What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Eggenfellner Engines - Technical Only

Status
Not open for further replies.
DanH

DanH

Legacy Member
Mentor
Ok guys, Milt said it best....let's have some restraint. Technical issues only, please. Same rules for everyone, including Jan. No discussion of business practice, resale value, the market acceptance of auto conversions, or any of the 1000 other things not strictly technical and mechanical.

Kahuna has agreed to moderate, which is pretty darn generous given the difficulty sometimes found in polarized subject matter. Let's make it easy for him.

Jan, an easy question to get the ball rolling. How do you prepare the base engine when it arrives at your shop, before adding accessory items and the gearbox?
 
This is one of the reasons I purchased an engine from Jan in the first place. He does not open up the engines to make any internal modifications, at least that has been my understanding. If the basic engine model from Subaru is rated to make the horsepower needed or wanted, all that is needed is the right packaging, care and feeding to make it a good candidate for an aircraft, plus a very reliable reduction gearbox of course.

Once a vendor opens up these engines the customer must place a tremendous amount of faith that the vendor's theories and machine practices will not result in an internal failure of some sort from internal modifications. In my opinion, if the engine is up to the task, it should not need internal changes.

It is a difficult enough task to get a package together that will work in an aircraft, without adding the additional complications of making internal modifications.

What is needed is a properly designed cooling system, a good supply of clean fuel with reduncy, a good flow of electrons to the ECU with reduncy, but most of all a good gear reduction unit so the engine can run in the RPM range where it makes it's best power.

I probably would not have purchased a Subaru engine from Jan or anyone else if it had internal modifications.

A person with a decent mechanical aptitude can make external changes to enhance the engines required operating environment with out too much difficulty. I would not want to be dealing with all the external stuff while wondering about what has been done to the inside of an auto conversion engine.

My technical .02 for this eve:)

I like the idea of this thread staying technical and leaving the personal and business practice junk behind. This is the fun stuff!

Randy C

PS. I guess to be complete here I need to throw in the fact that I did need to defeat the variable valve timing function on my STI as part of an upgrade program. This involved replacing the intake cam shaft sprockets and plugging some oil passages with tapered pins. I did not know about this requirement when I made my purchase and it would have effected my purchase decision. I guess this is not very deeply internal but close enough to be of concern.
 
Question: How do you prepare the base engine when it arrives at your shop, before adding accessory items and the gearbox?

Answer: My answer is identical to that of Randy. We do nothing to the original engines other than the removal of anything that looks like it has any weight :) When we use engines that have run before, we run compresion and oil sample checks.

Jan
 
Jan, Can you address the technical reasons for the Gen 3 gear box? What makes it better than the Gen 1 & Gen 2? I helped a buddy replace his Gen 1 and I must say the Gen 3 in an impressive "chunk of hardware" (technical term ;)) The biggest change we noticed was going from an external input shaft to female set up with the shaft "adapter" bolted to the flywheel.

1. A new bearing was added to the center of the flywheel to accommodate a stub shaft from the "adapter shaft" for what appears to be side loading? Was / is that an issue with the old box? What bearings did you use inside the new box, ball or roller bearings, both?

2. While the "engine mounting plate" appears stout enough, have you measured for "flexing" under the sever gyro force load of the spinning prop? Is that where side loading may be developing?

3. What temperatures rises (over ambient air temp) should a new Gen 3 gear box see during break-in and normal operation?

4. Do you suggest having an oil analysis done at regular intervals? Do you have any base lines for "normal wear".

Thank you in advance for your comments.
 
Last edited:
1. A new bearing was added to the center of the flywheel to accommodate a stub shaft from the "adapter shaft" for what appears to be side loading? Was / is that an issue with the old box? What bearings did you use inside the new box, ball or roller bearings, both?

The bearing you refer to is identical to a clutch pilot bearing used with a manual transmission car. This small sealed bearing does not turn but does move slightly when the torsion damping flywheel is exercising it's damping function. Its function is just to center the splined drive shaft perfectly on the crank centerline. From your description, it sounds as though you were working on an engine with a solid flywheel, and in that case, the bearing is only used for centering purposes during assembly. The new drive exclusively use ball bearings. To support the input gear, G1/G2 drives had a single input bearing and an internal roller bearing to the gear. G3 drives have two different sized ball bearings, spaced apart, on the outside of the input gear and a single ball bearing internal to it. Side loading is not the issue in this area. Creating a system with plenty of support, easy oiling and large enough bearings to handle the impulses from the engine, yet small enough for the RPM is the key. These are the fastest tuning bearings in the drive unit.

2. While the "engine mounting plate" appears stout enough, have you measured for "flexing" under the sever gyro force load of the spinning prop? Is that where side loading may be developing?

You are confusing the entire plate with what happens inboard of the bell housing perimeter. The plate is bolted in 10 places around the bell housing, effectively making it one with the housing. Then, only 1" inboard, the structure from the aft drive section, further reinforce the integrity and assure consistent alignment of the drive input shaft to the crank centerline. We have tested this by flexing the tip of the plate forward 4" with less than 0.001 deflection inboard of the bell housing perimeter. Many mistakes this plate for being very heavy. In fact it is not. The weight is only 7.5lb and provide the engine with an engine mount, gearbox mount, starter mount, oil cooler mount and cooling system mount. the plate is reinforced by traditional aircraft triangulation, using 4130 N steel tubing.

