2006-10-03

I'm starting to see a pattern here... do a rocket test, get a new job and move. That sure interrupts the rocket progress for a while. Anyway, I've been able to spend some time analyzing the data from the last runs so not all the time has been wasted. I'm planning on running the regen engine again in another month or two assuming the work schedule cooperates.

Here is a summary of the recent tests with the regen engine:

• Design thrust = 100 lbf, actual = 68 lbf
• Design chamber pressure = 200 psia, actual = 142 psia
• Design Cf = 1.21, actual = 1.17
• Design Isp = 217, actual = 147
• Design c* = 5464 ft/s, actual = 4057 ft/s

Note that the design Cf, c*, and Isp above were recalculated using the actual chamber pressure and mixture ratio. In general, Cf is a indicator of nozzle efficiency and c* is an indicator of combustor performance. The nozzle efficiency only looks low by about 3% so I'm suspecting poor mixing and combustion. With this new injector, I kept the spray pattern in the center to keep it away from the walls so maybe I need to spread out the injector orifices and redistribute them more across the face. However, the holes are already pretty small as it is so it will be hard to use more holes for a 100 lbf engine.

For the next set of runs, I'm going to make the following changes:

• Use helium instead of nitrogen for pressurization. Some people have reported significant performance gains when switching from nitrogen to helium because LOX tends to absorb the nitrogen, making it "fluffy". I'm really hoping this is a large part of the problem.
• Use a flowmeter to measure the fuel flow. I had such a huge pressure drop in the cooling passages (230 psi) that just the variation in temperature could have had a significant effect on the mixture ratio. I'd like to also measure the LOX flow but a cryogenic flowmeter is a bit out of my price range right now...

I'm also in the middle of doing some upgrades to the test stand. I'm replacing the homemade air actuated fuel valves with some "real" actuated ball valves because of some stickiness that has been bothering me for a while.

2006-02-28

Successfully fired my first regeneratively cooled engine last weekend! All the pictures and videos are in the Regen Chamber Tests section.

The thrust was a little low (70 lbf instead of 100 lbf) but it ran twice for about 30 seconds each without melting. I must have overpredicted the gas side heat transfer coefficient because the measured wall and fuel temperatures were well below my design point. I'm a bit confused on the thrust because I thought I did a better job this time predicting the required tank pressure to give the proper propellant flow rates. By the way, I remeasured the injector discharge coefficient as 0.84 for the LOX holes and the combined pressure drop across the fuel cooling passages as a whopping 208 psi. Since I have plenty of margin for cooling, I should be able to cut that down next time and get a more reasonable pressure drop.

For this test, I used Jet-A where last time I used K1 kerosene but I was under the impression that they were essentially equivalent. I went back and looked at my equations for the uncooled engine and I came up with a value for the density of kerosene of 36.21 lbm/ft^3. I have no idea where I got that value - as a result, for the uncooled engine, I pumped 8.06 in^3/sec of fuel. This time, I had good data for Jet-A at different temperatures and the equations resulted in a target of 6.53 in^3/sec. If you recall, on the uncooled engine, I had pressure droop in the regulators which caused a proportional drop in the chamber pressure and thrust. This time however, I figured out the droop in advance and corrected for it. So, the only thing I can think of is that last time my design was far enough off that it worked out but this time, I hit my flow rate targets. That tells me that there might not be enough volume in the chamber for complete combustion. For the uncooled engine, the cylindrical chamber volume was 15.6 in^3 and for this regen engine, I reduced the chamber volume to the minimum (12.7 in^3) to try and help out with the cooling problem. Next time, I'm installing a flow meter in the fuel line to so I can measure the exact flow rate.

I ran 3 total runs, one false start (Run 0) and two actual firings (Run 1 and Run 2). On the false start, I had too much of a fuel lead and it burned itself out after ejecting the ignitor from the chamber. It then proceeded to fill up with fuel and LOX - it's a miracle I didn't have a really bad hard start. I switched the valve sequencing around to lead with the LOX first and the next two runs started up nicely. That is the one disadvantage of an external ignitor - it is simple but once it gets ejected, you better hope the fire inside is going.

