Saturday, 28 July 2012

Peroxide Voodoo

Last Night we broke in and tested the two remaining flight engines. I was not entirely happy with their performance but considering that the controller can compensate for differences in thrust they should be good enough for a basic flight test.

Our only goal was to break in the engines so we used lots of quick pulses so its difficult to see the results.

I am continually surprised with the sensivitivity of peroxide catalyst packs to extremely small differences in make-up. Both engine 2 and 3 had an identical pack but the injector piece on engine 2 was about 3mm longer than that of 3 so it had slightly more compresion. The packs are so springy that I wouldn't have thought this would make a difference but engine 2 started producing thrust after we had run about 3L of peroxide through it whereas engine 3 started producing thrust almost immediately. From my experience with the first silver pack I made which had way too much compression I would have thought that the engine 2 should have behaved like engine 3 (less thrust but smoother and more catalysed) but it was the other way around.There were a few other differences we noticed between the 2 engines. Engine 3 seemed to be smoother (from its sound) and produced a clearer exhaust but looking at the results engine 2 appeared to be smoother and produce more thrust.

Engine 2:

We did three runs with engine 3 with 2L, 1L and 1L. The first 2 produced little thrust. This is a plot of the third test. Towards the end of the test I think the propellant was running out and nitrogen was coming out with the peroxide

 Engine 3:

Engine 3 started producing thrust straight away. This test was with 2.5L. You can see the first little bit with duty at 1/8th produced no thrust but after that the engine started producing thrust.

Engine 3 thrust seems to not really correlate well to varying flow rate producing about the same amount of thrust over the entire test. I am hopping that this is just because appearing to work well with smooth with a clear exhaust it was not fully broken in as engine 2 had about 2L more peroxide through it when we did the same test.

I have been thinking for a while about writing a program to vary duty automatically so we can conduct the same test on each engine. Ariel had a good idea to let the duty cycle (flow rate) be a sin wave and see if the thrust response is also a sin wave.

I would also like to experiment with changing the PWM frequency. At the moment I am using 31Hz because frequencies around that number seemed to work for Armadillo and because 31Hz its the only frequency I can make the hardware timers work at by changing their pre-scaller. The next frequency would be 15.5Hz which is too low. The other option is just to use software timers but I don't think I will be able to use them with lab view. I would also prefer to use hardware timers than software ones because its a huge waste of micro controller resources.

Tuesday, 24 July 2012

Test Results and general discussion

Tonight we tested some modifications to the catalyst pack in one of the flight engines. The modifications included a new bottom retaining plate/disc with a higher open percent, top disc with fewer holes and anti channelling rings distributed through the pack every 3 sets of 5 silver and 1 stainless mesh. I hadn't bothered with the rings before as I was only running the packs for short durations and the silver mesh fit tight in the chamber so I didn't think channelling would be an issue. We put stainless mesh on each side of the rings to stop them sticking to the silver. I am not sure exactly sure what the percent open area of the new bottom spreader is as Buren went to town and drilled as manny holes as was possible with some being different sizes. My guess would be at least %50. From our highly scientific  "blow test" (we blow through the assembled engine) the new pack seemed to have a lower pressure drop. Unfortunately I didn't get a lot of photos from tonight as we were rushing to get everything done. We also installed a jet piece with 3 1.5mm holes.

Buren came up with a really good system for pre-assembling the new pack which will save time for the next 2 we have to make. He made a jig which has a hole in it the size of the chamber which the pack is assembeled in and compressed. The jig has a smaller hole in its other end which is used to push the pack out. The pack is transferred directly into the engine and engine is then placed on the press where it is compressed closed and the bolts engine bolts tightened. We found it much easier to tighten the bolts on the press and it also means we don't have to worry about the gasket moving.

The test went quite well but produced some unusual results. We broke the the new pack in with 2L of %85. It was only towards the very end of the run (about after 1.7L) that the engine started producing thrust an there was allot of un-catalyised peroxide that came out. Here is the thrust plot of the break in run with flow rate normalised between 0 and 6 for the purpose of illustration.

Although it looks like it made thrust early on the engine would build up pressure then flooded and spray out un-catalyised peroxide. This was really starting as I would have thought that as the pressure was increasing peroxide flow would decrease but it seemed to be the opposite, as soon as it seemed to get going it would get floded.