3. What temperatures rises (over ambient air temp) should a new Gen 3 gear box see during break-in and normal operation?

There is no brake-in period. The temperature should remain 15 F below the engine coolant temperature. Max engine coolant temp. is 220 F. Enough air cooling of the drive should be provided to maintain this ratio. The spinner to cowling gap is directly related to the temperature of the drive unit and a 1/2" gap is preferred over a smaller gap. This also allow for easy cowling installation / removal.

4. Do you suggest having an oil analysis done at regular intervals? Do you have any base lines for "normal wear".

Yes, send in a sample every 50 or 100 hr. The company you send it to will provide a guide for normal wear. What you are really looking for are trends, comparing your first analysis to the next, to the next.

Jan
 
With regard to GEN3 operating temperature, my unit has about 40 hours since last summer and it has not been above 180F on the hottest days. Last time I flew in January, it barely got up to 145. This unit runs cooler than GEN1-2.
 
<<G3 drives have two different sized ball bearings, spaced apart, on the outside of the input gear and a single ball bearing internal to it.>>

Here is a photo borrowed from Jan's website....picture worth 1000 words:


Jan, do you have a photo with the components spread out like an exploded drawing? Safe bet that not everyone understands the power transmission path or the bearing support scheme by looking at the asssembled components.

<<2. While the "engine mounting plate" appears stout enough, have you measured for "flexing" under the sever gyro force load of the spinning prop?>>

The MT prop has a low MMI; 0.36889 slugs-ft^2. If you assume airframe pitch or yaw at 360 degrees per sec and 2500 prop RPM, the gyro moment is only 101 ft lbs. In the big picture, it is one of the little loads. Jan's bend test is very sensible.

STUPID ALERT! (2-7-08)

The above calculation is wrong. I skipped one multiplication and used prop RPM in revolutions per second. The correct value is prop RPM in radians per second. The equation is (yaw in rad/sec)x(RPM in rad/sec)x(prop MMI), and if the MMI is in slugs-ft^2 the answer is in ft-lbs.

Yaw = 360 deg/sec = 6.28 rad/sec
RPM = (2500/60) x 6.28 = 261.7 rad/sec
MMI = 0.36889 slugs-ft^2

So, 6.28 x 261.7 x 0.36889 = 606 ft-lbs

Duh!
 
Last edited:
janeggenfellner said:
2. While the "engine mounting plate" appears stout enough, have you measured for "flexing" under the sever gyro force load of the spinning prop? Is that where side loading may be developing?

You are confusing the entire plate with what happens inboard of the bell housing perimeter. The plate is bolted in 10 places around the bell housing, effectively making it one with the housing. Then, only 1" inboard, the structure from the aft drive section, further reinforce the integrity and assure consistent alignment of the drive input shaft to the crank centerline. We have tested this by flexing the tip of the plate forward 4" with less than 0.001 deflection inboard of the bell housing perimeter. Many mistakes this plate for being very heavy. In fact it is not. The weight is only 7.5lb and provide the engine with an engine mount, gearbox mount, starter mount, oil cooler mount and cooling system mount. the plate is reinforced by traditional aircraft triangulation, using 4130 N steel tubing.


Jan
Click to expand...

I did notice the inboard bolts and the area where the "bell housing" would be in the automotive application. That is one stout plate!

Jan, one more question. Is there a need (like the Rotax 912 / 912S) to adjust / check the preloading tension of the Gen3 gear box? I'm not sure if that is the correct term, but Rotax needs to have their gearbox serviced to check the "breakout force" to get the gear box to turn on it's own every 100 hours or so. Is there a need for that or any other servicing with the Gen3 as parts wear?

I greatly appreciate your responses here and to everyones questions. You are to be commended for your contributions to experimental aviation. Thank you for your dedication and hard work.
 
Last edited:
1) Pictures of the reduction drive unit laid out: I will take some pictures of the input section next time we assemble the units.

2) "Breakout Force": I believe they use a cam / Belleville disk spring system between the engine and drive unit. With this system I imagine you could turn the propeller, and hold the engine, until the shaft is passing over the internal cam lobe, then back down on the other side. Our system does not use this type of damper. To read about, "dual mass" flywheels, just Google it. That is what we use. As a preflight test, gently pull on a propeller blade and you will be able to pull it for an inch, prior to the engine turning.

Jan
 
<< I will take some pictures of the input section next time we assemble the units. >>

I'm sure everyone will find that helpful.

<<"dual mass" flywheels>>

Ahh, that kind of dual mass flywheel. "Dual mass"" had me thinking perhaps it was a suspended ring type. Not a damper of course. Can you share the inertia figures for the two wheels and the connecting spring rate?
 
Can you share the inertia figures for the two wheels and the connecting spring rate?

We don't share everything. Some of the many hours of testing and development is kept in hose.

Jan
 
<<We don't share everything. Some of the many hours of testing and development is kept in hose.>>

Sure, I understand. How about just the spring rate then? Anyone can measure that in less than 10 minutes, and 5 of them would be used removing the upper cowling. Need a fish scale, a digital level, and a screwdriver <g>.
 
How many test hours has the driver and flywheel system been exposed to and other what conditions?

I understand that the prior drives were tested for hundreds of hours using heavy metal props. What testing was done on the new G3?

I guess when you said it was a dual mass flywheel, I assumed it was the factory Subaru flywheel. Am I correct that it is actually made in house?
 
Leonard,
I don't think Jan is making any dual mass flywheels as they are readily available in the after market. If you do a search on them you will see several brands available. I think Jan has changed brands from the one in use on my STI engine.

