November 13, 2004
Turbo Power, Part 3
KITPLANES Magazine, May 2001
Turbo Power, Part 3
The builders experience the agony and ecstasy of R&D.
By Tom Wyatt III
In Part 1 of this series, I covered the initial fabrication of most of the major components (engine mount and turbo plumbing) for the Subaru EJ-22 we would install in Bob Goodman's Van's RV-4. In Part 2, we built the systems and got the engine running, and then the fuselage left my shop to go to the hangar to be finished. Part 3 concludes the story with the R&D period required to turn N311U into the flying machine it is today.
I have been turbocharging engines for more than 25 years, but this was one of the most intense shakedowns of my life. As the FAA intended with its Experimental category, this project was indeed a learning experience.
Making it Airworthy
Before we could have the FAA inspector come by, we had to make the aircraft absolutely ready for flight. That included installing the wings and propeller, rigging the controls, and performing a series of engine test runs. Those completed, we were signed off after a morning's worth of paperwork. The inspector found no problems with the aircraft. Our weight and balance was within 5 pounds of the O-320-powered RV-4 in the hangar next door.
We flew the airplane an hour later. That began six months of intense R&D. Our inexperience caused us to make several poor design decisions, and they would have to be rectified at considerable expense in both time and money.
Without a doubt, the biggest challenge faced by experimenters using liquid-cooled engines is cooling. Without resorting to large car-size radiators that cause unacceptable amounts of drag and can be difficult to package, the builder must be clever and resourceful. In most ways, we were.
Our initial flights proved disappointing, both from a performance standpoint and a cooling standpoint. After five flights, Bob made up some datacards to take along on each flight to record temperature and performance data. Using these data cards throughout the test period was instrumental in developing solutions to our problems. The first thing we learned by comparing our speeds to the speed-vs.-hp graph provided by Van's was that we weren't even making 120 hp. We also learned that anytime we pushed the manifold pressure up over 30 inches manifold absolute pressure (MAP), the engine got hot--both the oil and the water. A major re-think was required.
To get a better idea of what was happening, we connected the airplane's engine management computer to our laptop and ran the engine under load (tied down) and data-logged the status of all the sensors. This revealed that we had charge-air temperatures (MAT) in the intake manifold of 200¡F. We thought because we were running relatively low boost, an intercooler would not be necessary. We were wrong.
We also noticed that the throttle could only be opened to about 12% without over-boosting. Obviously, some changes would have to be made to the turbocharger. Our goal was to adjust the compressor size (called sizing) and turbine size (called matching) to achieve our desired boost without the use of a wastegate, which is an adjustable/controllable turbine bypass. This was a bigger challenge than I realized, but we eventually achieved our goal.
Another of our grand experiments was using an oil-to-water oil cooler instead of the typical oil-to-air cooler. We reasoned that we would have excessive in-flight water cooling and could use this surplus to cool the oil without having to add another air inlet to the cowling. This theory proved unworkable.
The First Round of Mods
First we had to get the charge-air temperatures under control. That was the factor driving both the lack of power and the high oil and water temperatures. Our first step toward lowering the output of the turbo was to remove it and send it back to have the TO4 compressor (a Super-S with a 3-inch wheel) replaced with a much smaller T3 (Std-60) compressor.
While the turbocharger was off the airplane, I modified an intercooler from a twin-turbo Nissan 300ZX to fit on the airplane. We also installed a few test instruments. One was an air/fuel ratio meter that ran off the oxygen sensor. Another was a digital temperature gauge and four sensors with a rotary selector switch. The sensors were installed in the cool side of the oil cooler and water radiator, and one on each side of the intercooler (in/out). A pressure gauge was installed to keep an eye on boost (the pressure of the air coming out of the turbo) so that we could compare that to what was actually in the intake manifold after being throttled. Bob added six columns to the datacards to record the new information.
When the turbo came back, we reinstalled it, and I spent a couple of days redoing the charge-air plumbing and hanging the intercooler under the propeller speed-reduction unit (PSRU). We roughed out a hole in the cowling under the spinner to clear the intercooler and sealed the intercooler to the cowling with duct tape for a ground run. It worked! Charge-air temperatures dropped from 200¡F to 160¡F, oil and water temps were down, and performance was substantially improved. We were starting to get somewhere.
However, we were shocked to see that with only 30 inches MAP, the turbo was making 48 inches of boost against the mostly closed throttle. Now, at least, with the new smaller compressor, we were able to open the throttle to 25%, as opposed to 12% open with the original larger compressor.
