How To Jet Engine
What’s going on here?
So in my career and hobbies I have had a lot of opportunities to work with, design and test piston engines. All types of piston engines I might add, and there’s good reason for it. The piston engine is actually quite efficient, easy to package, and very common (parts and standard practices are everywhere). However I purchase an unassemebled Monnett Monerai sailplane kit some years back, with the intent of making it self launching. After many layout studies, the piston engine would just break the actually quite elegant design of the glider, so I turned my focus to Jets.
Prior to this project I had limited experience with Jets, basically what I had learned in school many years prior. It’s not to say that my education was less than satisfactory, it’s that you can never truly understand a system until you try operating it. To do this, I got my hands on a couple of small jets, quickly excised the equipped controller, and got to work.
The JetCat P240 comes with a propane start, the highly flammable gaseous fuel is introduced to the air mixture and ignites by a common RC glow plug. This then ignites the main fuel charge, and off we go. However… this would not be practical for use in my self launch glider, I’m looking for Jet-A only operation. Although JetCat does offer a kerosene start package, what would be the fun in using it. Time to make my own.
Phase one: will it blend?
The first step was making my jet-A based ignition system. The Igniter itself is conventional, fuel is introduced around glow plug, passes through a calibrated orifice, enters a mixing and ignition zone near the tip of the glow plug, and after ignition passes into a flame holder area intent on anchoring a stable flame jet. Simple enough but fitting it in the existing tiny hole was a bit of a challenge.
My “machining center” is humble, usually just used for modifying tools or making small fixtures, but enough to get this job done. I made a series of three igniters out of 6061 since it’s easy for my mini lathe to handle, and good for one or two test ignitions before dissolving, and once I got the design I was looking for, made it from stainless. I rigged up a bench test as shown in the slideshow above, and quickly got the desired ignition, characterized by a stable flame jet exiting the outlet of the igniter. Even though in the fully developed phase a significant portion of fuel exited the vent holes, I elected to proceed to make one from stainless and move onto in turbine testing.
Turbine Test Setup
I haven’t done software in better than 10 years so to conduct this first phase of test I hacked together a series of almost analogue devices to make the various things work. At the front end I used a Pololu Maestro 12 and it’s time based sequence controller to script the start routine. The controller ran two motor controllers (the starter motor and the fuel pump) as well as a series of RC relays to activate things like the fuel shut off valves and a high current relay for the igniter.
The stand itself was pretty basic, I found a couple of semi-truck exhaust clamps that happened to be the correct diameter for the baby turbine, and welded together a small frame to hold all the equipment. Home Depot had a small Huskey “work-cart” on sale, which is no the my turbine test cart (no I did not read the safety manual)
The instrumentation was minimal, a hall-effect sensor glued to the starter motor housing to pick up RPM, and oscilloscope to read it, and a thermocouple in the exhaust nozzle for EGT. Basic but all that I needed to get things going. The Pololu would go through a series of preheating, ignition spooling and fueling to get the engine up and onto self sustaining idle… and just about that fast my little fuel tank would be empty, time to refil, readjust and do it again. A handful of starts and it was dialed in, about 60s from hitting the button till the turbine was ready to take power.
The Hot Start
Although it makes for very dynamic still images, the hot start isn’t exactly good for the engine. On my first several ignitions I stumbled into this condition, aborting the start and reattempting. I worked my way through it and this is what I learned (in terms that anyone should understand:
The hot start is caused by the time duration of the combustion process taking too long, such that it exits the turbine section while still combusting. The net result of this is the energy from the combustion is not delivered to the airflow ahead of the turbine, then not utilized to drive the turbine, which in turn does not drive the compressor, which does not add more air to the system to fuel the combustion event and cool internal components. The outward appearance is the exhaust gas temperature will continue to rise with fuel flow, but the turbine shaft speed will not. If left in this condition too long the turbine will exceed it’s operating temperature causing permanent damage… don’t do that.
This is in part due to the interesting vaporizer combustion system of the JetCat, that requires waste heat to vaporize the incoming fuel such that it can ignite quickly. This vaporizer tube system has some advantages in operation, but has started another design effort to replace this system with a more capable mechanism of fuel delivery.
Now you would think i could have learned this from a book and got to a good starting sequence faster with less risk to damaging components, and you are right. But what I learned in this process are the sights and sounds of a hot start and how to work through corrective actions. This (well not this example exactly) is what I wanted to learn when getting engaged in this project.
Phase 2: Time to Get Serious
So the first attempt was just to get a scope on what needed to be done. It’s cumbersome to operate, doesn’t record any useful data, and can’t really be used for any future development work. Take these negatives, and make them the objectives of the next version.
A new stand and control system are currently being designed, the stand capable of measuring thrust and fuel flow, such that meaningful test data can be taken. As for the turbine control, away with the Pololu hobby controller and onto more proper National Instruments equipment. Now it’s been a long while since i’ve done software and at the moment I’ve gotten as far as being able to read and record temperatures and make a light flash. Next steps are hardware selection for the different instruments and design of the test stand. Pics below for reference. Outside of Covid related hardware delays, I would expect in the coming few weeks to have some meaningful progress to share.
