I hate solid igniters.

They do not work with nitrous.

Maybe if you're at FAR and your nitrous is at like 90 degrees it will, but for us in the East Coast we've found that using a solid motor to light the main engine is a lot less reliable. This led us to design an Augmented Spark Igniter (ASI), which uses two gaseous props and a spark plug to generate a very hot flame that reliably ignites the main mixture. Although our lander is Nitrous/IPA, we chose gaseous methane and oxygen for our ASI propellants to maximize ignition reliability for our ground testing campaign. 

Injector

(NOTE): The fittings that interface with the injector are NOT the fittings we used. These fittings were COTS fittings from McMaster that were initially had YorLok nuts on them, but were unscrewed. THESE FITTINGS WILL SHEAR. They were 10-32 threads, we scaled up to a custom fitting with a 1/4-32 male thread and we haven't experienced shearing when compressing the copper crush.

The ASI contains a micro single element coaxial shear injector. The injector is ox centered, meaning that oxygen comes down the central tube, and fuel comes through the annulus. With an ASI, we don't really care about efficiency, just flame temp and chamber pressure, so it doesn't really matter if the injector isn't that efficient. We determined that a coaxial shear injector is a good trade between something overly complex (like a coaxial swirl) and something overly simple (two impinging inlets at 90 degrees). 

Each inlet consists of a custom-machined fitting and a COTS fitting that adapts to a braided hose. The custom fittings interface with the chamber via a thread and copper crush. Copper crushes were chosen because they allowed us to make the threads straight, and thus quite small. We could've done this with an ORB porting tool, but we didn't want to spend 300 dollars for one lol.

The tube is a micro piece of SS hard line (0.05" OD) that you can get from McMaster, and is brazed to the inside of the custom fitting. This joint actually wasn't that hard to braze – you can put flux and filler metal into the larger hole on the inside of the died 1/8 NPT side. Then, you can slide the tube inside and heat up the fitting from the outside with a propane torch. IMPORTANT: In the photo, the needle tube is only barely sticking out above the surface of the fitting on the inside; this is supposed to be mOsT oPtiMaL but in reality it needs to stick way above that to allow the flux to sit below and not clog the inside of the needle. Once the fitting gets to a rose red, the joint has probably heated up enough on the inside to melt the flux and seal it. Example below!

Chamber

add cross sections

The chamber is a 316 SS heat sink, which will survive for a second or two before melting. It contains a pressure transducer port, which senses ASI combustion pressure (and later main chamber pressure), and a spark plug port, which is located further down the chamber to allow more thorough mixing of props before ignition. The L* of the chamber is pretty low, but again, it doesn't really matter since we are not going for efficiency here. If we wanted a 40" L*, I think the ASI would be over six inches long lol.

Simulations of heat transfer were done on a spreadsheet (link it here). Since the material is 316 SS and quite thick, you cannot assume that the heat sink is a lumped mass. Instead, you need to divide the heat sink into a bunch of smaller nodes, and assume lumped capacitance for each of the small nodes. You can then write out a balance of heat flux between convection from the chamber to the outer node and conduction from the outer node from the inner node, and conduction through the inner nodes, and relate that to the heat stored by the heat sink. You can then discretize these formulas and iterate across time.

One thing that I still need to implement in the spreadsheet is radiative heat transfer, which tends to become the predominant heat transfer effect if the dT is above 2000K. However, as the heat sink heats up, radiative heat transfer will become less of a problem quickly, and the hot fires so far have not melted the ASI.

The spark plug is recessed from the inside of the chamber to prevent the tip from melting, although we have been having ignition problems lately. However, we are 95% sure that our ignition problems are due to the spark plug coil not being powerful enough (when we swapped the coil for a more powerful one, the ASI lit immediately). The PT and spark plug also have custom fittings that connect to the chamber with copper crushes. Here, the copper crushes were predominantly chosen due to their reliability at higher temperatures. Thermal cycling the copper crushes many times also has the potential of being a problem, but the outside of the heat sink does not get hot enough for them to thermally expand/contract much. However, we may consider swapping them out after X amount of firings in the future.