The second test went well and we finally got a smooth thrust plot:

 I didn't record chamber pressure as the flight engines don't have pressure ports. I improved the system by which the engine is mounted in the test stand so it is now fully restrained and can't move around producing which would have affected thrust readings. I try to get a photo of it next time I am at the workshop. I had previously thought that this was the some cause of rough thrust as the pressure plots didn't show occultations lining up with the thrust variations but now I am not sure it was entirely responsbale. In the above plot you can see that for the low flow rates the thrust is really rough but for the high ones it is quite smooth. The exception is at the end when the thrust gets rough briefly but I am fairly sure this is because the peroxide ran out and gas was coming out with it. Also the end roughness doesn't look the same as that initally. In this test I used a large "super pro shot" solenoid  (nitrous oxide accessories have silly names to appeal to their target market). This had a much higher flow rate than the small ones we are using on the vehicle which I should have used but we don't have any spare. I have noticed that at low flow rates the water coming out seems to pulsate and the solenoid makes a much louder pulsing noise so I think that the pulsating flow meant that the chamber never reached a steady state pressure. 

When we are breaking in the other 2 engines I will use the flight solenoid and try to figure out at what point the thrust becomes rough. If I can't get a good throttle range above the rough point I will have to add a jet to the solenoid to meter the flow or make the jet on the injector smaller. We probably should have made these smaller originally but Buren wanted to put 3 in as he though it would spread the flow out more evenly.

Sam also came along tonight and helped out which made setting and packing up much easier.  I am going to try to push the group side of things in the future as although it takes longer to get people up to speed which can be frustrating at times the long term gains rewards ave been well worth the investment. Overall  have also found working in a group to be much more satisfying than working alone. Sam has a  2nd year electrical engineering friend who might be interested in helping who would extremely useful with the electrical side of things. Now if I could only find a control systems engineer! I have always dreaded being a leader because I thought it would somehow give me less time to work on in-depth things myself but I have been pleasantly surprised to find that it it much more challenging and rewarding than any technical problem.

 Sam also finished the remote quick disconnect which despite looking a little dogie with the zip ties playing a structural role works every time. We decided to use a solenoid instead of a pneumatic ram. I didn't get a photo of it but will do so.

I also finally claimed victory in the troublesome battle for the pressure-gauge-in-bottle-jack! In the end I just put the jack in the lathe and settled a hole in the bottom on the jack. I did want one in the side so it could still rest on its base. This meant the jack had to be raised in the press to give the fitting room. The springs made this difficult as they are quite stiff and the press resisted  but eventually we got some spacers in. The press then fought back by bending the main plate which transfers force to the press-rod because the force was now off to each side and not directly on the rod. This caused the catalyst punch to fly out narrowly avoiding my head. Buren and I responded by reinforcing the plate.  In the end the victory was short lived as the high pressure gauge I used (reads between 0-10000 PSI) only really starts moving after the pressure is about about 500PSI which is useless for catalyst pack compression.

The rest of the photos are self-explanatory but this one is me trying to get the now longer jack back in the press.

Sunday, 22 July 2012

Slow progress.

I spent the last few days machining the remaining 2 engines, mounting them and bending stainless tube. The engine mount design is not optimal and looks a bit clumsy but works.

The stainless tube is turning out to be a huge pain to work with. I was tempted to ditch it and use teflon stainless braided hose instead but I don't think the ends would have fit in the small space bellow the bottom of the solenoid. It took me about 10 attempts to make one stainless line. Now that I have one right it is much easier to make more. 

One issue I have run into is how to assemble the motor and plumbing considering that nothing can rotate freely once tightened. I am hoping that once I tighten the compression fittings I will be able to take the fitting apart to work on the engine.

We also need to test the final configuration of the thruster. Previous versions didn't have anti-chaneling rings as the catalyst fit tight and we went running them for long periods of time so it will be interesting to see if they make any difference initially. I am also adding a metering jet which should hopefully smooth out some if the pressure oscillations in the chamber. I was thinking of testing today but  Buren couldn't make it. 