After Bob flew the airplane a few times with the intercooler, I noticed it was sawing while cranking with the starter instead of spinning over smoothly and evenly. This is usually a sign of trouble, so a compression check was in order. Sure enough, cylinder No. 4 had 60 psi of compression. Something was going up and down in there, but it was less than half the compression of the other cylinders. Within an hour, we had the head off and discovered the source of our problem: A small crescent of the edge of the piston was burned away.
Bob remembered the very flight during which the piston had been damaged. On the flight before we installed the intercooler, we had allowed the oxygen sensor to control the air/fuel ratio, and it had over-controlled to the point of killing the engine in flight--never a pleasant experience. Bob had slammed the throttle full forward disabling the oxygen sensor's control, but this also over-boosted the engine. The effects were not immediately apparent. In fact, the engine was performing better than ever with the new intercooler. It was only the uneven cranking that caught my attention.
We pulled the engine off and within a week had installed four new pistons that we had ceramic coated with a thermal barrier coating on the tops and a slippery poly coating on the skirts. Total cost including four new pistons, rings, gaskets and ceramic coatings was less than $700.
Round Two of Mods
At this point, we suspected that the propeller pitch was too coarse. We had started on this project expecting to turn the engine at a conservative 4500 rpm in cruise, redlining it at 5500 rpm. Even in a shallow dive, we had trouble getting past 4400 rpm. It seemed to run into a wall there. I called Lonnie Prince, who had made our first prop, and discussed the possibility of re-pitching it. It was impossible to make the magnitude of change we needed, so he sold us a second prop at a discount. We had it on the airplane in two weeks. That's service!
Our original prop was a 70x91 Prince P-tip. The new prop was a 69x80 P-tip--a significant change. It transformed the airplane. With the 91 prop, we could only get 4000 rpm for takeoff. With the new prop we could get 4500 rpm, and the rate of climb increased as well. The engine just generally seemed happier.
Back to the R&D
The intercooler helped control the temperatures, but the turbo was still turning way too fast and putting out far too much boost against the mostly closed throttle. We still could not open it more than 25%. The obvious solution was to slow the turbo. The usual method of enlarging the turbine A/R (nozzle size) by swapping the turbine housing for a larger one was not going to be possible because we already had the biggest one (0.83 A/R) installed. Our only choice was to have the turbine housing machined to accept the next largest turbine wheel. We would replace the O-trim wheel with a P-trim wheel.
The change to the P-trim wheel made only a minute change; it was hardly noticeable. This was going to require some serious effort.
Matching the Turbo to the Engine
As I briefly explained earlier, matching is the process of adjusting the size of the turbine to achieve the desired results from the compressor. The turbine powers the compressor, and matching can be thought of as similar to changing the pulley ratio on a belt-driven supercharger. To slow the compressor, one would change to a larger turbine nozzle or a larger turbine wheel. This lowers the operating pressure of the turbine and slows it down. Since the turbine inlet pressure is also engine exhaust back pressure, this has the additional benefit of reducing the temperature and pressure in the exhaust system.
To better understand the concept of turbine matching, picture squirting a pinwheel with a garden hose. If you use your finger on the end of the hose to reduce the nozzle, the water comes out with more velocity and the pinwheel turns faster. What you might not realize, though, is that the pressure in the hose (back pressure) has also increased. To slow the pinwheel down, you loosen your finger a little (larger nozzle). The pinwheel slows down, the pressure in the hose is reduced, and the flow through the hose is increased. If you attach a shaft to the pinwheel and make it turn something useful, you have a turbine!
Because we had already machined the turbine housing to accept the biggest wheel that would fit in it (a P-trim), and we were already using the biggest exhaust housing available (0.83 A/R), we now tried putting smaller turbine wheels inside the machined-out housing. Making a fixed internal leak or built-in turbine bypass called a wastegate were our only options to slow the turbo.
Wastegates are necessary in many applications such as in cars that must build boost rapidly and then have the boost level off at a predetermined level, or in airplanes that use the turbo to climb to very high altitudes. Our installation was different, and I believed we could do the job without the weight and complexity of a custom-built wastegate. Fortunately, I was right.
We ended up changing the turbine configuration five times to get where we wanted to be, but it works perfectly now. We just kept putting smaller and smaller turbine wheels into the housing that was machined out for the big wheel, and eventually we got the right amount of boost. If you were to go by the catalog application guidelines, we used the compressor recommended for a 100-150 cubic inch engine and combined it with a turbine recommended for a 300-350 cubic inch engine (if we had kept the P-trim wheel in the P-trim housing). The Subaru is 133 cubic inches.