The first good run (Run 1) was very stable. It actually only ran for about 25 seconds because it ran out of LOX 5 seconds early. I guess I spent too much time after filling the tank before I fired it and too much boiled off. The second good run (Run 2) was not nearly as stable. It ran for 30 seconds but had several steps in thrust during the run. I didn't tear down the engine between the runs (I only cleaned the LOX lines) so I suspect some of the sooty junk migrated up into the fuel injector and clogged one of the holes. During the teardown, I found black soot and small particles in the fuel injector face cooling passages. About halfway through Run 2, the flame changes shape and the thrust picks up by about 1/6th of the total thrust (there are 6 fuel injector holes).

Even after almost 60 seconds of hot firing, the engine shows no visible signs of wear, even around the injector orifices or chamber pressure tap. In fact, you can still see the tool marks on the walls. Assuming I can clean out the cooling passages, I should be able to run it again and try to figure out why the thrust was so low.

The video cameras I used were very sensitive so I purchased a 2-stop neutral density filter to try and cut down on the intensity but the CCD was still saturated. I wasn't too pleased with the video quality but the still images came out good. I was caught off guard by a problem I ran into with the IRIG video overlay devices. When run off of a generator, they had trouble locking on to the video signal and the IRIG time code just rolled vertically. Since I discovered this the night before the test, I hooked up a 300 W car power inverter to supply 120 V to the time code power brick. Then, I connected the power inverter to a 12 V ham power supply which was connected to the generator. I guess that crazy combination provided enough filtering to fix the problem. Next time, I'm getting one of those fancy UPS units that have power conditioning built-in.

2006-01-16

Performed cold water tests on the injector this weekend. The measured discharge coefficient was Cd=0.78 but I'm going to repeat the tests with a better pressure measurement. I took measurements at tank pressures of 10, 20, 30, and 40 psi but at these low pressures, the cracking pressure of the check valve causes the tank pressure to vary by about 5 psi during the run. I think it will work better if I just hook up a garden hose and a pressure gauge/transducer and bypass the test stand hardware. I suspect the actual Cd is about 0.8 to 0.82 but I have to prove it.

The flow from the orifices looks pretty clean but I was puzzled by some strange behavior from the central LOX holes. All 6 holes seemed to behave like they had a misalignment of about 15 degrees in the same direction and completely missed the fuel jets. I initially thought maybe there was a vortex forming in the small plenum just above the LOX holes but you'd think the holes would "straighten" that out. A visual inspection showed the holes were much straighter than the flow coming out of them. Just for kicks, I took the same drill bit I used for the holes and ran it in and out by hand a few times and presto! The holes flowed straight. I'm still puzzled by that one. I guess I didn't clean up the holes properly after drilling them.

Calculating the pressure drop in the injector coolant passages took a little bit of thought. Since I couldn't separate out the drop between the coolant passages and the fuel orifices, I assumed the Cd was the same for the LOX and fuel holes. Since they are about the same size holes and drilled from the same setup, this should be close. I then calculated the estimated pressure drop across the orifices and from that, I was able to back out the pressure drop in the injector coolant passages (~22 psi vs. my design prediction of 20.6 psi). Again, these numbers may change when I re-run my Cd measurements in a few days.

I also went ahead and cut the Grafoil gaskets that I'll use to seal between the injector and chamber, between the chamber and end plate, and to close out the holes for the cross-drilled coolant passages. The Grafoil is really cool to work with - I used a crafter's circle cutter to cut the outer/inner diameters and a regular hole punch to cut the bolt holes.

Part of the time this past month was spent working on the electronics portion of the test stand. I replaced the little switching power supplies with some linears and put all the remote equipment (PXI chassis, SCXI chassis, relay box, and power supplies) in a portable rack mount chassis to make it easier to move everything around. I still need to make some cables to connect the PCI IRIG card with the PXI chassis so I can synchronize the IRIG time video overlay with the DAS data. Last time I used a CompactPCI IRIG card in the remote chassis but I don't have access to that card anymore.