I also am interested in learning about the testing process with the Gen 3 drives.

Randy C
 
Slow afternoon, so I spent a little time on the web looking for good information on dual mass flywheels. Most of what you'll pop up with a Google search is garbage. Here's the gold; a technical white paper from LuK:

http://159.51.238.62/remotemedien/m...ry/downloads/04_DMF_simulation_techniques.pdf

Excellent illustrations, correct and incorrect applications, design methods, and simulation. The exploded DMF illustrations are worth the trip even if you find the rest unfathomable.
 
Take two asprin, and call me if you understand it.

Dan, just read the info you linked.

My head hurts.
 
Dan / Jan, I need to understand where the dual flywheels are in this set up. I could only see one fly wheel on the engine and it is stock. Is the other one the prop?
 
Last edited:
Geico266 said:
Dan / Jan, I need to understand where the dual flywheels are in this set up. I could only see one fly wheel on the engine and it is stock. Is the other one the prop?
Click to expand...

Larry,

Take a look at the pdf file Dan linked the thread to. There is a diagram of a dmf on page 13.

I had the thing installed at the EGG hangar in April of 2006 when we upgraded GEN 1 to GEN 2. It's pouring down rain here right now or I'd wander down to the hangar and check the log, but I'm guessing it has operated about 80 hours. So far it is working fine. It is a bit weird in that when a prop blade is moved, the engine does not follow it immediately. The springs are in compression and at some compressed value, the engine turns. I should have the engine back together in a week or so as GEN 3 is due back tomorrow after a seal change and other improvements. It would be sooner but we are leaving town for 5 days. I will measure the break out force as best I can and report what it is.
 
DMF & Boost Control

Dan,
My head hurst just from looking at all the pictures of the Dual Mass Flywheels, but they really do make me curious about the design of the DMF on my STI engine.

I would think the spring rate would be important, but mine must be working OK. The reason I say this is even at a low idle speed I hear no rattle up front. In fact I hear very little rattle at shut down. At Jan's suggestion, I am keeping my idle speed high, at around 800 prop RPMs to take it easy on the gear box and the DMF, but I am unsure if this is needed. I find that 800 RPMs is just a bit too much for taxi on pavement and results in more brake dragging than I like.

I have about 15 hrs on my Gen 3 so far and it has been a good experience. The airplane feels stronger, (kind of hard to explain that but it does feel different, and better), and the new gear ratio does seem to help the STI get more in to it's power curve at lower and more efficient prop RPMs. In my case the ratio has had some domino effect in that I now need to get a larger supercharger driven pulley to slow it's speed down. The new gearbox drive ratio results in hitting supercharger redline (14,000 RPMs!) at about 2300 prop RPMs.

At first I thought that restricting my prop RPMs to 2300 would not be acceptable, and actually it isn't, but my experience has been that the airplane still performs very well. At 2300 prop, I am hitting 4,646 RPMs on the engine. This is above the peak torque speed of the engine so I can feel free to apply full boost at this engine speed. I will still proceed with changing the SC pulley to get those RPMs between 2300 on up to 2700 available to me but I can see that I won't be using this engine in that zone very often.

A note on my flying experience with the supercharger. There are many pros and cons between superchargers and turbo chargers and I have the general feeling that a good intallation of a turbo charger would have a few more advantages than the supercharger I have, but...

I have been finding it a real pleasure to use the supercharger, now that I have a seperate lever for controlling boost. I recently changed out a single push pull vernier throttle arrangement that would close the waste gate at the end of it's travel, to a two lever quadrant, one for throttle and one for the waste gate. I can cruise around at very economical fuel flows with the waste gate open and there is quite a range of power level available to me simply by pushing that lever foward.

Using the two lever system adds a little more pilot work load but I find having a seperate lever to operate the waste gate makes controlling the power output of the engine much more precise. With the waste gate open I see about 29" at lower altitudes and closing it brings the MAP right on up to a max. of 52" hg at my field elevation of 460'. 52" really makes her groan. My normal take off boost level has come down some (maybe around 40") with the new two lever control.

Anyway my point here is that I still think a turbo charger would be a better method of boosting this engine but my flying experience is telling me that what I have is really working pretty well.

Sorry to swerve off into another subject, perhaps better to have added a seperate post...

Randy C
 
DanH said:
Jan, an easy question to get the ball rolling. How do you prepare the base engine when it arrives at your shop, before adding accessory items and the gearbox?
Click to expand...

Hi

I thought I would respond to the first post in this thread. In early January I visited Jan's to say hello and visit my new engine. It was in a crate awaiting Jan's / Gary's / Sal's gentle touch.

In the space of a few minutes, the engine was uncrated and most, if not all, the acessories were removed and scrapped. It was painful to see a new alternator go in the scrap bin but, according to Jan, there wasn't a market for the *new* removed parts

Photos of the before / after can be found at:

http://www.flickr.com/photos/23527073@N03/

Cheers

Les Keaney
RV10 - "some assebly required"
 
Some questions

Jan,

If you have time to reply, I'm hoping you might be able to answer a few questions:

1. What was the thought process that led to a dual helical gear design for the gearbox rather than a planetary arrangement? The reason I ask is that a planetary gearbox has the nice property of containing all pressure angle forces within the planetary cluster rather than being reacted through the housing.

2. Likewise, what was the thought process that led to using single cut helical gears rather than dual opposed helical (i.e. herringbone) gears which could eliminate the thrust loading of the shaft support bearings?

