2012-08-04

I was able to finish up machining the new v4 injector (drawings, photos) and matching v6 igniter (drawings, photos) this weekend. The v6 igniter threads into the injector body and seals with a copper washer. The extended threads on the igniter are to allow a nut to hold the lox cap in place. The inner LOX seal is a spring loaded PTFE shaft seal and the outer one will be either a Creavey Astra Seal ($82/ea) or an OmniSeal face seal ($125/ea). I may try a PTFE o-ring to see if I can get away with it. The only differences between the v6 igniter and the v4 tested previously are slight changes to the threading surface and dimensions to get it to fit the new v4 injector.

I ran into two issues while integrating the igniter and injector. First, after I installed the Helicoil thread insert into the injector body, when I broke off the tang it damaged the aluminum seat surface so I had to put the injector back on the lathe and take another 0.020 off the seat surface. Next time I'll need to use a tangless insert or put something down in the hole to protect the seat. But then, I realized the copper washers wouldn't fit with the thread insert in place (just about 0.015 too large - oops!). I got a quote for some custom stamped copper washers from Boker's (they can make pretty much any size) but it came out to essentially a $170 minimum order. Since I already had a whole bunch of the existing inexpensive washers on hand from McMaster, I made a small fixture that allowed me to turn down the washer OD on the lathe.

Next up are cold water tests to measure the injector orifice discharge coefficients and assess the mixing along with the spray pattern. For the test stand, I'm changing the main LOX and fuel valve actuators over to a better setup using "real" ISO 5211 rotary actuators. After that, I plan to get a dewar of LN2 to check out all the new hardware. One thing I've never had a good handle on is the amount of pre-chill (if any) needed before engine start. I'd like to not have to do any so one of the tests will be to see how long it takes for liquid to come out the injector with the plumbing at ambient temperature.

2012-06-08

I finally got a good set of tests on with the v4 igniter! I lowered the chamber pressure down to about 150 psi and made some changes to the GOX jet to slow it down. Now it will run for 2 seconds with a nice plume without eroding the inside wall. There were a few issues that had to be addressed:

• The GOX restricting orifice was right at the igniter so the GOX entered at a high velocity and contributed to scrubbing of the opposite wall. This prevented running at higher mixture ratios.
• Due to a spreadsheet error (wrong GOX temperature), all the previous igniter runs were at a much lower mixture ratio than expected, contributing to a weak exhaust plume.

Although having a high GOX velocity should help mixing the gas with the liquid fuel, moving the GOX orifice just a bit upstream (using a blank #10-32 fitting) will slow down the jet quite a bit. After that change, I could not detect any new hot wall erosion in the body of the igniter. The design spreadsheet assumed a GOX temperature of 288 degR (not sure how I came up with that...), much colder than ambient since the GOX would be generated by a heat exchanger from the main LOX supply. Since I was using GOX from a gas cylinder at ambient temperature, the change in density is significant and led me to run all the igniter tests a lower mixture ratio. Of course, I didn't realize this until after all the runs but I was able to get several runs in with a good plume. For example, when I thought I was running at a mixture ratio of 2.2 based on supply pressures, it was really closer to 1.5. So that completely explains the poor flame performance on all the previous runs - the mixture ratio was just too rich. The good news is that the exhaust plume looked great at 1.2-1.3 so I don't necessarily need to run it leaner with the associated higher temperatures. I didn't measure the GOX temperature so I don't exactly know the injection temperature but assuming ambient is sufficient to prove the design.

For the most part, other than the issues above, this latest round of runs was uneventful. I did try a different design (v5 igniter) in an attempt to angle the GOX and fuel jets at a 45 degree angle to reduce the hot wall impingement but that igniter wouldn't start - it only generated 20 psi and I don't think it ever actually started combustion. On some of the runs (5/19 Runs 5 & 6), there was a popping noise that sounded like combustion instability. Because all of these runs were not at the design condition (due to the wrong GOX temp used in the spreadsheet), I think the popping sound was the hot jet flipping back and forth between two different states. The video shows this too as you can see it change between a small plume and the larger plume as seen on the later runs.

