What's new
Van's Air Force

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

Viking 110 notice

Ross , you are so CORRECT ,my subaru 2.5 idled at 750 rpm, with NO SHAKING,
and my 1.8 Honda civic in my Zenith 701 does likewise. A friend with rotex is
constantly fussing with carbs even with 1400 rpm idle. Eventually someone will address this with not only gearbox, but good flywheel/coupling system. I don’t want to get in trouble listing business that unfortunately doesn’t advertise here,
but they supply 3 different gearboxes, depending on engine size and HP, as a
result have been very successful. Tomcatrv4
 
Dan, potentially they are designed to a price point and reliable "enough". How often to they require the gearbox to be overhauled ? Things like that can be quite profitable and keep A&Ps in work too... Rotax doing spur gears vs helical. Cheaper to make, stronger but noisier. Potentially allowing for bushings vs roller bearings.

I dunno gents...does anyone in the alt engine world really have standing to knock a 912? It has been a rather successful design.

Propeller inertia is the relatively immovable object against which the rest of the system oscillates. Given the same shaft load, low engine mass means the block oscillates more than a Subaru with a 50 lb gearbox.

The metal-to-metal rattle in the resonant range is the ramp-type dog clutch doing exactly what it was designed to do. There are several styles, and there is also an available friction clutch to further limit maximum shaft torque in some versions.

Photo here, two different dog clutch hubs. Driven side stacked on driving side, which is integral with the spur gear. Left is no-clutch, right is driven hub with external spline for overload clutch plates:

http://contrails.free.fr/images/rotax/DSCN5156.sized.jpg
 
Last edited:
I bring it up only to illustrate that even with all of their engineering and funds available, Rotax didn't solve the TV issue down low, they do what all of us are doing- avoid idling in the resonant range with the recommendation to be over 1400 rpm.

Remember, design is the art of considered compromise. Rotax incorporated soft elements on the C and E boxes, but ultimately elected to go with the ramped dog/overload clutch system for the 912. It is, for practical purposes, a torsionally stiff system until excited at its fundamental natural frequency, when it loads the dog springs and slips the clutch. The result is noisy in resonance (the dogs) and results in a higher frequency. On the flip side, it probably pushes the F2 above the operating range. I only say "probably" because I have not instrumented one, but it is what the theory suggests. And doesn't require a flywheel, which brings us to...

The 912 I ran on my test stand was much worse than my Subaru ever was below this rpm.

I am reminded of the old joke which goes "Doctor, it hurts when I do this", to which the doc says "Well, don't do that".

Rotax clearly asks operators to stay above the resonant RPM. With 2.43 or 2.27 ratios, 1400 at the crank is 600 at the prop.

Hey, I'm not here to play fanboy for Rotax. I'm just sayin' the design has reasons, and they may not be real obvious.
 
Last edited:
I suspect the Rotax case was pretty challenging without flywheel inertia. The gearing helps to keep residual thrust low but some folks still wish it could be less. I have some time in a CT and it was a real floater in ground effect due to the high residual thrust.

Yes, a compromise and not ideal. I just expect something better from a mass produced aero engine.
 
I downloaded Freecad last night and was able to get a few FEA iterations run in 3-4 hours, learning how to make a model and then go through the steps going from model to mesh, applying loads etc. Im still learning how to drive it, but it is a heluva tool for the price ($0)...

I dont have the tangential nor radial loads applied at exactly the right angle yet, I probably have to add some wireframe elements to align the loads to, but the location of the stress concentrations and the magnitude is starting to look very close to the failed parts...

Peak Von Mises stress is at 417MPa or 60.5ksi Certainly high enough to cause fatigue without the benefit of TV problems, misalignment etc yet Im sure the Viking 110 has some of each of those, just like any other engine. I also think that at these levels one needs really good surface finishes and some of the flanges are clearly horrible in this regard.

I need to finesse the design some more, figure out the wireframe elements and correct the load angles and then look at some larger blend radii, thicker web and the original design for bolts to see how they compare.
 

Attachments

  • Freecad-01.png
    Freecad-01.png
    187.2 KB · Views: 64
  • Freecad-02.png
    Freecad-02.png
    23.5 KB · Views: 42
  • Frecad-03.png
    Frecad-03.png
    278.2 KB · Views: 48
  • Freecad-04.png
    Freecad-04.png
    205.1 KB · Views: 62
The quote finally came in for the Centaflex coupling. Will order it tomorrow.
 

Attachments

  • Centaflex quote.JPG
    Centaflex quote.JPG
    33.9 KB · Views: 112
Great start, Keith on the FEA. Would take me days to get that far from a cold start.
I wish there was a free engine dynamics program as well. I have used an electrical circuit simulator LTspice to do it, but the results are hard to interpret. It involves converting mass, elasticity, friction into capacitance, inductance and resistance. The resulting circuitry is fairly simple, but as I say, hard to interpret and no real visual output.

