December, 2009 The project was completed (behind schedule but within budget). Here's the final result:

Hull: 11' X 7' fiberglass Rigging: Brazed 304 stainless tubing Engine: Tiernay TT10 turbine Prop: Woodcomp SR2000 Rudders: NACA 0-0-12 airfoil Trailer: Welded steel

At the first swamp-trial.

Read on for the humble beginnings...

Why:

No other kid on the block has one. Learn a new set of skills. Most airboats can’t back up. Share a project with my son. He likes things that make noise and go fast.

Basic Concept

1. Find a turbine engine between 140 and 260 shaft horsepower and weighing no more than a piston or rotary engine of similar horsepower. Familiarize myself with the proper care and feeding of the engine by running it on a test stand in my garage.
2. Find a variable pitch reversing propeller no more than 72 inches in diameter and appropriate for the engine’s horsepower.
3. Mate the engine to the propeller and measure static thrust on the test stand.
4. Locate or build an airboat hull of proper size and weight for the thrust produced. Mount engine and propeller in the hull. Attach seats, fuel tank, throttle, propeller pitch and rudder controls (plus equipment required by the Coast Guard to make it legal).

August, 1999 - Purchased a Garrett GTP 70-52 APU through barnstormers.com for $500 plus $250 shipping and handling. This engine is rated for 160 shaft horsepower at 6000 RPM after gear reduction. It had not been run in 15 years and was missing its oil cooler, fuel pump and ignition system. At 200 pounds, it is on the heavy side, but still within the stated requirements. I mounted it on a test stand, attached a $49 transmission cooler and hooked up the oil pressure gauge and tachometer. To get the 24 to 28 volts DC required by the starter motor, I used 2 marine batteries out of my wife’s boat, a Perko switch and a Ford starter solenoid. Oil pressure was 95 PSI at 30% of operating RPM. To determine the correct voltage for the igniter, I attached a large neon sign transformer and varied the primary voltage. An acceptable arc in the combustion chamber was observed between 6000 and 8000 volts.

Before shipping What a mess!

On a test stand

Instrument Panel

September, 1999 - With the required voltage known, it was fairly easy to construct an ignition system. Version 1.0 consists of a Delco capacitor, coil, Radio Shaft DPDT relay and a toggle switch. Input is 24v DC to one terminal of the coil and the normally closed contact of one relay pole. The central contact of the same pole is wired through a 150 ohm resistor to cause the relay to vibrate. The other pole of the relay is used to oscillate ground to both the capacitor and the other terminal of the coil. It is much like the early ignition circuits used by Nicola Tesla and Henry Ford 100 years ago, but effective. In version 1.1, power transistors will replace the relay and the 24v DC input will be wired in parallel with the starter motor to replace the toggle switch.

Ignition system

Schematic

To test the injector, I made a fitting to adapt the fuel line to an electric paint sprayer filled with Coleman fuel. After spraying some fuel into a jar to verify that the line was clean, I coupled the line to the injector and measured fuel flow at 5 gallons per hour. To run at full load, the engine needs fuel at 34 GPH. The low flow rate is likely due to the relatively low pressure produced by the sprayer. To test the ignition system, I spun the engine up to 25% on the starter motor, turned on the ignition and fuel (in that order!) and the engine accelerated to 40%. No flames or evidence of overheating were observed at the turbine exhaust. Disengaging the starter caused the engine to decelerate slowly, so I turned off the fuel. This test shows the injector and ignition to be operational, but the engine needs more than 5 GPH of fuel to achieve and maintain idle RPM. The next test should be done with JET-A fuel and a larger fuel pump.

Though functional, the first ignition system proved unreliable. As predicted, the relay contacts did not last very long. Even after replacing the relay with solid state circuitry, the coil became hot after each use. Version 2.0 of the ignition system consists of a 300 watt inverter and the transformer previously used to determine ignition voltage. Because of exposed terminals and a grounded center tap on the transformer, the components had to be mounted in a plastic box with nylon nuts and bolts. This is bulky, but should suffice until an original exciter box can be located.

Ignition 2.0

Fuel Pump

The original equipment fuel pump arrived from Avon Aero. It is a gear type pump with pressure to the injector regulated by engine RPM, compressor pressure and exhaust temperature. The governor trim control is electric. This is an interesting piece of 1960's technology that mechanically performs functions that most people would not attempt today without an embedded microprocessor. However, it does require fuel to be supplied to the inlet at 15 PSI. Self priming does not seem to be one of its design requirements. Once the fuel pump is installed and tested, I expect to be able to run the engine up to self sustaining speed.