3. What type of analysis or testing has been done to determine the resonance frequency (frequencies?) of the engine / fly wheel / gear reduction / propeller system?

4. Does the DMF have internal damping? If not, what's the advantage of using a DMF rather than using a single mass flywheel with a tuned torsional spring (rubber coupling?) and essentially using the propeller as the secondary mass? Perhaps the propeller inertia reflected through the gearbox wouldn't be sufficient?

5. Can the engine output enough torque to cause the DMF to "go solid" resulting in direct transfer of engine torque to propeller?

6. Were the springs in the DMF selected specifically for your application or is this a stock DMF that's used for automotive applications? The reason I ask is that since spring life is a function of both mean and alternating stresses, even if the max deflection is identical, there can be a huge difference in spring life depending on the case. What I mean to say is that a spring with a 1% mean load and a 20% alternating load will fail far sooner than one with a 20% mean load and a 1% alternating load. Thus, I'm wondering if the mean vs. alternating load expectations of this particular application were taken into account when designing the DMF?

7. What sort of FMEA was performed for the various components in the drivetrain and which component(s) have the greatest likelihood of failure?

8. How much, if any, FEA was performed of the various components in the drivetrain and where are the peak stresses expected? What are the expected peak stress values and types of stress (primary / principal / VonMises)? OT: As a curiosity, what software do you use (if applicable)?

I know your very busy, Jan, so I appreciate any time you're willing to spend answering my questions.

Best regards,

Mike
 
Randy,
<<looking at all the pictures of the Dual Mass Flywheels, but they really do make me curious about the design of the DMF on my STI engine.>>

One of the reasons I posted that particular .pdf link was that it illustrated three different variations of a DMF. There is a fourth common variation too, a DMF with a slip clutch to limit maximum transmitted torque. None of us know what type Jan has selected, but hopefully he will tell us.

<<I would think the spring rate would be important..... At Jan's suggestion, I am keeping my idle speed high, at around 800 prop RPMs to take it easy on the gear box and the DMF, but I am unsure if this is needed. >>

The torsional stiffness of the DMF is probably the lowest connecting stiffness value is the system. As such it would be a key factor in setting the system's lowest natural frequency, and by extension the RPM at which the system will resonate. I'm pretty sure it is below your 800 prop RPM idle speed, but I can't say exactly how much without a detailed analysis. A detailed torsional model for the MT/G3/6-cyl would include at least 11 stiffness values and 12 inertia values; nobody seems to have all of them. I can make a quick approximation of F1 by consolidation of the above into just a few stiffness and inertia values (some known, some guessed, some gathered), which is why I asked Jan for the DMF's spring rate. Without the above detailed stiffness and inertia values, concern for the proprietary nature of the DMF's spring rate is, well, rather moot.....and as mentioned, any Egg owner can measure it for himself.

Mike,
Excellent questions!

Jan,
Looks like you have a great opportunity here. Almost 2000 views on a thread less than 3 days old. Sure hope you can find time to cover the questions in detail; a lot of people are really interested in the engineering you put into your drive systems.
 
Jan and Dan too, thanks.

Jan, Thanks for your participation. Dan thanks for your links, while I'm a mechanical engineer most of my friends aren't and the postings allow me to easily show them what we are talking about. The picture worth a thousand words senario. Mike, I believe that if you watch the video link Dan provided it will show you that, yes there are outputs at which the dual-mass flywheel can be limited in travel. The design engineer has the responsibility of selecting a flywheel system with adaquate springs to handle torques equal to or greater than the output of the engine.
Thanks again Dan
Thank you Jan for participating in a reasoned discussion providing us with facts about the product!
Bill Jepson
 
Bill I think the question was not whether any DM flywheel could be designed in the way you describe. The question is whether the Egg DMF is single or dual stage and if it is designed to float or go soild during normal use. Collateral equations include within which range of prop weight are these calculations expected to hold true.
 
Jconard said:
Bill I think the question was not whether any DM flywheel could be designed in the way you describe. The question is whether the Egg DMF is single or dual stage and if it is designed to float or go soild during normal use. Collateral equations include within which range of prop weight are these calculations expected to hold true.
Click to expand...

Correct. I am curious about the construction of this particular DMF including spring sizing, over compression protection, natural frequency, mean and alternating stresses, etc.

I didn't take the time to read the full LuK document, but the concept is simple enough...insert a device into the powertrain with a low torsional spring coefficient to lower the natural frequency. However, as anyone who's done design knows...the devil is in the details. :)

This thread has actually helped me recollect a tour I had of LuK's facility in Wooster, OH some 10 years ago. One of their engineers was kind enough to give me the full $.25 tour including their work on clutches and flywheels. OT: The thing that made the biggest impression at the time was a stop-motion video of valve spring surge in an indy-car engine. That and the fact that the valve springs are designed to incrementally rotate the valves within their seats to even out the wear. Very cool.
 
<<The design engineer has the responsibility of selecting a flywheel system with adequate springs to handle torques equal to or greater than the output of the engine.>>

.....which would be peak instantaneous engine torque (not mean), or the peak resonant vibratory torque, whichever is greater.

<<However, as anyone who's done design knows...the devil is in the details.>>

Hear, hear! Bring on the details!

Guys, while we're waiting on Jan let me kick something around with you.