Why the big difference? Our application is steady state, and fast response is not an issue. Published guidelines are meant for applications where response is everything. This had never been done before, and each time I sent the turbo off for modifications that would reduce the boost, I worried that it wouldn't make any boost at all. The folks at our turbo shop are still scratching their heads.
Bob was curious about how the airplane would fly without the turbo, so one time when the turbo was shipped off to get another turbine wheel installed, I made a little plate to cover the opening where the turbo connected to the housing and flew it for a week without the turbo. It was interesting and informative.
With the new prop, the airplane was quite flyable. Bob could climb at 700 fpm and topped out at 155-160 mph and 5100 rpm. The performance was equivalent to an O-235 airplane engine even with our less-than-optimum propeller. With a finer pitch prop that would allow the engine to turn a little faster, the performance would have been much better without the turbo; we flew it wide open all the time. The fuel burn averaged 5.5 gph, and it felt like it would run forever. Even without the turbo, this engine made a mighty fine airplane if all you needed was 120 hp.
We also learned that even without the turbo, cooling was marginal. We were flying in the hottest part of the year, July and August in Georgia. We installed a conventional aircraft oil cooler to replace the oil-to-water unit. The bottom cowl was sprouting air inlets--one for the intercooler right below the spinner, and another just below that for the new oil cooler. I designed and built a one-piece cover/inlet duct to neatly finish off the two openings.
By this time, with the turbo working correctly, even the local airport skeptics were starting to be impressed. We got climbs away from the airport that couldn't help but be noticed. (Our 2000 fpm VSI was pegged hard.) One day a Lear Jet pilot even commented on frequency to the tower. The addition of the conventional oil cooler had brought our water and oil temperatures down to normal. Life was good!
The Agony Revisited
Finally, with 40 hours on the airplane and everything working well, I was going to get my turn to fly it. Less than 20 minutes into my flight, I noticed the oil temperature rising and the oil pressure dropping. We were flying from a small uncontrolled field, so I just pulled the power and made a quick landing. An on-the-spot inspection of the oil filter confirmed the worst. Something was seriously wrong.
Wayne Parks, the owner of the repair shop at the Monroe, Georgia, airport, literally gave us the keys to his hangar and tool chest. We removed the engine on his ramp and took it to my shop for a teardown and inspection. We found the No. 4 rod bearing had worn out, probably damaged when the No. 4 piston was damaged by detonation.
We bought a brand new turbo short block from the Subaru dealer to replace our high-compression, naturally aspirated assembly. The primary changes from the NA block are that it has a solid deck where the heads bolt on, and it has lower compression HD pistons. We also bought the larger turbo oil pump. In two weeks we were flying again, and the engine was running better than ever. The oil pressure also went up by 15 psi. The total cost for this repair was about $1700.
With the new turbo block, the engine has been essentially trouble free. We have spent the months since tweaking the fuel injection and ignition and making small aerodynamic modifications to improve the cooling. (It was summer again.) Ideas for improving the cooling air outlets came from Barnaby Wainfan's "Wind Tunnel" column in this magazine.
We entered the airplane in the Sun 100 race at Sun 'n Fun and finished third in class with a speed of 203 mph. Bob ran the entire race at 6100 rpm and 40-inch MAP. We consistently climb at more than 2000 fpm, even with two people on board. It climbs at 1500 fpm at 14,000 feet at 40-inch MAP with two on board.
We flight-plan for 8 gph to be conservative and cruise at 5500 rpm and 32-inch MAP. That usually works out to 160 knots on the GPS. We climb at 5500+ rpm with the MAP between 40 and 50 inches. The engine just loves 5500 rpm.
With one exception, we have equaled the performance of an O-320 without exceeding its weight. What is the exception? The climb rate! This engine produces climbs far superior to any RV powered by an O-320. The sound this engine makes is also lovely, much like a P-51, especially with the turbo whistling at 5500 rpm. It's vastly smoother and quieter than an O-320.
Our plans for the future? We're going to fly the pants off this airplane! We have nearly 200 hours on it now, well more than 100 since we replaced the engine. Are we having fun? Absolutely! KP
Tom Wyatt III has been turbocharging engines for 25 years, but this is his first airplane engine. He is a glider flight instructor, glider towpilot, and he is currently building a Europa. He works as a turbocharging consultant and software application specialist for computer-controlled management systems. (Tom Wyatt died as the result of a traffic accident early in 2006. Ed.)
IMPORTANT CUSTOMER SERVICE NOTICE TO KITPLANES SUBSCRIBERS:
All mailed correspondence, including subscription invoices, renewals, and gift notices, will bear our address:
PO BOX 8535
Big Sandy, TX 75755
Third parties claiming to be selling KITPLANES subscriptions are not legitimate.