# Just moozing

Before you can check your notes, you must make them…

## Fiat lux aka. doing high school physics experiments

with one comment

Having built a solar air heating device, we want to measure the energy it collects and the flow of air it produces. With a microprocessor it is simple and cheap to get a temperature measurement. Flow is more expensive – so we are trying to built one from light bulbs.

## Introducing the idea

In essence we are doing a hot wire anemometer using a light bulb. The basic idea is to break the light bulb, heat the filament and measure the change in temperature when an air current blows past the filament. An example from YouTube.

We have tried it using a variable power supply with current limitation. When the filament is heated until it glow, blowing on it show an obvious reduction in illumination and fairly large changes in the resistance of the filament. So we are certain that the idea is sound.

Bonus information: Have light bulbs in stock. They break if they are heated too much and they burn (as in reacts with air), which reduces their lifetime.

## First experiments

In order to refine our way of measuring the resistance of the filament, we decided to not break it, and use a normal function light bulb. The ones we are using are common 12V/5W light bulb used for cars and trailers.

The setup is simple enough to be a high school experiment. Measure voltage drops and calculate currents, power and resistances. The power is supplied by a variable voltage (Vin), which is current limited to avoid surprises.

1. a1 and a2 are the voltages measured using a Velleman k8055 board. It has ADs which return an integer value between 0 and 255.
2. First conversions: $V_L = \alpha_1 a_1$ and $V_{in} = \alpha_2 a_2$ where a simple calibration gives that $\alpha_1 = \alpha_2 = 1.93 mV/inc$
3. Voltage drop is measured over a known resistance: $I = \frac{V_L}{R_L}$ with $R_L = 12 \Omega$
4. Lamp resistance Rf: $R_f = \frac {V_{in} - V_L}{I}$
5. Power dissipated in lamp: $P_f = (V_{in} - V_L) \cdot I$
6. Estimated temperature based on resistance: $T = \frac {1}{c} (\frac {R}{R_0} - 1) + T_0$ where c is the temperature coefficient of resistance (usually referred to as $\alpha$). For tungsten $c = 4.5 \cdot 10^{-3} K^{-1}$

The experiments are done by manually increasing the power until a2 is saturated and the reduce power again. The Velleman board is well supported by Python, so I did a quick adaptation of an example script and piped the data into a CSV file.

## Results

We did a couple of runs and the first results are shown in the graphs. The others were similar.

I have include the R2 value, and without too much data fiddling, I get a good fit. We saw a the lightbulb just starting to glow faintly – that should be a bit less than  750 K – so we not that wrong in our estimates, but it should be checked further.

• The PSU used has a settling time of around 10 seconds. That pollutes our measurements.
• I have done some calculations on the uncertainties, and – surprise – having a measurement of 1 +/- 0.5 that is used repeatedly gives ridiculous uncertainties.
• The nice linear relation between temperature and power, surprises me, I had anticipated T^4 relation due to radiation.
• Latex in WordPress is cool. I use it for equations.
• Fiat lux I got from reading this novel.
• Next step is to break the light bulb and see if we can recreate the results. We also need to decide if we want to measure using a fixed power and work with the change in R, or if we want to work with a fixed R and estimate flow from the power.