(1) Seems to me the fundamental idea behind the DMF is the placement of a soft spring at a larger radius from the rotational center, ie, it allows a softer spring rate while maintaining the same torque capacity. This a likely a good thing in the context of a PSRU.

(2) I'm thinking the "dual mass" part is a secondary benefit, and mostly a vehicle NVH thing. The driven flywheel adds inertia to the gearbox input gear, across the short, probably stiff splined shaft. That should lower that particular element's frequency, ie reduce the angular velocity of gear clatter at some vibratory orders. I don't see much benefit to the second flywheel mass in the context of reducing resonance at F1 and F2.

Thoughts?
 
Last edited:
How many test hours has the driver and flywheel system been exposed to and other what conditions?

The new drive system had to survive 200 hr full power on our test stand. We do not test for propeller loads, other than thrust, because this part of the drive has adequate bearing support. We are mainly interested in the interaction of the propeller / reduction drive / engine. We also ran two destructive tests to instant stoppage with no damage (other than the propellers) from 2700 RPM.

Jan
 
How many test hours has the driver and flywheel system been exposed to and other what conditions?

The new drive system had to survive 200 hr full power on our test stand. We do not test for propeller loads, other than thrust, because this part of the drive has adequate bearing support. We are mainly interested in the interaction of the propeller / reduction drive / engine. We also ran two destructive tests to instant stoppage with no damage (other than the propellers) from 2700 RPM.

I think many have now looked at the DMF information. as you can see, the benefit of the dual mass is less at high rpm and the additional mass is not desirable in an airplane. So we use only the aft portion of this setup. Someone asked about testing. We had MT do a vibration analysis with strange gauges to see how the #'s turned out. Results were very low compared to a direct drive engine.
 
1. What was the thought process that led to a dual helical gear design for the gearbox rather than a planetary arrangement? The reason I ask is that a planetary gearbox has the nice property of containing all pressure angle forces within the planetary cluster rather than being reacted through the housing.

Satellite gears have to spin very fast. We do not have any bearings spinning faster than the engine.

2. Likewise, what was the thought process that led to using single cut helical gears rather than dual opposed helical (i.e. herringbone) gears which could eliminate the thrust loading of the shaft support bearings?

Easier manufacturing and the ball bearings at each end have no trouble with this force.

3. What type of analysis or testing has been done to determine the resonance frequency (frequencies?) of the engine / fly wheel / gear reduction / propeller system?

Strain gauge testing / accelerometer tests.

4. Does the DMF have internal damping? If not, what's the advantage of using a DMF rather than using a single mass flywheel with a tuned torsional spring (rubber coupling?) and essentially using the propeller as the secondary mass? Perhaps the propeller inertia reflected through the gearbox wouldn't be sufficient?

We use what you suggest above.

5. Can the engine output enough torque to cause the DMF to "go solid" resulting in direct transfer of engine torque to propeller?

Only if the propeller hit something.

6. Were the springs in the DMF selected specifically for your application or is this a stock DMF that's used for automotive applications? The reason I ask is that since spring life is a function of both mean and alternating stresses, even if the max deflection is identical, there can be a huge difference in spring life depending on the case. What I mean to say is that a spring with a 1% mean load and a 20% alternating load will fail far sooner than one with a 20% mean load and a 1% alternating load. Thus, I'm wondering if the mean vs. alternating load expectations of this particular application were taken into account when designing the DMF?

We tested the unit before and after the 200 hr full power test. Propeller deflection was identical before and after the test. The ultimate spring life remain to be seen. The springs are arranged in such a way that they can not fail such that the engine would disengage from the propeller. We also have 3 different spring rates in the unit.

7. What sort of FMEA was performed for the various components in the drivetrain and which component(s) have the greatest likelihood of failure?

Destructive test show the propeller to give first. Prolonged running showed slight wear on the inner race of the flywheel pilot bearing.

8. How much, if any, FEA was performed of the various components in the drivetrain and where are the peak stresses expected? What are the expected peak stress values and types of stress (primary / principal / VonMises)? OT: As a curiosity, what software do you use (if applicable)?

I do not use any stress analyses software. We test for this on completed assemblies only. We use personal judgement and automotive converted parts. Hence the term "Auto conversion engines :) Our philosophy has always been: Small displacement engines, heavy flywheels, high rpm, sturdy drive unit and light propellers. The combination has worked for many years. also, keep in mind the spring supported flywheel is not new. It has been on engines since 2003.

I know your very busy, Jan, so I appreciate any time you're willing to spend answering my questions.

No problem
 
A little more on your question about the dual mass. Yes, we only use the spring portion of the system, not the secondary mass. (other than the weight of our drive disk and splined shaft / reduction drive and propeller) It is defiantly the aerodynamic loading of the propeller that is loading against the springs.

Jan
 
<<Yes, we only use the spring portion of the system, not the secondary mass.>>

Ahhh, that answers a lot.

What Jan is saying is this particular DMF has a removable friction face. In the OEM application it is the contact surface and mass heat sink for the manual transmission's clutch as well as being the secondary flywheel mass. Jan unbolts it, throws it in the scrap heap, and bolts on his splined drive adapter.

Of course it is no longer a dual mass flywheel, which doesn't matter (other than creating name confusion). It does package a soft stiffness element into a space no greater than a standard flywheel. Very clever in that regard. Good thinking Jan!
 
DanH;196489Of course it is no longer a dual mass flywheel said:
This why I was confused. I did not see anything dual about the flywheel I was looking at.