October, 1999 - When connected to the fuel supply, the fuel pump leaked from several places that shouldn't have fuel in them. Upon disassembly, it was found to have a blown seal around the shaft between the gear pump and the governor and 2 ruptured diaphragms in the acceleration limiter assembly. Starting the engine under those circumstances would have been a very bad idea. While waiting for an exchange pump, I tested the accuracy of the instrument panel. The original electric tach was 99% accurate by comparison to an infrared digital tach. The thermocouple for the EGT (Exhaust Gas Temperature) was found to be defective. I replaced the EGT thermocouple, wiring harness and gauge with a $153 kit from VDO instruments.

New EGT Gauge

Thermocouple

Phototachometer

The new fuel pump arrived and passed its leak test! Before it gets installed on the engine, I need to measure fuel pressure and flow over a range of speeds and simulated compressor pressures. Bench testing is tedious, but a far cheaper way to detect problems than during the first start cycle.

During initial bench testing, the fuel pump did not produce sufficient flow to start the engine until 27% of operating RPM. With the pump installed and all the compressor air bleed lines connected, the engine only spins up to 25% on the starter motor. Even lower starter speed can be expected after an alternator is installed and a load is applied. Further bench testing showed the pump to be producing good internal pressure at low RPM, but the fuel control unit refused to output more than a dribble below 27%. By installing a bypass valve from the gear pump pressure port to the injector line, I plan to start the engine with the valve open and close it after RPM exceeds 27%. If successful, this procedure could be automated.

Bypass Valve

November, 1999 - With the fuel pump installed, the engine started easily and quickly accelerated to 80%. At that point the bypass valve started to leak and the fuel had to be shut off. Though certified for fuel on an airboat engine, the valve was not rated for the pressure generated by the pump (pegs my 300 PSI gauge). The next start will be with a properly rated bypass valve.

December, 1999 - By spending most of my spare time with family and friends during the holidays, I did not make much progress on this project. With a new bypass valve, the engine now starts reliably. The next run is scheduled to start at 23:59 31 Dec 99.

Because of their tendency to make foam, synthetic piston engine oils (Mobil 1 and Castrol Syntec) are unsuitable for test runs over a minute. I am now using Mobil jet oil II at twice the price, but worth it.

Turbine Oil

The cost of reversing propellers has caused me to reconsider my original strategy. By using the engine to drive a variable displacement hydraulic pump and hydraulic motor, it should be possible to use a $700 fixed pitch prop instead of a $7000 variable pitch prop. The major drawbacks are additional weight and less horsepower to the prop.

March, 2000 - I got busy with revenue producing activities for a couple of months and didn't have much time for fun stuff, but now it's back to the project. The engine had a bad habit of flaming out at 100% and restarting with a loud bang at 80%. By installing a 400 PSI pressure gauge on a tee fitting near the injector, a sudden loss of fuel pressure could be observed at flame out. Tracing the problem back through the solenoid fuel cutoff valve and the wiring harness showed the oil overpressure switch to be activating as turbine speed approached 41,000 RPM. Oil pressure at the manifold next to the oil pump never exceeded 110 PSI. Two possible causes of this problem are restricted oil flow through the bearings or a defective overpressure switch. Measuring oil pressure and flow through both sets of bearings indicated that the switch was causing the problem.

April, 2000 - We took a field trip to the Sun-N-Fun fly-in at Lakeland, FL. to check out some other people's projects. NASA had the new Williams FJX2 on display. This engine has a 4:1 bypass ratio, weighs about 100 pounds and produces 750 pounds of static thrust. I want one!

My son concluded that a seaplane with retractible gear would be more versatile than an airboat. I can see the beginnings of our next project 5 or 10 years down the road.

Not one of my projects (yet)

August, 2000 - The fuel control unit has defied all attempts to get it to act as a governor. Each start accelerates straight to 110% where the safety mechanism on the centrifugal switch cuts off fuel completely.

To get some manual throttle control, I installed a ball valve between the injector line and the return line to the fuel tank. This worked, but was difficult to control. Installing a Helmco stainless steel needle valve (not available at my local hardware store or parts counter) produced very fine control of engine speed under no load conditions. If this valve can be equipped with an actuator and controlled using the the signal from the tachometer generator, it could compensate for the shortcomings of the fuel control unit.

Needle valve throttle

November, 2000 - Being temporarily stumped on how to economically attach a prop to a single shaft engine, I started rebuilding a Fairchild J44 turbojet. This was the original "two turning, two burning" engine on the C123J before being replaced by the more powerful J85 on the C123K of Air America fame. I have not been able to locate a manual and am not really sure what to do with it when finished, but it should make an interesting project.

