🚨 Alley Lamp

What you're building

A night light that also acts like an ordinary lamp. During the day the sensor sees sunlight and keeps the LED off — standard Automatic Night Light behaviour. At night the sensor sees darkness and the LED fades on by itself. The new trick: if you'd rather have the lamp on now, in the middle of the day, hold the override button. The LED comes up as if it were midnight. Release the button and the lamp goes back to listening to the sensor.

Alleys and stairwells use this exact pattern. The light runs on a dusk-dawn sensor so no one has to flip a switch at sunset, but there's a manual button next to the door for when the sun's still up but the hall is dark enough that you want the light anyway. One lamp, two reasons to turn it on, joined with an OR.

The big idea: two paths to the same base

So far every L3 project you've built had one input moving the transistor's base — the LDR (or the thermistor, or a trip-wire pull-up). Alley Lamp adds a second input: the push button, through its own 1 kΩ resistor, pulling the base-divider's middle node up toward +.

Both paths push the same middle node toward +. Either one is enough to saturate the transistor. The circuit doesn't know (or care) which one did it — current is current, and voltage above the 0.7 V wake-up line is above the 0.7 V wake-up line.

The result is a hardware OR gate — the LED fires if dark OR button-pressed. No logic chip, no code, just a resistor summing two currents into one node.

The sensing half — unchanged from Automatic Night Light

Divider: 10 kΩ on top, LDR on the bottom, middle node drives the 1 kΩ base resistor. If this is new, read Automatic Night Light first.

• Bright room, LDR ≈ 1 kΩ: V_middle = 4.5 × (1 ÷ 11) ≈ 0.41 V. Below the 0.7 V wake-up. Transistor off. LED dark.
• Dark room, LDR ≈ 100 kΩ: V_middle tries to climb toward 4.5 × (100 ÷ 110) ≈ 4.1 V. Base-emitter junction clamps it to ~0.9 V and drinks the extra push. Transistor saturated. LED on at ~10 mA.

The new piece — the override branch

Add a second path from + to the middle node, this one gated by a push button and sized by a 1 kΩ resistor:

+ → push button → 1 kΩ → middle node

When the button is up, that branch is broken — the circuit behaves exactly like Automatic Night Light.

When the button is down, the 1 kΩ sits in parallel with the 10 kΩ above the middle node. Two resistors in parallel combine as smaller than either alone:

R_top_effective = (10 kΩ × 1 kΩ) ÷ (10 kΩ + 1 kΩ) ≈ 909 Ω

The LDR is still 1 kΩ in bright light. New divider arithmetic:

V_middle = 4.5 × (1 ÷ (0.909 + 1)) ≈ 2.36 V

That's well past the 0.7 V wake-up line. The transistor saturates. The LED lights at ~10 mA — same brightness as the automatic nighttime glow, triggered by hand instead of by the sun.

Press the button at night: V_middle climbs even higher (two pull-up paths from +, one single tiny pull-down through the 100 kΩ LDR). The lamp was already on, so you see nothing change. The circuit's answer is "I was already going to do that, thanks".

The truth table — OR behavior

Four input combinations, three "LED on" outcomes:

LDR says	Button	V_middle	Transistor	LED
bright	up	0.41 V	off	dark
bright	down	~1.9 V (clamped by base)	saturated	on at ~10 mA
dark	up	~1.2 V (clamped by base)	saturated	on at ~10 mA
dark	down	even higher, clamped same	saturated	on at ~10 mA

Only the first row — bright AND button-up — keeps the LED off. Any one input is enough to light it. That's an OR gate, in resistor- and-transistor form.

Why 1 kΩ in the override, not 10 kΩ?

The override path needs to pull the middle node high enough to cross the 0.7 V wake-up line. If you used a second 10 kΩ as the override resistor, parallel with the top 10 kΩ, R_top_effective would be 5 kΩ, and in bright light V_middle would be 4.5 × (1 ÷ 6) ≈ 0.75 V — right at the wake-up edge. The LED might flicker on, might not. Reliability is 1 kΩ's job here: a strong enough pull-up to leave no ambiguity about which side of the threshold the circuit lands on.

If you used something smaller than 1 kΩ (say 220 Ω), the base current when the button is held would climb — (4.5 − 0.9) ÷ 220 ≈ 16 mA into the base. Enormous, inefficient, and makes the button hot after a while. 1 kΩ is the sweet spot.

Where OR gates on a transistor base show up

• Hall lamps with a dusk sensor + "turn on now" wall button — this exact circuit, in every modern US apartment building that predates smart bulbs.
• Garage lights that fire on either the motion sensor or the door-open switch — two inputs ORed into a single relay or transistor driver.
• Car headlights that come on either because the auto-light sensor says it's dark or because you pulled the stalk — the manual switch and the sensor both drop the same signal to the lamp relay.

An AND wants everything to agree; an OR is permissive — "any reasonable input is enough". This is the first project on this site where you've built an OR without needing more than resistors and a transistor you already own.