Monday, 20 August 2012

First teathered flight test

Yesterday we flew the hovering rocket for the first time. I am still working on the stabilisation program so it was controlled just by pulsing all the thrusters at maximum which resulted in a crazy looking flight.

We actually did 3 separate tests. The first one was a practise run with water to make sure everything was working. Then we did a test with peroxide but after the first hop (if you can call it that) the battery came loose so we stopped the test. We then replaced the hose and did a second test. It was quite dark by this stage so it is a bit difficult to see what it happening in the video. Here is short video showing some of the test:

My objectives for the test were:

  1. Test the ground, flight command and data logging systems.
  2. Demonstrate that the vehicle could actually liftoff.
  3. Record orientation data to see how the ArduIMU coped with the flight
I had a fairly hectic saturday trying to get everything finished in time. I decided to redesign the disconnect system as it doesn't quite work as well on the vehicle as it did on the bench tests. The reason for this is that if there is any torsion on the disconnect it takes much more force to actuate and the solenoid isn't able to provide enough force.

Instead of the quick disconnect I decided to go with a more reliable system which involves cutting the pressurisation line. To cut the line I used a set of electric shears and modified them to be actuated remotely. I was really surprised by how complicated the control electronics of the shears were....

Test of the system:

For this feathered test we cut the line on the ground. The idea was for the coiled air hose to recoil so there wasn't much hanging off the rocket. Unfortunately the hose didn't recoil much so there was hose flailing around on the side of the rocket during the test. For the next test  plan on cutting the hose just above the solenoid with the cutter on a trolley leaving minimal hose remaining on the rocket.

For this test we only pressurised the vehicle to 450PSI as the cheap hardware air hose we used burst at around 500PSI. Unfortunately  I was at work and Buren didn't get to the fitting store before it closed to get thicker nylon hosing.

There were quite a few issues with the command system. A few times when I would command a pulse of all engines one would stay on after it should be off for a few seconds which was quite concerning. Also the vehicle didn't respond to a few of the pulse commands which I think is because it is trying to send too much data (every time it gets a new IMU update) so misses the ocational command. I think that this might be the cause of one engine staying on too long as it gets stuck sending data. I either need to get a separate radio for commands than for data or log the sensor data in HD onboard and only send data back occasionally. On a side note I was rushing on saturday truing to get everything ready and I fried the Arduino by reversing the voltage in. Luckily I could get another one locally but they didn't sell the exact same one, instead I got one based on a Atmega2560 which also has a micro-sd port and a lan port. I didn't have time to learn how to use the SD port but should be using it next test.

All the ground systems worked really well. I edited this tool which displays the orientation of the rocket   (currently a picture of a plane) in 3d and also pressure in real time and allows me to send text commands I type in (I didn't have time to put in buttons....). The pressure transducer didn't work but I didn't have time to figure out why. It also logs the data and displays them in the program. This was the first time I have used the processing language and maybe its the mechanical engineer in me but I love it! I haven't looked at the orientation data propelly (need to make a program to re-play it from the log) but from looking at the screen during the test and the text log it seems quite smooth and doesn't jump around much. The vent solenoid seems to interferer with the yaw (as it uses the compass sensor) but I don't need yaw. I need to have a better look at the arduIMU firmware. I increased the arduIMU's output rate from about 5hx to 50hz but I think the way I did it might have messed up something else as even when I am holding it still the roll and pitch seems to be rougher than it was before.

We were all a bit disorganised at the test and there are a bunch of things to improve for next time. I wasn't particularly happy there were things we hadn't thought of in advance. Sam started to do an operations manual for the test but it wasn't finished to the point of being useful and we didn't all have proper roles. 

I am happy with our filling and pressurisation system, its the safest system yet. Briefly:  The vehicle is filled using a funnel in the clean space of the workshop and then is carefully carried outside without the pressure relief plug installed so it is open to the atmosphere. We then set up the disconnect, cameras etc. and everyone but me goes away. I then turn on the computer, test it and arm the failsafe vent solenoid. On a side note I put a really bright light on the vent solenoid so I could see when it was on so there is less chance of leaving it on and burning out. Lastly I install the pressure relief plug so then vehicle is sealed. 

So there are a heap of things to improve before the next test. I cant see us being ready for an active stabilisation test this weekend but we should be ready the week after.

1 comment:

  1. Nice job guys! Keep at it, look forward to seeing a stable hover!