Dan, what is your opinion about no testing of side loads, only thrust? Seems to me a spining prop exerts a "ton" of side force on the gearbox.
Click to expand...
 
Last edited:
Dan, what is your opinion about no testing of side loads, only thrust? Seems to me a spining prop exerts a "ton" of side force on the gearbox.

Let me clarify. The bearing support in the G3 is 3x that of the G1-G3 with thousands of accumulated combined hours. G3 drives have a heavier output shaft, much larger spline connection and more bearings. The only item to remember regarding the front bearings and housing is to supply adequate cooling for the best possible bearing retention in the housing. The cooler the drive can operate, the more true the internals of the unit is operating. A 2 - 2.5" blast tube, right on the nose section does a lot of good.

Jan
 
I am wondering where the heat comes from. Is the drive simply absorbing heat from the running engine, or is the heat from drive operation. It would be interesting to calculate how much thermal energy is generated. Jan...is the drive just absorbing heat from the engine throught the big plater, or is the heat from the internals?

Second, am I reading correctly that operating temperature is mainly a concern for bearing retention...ie the bearing is a press fit, and because its an aluminum housing there is some concern or necessity to manage temperatures for that reason? What is the optimal temperature of the drive?
 
janeggenfellner said:
1. What was the thought process that led to a dual helical gear design for the gearbox rather than a planetary arrangement? The reason I ask is that a planetary gearbox has the nice property of containing all pressure angle forces within the planetary cluster rather than being reacted through the housing.

Satellite gears have to spin very fast. We do not have any bearings spinning faster than the engine.
Click to expand...

Hard to understand why this would be a problem-it's not a problem on turboprop engines that are geared down from around 35,000 RPM via planetary gearsets. Having lots of gear teeth in contact at all times would seem to be a more reliable way of transfering torque, while reducing speed and all in a very compact package. Am I not considering something important here?
 
Note from Moderator

I want to say how proud I am of all the readers and posters for their restraint. As this social message board experiment continues (moderated thread to stay on topic with factual discussions of engineering issues), Im impressed with the way it has progressed WITH OUT me having to do anything.

While I care very little about the topic, Im enjoying the reading just the same as Im learning stuff and not getting annoyed with opinion, retoric, and hyperbole.

Keep up the good work team.
And please do NOT comment on this message. Feel free to start a thread on the topic of highly moderated discussions if you like.

Best,
 
Planetary Gears

The rotational speed of the planetary gears is a potential problem and has caused some design changes in other vendor's transmisions. I believe I read about them needing a good clean supply of pressurized oil. They can obviously be made to work but there are always trade offs.

Regarding temperatures of the G3, mine has been running quite a bit cooler than the previous Gen 2 drives, 30 to 40 degrees F cooler. In the past it was thought that the reason the coolant, oil, and gearbox temps stayed so close together was conduction from the engine but I think that theory has weakened with the Gen 3 data. My Gen 3 generally runs 20 to 30 degrees F cooler than coolant and oil temps now. This is probably due to the quite large spinner gap I am have at present. As I close it up I will have to direct some cooling air to replace it.

Now I know that it is probably best to direct that cooling air right up near the front bearing support area.

Randy C
 
<<Dan, what is your opinion about no testing of side loads, only thrust? Seems to me a spining prop exerts a "ton" of side force on the gearbox.>>

Well, for starters you should go back and look at post #7 in this thread; I just edited it because I made an error in calculation. The prop moment isn't a ton in this case, but it is about 600 lbs.

Even so, actual measurement of side load issues would be low on my priority list. It might be interesting to check if you think (1) gearbox case fixation in relation to the engine case is weak, or (2) you're concerned with fixation of the entire engine assembly, ie, motor mount and mount plate issues. Those would be fairly easy to measure.

The real issues relating to prop moment are reliably calculated (if you check your figures at least twice!). Prop moment would be an input when calculating an L10 life for the front bearings, and even more important, a material stress value for the propshaft.
 
Last edited:
Spinner gap

Randy said:
The rotational speed of the planetary gears is a potential problem and has caused some design changes in other vendor's transmisions. I believe I read about them needing a good clean supply of pressurized oil. They can obviously be made to work but there are always trade offs.

Regarding temperatures of the G3, mine has been running quite a bit cooler than the previous Gen 2 drives, 30 to 40 degrees F cooler. In the past it was thought that the reason the coolant, oil, and gearbox temps stayed so close together was conduction from the engine but I think that theory has weakened with the Gen 3 data. My Gen 3 generally runs 20 to 30 degrees F cooler than coolant and oil temps now. This is probably due to the quite large spinner gap I am have at present. As I close it up I will have to direct some cooling air to replace it.

Now I know that it is probably best to direct that cooling air right up near the front bearing support area.

Randy C
Click to expand...

We've used a .400 inch spinner gap from the start with our Marcotte drive and gearbox oil always runs below coolant temp. CFD, actual pressure measurement and tuft testing in flight confirms that the gap area is a source of high pressure and is ideally located to bathe the gearbox case in cooling air. Cowling removal with 3 blade props is also much easier. My OAT probe is located just above the bearing casing and reads true when compared to externally mounted probes.

Speculation by some that air is flowing out the gap are not supported by my measurements. Spinner gap is probably the best/ easiest way to cool a PSRU in an RV.
 
Last edited:
Guys, I've been reflecting a bit on what Kahuna said. He is right, we're doing great and I thank you all.

Let's keep a tight technical focus on the system of interest, Eggenfellner.