Another interesting problem was that the same conditions on different days didn't necessarily produce the same visual results. For example, Run 7 on 5/19 used the same conditions (regulator settings and pressure readings) as Run 1 on 5/20 yet the visible plume on 5/19 is brighter. There may have been some temperature effects which would affect the GOX flow rate more than the fuel rate. Another example on 5/20 is that Runs 6 and 9 were essentially the same conditions but the plume on Run 6 is barely visible while Run 9 is much brighter. Run 9 used the "slower" GOX jet so maybe there was better mixing which would lead to better combustion. Also, the igniter chamber pressure on Run 9 was about 10 psi higher but alone wouldn't explain the dramatic change in the plume.

I'm currently working on integrating this new igniter design into the injector for the 250 lbf motor so I can test it again. I don't have any welding capability and spring loaded cryo seals are hard to get in small quantities so I may just try o-rings for the face seals. It's a metal-metal seal and some leakage can be tolerated.

2012-02-11

I tested two new igniters last weekend, a short and long version using the NGK ME-8 spark plug. The short version is designed to thread into the side of the main 250 lbf motor and the long version is designed to thread into the top of the injector. The short version was generally successful and I think I'm narrowing down the range of conditions where it put out a big enough plume without melting the igniter body. It seems that for this stainless steel design, the "short" version running at a mixture ratio of around 1.4 seems to be the optimum. With lower mixture ratios, the plume is barely visible and at higher mixture ratios, the chamber starts to melt.

Everything went pretty smooth the first day of testing (2012-02-04) up until the last run. Runs 1-2 were a repeat of last month's runs at a mixture ratio of 1.0 and didn't show a very bright plume (as expected). Runs 3 and 4 were at a mixture ratio of 1.4 which seems to be the sweet spot for this design. The chamber pressure was only about 240 psi vs. a design of 285 (psig) but that may be expected for a non-impinging gas/liquid injector. Run 5 was supposed to be a repeat of Run 2 but at the end of Run 5, there was a small shower of sparks indicating some erosion. Upon disassembly, I found some melting of the small hole at the bottom of the pressure tap fitting but otherwise looked pretty good. For Run 6, I installed the long version of the igniter and I think there must have been some FOD in the chamber because the run was very unstable. When I disassembled the fittings afterwards, I found some debris in the #10-32 threaded hole for the fuel fitting. It's unlikely that it came from upstream because there is a 40 micron filter right in front of the 0.010 inch diameter fuel orifice. My theory is something clogged the 0.050 inch hole at the bottom of the threaded fuel hole (insufficient cleaning of the fitting?) which caused the chamber temperature to spike momentarily, melting some of it. There was some o-ring erosion on some of the runs so maybe a piece of the o-ring was floating around in the threaded hole.

The second day (2012-02-05) was intended to be a repeat of Runs 4 and 5 at a mixture ratio of 1.4 with the short version of the igniter. I didn't originally machine the entire profile on the body of the igniter to save time so for these runs, I turned down the diameter of the injector to the OD of the thread to get it closer to the final profile. The idea was to see if less metal around the nozzle would make it more likely to melt. The chamber pressures and mixture ratio were pretty close but the plume wasn't as bright as it was the first day. The only thing I can think of is maybe there was some nozzle erosion (I didn't measure it in between runs). For Run 3, I had intended to run it at a mixture ratio of 1.8 but I must have set the pressures wrong and I got 2.2 instead which was way too hot for this uncooled igniter so it melted. I also noticed afterwards that the Beswick mesh filter in the GOX line was dislodged slightly (still partially attached) but I don't think was related to the run conditions. During Run 4, the igniter had a strangely shaped plume and it sounded like an air horn on an 18-wheeler. I think it must have had a "no-start" and the only combustion was external to the nozzle. On my 100 lbf regen motor, I had a similar occurrence during Run 2 on 20070225. The ME-8 spark plug I was using was in pretty bad shape at this point so that was probably the cause. For Run 5, I changed the spark plug over to a cheap overseas knockoff of the ME-8 and finally got a good run out of the long chamber version of the igniter. Well, sort of. The mixture ratio was 1.38 and the chamber pressure was about 230 psi, both in the reasonable range, but at the end of the run, the nozzle melted. The combustion chamber of the igniter was 0.75 inches longer than the short version so maybe the additional surface area helped to heat up the whole body too much. I didn't perform a thermal analysis of the igniter so I'm not really sure how long it was supposed to last.