As an aside, I flew behind a geared engine decades ago for many hundreds of hours, a Lycoming GO-145, as smooth as an electric motor. It had problems, but nothing to do with the gearing. Loved it. For the TV solution, it had four enormous pendulous dampers on the crank. The crank gear ran inside the prop gear at a ratio of 17 to 27. No shutdown shudder, no soft components, but a lot of clunk sounds starting and stopping.:)

For an auto conversion using tuned dampers on a four cylinder, the 2nd order frequency is the big one, and therefor heavy damper weights. And since they would be busy all the time at all RPM's, would need to be well lubed. So, maybe they could be inside a hollow, lube containing flywheel. All 2nd order vibration to the gearbox could be cancelled. The idea could be carried to the 4th order at which point other frequencies would require very little mechanical consideration. Basically just DC twist going to the prop.

As I typed this out, I have begun to think this might actually be very feasible. hmm

If I where younger and had the skills and hands-on ability of Dan, I'd be on my way to the junk yard to get a good motor.:)

A bit OT, I know, but maybe a worthy idea. It seems Keith now has the bull by the horns solving the serious problem addressed in this thread..

Ron
 
Last edited:
I dunno gents...does anyone in the alt engine world really have standing to knock a 912? It has been a rather successful design.

No knocking on the 912 intended from my side either Dan. I like that engine. I learned to fly behind one. They are good, solid engines and have done the LSA/ VLA/ EU Microlight aviation a tremendous service, basically single-handedly enabling these categories. If they'd do a 200 hp version (6-cylinder 915i?), then Lycoming might be in for some hard times.

Like Ross, my remark was only to confirm that they too have to content with a significant rattle at these lower RPMs.
 
I did an analysis of the original bolted design too. Here is how that came out.

Von Mises stress at 454Mpa, not surprising given the 1/3 reduction in cross section caused by the counterbore. Highest stress concentration at the OD of the giubo bushing. Unlike the welded pin design, a much lower stress concentration tangent to the central boss. I think the reason why the sample lasted as long as it did was because of a very good surface finish, the best of all the samples in my possession.
 

Attachments

  • Bolted giubo flange 5800 bmep-max.JPG
    Bolted giubo flange 5800 bmep-max.JPG
    99.3 KB · Views: 47
  • Bolted giubo flange 5800 bmep-max02.JPG
    Bolted giubo flange 5800 bmep-max02.JPG
    145.3 KB · Views: 62
  • Bolted giubo flange 5800 bmep-max03.JPG
    Bolted giubo flange 5800 bmep-max03.JPG
    131.9 KB · Views: 53
What the actual part looked like
 

Attachments

  • IMG_0400.jpg
    IMG_0400.jpg
    54.6 KB · Views: 54
  • IMG_0408.jpg
    IMG_0408.jpg
    409.9 KB · Views: 61
This is what the new flexible coupling system is looking like that is being proposed.

First the existing flex plate, flywheel side drive coupling, giubo and gearbox input flange all get removed.

Then a new CrMo flywheel is mounted to the crankshaft. The flywheel will help damp the vibration pulsations from the engine thus making a reduction at the source. Onto the flywheel gets bolted the Centaflex coupling element with 3x M12 bolts. On the gearbox side the splines get cleaned up and coated with high pressure copper slip grease (after pressing off the drive flange and putting the gearbox back together with potentially new bearings and seals)

A new centering bushing is installed in the crank. The far end of the gearbox input shaft, if it has a square edge like some do, it will get a chamfer ground into it to remove the sharp edge and make installation easier. The ID of the bushing likewise gets some high pressure grease because there will always be some relative motion as the flexible coupling does its job.

The new tubular cross section drive coupling is slid onto the splines of the gearbox input shaft. It has 3x M12 tapped holes that align with the radial fastening locations on the new coupling. The gearbox is placed with the input shaft engaged with the alignment bushing and the gearbox mount bolts are tightened. Now the drive coupling can be slid in the axial direction and the prop shaft rotated till the holes align allowing 1 fastener at a time to be installed hand tight before they are all tightened. Blue loctite will be used on all the fasteners.

Special thanks to DanH for his experienced advice on this application and Mark Kettering for playing devils advocate...
 

Attachments

  • Centaflex coupling config.JPG
    Centaflex coupling config.JPG
    62.3 KB · Views: 64
Last edited:
This is what the new flexible coupling system is looking like that is being proposed.

...Onto the flywheel gets bolted the Centaflex coupling element with 3x M12 bolts. On the gearbox side the splines get cleaned up and coated with high pressure copper slip grease ...The new tubular cross section drive coupling is slid onto the splines of the gearbox input shaft. It has 3x M12 tapped holes that align with the radial fastening locations on the new coupling. The gearbox is placed with the input shaft engaged with the alignment bushing and the gearbox mount bolts are tightened. Now the drive coupling can be slid in the axial direction and the prop shaft rotated till the holes align allowing 1 fastener at a time to be installed hand tight before they are all tightened. Blue loctite will be used on all the fasteners..

Keith, sorry, but that assembly process will need some modification.

The radial bolts, installed last in your description above, are intended to compress the entire Centaflex radially when tightened. The goal is to place all the connecting rubber elements in compression, a key feature of the element. Prior to compression, the diameter of the axial bolt circle is larger than stated in the spec sheet. Without radial compression, it will be difficult to bolt the element to the flywheel...the holes won't line up.

