In this article, we’ll learn a bit about diacs and triacs while repairing a dimmer switch.
Initial conditions
A while ago one of my house’s lamp dimmer switch died, which meant I had an excuse to dabble in electronics again, yaye!
Opening up reveals an interesting circuit:
And quickly our attention turns to a likely failure point: a BTB600. Is it a mosfet? It’s not!
The triac
A quick look at this component’s datasheet reveals it’s a triac.
It’s a really fun component:
- It’s a thyristor:
- It will conduct when current is applied to its gate
- Unlike the transitor, it will keep conducting if current is removed from the gate – shutting it down requires stopping the current flow between its anodes.
- It’s a double thyristor, made to work in alternating current, so its inputs are named Terminal (or Anode) 1 and 2.
As a thyristor, it’s easy to test out:
The gate is the right-most pin: when connected, the LED lights up, and will stay on even if we disconnect the gate.
With this simple setup we test a freshly-bought triac against the one extracted from our device: the latter does not work, so it’s our failure point.
Simplified dimmer circuit
The principle of a traditional light dimmer is that on each rise of the AC sinusoidal, a delayed trigger will start conducting out of phase, limiting the amount of total power sent to the lamp.
Note: this goes to show why such dimmers cannot work on LED bulbs as-is: while a lamp will absorb the on/off periods and simply shine less bright the smaller the triac’s on-time is, an LED transfomer will not magically reduce current based on the form of its input’s wave form. Thankfully, there exists LED bulbs that do work on such dimmers, requiring more circuitry.
Such a typical circuit looks like this (simplified):
This introduces a new component: the diac. It’s an AC diode that starts conducting at large-ish breakdown voltages like 30 V.
Hence the above circuit makes sense:
- When the sinusoidal starts rising, the triac is off.
- Current flows through C1 limited by R + R1: the voltage across C1 rises at a rate that depends on R1.
- The voltage across C1 eventually reaches the diac’s breakdown voltage: C1 discharges rapidly in the triac’s gate and triggers it on.
- For the rest of the half-period, the triac stays on, and virtually no current passes through C1 anymore.
- the sinusoidal reaches 0v, the triac turns off, and the cycle repeats.
And how much of the period is cut off depends on the variable resistance’s value. Neato! You can see the circuit in action over at falstad.com: select “Circuits” / “Misc Devices” / “DIAC TRIAC Dimmer”.
Something feels weird though: what happens when T2 is lower than T1 (respectively -120 and 0 V)? What is the potential of the gate? If it was related to T2, you would immediately get a large potential across the diac (from -120 V to 0 V at the start, since C1 starts with a voltage drop of 0), triggering the diac instantly, so that can’t be it.
Instead, my understanding is that the gate’s potential is always related to T1, hence why the capacitor + diac group is across the gate and T1, and it can trigger with either a negative of positive potiential. So a triac is not exactly the same as two thyristors glued together, and T1 and T2 are not interchageable. In our case, when T2 < T1, we enter what is called quadrant 3, because the gate is also < T1.
Dimmer’s actual circuit
Note we’ve looked at a simplified circuit. In the photo above we see at least one large inductor and two capacitors. The basic intention is apparently to filter the circuit, with these additions:
I reversed the actual circuit, which gave me something like this:
We see:
- The filtered dimmer circuit above
- A fuse F1
- TS1 which I believe is a thermal switch or thermal fuse of some kind - it is glued to the coil.
- An LED indicator (2mA with its 120 K resistor)
The repair
Anyhow, replacing the triac with a newly-ordered one fixed the circuit – success!