The good news is that at mixture ratios of around 1.4, the NGK ME-8 doesn't show any erosion of the center terminal which I ran into last month. However, after several runs, enough carbon built up around the center conductor and the spark tended to jump the gap farther down in the plug. I went ahead and fabricated an igniter body for the NGK R847 but when I got ready to test it, the RCEXL ignition didn't have enough juice to drive it. I could hear a spark down inside the plug but I couldn't see anything at the tip. I didn't want to cut off the metal shell from the end of the cable for the ME-8 so I just ran a HV wire up into the metal shell. However, the ME-8 is a non-resistor plug so the pre-packaged shell for it from RCEXL has a 1k resistor built in. The R847 has a resistor type plug with an additional 5k of resistance so even if I had removed the 1k resistor from the existing setup, it still may not have been enough. I could go with a standard automotive style coil on plug (COP) unit but the RCEXL is so compact and has a small attachment that generates pulses for it. Given that the ME-8 seems to survive a mixture ratio of 1.4, I'll stick with that for now.

I need to conduct one more set of igniter tests to very repeatable and reliable operation before I'm ready to test the 250 lbf motor again. I'm leaning toward attaching the igniter directly into the side of the main motor because that is an easier installation than redesigning the injector to accept the igniter in the center. I'm somewhat worried about local heating around the 0.190 inch diameter hole however.

2012-01-21

I tested the new Augmented Spark Igniter (ASI) last weekend (see archive notes from 2011-09-29 for details). Four runs were performed, the last of which melted parts of the igniter so I'll have to make a new one. But that's not a big deal because it was fairly easy to fabricate and I want to change some aspects of the design anyway. Runs 1 and 2 were performed at the design mixture ratio (design r=1.0, actual r=1.03-1.05) but I didn't see a visible plume so I initially assumed that it wasn't working. I measured a reasonable chamber pressure but didn't see anything interesting in the video. Close examination of the still pictures for Run 2 show a slight darkening of the image in the exhaust so it appears that the igniter was working. However, the combustion reactions were probably complete before the exit due to the long nozzle so nothing was visible. After the run, as the pressures were dropping, the mixture leaned out enough where I saw a brief flash of flame. For Run 3, I adjusted the pressures to get a leaner mixture (r=1.60-1.75) which resulted in a more "normal" looking plume but was much hotter than I had designed it for. For Run 4, I leaned it out even more (r=3.54 oops!) and got a really impressive plume but it was so hot that it melted parts of the stainless steel chamber. Interestingly, all the of the damage was confined to the upstream portion of the chamber - the throat was unaffected. I think I also had a leak at the pressure port because you can see the tube glowing red hot almost immediately. For the next version, I'm going to move the throat down closer to the end of the igniter so the the body doesn't draw out all the heat from the gas flow.

It looks like the NGK ME-8 spark plug may not be the best choice for this application. The center electrode is very small and indeed, after Run 3, I noticed it was mostly gone so I changed it out prior to Run 4. The new one didn't survive Run 4 either. Of course, both Run 3 and Run 4 were much hotter than I intended so the plug might have been fine for a mixture ratio of 1.0. For the next version, I'm considering using an NGK R847 or R0045J NGK surface gap plug, both of which have larger center electrodes. They are also larger in diameter (8 and 10 mm) and have a much longer reach (19 mm) so the igniter body will have to be longer to hold the threads. The RCEXL ignition worked great so I'll keep using it.