A thermistor and a divider

You need the light divider working or its parts to hand: one temperature sensor, one 10 kΩ resistor, the meter and the power module. As in the last article the numbers are for a typical part. The listing does not name the sensor, so treat 10 kΩ at room temperature as a guess until you have measured yours.

A resistor that listens to heat

The temperature sensor is almost certainly an NTC thermistor, the small black bead on two wires. NTC means negative temperature coefficient: warming it lowers its resistance, because heat frees charge carriers in the ceramic. Typical beads sold at 10 kΩ fall about four percent for every degree.

The curve is not a line. Resistance falls faster per degree when the bead is cold than when it is warm. A meter that reads resistance still gives you the temperature, if you have the curve, and the curve is the part you have to find.

Measure the bead

With the bead out of any circuit and the power off, put the meter on the resistance range and a probe on each wire. Either way round is right.

• At room temperature, held in nothing, left to sit: write it down. This is your bead's R at room temperature, and what the whole calibration anchors on.
• Pinched between two fingertips for 20 to 30 seconds: the reading falls. Write the lowest number down.

Skin is warmer than a room by about ten degrees, so the fall is large: about a third. That change is the thing you can trust. The absolute number is not yet.

Read it with a divider

Resistance does not read easily on a live circuit, so put the bead in a divider: the bead from the plus rail to a node, a fixed resistor from the node to ground. The node sits at the fixed resistor's share of the rail, and it rises as the bead warms and falls in resistance.

Choose a fixed resistor equal to the bead's room-temperature value. Then the node is half the rail near room temperature, where the curve is steepest and the reading changes most per degree. For a 10 kΩ bead that is the kit's 10 kΩ.

Build it

The power module is on its 5 V side, so the rail is about 4.8 V.

1. A jumper from the plus rail to column 2, row a.
2. The bead's legs in column 2, row d and column 4, row d. The strips are different, so the bead is the only path between them.
3. The 10 kΩ from column 4, row c to column 9, row c.
4. A jumper from column 9, row b to the minus rail.

Measure the divider

Meter on DC volts, black lead on the minus rail. First measure the rail with the red lead on the plus rail: the divider is a share of whatever the rail really is, and the rail moves a few hundredths from one minute to the next.

Then put the red lead on column 4, the node.

• Room temperature: about half the rail, 2.2 V on the modelled bead, with the rail at 4.8 V. If your bead is not 10 kΩ it will be higher or lower.
• Pinch the bead for 30 seconds: the reading rises by about half a volt, a little over 0.05 V per degree. Let go and watch it fall back over a minute.
• Divide the node reading by the rail reading. That ratio is steadier than either number alone, and it is the number to put in the formula.

Turn a reading into the bead's resistance with the fixed resistor: R = fixed × (rail ÷ node − 1). The same bead measured this way and measured directly on the resistance range should agree to within the meter's error. If they do not, the divider is miswired.

From volts to degrees

Resistance to temperature uses the beta equation. With your bead's resistance R0 at a temperature T0 you know, and its B, the temperature is

T = 1 ÷ (1/T0 + ln(R ÷ R0) ÷ B), with every temperature in kelvin, which is Celsius plus 273.15.

The datasheet's typical B is about 3950 K, and your bead's is a few percent from that. So calibrate: read the bead's resistance at two temperatures you know, and B = ln(R1 ÷ R2) ÷ (1/T1 − 1/T2).

Use a room thermometer for the first point. For the second, warm tap water in a cup works: put the bead in a small zip-top bag, with the bag in the water and a kitchen thermometer beside it, and read the resistance on the meter's resistance range with the power off and the bead out of the breadboard. Use warm tap water only: never hot, never boiling, and keep the breadboard and the power module dry.

Take the second point much warmer than the first. B is a ratio of two small differences, so two points close together turn a thermometer's error into a big error in B.

How good is it

Not very. A bead warmed by a finger, a meter reading and a cheap thermometer to calibrate against give a temperature good to a couple of degrees at best. The part's own tolerance, the thermometer's error and the air around the bead all count. It is a good way to see temperature happen on a meter. It is a poor replacement for a thermometer.

Part of the error is the bead itself. The divider current flows through it and warms it a little, and the smaller the fixed resistor the more it does.

Measure what changes

Pinch the bead and watch the node rise. Blow on it gently and watch it fall back toward the room reading. Hold it close to, not touching, the power module's regulator or a warm hand and watch the reading follow. Each of those is a few tenths of a volt and a few degrees, which is the whole instrument: a resistor that does one thing, a second one that does not, and a meter between them.