Ever since I completed my Engineer’s Hookah, I’ve been interested in quantifying the smoking experience – what the flow profile of the smoke is, what the temperature of the various parts of the hookah are, how the smoke’s density changes over time – and correlating those metrics to the “quality” of the smoke. Thus far, I’ve focused primarily on the first metric – smoke flow profile – and have to date designed and built three different flow meters to measure it.
Measuring the flow of hookah smoke is not a particularly straightforward task. On average, the flow of smoke through a hookah is quite small (2.15 x 10^-4 m^3/s) with a low Reynolds number (Re = 1000-3000). Additionally, it’s an extremely “dirty” flow – although a gas, it carries a significant amount of suspended particulate material, which is a) sticky, and b) corrosive. Finally, any flow meter used should not impact the user’s smoking experience, with the total head loss caused by the meter remaining under 0.1 kPa.
The current iteration of this project is shown above. It’s a wedge flowmeter, designed and built as a final project for my Fluid Mechanics class. Wedge flowmeters are differential-pressure meters like venturi or orfice plate flow meters. They are particularly suitable for this application for a number of reasons: they function at extremely low flow rates, are relatively simple to construct, and are self-scouring (the obstructing surface is automatically cleaned by the flow. I manufactured both the meter body and the wedge element from 303 stainless steel. The pressure differential is conducted through food-safe tubing and fittings to a Sensirion 1000-L025 differential pressure sensor, which is read by an Arduino Pro Mini. Thus far, the meter has performed well, with no evidence of fouling either in the meter or in the sensor.
Some more images of the flowmeter:
As I mentioned, I’ve been working on this problem for a while – unfortunately, my previous designs were not so successful. My first attempt at this project (made in 2009, after I’d finished the hookah) was a simple drag-based flow meter – flow through the chamber (top to bottom) exerts drag on the stainless pendulum suspended inside, which is resisted by the spring supporting the pendulum. While visually attractive and (theoretically) corrosion-resistant, it became immediately clear that this system would not work. The small gap required to produce drag on the pendulum clogged easily, quickly interfering with the meter’s operation. Additionally, the “bubbling” behavior present in the flow caused the pendulum to bounce wildly while in operation, making it hard to take a reading from the meter.
My second attempt, made in Fall 2010 as the final project for my Fundamentals of Digital Systems course, used a paddlewheel to measure smoke flow. The Teflon paddlewheel was enclosed in an aluminum frame which directed flow over the wheel; a simple encoder pattern was printed on the top of the paddlewheel, and was read by an image sensor taken from an optical mouse. I used an Arduino Mega, and borrowed heavily from Martijn Thé’s Optimouse libraries to read from the sensor. Unfortunately, this design was also unsuccessful. The paddlewheel element rotated unreliably under smoking conditions, and when it did rotate, the encoder signal aliased quickly.