I previously suggested using the "S" (think "slide") mounting pin configuration, with the pins on a nice thick section of the flywheel center. The pins are actually sleeves with the same 12mm bolts. This configuration would mount the splined hub and coupler on the gearbox input shaft, and then the gearbox assembly would then be slid onto the flywheel pins. Note it allows the use of a clamped spline, and the pins accommodate any minor axial inaccuracy due to fabrication or temperature change.

If you really want to bolt to the flywheel as described, there is a work around. You'll need to install and tighten the radial bolts into the splined hub on the bench, then bolt to the flywheel, then either (a) install the gearbox by inserting the input shaft into the splined hub, or (b) remove the hub, install it on the input shaft, install the gearbox, then re-install the radial bolts.

I should note a potential drawback to using this coupler; there is no convenient way to safety wire the 12mm metric allen cap screws. Be sure to use a torque wrench and a good threadlocker compound. I never had any sign of a problem.
 
Last edited:
Dan, thanks for the feedback from practical experience of using the coupler. I didnt want to make the splined adapter a press fit and so far I doubt that a standard spline profile is being used. The original bolted input flange was a sliding fit on the splines and going that way with the design means that the gearbox doesnt have to be disassembled to press on the new coupling after removing the old one.

So I would rather constrain the position of the flexible element and allow the opposite end to float. If one needs an assembly aid then one could certainly use the tubular coupler to get the part set up on the flywheel.

I'm attaching a picture of the Aeromomentum flywheel which is close to the sort of design I think would work best, please ignore the 3 giubo attachment holes which fall in a horrible position.... The aeromomentum piece is sold for $200 I believe with the ring gear being separate. I do like the fact that a lot of the racing 4140 flywheels have the starter gear teeth cut directly into the billet before heat treatment. It would obviously have to cost a little more.
 

Attachments

  • flywheel.jpg
    flywheel.jpg
    947.9 KB · Views: 70
I didnt want to make the splined adapter a press fit and so far I doubt that a standard spline profile is being used. The original bolted input flange was a sliding fit on the spline...

So why the change?

I'm attaching a picture of the Aeromomentum flywheel which is close to the sort of design I think would work best...

Good concentration of mass at the outer rim for maximum inertia with least weight. The big holes would make me nervous on a non-balance-shaft 3-cyl due to the vibratory block wobble, but may be fine on a 4-cyl. If you like 'em, plug it into your new FEA tool, and apply pitch or yaw at about 1 radian per second.

I do like the fact that a lot of the racing 4140 flywheels have the starter gear teeth cut directly into the billet before heat treatment. It would obviously have to cost a little more.

Is that done for high RPM use? I'd use the money elsewhere. It's easy to shrink a ring gear on a custom flywheel. All you need is an oven, a freezer, and a wife who isn't home ;)

From the way back machine (1999), custom flywheel cut with a flat face to accommodate a viscous disk damper. Stock Suzuki ring gear.

The disk operated in parallel with the Centaflex, oscillating between the flywheel face and the inside of the blue anodized shell, clearance being about 0.010" on each side. Here it's being pumped full of liquid silicone. Compare vibratory torque plot with the undamped runs posted previously.
.
 

Attachments

  • Viscous Damper.JPG
    Viscous Damper.JPG
    82.4 KB · Views: 119
  • Fig 3 Damped.jpg
    Fig 3 Damped.jpg
    30.1 KB · Views: 64
I didnt make the decision to change the fit of the input flange on the shaft, Jan made that back when the service bulletin was issued for the gearbox or even prior to that, because a lot of gearboxes that came back to get the drive flange changed for one with the welded pins already had a pressed on input flange.

Then he released a 2 part video of how to change the drive flange, yet if you actually touched the gearbox your warranty was voided and you were flagged as being non compliant. Apparently the Eric Miller engine passed through 2 sets of hands prior to even getting to him and apparently one of the 2 previous owners changed the gearbox flange or had his A&P do it. That fact alone allowed Viking to say that his engine was not in compliance with the SB, thus nothing that happened with it was the fault of Viking.

So why the change?

.
 
See the 2 viking flywheels. I know that the aluminum one with all the pockets in it cracked and the next one is a modified flex plate. Not sure why the center was plasma cut on the shown sample, maybe to prevent it being used again ?? The aeromomentum one looks pretty bomb proof compared to either of these.
 

Attachments

  • Viking Honda 110 AL6061-T6 Drive Plate Weight 2lb 13oz.jpg
    Viking Honda 110 AL6061-T6 Drive Plate Weight 2lb 13oz.jpg
    830.5 KB · Views: 95
  • Viking Honda 110 Steel Drive Plate Weight 3lb 6oz.jpg
    Viking Honda 110 Steel Drive Plate Weight 3lb 6oz.jpg
    815.6 KB · Views: 98
Last edited:
If there is someone reading this forum who owns a Viking 110 powered airplane who would be willing to fly a prototype new coupling system next spring/summer, please contact me. The intention would be to supply the parts for free and provide on site help with the install. In addition, I would like to inspect the customers existing drive flange after which all original material will be returned to the owner.

