I tried using a car compressor for blowing the pipe, but it made a racket and destroyed the sound I wanted. I asked on instructables and someone suggested using an air duster (a can of compressed air basically). I bought one and it worked! It adds its own hissy flavour to the sound but in a good way. I'll upload a quality sample when I get one.
It was the first time I've asked a question on instructables and I'm grateful for the answers! Thanks everyone!
I was initially looking for a solenoid but couldn't find one so I used a servo. I wanted a solenoid because I wanted a simple circuit (I'm lazy!) Using a servo gives me better control but I had to use a microprocessor to do it.
I mocked up a frame for the can, the servo, a lever for mechanical advantage (the can nozzle takes quite a bit of pressure to press. Incidentally this frame would work for something that needed to spray with a spray paint! A la HEKTOR!). I worked up a basic assembly and tried different things until I had something reliable and as small as easily possible. The setup in the video is this version. The photos show the finished version.
I wanted wireless, so I went to a pound shop and bought a £6 wireless doorbell. Not very good for sustainability I know but I wanted it quick and didn't want to have to make the button too! Cracking it open and using a scope I found the pin that controls the musical bit (the musical bit is a separate PCB soldered onto the wireless receiver board). I just soldered a wire onto that and tested I could pick it up with an Arduino - no problems! Looking back I think I should have cut the trace or removed that board, it might have solved some problems later on.
I started off with an Arduino to prototype the electronics. I used the servo library which easily let me get the motion right for the lever nozzle presser. Then I added the wireless receiver and checked the whole system.
For the final electronics I didn't want the expense of an Arduino, and I had a few tiny2313s about so I decided to give them a go (been avoiding delving into that for ages!) I got them working by modifying the Arduino environment to add the tiny2313 as a target and my Olimex in circuit programmer as a programmer. I used Arduino-tiny, for the core (thanks to the developers) and followed instructions with Arduino-tiny and searched the forum a lot!
A problem I had with the coding was that I used the servo library with the Arduino, but the servo library doesn't work for the 2313. I didn't want to get into library stuff / data sheet so I googled and found some code at avrfreaks. It's a matter of programming the timing registers to get PWM to work at the right frequency to control a servo. Simple in hindsight!
I had it all setup on the bread board but then the signal from the wireless started behaving really strangely. Before it was cleanly triggering the Arduino but now the signal seemed really noisy on the scope and was triggering the servo multiple times. I couldn't work out what the difference was so after sprinkling some capacitors about I added a transistor to deal with the signal (sorry no circuit diagram if you want please ask), and a delay in the software to wait for the noise to end. I also added a 3.3v regulator because I wanted to run everything off an old phone charger but the wireless receiver needed 3v.
Then I transferred the circuit onto a small veroboard and made sure everything still worked. Success!
My mate Lou is a great carpenter so the next step is to put all this in a nice box and have some sort of mount for the pipe.
I used up most of a can of compressed air (I think, judging by the weight of it) in about 40 tests I'd guess. I hope a 200ml can will last at least 100 rings before it runs out. That is definitely an area that could be improved. Would have been nice to build a simple and elegant bellows for it.