Constructing a wind tunnel in theory is not a hard thing to do. I first did some internet research as to how I might build a simple wind tunnel, starting on the Internet.
NASA has some links up for educators and I looked through them but didn’t really find much at the level I needed.
Next I used one of my preferred Internet search techniques – doing a Google search, viewing the “images” option, then clicking through to the pages which display promising images: https://www.google.com/search?q=wind+tunnel+construction (The preceding link may be browser specific and may not work for you.)
Stanford had some of the most useful material for wind tunnel theory and design
The outline and engineering drawings and photos all started to look fairly familiar, although I realized from the photos of these things that the scale could vary hugely.
After reading through some material I started to categorize wind tunnel parts, and there were not that many functional categories.
- Some kind of fan or blower, sometimes termed a power plant, to power the wind tunnel
- Tuning vanes in the corners of the wind tunnels, to help reduce static pressure drop and even out the airflow.
- Some sort of honeycomb or turbulence “straightener” to help eliminate turbulent flows.
- Similar to the items above, screens introduce some static pressure drop and help even out uneven air flows.
- A contraction to increase wind speed (and even out the ratio of air flow even further).
- Longer lengths of tunnel, sometimes called diffusers or settling chambers allow turbulent air to settle.
- Since I would need temperature control in my wind tunnel, a heater or combination heater/cooler would be required to provide the necessary temperature controls.
I also had to decide on an open loop or closed loop tunnel. Originally I had decided that a closed loop tunnel would be necessary for temperature control. As I got into though I saw that an open loop variety tunnel could still work with temperature control, and perhaps be more simple to construct. If I needed to close the tunnel loop later for even tighter temperature control, this would be something that was not would not be that hard to do.
Most authors seem to recommend deciding on a test chamber and working backwards from your specifications at the test chamber. Because my sensor was very small (on the order of 2″), I decided that a 1.5″ test chamber would be sufficient. For the top wind-speed specification, I somewhat arbitrarily decide that around 40 mph would be good enough.
As to the shape of wind tunnels I was seeing, there didn’t seem to be a lot of preference for either square/rectangular tunnels or round tunnels. I figured that, on a large scale, square tunnels would be easier to construct. However for me, PVC pipe from a home improvement store would be a fairly appealing build material. It also had the advantage of being modular, so I could take the thing apart once the experiments were done and it wouldn’t be taking up a lot of space in the shop.
The contraction of a wind tunnel seems to be one of the crucial design features. The contraction’s job is to increase the wind speed, without creating turbulence. This allows the fan to run at lower speed and allows the pressure drops generated by the screens to be lower. The contraction also “further reducing percentage variations in velocity”, according to the Stanford site so in layman’s terms it “helps smooth out variations in wind speed” which would lead to turbulence. Although, several design guides recommended “as large a contraction as possible”, I decided, completely arbitrarily, that 3″ pipe, hence a 4 to 1 contraction would be good enough. After all I wasn’t building a high-speed wind tunnel and didn’t expect fan power to be an issue.
I’d like to be able to show you my beautiful working drawings, but the truth is that I went to my local “big box home store” and bought some 3″ PVC, and started building. I item I found immediately useful was a coupler called a “repair coupling”. Regular PVC couples have a stop in the middle, so that the pipe is force fitted in close at the stop. A repair coupling can slide right over two pieces of pipe, facilitating easy assembly, disassembly and adjustment.
I debated whether I needed turning vanes in the pipe elbows. All the wind tunnel plans always seem to show turning vanes, so I made one elbow with some vanes then did a test with a wind sensor before making a decision about making more. The results of the wind speed tests seemed to show about 20-30% improvement in speed variation (remember the results are non-linear so this is a bit of guess) with the vanes installed so I decided that they were worthwhile.
Below are some photos of how I made the vanes.
First I bandsawed an elbow in two to measure the radius, then calculated the rectangular section that would make my vane.
Next I laid out the rectangle on some flat galvanized metal ductwork I had picked up for an earlier project at the home improvement store. Below is a photo of laying out the drawing on the metal. Professional machinists often use some kind of “marking blue” for this kind of layout but putting down a patch with a Sharpie, letting it dry for a few seconds and scratching a line in the Sharpie patch seemed to work fine.
I then just out the rectangles with tin snips and bent them to radius by hand. The results were far from perfect – they still needed some trimming in length from my calculations, and gluing them in was a bit of a trial. (I discovered something called “epoxy gel” that was useful.) Still over the course of half day or so I had two elbows with turning vanes. I believe that is all that I will make as I think the turbulence from the fan will probably dwarf any other unevenness in the return air flow if I decide to turn the wind tunnel into a closed loop. Here’s photo of the end of a pvc elbow with the turning vanes installed.
The power-plant (aka fan) I found (actually I ended with a collection of fans) was a nice fan from Digikey that ran on 12 volts, at up to 4 amps and had a PWM and tachometer pulse output built into the fan all for around $35. This saved all the hassle of having to build a separate control system, which wouldn’t have been a big deal, but I still am glad to have purchased one with the control system built in.
The nice modular heater was nothing more than the “guts” from a blow dryer I bought new, around $20 I believe. I had some more photos in process before I threw away the shell and deleted the motor but iphoto decided to eat them on import. It’s probably been a good lesson in why not to check the “delete after import” checkbox. In any case the blow dryer heater section was engineered to go in a tube – albeit a smaller one, and has a nice engineered heatproof structure.There are also some built in safety features such as a bi-metal over temp switch, and a thermal fuse, which I unfortunately blew out during software development and had to eliminate. I thought about replacing it but it was riveted in with some system that I couldn’t immediately get my hands on.
This photo kind of makes it look the tube is growing algae but it’s really just epoxy. The control interface for the heater is just an 8 amp SSR that we sell in the shop. I also re-purpopsed the on-off switch, and plug from the blow-dryer, which handily has a ground fault interrupter built into it.
In the next post I’ll go into the screens, turbulence control and home brew contraction (cone). And in the final post I’ll go over the software and control system.