I've deleted a reference to Lycoming from my previous post. I'd like to suggest we also take care in introducing discussions of alternate design choices. I'm not suggesting we ban such discussions. I am suggesting there's a lot to cover in the design we have in front of us.
 
While we're thinking good community thoughts......

Jan,
You've been both cautious and a brief. I understand both. You've not always been treated well, and you're busy.

Please understand what we have here. Kahuna said it well; a whole lot of us have been annoyed with opinion, rhetoric, and hyperbole regarding your work. We want facts. Everything else is just....noise.

I'd like to ask you to be a bit more open with details and data. Take a few extra minutes to explain things. Give us some numbers. Sure, some things should be proprietary, but less than you might think. Nobody is asking for part numbers. We want operating principles, descriptions, design assumptions, and test data....not the keys to your business.

Can you make it a little easier for us? Please?
 
Planetary gearheads

captainron said:
Hard to understand why this would be a problem-it's not a problem on turboprop engines that are geared down from around 35,000 RPM via planetary gearsets. Having lots of gear teeth in contact at all times would seem to be a more reliable way of transfering torque, while reducing speed and all in a very compact package. Am I not considering something important here?
Click to expand...

Please consider:
  • All (or nearly all) automotive automatic transmissions (not including CVTs) use planetary gears for the gear reduction.
  • The Egg is a Subaru auto conversion engine.
  • Subarus come with automatic transmissions.

Therefore, it's well proven already that planetary gear reductions will work well and very reliably at the speeds the engine will be turning.
 
Therefore, it's well proven already that planetary gear reductions will work well and very reliably at the speeds the engine will be turning.

Not so. It has been tried in this application before. Turbine engines have none of the problems we work with. They operate silky smooth. Oiling is critical for these designs and anything delicate, that has the internal bearings, or needle / roller bearings should be avoided. They will not last, unless the oiling system is very carefully designed and pressure is applied. I never liked the idea of running engine oil through the drive unit. Very hard to keep the drive 15 F below engine temperature when hot and thin engine oil is sprayed into it. This is a place where high quality, large and completely open bearings have a place. The drive has it's own oils supply of adequate viscosity and no chance of contaminating the engine by returning oil to it. No pumps and no coolers. The design we use has enormous gears, designed for the load and shock of the engine. If you look at the gears in the C-175 Continental GO-300 engine, they do not resemble those in a PT-6 Turbine. They have the disadvantage of operating in engine oil, compared to our drive.
 
Good thinking

janeggenfellner said:
Not so. It has been tried in this application before. Turbine engines have none of the problems we work with. They operate silky smooth. Oiling is critical for these designs and anything delicate, that has the internal bearings, or needle / roller bearings should be avoided. They will not last, unless the oiling system is very carefully designed and pressure is applied. I never liked the idea of running engine oil through the drive unit. Very hard to keep the drive 15 F below engine temperature when hot and thin engine oil is sprayed into it. This is a place where high quality, large and completely open bearings have a place. The drive has it's own oils supply of adequate viscosity and no chance of contaminating the engine by returning oil to it. No pumps and no coolers. The design we use has enormous gears, designed for the load and shock of the engine. If you look at the gears in the C-175 Continental GO-300 engine, they do not resemble those in a PT-6 Turbine. They have the disadvantage of operating in engine oil, compared to our drive.
Click to expand...

Although we disagree on the pros / cons of planetary gearsets, Jan makes an excellent point about not sharing gear oil and engine oil. Years ago a British auto manufacturer (not sure which...maybe MG or Mini or Triumph...one of those three) had a design in which the engine oil and gear oil were shared. After engines started breaking down, they learned that the gearbox was shedding small bits of metal which were being carried into the engine and galling the fluid bearing surfaces. The solution was to put a strong magnet in the oil path between the gearbox and the engine.

Another issue:
As I was working on the fuselage tonight, I was contemplating the DMF concept. As previously discussed, this should lower the resonant frequency of the system. However, the springs should also work as a low pass filter. I'm not very familiar with the MT propellers, but believe they're wood wrapped in composite. Such a design would be expected to have a fairly high resonant frequency. The springs in the DMF could act to filter out any engine-produced frequencies that would tend to excite the prop. Whether that benefit is required or achieved, I have no idea. Perhaps Jan can comment on whether or not that was a design consideration.

DanH said:
While we're thinking good community thoughts......

<snip>

I'd like to ask you to be a bit more open with details and data. Take a few extra minutes to explain things. Give us some numbers. Sure, some things should be proprietary, but less than you might think. Nobody is asking for part numbers. We want operating principles, descriptions, design assumptions, and test data....not the keys to your business.

Can you make it a little easier for us? Please?
Click to expand...

Excellent point, Dan. It's the reality of mechanical systems...reverse engineering for anyone so motivated is exceedingly simple. Even part numbers are laser etched onto the bearings and other components.

There's a software development methodology known as "open source" software. The concept is that everyone can evaluate and contribute to the design of particular piece of software. Although it may seem counterintuitive, this design methodology leads to some of the most secure software available. Take the Apache web server, for example. It powers ~50% of the world's websites ***, but worms / viruses for it are virtually unheard of.

It would be wonderful if engineering was done the same way. I wonder how many of the Engineering Disasters could have been avoided if the calculations had similar transparency?

Although I'm a proponent of an analytical engineering approach in cases where it's warranted, there have been many wonderful designs created without all the calculations and computers. The primary benefit of the analytical approach is a reduction in the number of iterations required to get to a robust design.