We would need access to an arbor press in the proximity of the owners hangar to remove the pressed on drive flange and inspect and re-assemble the gearbox. Obviously it would be most beneficial to the community if the person volunteering was someone who flew regularly and put some hours on his airplane.

If this interests you, please send me a PM in this regard. If there are several volunteers we could consider making a small batch of parts from what would potentially be the production vendor, this would be even more interesting than running a prototype assembly.
 
Last week I had a conversation with Jan in which I tried to convince him that the giubo is a very strong composite component packed full of synthetic fiber and it probably wouldnt fail at much less than 10 000lb. Therefore was more than strong enough to fatigue steel parts like the drive flange. He kept referring to them as "rubber bands".

So he did a little experiment with a single segment and a pair of steel cables and his 2500 series diesel truck and a stout tree and even after breaking the steel cables, bending the 12mm bolts he was still commenting on what a nice flexible rubber part it was... Comical.
https://youtu.be/dgBHff4pVNo
 
I didn't make the decision to change the fit of the input flange on the shaft, Jan made that back when the service bulletin was issued for the gearbox or even prior to that...

That's what I was asking...why did Jan make the change? I think you're saying he did not share his reasons. Given switching from a slip-on to a pressed-on spider added significant effort to the fabrication and assembly process, it seems like a flag.

The aeromomentum one looks pretty bomb proof compared to either of these.

Might be, but you have an analytical tool. No doubt Jan thought the aluminum flywheel and welded spider looked pretty bomb proof.
 
That's what I was asking...why did Jan make the change? I think you're saying he did not share his reasons. Given switching from a slip-on to a pressed-on spider added significant effort to the fabrication and assembly process, it seems like a flag.

Might be, but you have an analytical tool. No doubt Jan thought the aluminum flywheel and welded spider looked pretty bomb proof.

Ah, but Jan couldn't pick out an SN curve in a suspect lineup, nor would he have any idea what such a curve would look like for aluminum vs CrMo steel....
 
That's what I was asking...why did Jan make the change? I think you're saying he did not share his reasons. Given switching from a slip-on to a pressed-on spider added significant effort to the fabrication and assembly process, it seems like a flag.



Might be, but you have an analytical tool. No doubt Jan thought the aluminum flywheel and welded spider looked pretty bomb proof.

Answering your question: The original aluminum flywheel had three threaded studs to which one end the flex coupler segment was secured with a nut. The other end on the flex coupler had a stud that went in the drive flange hole and was secured with a nut. The drive flange could float on the pinion drive shaft because the nut/studs holding the coupler were keeping the drive flange in place on the pinion shaft. While I never experienced cracks with my aluminum version, I can understand why they occurred with steel stud inserts in the 6061-T6 threaded holes. The answer to aluminum cracks was the change to the steel flywheel. For a short time, there was a steel disk version that was replaced with the Honda drive plate version now being used.

The change to the steel Honda drive plate replaced the stud/nut with a floating pin design on both the crankshaft drive flange and the pinion flange. The press fit on the pinion drive shaft was necessary to keep the two sets of pins aligned while the flex coupling segments float.

There was never a combination of the aluminum flywheel and welded pin drive flange, although I suppose someone could have tried it. If I understand correctly, one of the failed "ears" on the drive flange was the original stud/nut design being used with the steel drive plate floating pin design. The other failure was the newer pressed on flange/floating pin on both sides design.

It the photo, it is pretty clear where the three-ear design came from, and it seems to be typical for this style of coupler with bolts/studs/nuts. With the floating pin design there is no reason a round disk could not be used on the drive pinion side, which is what the crankshaft side uses.

John Salak
RV-12 N896HS
 

Attachments

  • Original Viking 110 Drive Coupler Design.jpg
    Original Viking 110 Drive Coupler Design.jpg
    501 KB · Views: 93
  • Viking 110 Gearbox, Gen 2 of crankshaft side drive flange.jpg
    Viking 110 Gearbox, Gen 2 of crankshaft side drive flange.jpg
    384.6 KB · Views: 75
The press fit on the pinion drive shaft was necessary to keep the two sets of pins aligned while the flex coupling segments float.

John, thanks for the input, but I'm not following. Could you expand a bit?
 
John, thanks for the input, but I'm not following. Could you expand a bit?

Dan,
The photo of the pin drive is the current configuration, those pins are mirrored on the pinion drive flange. The flex coupler segments float on the pin pairs, so the pinion drive flange needs to be fixed on the shaft by pressing it on to control alignment. If you float the segments and the pinion drive flange, the drive flange could strike the back of the gearbox housing when running. The original bolted together pinion drive flange slid on the pinion shaft splines and the bolted segments kept it place with some very minor fore/aft movement potential. When I removed the gearbox the first time to replace the aluminum flywheel, the pinion shaft came out of the original flange without using a press.