The PT is also separated a good amount from the chamber by around 6 inches of hard line; there are some graphs online that have correlations between flame temp and how far away your sensors should be. The air inside of the PT hard line will act as an insulator that will protect the PT from the hot flame in the chamber. A lot of designs on the rocket side don't recess flame-probing sensors like this and are fine, but since our reliability and test count requirements are much more stringent we think this is necessary.  

Every fitting connecting to the ASI is a copper crush except the fitting that connects it to the main injector, which is an ORB fitting. We chose an ORB fitting here mostly due to the fact that it'd make installation and clocking way easier, and also because we already had an ORB porting tool of the right size lol. However, we recently de-integrated the ASI and observed some O-ring melting. Currently our potential solutions are buying FFKM O-rings which are rated to 600 F (previous high temp silicone O-rings are rated to 450 F) or just burning for less time. However, we are concerned that the FFKM can decompose into toxic chemicals (HF) if it still melts.

Nozzle

add cross sections

The "nozzle" is not really a traditional rocket nozzle, as it does not have a diverging section. Its converging angle is also not optimized because we wanted it to be easy to manufacture (just drill with two different sizes to get a converging geometry). There isn't really a reason why the igniter flame needs supersonic – if they come in sonic that's fine. So, we didn't see a need to drill on the other side for a diverging section.

The flame travels through the nozzle and goes through the main injector before coming out into the main chamber. The whole ASI assembly is at an angle so that it can shoot out the flame closer to where the nitrous exits the pintle orifices to maximize ignition probability. The nitrous itself should regen cool the pintle enough that it doesn't melt due to the ASI flame temp, but we do want to test whether or not the ASI plume would melt the pintle if the valves failed to open for whatever reason. This could be a major problem, as if we have a failed ignition during a hot fire attempt where the ASI fires but the throttle valves don't open, we may not know whether the pintle melted or not. We have an extra pintle tip that we'll try to test this on ASAP.

Flow Rate 

Since this is an igniter, the flow rates of GOX and CH4 are very small, which presented some issues with sizing the annulus. To avoid tolerancing issues here, I opted to decouple flow rate from the ASI by having upstream orifice fittings with an area less than the annulus and GOX needle. McMaster orifices have flow data for water, which you can use to back out a discharge coefficient (Cd). Then, you can plug this Cd into the equation for choked flow through an orifice to obtain the mdot that this orifice sets. 

In reality, you are uncertain about the pressure right before the orifice fitting, so if you actually want to characterize your pressures well, you should have PTs reading the upstream orifice fitting pressures and the chamber pressure, and then use the pressures that you see in the data to iterate towards your desired mdot. 

P&ID

This image shows the P&ID for the ASI, which is pretty standard. Each propellant line has a regulator and a solenoid valve to precisely time propellant delivery to the ASI. At the test site, FK-3 and OK-3 are ~10 ft away from the test article and are behind separate concrete walls to isolate the system sufficiently. However, NK-3 is next to the test stand.

For a more in-depth ASI testing campaign, there would also be PT's before the orifice fittings on the propellant lines.

RF

During our testing campaign, we encountered significant RF issues. Initially, we observed that whenever we turned on the spark plug, it reset Papyrus (our avionics system). Sometimes, it would even cause the huge speakers above the hangar to static. It was really spooky. To mitigate this, we bought a bigger spark plug that had a resistor inside of it, which seems to have solved the issue. Additionally, we insulated all wiring to the spark plug heavily. We initially chose a non resistor spark plug because it was very small and thus easier to fit in the assembly, but I think that getting the bigger one and fitting it in will save a lot of headache.

Purging

The ASI initially did not have purging capability, but we realized that in the event of a failed ignition, there would be some residual methagox mixture left in the chamber, which is a hazard if approached. To make the system safer, we added purging capability to the ASI by having some tee fittings that inletted nitrogen from a separate nitrogen cylinder (not the same cylinder that feeds nitrogen to the rest of the engine). There are manual valves and check valves to prevent backflow into the lines during ASI operation; additionally, we set the nitrogen reg to be at a higher pressure than the ASI operating pressures so that it is physically impossible for the gas to backflow into the nitrogen lines unless there was some significant over pressurization event in the chamber.

Fire Videos