I decided to include a pressure transducer on the vehicle as it doesn't add much weight and knowing the tank pressure will be useful when pressurising remotely and also when flying as you can get a fairly good estimate of how much propellant is left from knowing the initial pressure, volume and current pressure. 

Today was the day I said I wanted to have a strap down test done by. I am a bit disappointed we haven't got that far but looking back it was not a particularly realistic goal.

Tuesday, 17 July 2012

Hovering Vehicle Progress

Burren and I have been working on rocket things full full time since saturday. There is too much to talk about in detail but here are a few progress photos.

I have made a new electronics box as the old one was too big and heavy for the current design. 

Burren has been machining engines.  They don't look quite as impressive with the small (optimal) nozzle.

I decided to move all the electronics to the top as the bottom was getting a little cramped with the engine spacers. My reason for having it on the bottom initially was to not have to run wires to the engine solenoids but an extra 20g of wire is worth the convince. Also the electronics will be much easier to work on when there are on top.

We decided to make a custom 4 way splitter as the 4 way cross was heavy and would have meant that one of the pipes came out the bottom. This would have caused an imbalance from the forces due to the liquid flowing through the pipes.

We also did another thruster thruster test on Saturday which was the best one yet. As it turns out the variable thrust produced by the engine is actually due to the motor rattling around the tube as it wasn't restrained properly. Max thrust was about 5.2Kg which I am happy with. I will upload the video/graph when I get time.

We also have another volunteer. Sam is 2nd year Mech and Aero student at the University of Queensland. Sam is currently working on the quick disconnect system which will closely resemble Paul Breeds quick disconnect

Friday, 13 July 2012

Swagelok - The Apple of fittings.

On thursday night Burren and I did some milling of the plates and practised bending the 1/4" stainless steel tube.

In the past I have avoided swagelok because their fittings are twice the price. Instead I would run around to lots of different stores and when I couldn't get things locally wait weeks for parts to come from overseas. I had heard good things about swagelok compression fittings (from a ex sales rep so i suppose that doesn't count) and they seemed moderately priced so I decided to give them a try. I have to say I was impressed. As it turn out, for the premium you get higher quality parts and good service (who would have thought...). They even got delivered the same day. Amazing!

I still haven't decided exactly how I am going to fly the vehicle. A RC remote would probably be more reliable than joystick through computer so that is what I am leaning towards at the moment. This is how most UAV's work so there is lots of code around. Another issue I haven't resolved is a robust failsafe so in the event of power failure we are not stuck with a pressurised vehicle that we can't de-presurise. All I can think of is a normally open solenoid but there must be a better (lighter) way than that.

Monday, 9 July 2012

Not a whole lot to report since the last test. Work work has been time consuming and I haven't had much time to work at the work shop. work.

Last weekend I tried adding a tapping to the bottle jack so I could install a pressure gauge. When drilling the hole I drilled too far and ruined the jack. I bought a replacement jack today but destroyed it in the process of opening it. I think I will just cut my losses and get a hydraulic ram to use with my hydraulic pump (the one i use for pressure testing) as it has a gauge in it.

Today Buren weighed all the components for the vehicle. Excluding wiring, propellant, quick disconnect and relief valve it is looking like the final vehicle will be about 5.9Kg. I am still not sure exactly sure how much I can push the 20mm cat pack thruster but assuming a modest 4Kg thrust the vehicle should be able to start with 5Kg of propellant which should give a decent run time. I have given a significant amount of worry that 3 thrusters wouldn't be enough to lift the vehicle so it was good to get some facts.

We now have most of the bits we will need to complete the vehicle. We have made a few small deign changes around the placement of the engines. It was going to be difficult to have them mounted to the base plate as it would mean lots of bending in the stainless tube. They are now mounted on extensions which are attached to the base plate. I was also worried about the plate heating up from contact with the hot engine so the new arrangement also fixes that.

The plan is for Burren to get most of the structure finished by the end of this week. On the weekend we are going to do another thruster test to make sure we can get smooth thrust then next week we can machine and assemble the 3 engines and install electronics. There are about 2 days of machining in the 3 engines which I will do. After thats done I will move on to programming and Burren will be installing and wiring electronics.

I doubt the vehicle will be at the point where I will feel comfortable doing a hanging test at the end of next week but I don't think a strapdown test is too much of a reach.