The other part of the puzzle is the pinion drive shaft moves forward a bit as the helical gear in the gearbox loads. There is a gap between the front and back pinion gear bearings that allows this fore/aft movement before the inner races take the thrust load. The fix for that is a newer pinion shaft that reduces the inner race gap to the bearing faces. You can plainly see the movement in this video as the prop loads and unloads. I regularly check the clearance between the flange and gearbox mounting plate and have seen no change in the past two years. This is a known phenomena and is clearly documented on the Viking website under the 110 engine service bulletins.

https://www.flickr.com/photos/187195560@N08/50035579356/

John Salak
RV-12 N896HS
 
The drive flange gets pressed on the input shaft until it makes contact with a shoulder. Jan did not allow enough clearance between the rotating flange and the gearbox housing to accommodate his tolerance stackup. To exacerbate the situation either the input shaft was designed wrong or supplied wrong meaning that the shaft had about 0.040" of axial float when assembled.

This apparently wasn't caught on assembly (not checking the clearance between flange and gear case) thus there were instances of the flanges making contact with the gear case when in operation. The solution is for the owner to buy a new $300+ input shaft, send the gearbox to Jan and have the input shaft replaced and the drive flange pressed off the old shaft and onto the new shaft. He could have fixed the problem by putting a shim behind the bearing, but I'm not sure people would have paid $300+ for a shim.

In the instances I have investigated of failed drive flanges, Jan is claiming that the flanges had been making contact with the gearbox housing and that contact was what was responsible for the drive flange failure. He is now claiming that publicly on Facebook. He refuses to acknowledge that a flange with 2 huge fatigue cracks in it is capable of deflecting and the stress applied to the pins makes it want to deflect towards the gearcase, at which time it certainly could have made contact, especially right before it broke off. See the FEA analysis of the deflection of the pins and the flange when loaded (magnified many times of course).
 

Attachments

  • Jan failure mode-1.JPG
    Jan failure mode-1.JPG
    72.6 KB · Views: 81
  • Jan failure mode-2.JPG
    Jan failure mode-2.JPG
    63.7 KB · Views: 76
  • Jan failure mode-3.JPG
    Jan failure mode-3.JPG
    34 KB · Views: 66
  • Jan-1.JPG
    Jan-1.JPG
    136.3 KB · Views: 68
  • Jan-2.JPG
    Jan-2.JPG
    199.5 KB · Views: 69
  • Deflection.JPG
    Deflection.JPG
    62 KB · Views: 52
  • IMG_0388.jpg
    IMG_0388.jpg
    111.8 KB · Views: 58
Last edited:
If you float the segments and the pinion drive flange, the drive flange could strike the back of the gearbox housing when running.

Odd. The Viking video linked earlier shows the driven flange seating against a step on the pinion shaft (video snips below). Press fit or slip fit of the spline would make no difference in terms of clearing the gearbox back plate. That clearance is a function of where the step is machined on the shaft, and the axial position of the shaft in the gearbox...

The other part of the puzzle is the pinion drive shaft moves forward a bit as the helical gear in the gearbox loads...You can plainly see the movement in this video as the prop loads and unloads.

...was was apparently kinda iffy.

You're certainly wise to keep an eye on the axial freeplay dimension shown in your video. The shaft and flange is hammering fore and aft every time the system passes through its first natural frequency, at a rate equal to that frequency.

BTW, what is the first resonant RPM with the existing cut coupler?
.
 

Attachments

  • ScreenHunter_1468 Dec. 09 15.23.jpg
    ScreenHunter_1468 Dec. 09 15.23.jpg
    33.2 KB · Views: 49
  • ScreenHunter_1467 Dec. 09 15.23.jpg
    ScreenHunter_1467 Dec. 09 15.23.jpg
    30.6 KB · Views: 52
Got thinking about John's video showing fore and aft freeplay at the input shaft. As he noted, it would cycle through a fore and aft motion at every torque reversal.

So what fixates the shaft axially? The flange pressing video Keith linked earlier seems to show a cylindrical roller bearing in the front case. Flange configuration of the bearing race appears to be what SKF calls an NJ type, and Torrington calls an RIJ and RJ. A bearing in this style can carry very heavy radial loads, but axial load is limited to 10% of radial (Torrington), or a max determined by equation (SKF). The axial load is all sliding friction on the ends of the rollers.

Under power, without torque reversal, the pinion shaft would move forward, loading the flange on the roller bearing. Are users seeing wear on the inner race flange and the ends of the rollers? It may be fine if load is less than that nominal 10% of radial. Mostly I'm just curious, as it's an interesting gearbox design detail.
 

Attachments

  • ScreenHunter_1469 Dec. 11 09.51.jpg
    ScreenHunter_1469 Dec. 11 09.51.jpg
    49.9 KB · Views: 69
Last edited:
Got thinking about John's video showing fore and aft freeplay at the input shaft. As he noted, it would cycle through a fore and aft motion at every torque reversal.

So what fixates the shaft axially? The flange pressing video Keith linked earlier seems to show a cylindrical roller bearing in the front case. Flange configuration of the bearing race appears to be what SKF calls an NJ type, and Torrington calls an RIJ and RJ. A bearing in this style can carry very heavy radial loads, but axial load is limited to 10% of radial (Torrington), or a max determined by equation (SKF). The axial load is all sliding friction on the ends of the rollers.

