™
Go UP young man!
On Monday we did another test with a new design for the hybrid grain. We finally have a design that held together for the entire burn. The component with the most wear was the phenolic insulator disk up by the injector. It didn't fail after 8 seconds of firing, but it looks like it will be the component that limits our burn time. We have some ideas for improving it with a thicker disk and more o-rings. At the nozzle end we have a graphite insert that showed no wear at all.
There was one anomaly in the test. I initially shut down the engine after 2 seconds. We wanted to do a short firing to make sure we got the chamber back in one piece to inspect it after having several tests where it came apart and we had to piece together the failure mode. But when I shut it down you may notice in the video that there was still a flame coming out of the nozzle. The ball valve was leaking through into the chamber. This valve has been through a lot in previous tests, and it will be replaced. At that point I opened the valve back up again to burn off the last of the peroxide.
The second burn was longer than I expected, 6 seconds. It looks like we were only flowing ~50ml of peroxide per second, not 100 that the injector was designed for. I think our throat is too small and it is the bottleneck, not the injector. If there wasn't enough injector pressure drop that explains why there was still some roughness in the combustion. Hopefully if we open up the throat it will give us more pressure drop in the injector and that will go away.
We now have a load cell physically mounted to the test stand, but we didn't have enough time to get a data acquisition system up and running for this test. So we still don't have any measured thrust or Isp. We did find out the fuel regression rate is about a fifth of a millimeter per second. The rule of thumb I had read for peroxide/polyethylene was half a millimeter per second so we're in the ballpark. In this configuration it gives us an O/F ratio of 25, about four times optimal. We knew this engine wasn't going to be optimized. Its main purpose was to get the data so we could design a more optimal engine in the next iteration.