Adding instrumentation, fuel, oil and ignition systems to the parts pictured here should produce an engine capable of 1000 pounds static thrust. Back row: Compressor housing, centrifugal compressor with shaft, combustion chamber and outer case. Front row: Stator, turbine and hot section bearing assembly.

Fairchild J44

December, 2000 - The rear bearings have a dry sump and are loosely sealed between the shaft and the turbine. Forward bearings had no sump and no seal between the shaft and the compressor. Fairchild's design apparently was to circulate oil in the rear bearings, but to loose oil through the forward bearings into the compressor. Presumably, this was to save the weight and complexity of oil cooler, deaerator and scavenging pump on an engine intended for limited use. By putting together a kit of new, custom machined and reworked parts, I have adjusted the position of the forward bearings to allow for a seal and a sump to be added to the assembly.

Parts kit

Assembled on a stand

With all internal parts cleaned, inspected and assembled, the turbine rotates freely with no noise. I even spun it up to 400 RPM (2.5%) using a combination of shop air and a 1 HP electric motor. It takes a respectable amount of power to spin a 24 inch compressor and turbine. Much more air or a larger motor will be needed to actually start the engine.

Lubrication System

January, 2001 - Concern that pressure differential between the front and rear bearings might cause all the oil to flow through the front bearing (and none through the rear) led to a more complex lubrication system than originally planned. The first of 3 electric pumps draws oil from the tank and pumps it through a filter to the rear bearing. The second scavenges oil from the rear bearing and pumps it through a second filter to the front bearing. The third scavenges oil from the front bearing and returns it to the tank. An oil cooler and deaerator can be added between the third pump and tank as needed.

February, 2001 - After learning all the wrong ways to build an ignition system with the GTP 70, it was time to take a different approach for the J44. AeroSource in Navarre, FL sold me a box of Champion FS-31 ignitors. No way to tell if these were original equipment on this engine, but they are clearly the right type and vintage. The length, diameter and bolt pattern match the igniter hole perfectly. I bought a GE exciter box from and traded it to a turboprop mechanic for a Bendix exciter box and a 20 inch ignition cable. Yes, ebay actually has a category for aircraft parts. It can be a little hard to find using their index, but you can always search for a term like "pitot tube" if you get lost. Applying 24V DC to the assembled components causes the ignitor to fire rapidly and make a loud popping sound with each discharge.

J44 Ignition System

Control Panel

In anticipation of a first start, I built a simple control panel on a piece of tinted Lexan. Controls are main circuit breaker, 3 individual oil pump switches, ignition, keyed fuel cutoff / starter, fuel pump and fuel control trim. Instruments are EGT, tach and LED's to indicate upper and lower trim limits.

March, 2001 - Thanks to E. J. Potter for some much needed J44 advice. The larger of the 2 ports in the rear bearing assembly is not an oil sump, but for compressed air to provide cooling. Both sets of bearings were intended to operate with a total loss lubrication system. Even with this port connected to compressor bleed air, the bearings and the sheet metal portions of their housing can overheat if the engine is operated for more than 10 minutes. This condition can quickly cause complete engine failure. Now, the engineering challenge is to figure out how to keep the rear bearings and the neighboring parts cool when the engine is running. Increased oil circulation, higher volume of compressed air, liquid coolant jacket / radiator, water injection, ...?

April, 2001 - The previously mentioned problems of attaching a prop to a single shaft engine got the better of me and I sold the GTP 70. I am now looking for an engine with a separate power turbine to continue that project.

Meanwhile, the J44 is still making progress toward a first start. By applying 60 cubic feet of compressed air at 2000 PSI to the air start fitting, the turbine spun up to 1000 RPM (6.5%). This still seems a little low for anything but a really hot start. However; without a transmission it's a choice of having a tach gen or an electric starter on the accessory shaft.

Yet another project is taking shape. My son wants to build a 58 pound class battlebot. The rules don't actually prohibit the use of turbine engines, but the fuel limitation makes it impractical. Besides, parts rotating at trans-sonic speeds don't get along well with an opponent bent on poking stuff into the works. Assuming a more conventional power source, look for a link to a new web page in the near future.

May, 2001 - Believe it or not, this assortment of aircraft, hydraulic and industrial parts is the beginning of my fuel control unit. There is a 3/4 HP electric motor driving a vane type hydraulic pump. The 2 valves are a 24 volt industrial solenoid valve and the original J44 throttle valve connected to a Whittaker DC gear motor. The Bendix fuel filter is also J44 original equipment.

Home Brew FCU