Under power, without torque reversal, the pinion shaft would move forward, loading the flange on the roller bearing. Are users seeing wear on the inner race flange and the ends of the rollers? It may be fine if load is less than that nominal 10% of radial. Mostly I'm just curious, as it's an interesting gearbox design detail.
.

There is a YouTube video discussing the bearings in the gearbox, if you look at 18:30-20:00 you will see the discussion on the axial loading on the pinion gear shaft and bearings. It agrees with your observation on how it works, although it is clearly states the bearing is rated to take 50% of the radial load in thrust (axial load). I have the OEM bearing PN somewhere and have not checked that number. {Nachi NJ206EG}

https://www.youtube.com/watch?v=SODvLtvmYbI

I have seen no measurable change in the flange to gearbox mounting plate clearance over the past three years. Any wear on the bearing roller faces or race flange should show up as decreasing clearance at the pinion gear moves further forward.

John Salak
RV-12 N896HS
 
Last edited:
Odd. The Viking video linked earlier shows the driven flange seating against a step on the pinion shaft (video snips below). Press fit or slip fit of the spline would make no difference in terms of clearing the gearbox back plate. That clearance is a function of where the step is machined on the shaft, and the axial position of the shaft in the gearbox...



...was was apparently kinda iffy.

You're certainly wise to keep an eye on the axial freeplay dimension shown in your video. The shaft and flange is hammering fore and aft every time the system passes through its first natural frequency, at a rate equal to that frequency.

BTW, what is the first resonant RPM with the existing cut coupler?
.

When I installed my pinion drive flange, I pressed it to a clearance distance between the back of the flange and the gearbox housing. I did not really pay attention to seeing if it was fully seated on the pinion drive shaft splines. The final clearance was 2mm according to the photos I took. The next time I have it apart to install the 1.5" pinion drive shaft I will base the clearance on aligning the two sets of pins on the flanges, which I think will increase the distance from the back of the gearbox housing.

Long ago when resonant RPM was openly being discussed, operation (idle) below 1,500 rpm was not recommended. A region of 3000-3700 rpm was also to be avoided. My tach has a yellow arc and alarm to remind me to stay out of the 3000-3700 region. A prop manufacturer supposedly did some instrumented measurements with one of their props on the engine, so there is some data in existence that may document that recommendation.

John Salak
RV-12 N896HS
 
Without pressing the flange against the shoulder on the shaft, it is highly questionable that it would be on straight and by design that is what one should be doing. But it only works if the person doing the design is capable of adding. Jan added a boss that was 7.5mm high to the flange where it previously had a counterbore that was 2mm deep. Given that the original nut was 7.5mm high, Jan effectively removed 2mm of clearance with this design change.

He could have fixed it by making the center boss 2mm longer towards the case to get the original clearance back, but he didnt. Now band aids need to be applied like pressing the flange partially onto the splines and if there is ever a scratch on the gearbox case, sorry, your out of luck because you allowed your flange to make metal to metal contact.
 
There is a YouTube video discussing the bearings in the gearbox, if you look at 18:30-20:00 you will see the discussion on the axial loading on the pinion gear shaft and bearings. It agrees with your observation on how it works, although it is clearly states the bearing is rated to take 50% of the radial load in thrust (axial load). I have the OEM bearing PN somewhere and have not checked that number. {Nachi NJ206EG}

https://www.youtube.com/watch?v=SODvLtvmYbI

Listen closely. Jan clearly says the front bearing is a "heavy duty ball bearing" (19:10). An axial load of 0.5 x radial is indeed reasonable for some types of ball bearings. However, the Nachi NJ206EG is a cylindrical roller bearing, as seen in the flange pressing video. Did Jan misspeak, or was the bearing type changed in later gearbox iterations?

Anyway, Nachi provides a simple equation for estimating axial load capacity. I've attached the catalog page below. I assumed 4600 RPM (torque peak) and 6000 RPM, oil lubricated continuous end thrust. Looks like 302 and 232 newtons (68 and 52 lbs), both less than 1% of the bearing's 39,000N basic dynamic load rating.

Ya'll check me here. I've never done any gear calcs, but the equations for the axial load of a helical gear seem simple enough:

Tangent force (Ft) in Newtons = (2000 * torqueNm) / pitch diameter in mm

then

Ft * tangent of tooth angle = axial force in Newtons

Ok, so assume 90mm as a pinion pitch diameter, 153 Nm for torque at 4600 RPM, and a 15 degree tooth angle. I get 911 Newtons, which exceeds Nachi's desired limits by a factor of three.
 

Attachments

  • Nachi.jpg
    Nachi.jpg
    149.3 KB · Views: 71
Some types being angular contact BBs.

Thrust load on a roller bearing could be within material limits but it's service life would be pretty finite. Guessing they don't use same evaluation as L10 life. If so, gotta be a large derating factor.

Would love to here the rationale for the application.
 
I believe it is this bearing from SKF
I dont know if the lubrication oil qty and temperature comply with the requirements, but as mentioned in the text under the right conditions they can support up to half the radial load in axial load. The Hubelbank gearbox that had 800 hours on it did not show a collection of fine metal from bearings and one cannot see any visible wear on the rollers yet.

