Building an air raid siren
When I experimented with building a blower for my small dust collector, I noticed some configurations made quite a howl. When building a blower, you try to minimize noise. But this inspired me to try building a blower that maximizes sound --- basically, a siren!
Building the impeller hubThe siren has a substantial size rotor, or impeller. To mount this hub on the motor shaft, I needed to figure out a precise way to make a hub, with a circular part that was exactly the diameter of my largest forstner bit.
Bandsaw 3 motor pulley.
But with its rectangular outside, I wasn't keen on having my hands or chisels
near it while it was spinning.
I previously cut most of the waste away with my bandsaw, and I was only taking off two millimeters at a time on the table saw. It worked out well.
I wasn't sure if it would work though, so I kept way back from the action. I couldn't see if I was cutting, so I was listening to it to know if I was cutting. The sound changed once I got past the cylindrical part of the hub, which told me that I had pushed the motor far enough.
Building the impellerNext, making the winglets for the rotors. I made a pattern out of plywood, and used that to mark the outline on a piece of spruce.
I'm not using the good birch plywood here. This whole thing was very much an experiment, so I didn't want to use up any "good" materials. In fact, All the plywood for this project was trash picked.
It spun up ok the first time. But then I spun it up again to get video, and disaster struck. One of the winglets flew off. This totally threw the rotor off balance. The vibrations from that quickly loosened the bar clamps holding down the motor. With all the vibrations, it didn't take long for the motor to fall off the bench.
Fortunately, the power cord was caught behind the vise, and by pulling on the power cord, I was able to save the motor from crashing to the floor.
Other than the missing winglet, the rotor was undamaged.
Based on power measurements before that disaster, I also figured the layers of the rotor should be thinner, so I passed the rotor over my jointer a few times to plane the winglets down to 29 mm thick.
I last used that same trick when I built a crokinole board
I removed the table saw insert for this step to try to suck more of the sawdust into the table saw.
The rotor has three layers of plywood sandwiching two layers of winglets. The first layer has six winglets, the second ten. The idea is that the siren will produce two notes simultaneously, just like WW2 air raid sirens in movies.
After gluing on the second layer of plywood, I realized --- oops, I forgot to lay out the geometry for the ten winglets. Without a center point, that's tricky to lay out.
So I wrapped a flexible tape measure around the perimeter, and it worked out to be exactly 95 cm around. So to do the layout for the ten winglets on the second layer, I just ran the tape around the edge, fixed it in place with some clamps, and then made a mark on the rotor for every 9.5 cm of circumference. You can barely see the yellow tape measure around the edge in the photo.
Balancing the impellerMy impeller turned out to be quite heavy by now. Considering the consequences of just one winglet flying off earlier, I knew it needed to be well balanced.
The "balancing it on a marble" method that I previously used for my dust collector impeller worked well enough, so I used it again to balance this rotor.
I jammed a marble, just slightly larger than the 5/8" hole into the middle of the hub...
It also turned out that the marble I was using wasn't completely round, so that made it very difficult to fine-tune the balance. Once I realized that and replaced the marble, it got easier, but by that time I had already drilled a few more holes than necessary!
This was a much easier shim job than when I shimmed some plumbing pipe up to fit into a 7/8" hole when I built my 14" bandsaws.
I also stuck a large, cut-off nail into the shaft's keyway. I had filed a notch into the side of the hole of the hub, so the nail acts as a "key" to prevent the shaft from spinning inside the hub.
As it turned out, 1181 watts is about the rated power for that motor, but then again, I'm not planning on running this siren for long periods of time, so it should still be within reason.