I second the request to share as much data and informaton as you feel you can, Jan. Ideally we will all learn more from the process (including Jan). :)

*** Source: Netcraft
 
Looking around at other available gearboxes, it seems that most of them are using some type of rubber elastomer type of isolation / dampening rather than the DMF.

I have not had the opportunity to compare the stiffness of the rubber type dampeners or isolators to the DMFs, but I would think they would be considerably stiffer than the DMFs spring rates.

Perhaps it really isn't that important how stiff the isolation is as long as it provides some "give" to protect the gears and other components from the hard hits?

I suppose if the spring rate was too soft they could become completly collasped with hard pulses and bottom out, becoming solid, and possilby deliver those hard hits.

This might be a good reason to monitor how much the spring rate might change or weaken over time from use and abuse to come up with a service interval for DMFs in this application?

Randy C
 
"Speculation by some that air is flowing out the gap are not supported by my measurements. Spinner gap is probably the best/ easiest way to cool a PSRU in an RV."

Ross, I am learning to listen to you about airflow after my experience with the NACA inlet turning into an outlet. I have also learned that air doesn't often do what we think it should and only measureing in some way will tell the story. I will find a way to measure which way the air is flowing in my large spiner gap before I start closing it up. Perhaps it doesn't matter as long as the PSRU stays cool enough and the spinner gap does not create an excessive drag problem.

Randy C
 
Perhaps it doesn't matter as long as the PSRU stays cool enough and the spinner gap does not create an excessive drag problem.

Randy C


Several builders upgrading to the G3 drive, report that some spinner gap is a great way to keep the drive unit cool. Also, it has not degraded the engine cooling, at least not measurably so. Everyone building wants the gap to be like a business card. Then they quickly start to hate it when the cowling removal becomes difficult and the paint is all scratched. A 0.5 gap to start is a much better approach. Also, who said that no gap is prettier? Lexus use gaps, just even :)

Jan
 
<<...I was contemplating the DMF concept.>>

For the record (and I know you already realize this Mike), as applied here it is no longer a dual mass flywheel. The secondary mass has been removed. That leaves the primary inertia and a torsional soft element (the spring set). Placing a soft torsional stiffness at the engine output inertia is quite conventional. However, applying the guts of an automotive DMF to a PSRU is something I've not seen before, and has a number of interesting advantages, some mechanical, some vibrational. I suspect Jan had one of those wonderful "Ahhhh!" moments a few years back. Satisfying, eh Jan? <g>

<<As previously discussed, this should lower the resonant frequency of the system.>>

Actually "frequencies", plural. Theory tells us the number of natural frequencies in a torsional system is equal to the number of inertias minus one; the soft spring will move all the natural frequencies down the RPM scale. It is likely that we're only concerned with the lowest two in the context of system resonance, and a few higher ones in the context of propeller resonance.

The stiffness-inertia model would look something like this:


J's are inertia values, K's are stiffness values. The largest inertia is usually the propeller; in this case we know it to be 0.36889 slug-ft^2 if the user mounted Jan's recommended MT. The flywheel inertia is usually the next largest individual inertia, but in the case of a PSRU it will be about an order of magnitude less than the prop, in the area of 0.03. Yes, this is a generalization and should be considered as such. We don't know the actual inertia value for the Egg flywheel. K3 is the spring rate of the (non-)DMF.

The other inertia values (J) are the gearsets, each individual crankthrow, and J11, the harmonic balancer (another grossly misleading name) combined with the accessory inertias. The stiffness values (k) are whatever connects them together. This model treats the propeller as a single lump inertia, which is not entirely accurate.

The engineers here know all this stuff, but perhaps the above will help others as we proceed. Quick reminder; when any of the natural frequencies are matched by an exciting frequency, the system (or some part of the system) resonates.

<<However, the springs should also work as a low pass filter.....The springs in the DMF could act to filter out any engine-produced frequencies that would tend to excite the prop.>>

I think you're right Mike. I suspect a soft spring rate would indeed filter out an awful lot of the high exciting frequencies that would match the high natural frequencies of the prop blades. It would nicely explain the low values rumored (hint, hint, Jan<g>) to have been recorded during the MT prop telemetry.

The big question regarding that prop telemetry is the strain values recorded near the root of the blades when the system resonates at the first and (perhaps) second natural frequency. F1 is probably below Jan's specified idle speed. Nobody knows where F2 lies, at least not right now. Yes, I hope Jan knows. Devil in the details, eh Mike?

<<I'm not very familiar with the MT propellers, but believe they're wood wrapped in composite. Such a design would be expected to have a fairly high resonant frequency.>>

May not be all that different compared to aluminum; not a huge difference in modulus. Here are some sample curves:



(post lunch edit: Whoops, brain f*rt! Forgot about the mass difference between wood/glass and aluminum; stiffness ain't the whole picture.)

I am pretty sure a wood/glass blade has a higher vibration decay loss factor, ie it would be more "dead" compared to aluminum.

<<I have not had the opportunity to compare the stiffness of the rubber type dampeners or isolators to the DMFs, but I would think they would be considerably stiffer than the DMFs spring rates.>>

A lot of the rubber couplers for diesel applications are quite stiff, but consider the Goetz coupler in a Rotax C or E box, or some in the Centaflex line from Lovejoy.

Great stuff guys, and thanks Jan.
 
Last edited:
Status
Not open for further replies.
Back
Top