While I dont agree with the choice of bearings or the way there very poor control of the axial position of the shafts, that does not have appeared to have been a factor in the 2 failures that I investigated. The prop shaft bearing is simply held in place with a snap ring, which is obviously not a zero clearance setup.
 

Attachments

  • Bearing-1.JPG
    Bearing-1.JPG
    36.8 KB · Views: 64
  • Bearing-2.JPG
    Bearing-2.JPG
    55.3 KB · Views: 66
  • Bearing-3.JPG
    Bearing-3.JPG
    39.8 KB · Views: 63
  • Bearing-4.JPG
    Bearing-4.JPG
    49.3 KB · Views: 69
...but as mentioned in the text under the right conditions they can support up to half the radial load in axial load.

I will give you up to one million dollars...

While I don't agree with the choice of bearings or the way there very poor control of the axial position of the shafts, that does not have appeared to have been a factor in the 2 failures that I investigated.

I hear 'ya. Was the front bearing changed to a ball type in later boxes, or was Jan just having a senior moment in the video?
 
From what I see, the prop shaft rides in 2 deep groove ball bearings, the one closest to the prop flange held in place with the snap ring. The rear rides on the inside of a second deep groove that is held in place by a shrink fit on the outer race of the bearing. That bearing is prevented from migrating out of the case by a shoulder on the prop shaft.

Based on the 2 samples I have, both bearings on the input shaft use the same roller bearing, with just the inner race pointing in opposite directions.

The 3 gear gearbox he was holding clearly did have a different bearing than what is in the Viking 110 gearbox.
 
Last edited:
Here is a photo of my gearbox cover, it has the Nachi NJ206EG bearing as originally installed. The parts list in the Viking 110 Engine Manual has the same roller bearing pn on both sides of the drive pinion shaft. It does indeed appear the 3-gear GBs are using a ball bearing in the cover and a roller on the mounting plate side of the pinion shaft. There has never been a notice to change the 110 gearboxes to a ball bearing version in the front cover. The pinion shaft is the same part for both the 2-gear 110 and the 3-gear 130/150 GBs. They may also be using it on the 190 hp version of the gearbox, so changing the front bearing from a 110 hp use to 190 hp use might be the reason.

The prop shaft gear is riding on single row Nachi 6009 ball bearings. I believe the 3-gear versions are using double row ball bearings because of the higher power.

John Salak
 

Attachments

  • Viking Honda 110 Gearbox Cover.jpg
    Viking Honda 110 Gearbox Cover.jpg
    423.4 KB · Views: 102
After finally resolving some bugs with doing a true centrifugal force FEA I managed to get it done today and the stress values come out higher than the static analysis I did previously. Peak stress 578MPa which is 84ksi. This is for the obsolete bolted giubo version (Hubelbank failure) Really up there. I will repeat for the welded pin version now that I know the workflow.
 

Attachments

  • centrifuge-1 bolted.JPG
    centrifuge-1 bolted.JPG
    128 KB · Views: 69
  • centrifuge-2 bolted.JPG
    centrifuge-2 bolted.JPG
    144.1 KB · Views: 76
  • centrifuge-3 bolted.JPG
    centrifuge-3 bolted.JPG
    133.3 KB · Views: 65
Here the pinned version. Underside is a low stress region. 2 major stress concentration on the upper side that faces the giubo. The principle stress is up to 638MPa or 92.5ksi. All that extra mass trying to pull the spoke off certainly heightens the peak stress at the central hub. I added the picture of the cracked part for comparison.
 

Attachments

  • centrifuge-1 pinned.JPG
    centrifuge-1 pinned.JPG
    119.4 KB · Views: 63
  • centrifuge-2 pinned.JPG
    centrifuge-2 pinned.JPG
    133.9 KB · Views: 58
  • centrifuge-3 pinned.JPG
    centrifuge-3 pinned.JPG
    131.5 KB · Views: 61
  • IMG_0388.jpg
    IMG_0388.jpg
    111.8 KB · Views: 68
Keith, I would suggest you use the loading vector direction as the line between the two pins carrying the loads, not tangential.
 
There has never been a notice to change the 110 gearboxes to a ball bearing version in the front cover. The pinion shaft is the same part for both the 2-gear 110 and the 3-gear 130/150 GBs

Got it, thanks. It's your box, but it it were mine, I'd forget the "50%" ad copy, do the actual calcs per the tech page above, and maybe retrofit a suitable ball bearing, with whatever spacer or shim is necessary to set end play to a small value. The current clearance (https://www.flickr.com/photos/187195560@N08/50035579356/) means it's hitting the bearing like a rivet gun when it passes through the resonant range.
 
Last edited:
Today the discovery was made that the flex plate was cracked too (340 hours total hobbs time on the engine).
 

Attachments

  • IMG_0434.jpg
    IMG_0434.jpg
    460.7 KB · Views: 87
  • IMG_0433.jpg
    IMG_0433.jpg
    443.7 KB · Views: 88
Keith, I would suggest you use the loading vector direction as the line between the two pins carrying the loads, not tangential.

