Subaru diesel EE20 aircraft RV-7

[rv6ejguy]

Thanks for the realistic weight and likely BSFC info on the EE20 Thomas.

So we're looking at around .33-.35 lbs./hp/hr. actually, within a fairly narrow rpm and load range range.

[tgmillso]

No worries Ross. I just wish I had the time on my hands to finish the EE20 build. You could have warned me that just building this aircraft from the standard kit was going to be such a massive undertaking!!! Then again... I think you may have and my overly ambitious nature may have brushed it aside. It's the over analysing and constant desire for perfection that gets in the way!
T.

[DanH]

Originally Posted by tgmillso

I see the key to making this a success being the use of much lighter flywheel and ditching the heavy DMF and using a wood core composite prop such as an MT that can handle the impulse loads. This is what is done on the SMA diesel. The Continental 2.0 diesel needs the DMF as it is still running through a gearbox. It would also be beneficial (and in my eyes possible) to trim the weight of the direct drive mount. The marcotte unit is designed for applications of over 800hp, so it's a bit of overkill in this situation.

The dual mass aspect of a DMF (Dual Mass Flywheel) brings an additional inertia to the system, which has an effect on some natural frequency or another. However, at the fundamental level the DMF is a low stiffness soft element.

It might be a major mistake to assume you can run this diesel with a prop bolted to the end of the crank, and assume there will be no torsional issues. We're talking old knowledge here; it's why pendulum absorbers were added to Lycoming and Continental cranks.

(BTW, I have not seen an SMA crank. Anybody know if it has pendulums?)

Remember, a resonant period runs crank stress way up. My guess is that the Fuji engineers already pared the crank down to what they felt was an absolute minimum strength given the careful torsional tuning possible in the car. Given the broken cranks out there, there certainly isn't any great excess. I doubt it will handle any significant resonant period for long. Only way to know for sure is to do the math, an/or strain gauge the shaft.

As for the Marcotte, the typical "pin and bushing" flywheel-to-drive connection is not nearly soft enough. Ross will tell you his Marcotte/gas Sube resonates quite badly at an RPM in the teens. Given the same coupler, gearbox, and prop, the diesel will hammer even harder.

Either way, up-front attention to torsional issues would make it a lot less like a dice roll. This is 2016. Isn't it time to bring a little engineering to the game?

It's the over analysing and constant desire for perfection that gets in the way!.

Sheesh.

[rv6ejguy]

As far as I know, Thomas is not using the Marcotte gearbox, rather a custom made bearing set to isolate prop loads from the engine.

Certainly experimental in nature.

[tgmillso]

Hi Dan. Ross is correct in that the the unit is direct drive and has no gears at all, just a massive pair of bearings (TRB if I remember correctly). I would not be coupled directly either. The clutch friction disc is currently connected directly to the Marcotte spline, and the friction disc face will be bolted directly to the flywheel through the existing rivet holes in the friction disc and into corresponding holes drilled into the flywheel. The friction disc is like most friction discs in that it has a set of springs anyway. This is used primarily in this application for negating any misalignment issues between the crank and the Marcotte shaft, but will also help to distribute the combustion impulse loads over a longer time period. I fully realize that torsional vibration testing is an absolute must for this setup, given that calculating any TV issues is a highly complex task beyond my modeling capability. I'm also aware of the history of the crankshafts and the multiple iterations to address the prior issue, and from the research I have completed, the latest versions have been without fault. In addition I spoke with the engineer many times at MT who developed the propellers for the other diesel applications, and they are extremely confident of the propeller they have specified for the EE20 on an RV-7. It is essentially the three blade O-320 setup with an 0-360 bolt/flanges for the higher impulse loads. I'll upload a picture of the setup on the Marcotte when I get home.
I just hope someone finishes developing this before me because I've been so slow!
Tom.

[DanH]

Originally Posted by tgmillso

The clutch friction disc is currently connected directly to the Marcotte spline, and the friction disc face will be bolted directly to the flywheel through the existing rivet holes in the friction disc and into corresponding holes drilled into the flywheel. The friction disc is like most friction discs in that it has a set of springs anyway. This is used primarily in this application for negating any misalignment issues between the crank and the Marcotte shaft, but will also help to distribute the combustion impulse loads over a longer time period. I fully realize that torsional vibration testing is an absolute must for this setup, given that calculating any TV issues is a highly complex task beyond my modeling capability.

Adapting a clutch center is an approach straight from the bad old days of PSRU development. Run it, break, it, "experiment" with another random selection, run it, break it, repeat. It's like trying to write "War and Peace" by arming a million monkeys with a million typewriters.

