Introduction
The bandgap temperature sensor is a very common temperature measurement technology. Applications include (but certainly not limited to) monitoring engine coolants, air conditioning, power supplies, and scientific instruments.
How It Works
If a diode is operated forward biased with a known current then the voltage across it has a predictable change with temperature. This change is in the order of -2mV/°C.
Fig 1 - Bandgap temperature measurement with a single diode
While the above circuit certainly works, it's somewhat crude. It's more usual and ideal to use a diode connected (functions only as a diode, no amplification) transistor (Fig 2) for a more stable circuit.
Fig 2 - Bandgap temperature measurement with a single transistor
Two transistors can be used together if more accuracy is needed. An example is the Brokaw bandgap reference circuit (Fig 3).
Fig 3 - Brokaw bandgap reference circuit
The two transistors are operated with different currents (IC1 , IC2). Each transistor's emitter-base will have a different voltage (VBE1 , VBE2). The difference between these two voltages (ΔVBE) is proportional to the temperature - see FIG 4.
From the equation, k and q are constants - Boltzmann constant and charge on an electron respectively.
Since IC1 , IC2 and ΔVBE are known it is a simple task to rearrange the equation to calculate the temperature in Kelvin.
Advantages
It has many advantages including:
- Very low cost.
- Can be individually calibrated for excellent accuracy.
- Can be built into an IC.
- More linear response than thermistors.
Disadvantages
Even good things usually have some disadvantages:
- Less linear than resistance thermometers.
- Limited temperature range, typically the upper limit is 250°C.
- Fewer options for measuring off-board temperatures.
- There is potential for measurement drift, particularly in humid environments.