before changing anything else.

11 · Troubleshooting

What you see	Likely cause	Fix
Nothing prints when you step	Voltage never passes THRESH; motor spinning too slowly	Check the belt isn't slipping; spin the pulley fast by hand to test; lower THRESH to 0.1
Lines print with nobody on the tile	Noise on A0	Twist the A0 and GND leads together; check C1; raise THRESH
Only half the expected voltage	Both wires of one coil in the same bridge as the other coil's wire	Recheck the pairs with the ohmmeter; one pair per bridge
A0 reads near 0 always	D9 backwards, or A0 lead on the wrong pin	Stripe of D9 toward the A0 line; check the Nano pin label
Joules far below the physics tab	Diode drops dominate at low voltage; loose end stops	More step-up; check travel with calipers; heavier walker
R1 warm after drill tests	Continuous spinning, not steps	Normal; keep drill spins under 10 s