Coupling diesel engines to driven systems of all kinds is a mature science with the support of an entire industry. The correct torsional coupler is selected based on required performance. Maybe you don't want to study five volumes of Wilson, but even the catalogs have enough selection information to start your project down a reasonable path. If you later need to make a change, the catalog allows a new selection based on specified torque and stiffness data. You can pick something a little stiffer or softer, and know what you changed.

Take some time and look around.

Centaflex offers excellent catalog and online selection help:

http://www.centa.info/?show=downloads&c=us#

The Lovejoy catalog includes pages with simplified selection calculations:

file:///C:/Users/wks/Downloads/Torsional2010.pdf

[tgmillso]

Hi Dan,
Below is a picture of the back side of the Marcotte direct drive housing. The friction disc has 16 rivet holes of approximately 5 to 6mm in diameter (hard to measure until I drill them out) which will be bolted to the flywheel face (so that the pressure plate etc. will not be required). I apologize about the quality of the photo, however you can also see the friction disc springs, which have a second black spring internal to the outer blue spring.
Tom.

[tgmillso]

Hi Dan. Interesting about the coupling setup. Guy actually steered me away from this path and toward direct bolting of the friction disc to the flywheel, as the springs would take up any potential misalignment. I will take a look further at the links you have provided.
Cheers,
Tom.

[DanH]

Originally Posted by tgmillso

Hi Dan. Interesting about the coupling setup. Guy actually steered me away from this path and toward direct bolting of the friction disc to the flywheel, as the springs would take up any potential misalignment. I will take a look further at the links you have provided.
Cheers,
Tom.

Bad advice. The ability to accommodate axial, radial, and angular misalignment is a key feature of most engineered couplers. For example, an ordinary Centaflex A will handle 1.5 to 2 mm of radial misalignment. And unlike a clutch center, there is a published spec for it.

If your design requires an input shaft spline, note the availability of clamped spline hubs, designed to eliminate freeplay and fretting.

Model overview, one manufacturer:

http://www.centa.info/data/content/p...11englisch.pdf

The massive Marcotte bearing support case may be quite a lot heavier than necessary. The thrust and moments are not huge, unless you plan on really, really hard acro. You can easily quantify the prop moment:

M = I x Q x B

where
M = propeller moment in ft-lbs
I = prop mass moment of inertia in slugs ft^2
Q = shaft rotation rate in radians per second
B = pitch or yaw rate in radians per second

Propeller mass moment of inertia is determined with 2 lengths of safety wire and a stopwatch ("bifillar suspension"), easily done in the home shop. To convert RPM to radians per second, multiply by 0.10472. Use 6.28 radians per second for pitch/yaw, a fairly radical assumption. With moment in hand, you can pick bearings, shaft size, and required case strength.

For reference, the magnesium nose case used on the M14 is very light, and handles loads way beyond anything you'll generate with this little diesel.



The nose case bearing support idea for an auto conversion has been around a while. Here is an Arden Hjelle design, the aluminum block Olds as used in Steve Wittman's Tailwind. The open bellhousing proved to be more than strong enough.


An earlier nose case, again Hjelle/Wittman, for the Formula V racer, circa late 60's. It's just a light sheet metal cone welded to end plates.

[drdrainb]

Originally Posted by BillL

Typically, the fatigue life may be ok if operated within its "normal" limits, but continuously elevated temperatures would be the area of interest. The light duty diesels in automotive are already hugely different (i almost said compromised) compared to an engine designed for heavy continuous service, so installation in an airframe and appropriate consideration to loads, torsionals etc seem like an intelligent path to evaluate this engine in this application.

Facts and data will need to be gathered here. I don't want to speculate on its success as there are many ways to mitigate issues and potential issues. This is a "development" project with no obvious prediction of the final result. I am sure the process will be interesting, and success would be a big positive step forward for a diesel engine option.

a good question indeed. I owned a 2Liter peugot diesel (turbo) for 5 years. made reqular trips of 240+ miles with 2 passengers & 200Lbs. of gear running the engine at around 2500 rpm for the ENTiRE trip.

this was HARD driving, interstate driving, so... continuous. steep hills, 4 to 6% grade.

NO break downs.

NO problems- not even MINOR ones.

38 MPG.

but- this has been my experience with EVERY foreign made diesel engine. ESPECIALLY japanese & german made.

so? YES. given subaru's WELL PROVEN history? I'd say he's on the right track.. putting a subie diesel in an aircraft. i look forward to the results :-)