Bill, its easy enough to change the direction but its going to have a minor impact on the results. I set the radial and tangential loads this way when I did the original static analysis because I had some issues getting the centrifugal analysis to work. The pins are 60 degrees apart so that load line will be 30 degrees off the tangential. Its the centrifugal loads that dominate currently.

I changed the load line and it does reduce the stress, but not enough to make it a good part, its still 389MPa or 56.4ksi Its a more than 200MPa reduction, so not so minor.. Will pay more attention in future.
 

Attachments

  • centrifuge-1 pinned 30 degree load line.JPG
    centrifuge-1 pinned 30 degree load line.JPG
    133.3 KB · Views: 67
Last edited:
Today the discovery was made that the flex plate was cracked too (340 hours total hobbs time on the engine).

Not a surprise. Recall I suggested an FEA with a 1 rad/sec pitch or yaw rate.

Is it an unmodified Honda OEM part? It looks like a Fit or Civic flex plate which has been modified by welding a ring gear on its perimeter, and drilling a bunch of extra lightening holes between the original holes.

In the car, the rim of the flex plate is bolted to the rim of the torque converter, which is supported at its center by the crank and the transmission. It is never subjected to a bending load. Weld a mass to an unsupported perimeter, then apply a pitch or yaw rate, and the web is bent with load cycles equal to RPM.

Not sure why the center was plasma cut on the shown sample, maybe to prevent it being used again ??

Now you know. Where did it it come from?
.
 

Attachments

  • Viking Honda 110 Steel Drive Plate.jpg
    Viking Honda 110 Steel Drive Plate.jpg
    100.8 KB · Views: 96
Last edited:
From what I see, the prop shaft rides in 2 deep groove ball bearings, the one closest to the prop flange held in place with the snap ring. The rear rides on the inside of a second deep groove that is held in place by a shrink fit on the outer race of the bearing. That bearing is prevented from migrating out of the case by a shoulder on the prop shaft.

Based on the 2 samples I have, both bearings on the input shaft use the same roller bearing, with just the inner race pointing in opposite directions.

The 3 gear gearbox he was holding clearly did have a different bearing than what is in the Viking 110 gearbox.

Hate to harp on the subject; but,

I can't help but think the previously listed literature's reference to the roller bearing supporting up to half of the radial load in thrust would be for a transient condition. What reputable bearing supplier is going to accept constant sliding contact as a design parameter? Angular Contact ball bearings are specifically design to support both loads. Roller bearings are typically applied on mechanisms that have very high radial loads (and limited support axial length).

Secondly, a bearing installed on a shaft without an interference fit/relying on a snap ring? I guess fretting wasn't a concern and things only get worse after that starts. The lack of proper application here is truly astounding. "It turns the right speed and direction and we found a way to mount it."

Marginally related. The internet is great but I still like books (insert your jokes here). If you ever find this book, any revision, buy it. Written by someone who excelled at their craft and was good at conveying their thoughts in writing. I loaned mine out to an intern and never saw it again.
 

Attachments

  • Screenshot 2022-12-13 091622.png
    Screenshot 2022-12-13 091622.png
    457.3 KB · Views: 69
  • Screenshot 2022-12-13 091329.png
    Screenshot 2022-12-13 091329.png
    337.9 KB · Views: 80
Last edited:
Not a surprise. Recall I suggested an FEA with a 1 rad/sec pitch or yaw rate.

Is it an unmodified Honda OEM part? It looks like a Fit or Civic flex plate which has been modified by welding a ring gear on its perimeter, and drilling a bunch of extra lightening holes between the original holes.

In the car, the rim of the flex plate is bolted to the rim of the torque converter, which is supported at its center by the crank and the transmission. It is never subjected to a bending load. Weld a mass to an unsupported perimeter, then apply a pitch or yaw rate, and the web is bent with load cycles equal to RPM.



Now you know. Where did it it come from?
.

I believe the flexplate/ring gear cracks are a sperate issue from the gearbox drive coupler failures. The flexplate is a stock Honda part with a starter ring gear welded to it. The welds are as good as any I have seen, they are uniform, and the plate shows no sign of warpage. There are no additional holes drilled.

The cracks appear to be limited to the 110 engines and not the 3-gear setup on the 130/150/190 engine which use the same part. Charlie Rosenzweig has the 3-gear version with the same flexplate/ring gear on his turbo 1.8L Honda, so it would be interesting to see if he has the same problem. If the cause is pitch/yaw induced bending it would affect every installation of this part.

So what is different about the 110 setup that could cause a bending vibration? The big difference is the bottom half of the flexplate is in the radiator airstream where it is subject to the prop pressure pulses. It may be the same phenomenon at work as cracked exhaust pipe-muffler joints because the pipe sticks out a bit too far from the cowling. The other engine installations use a horizontal radiator mounted under the footwells, vs the 45 deg angled 110's which is about 10" behind the prop. It could also be turbulent flow through all those holes are causing some resonance to occur.

Just one more thing to check during the CI.

John Salak
RV-12 N896HS
 
I believe the flexplate/ring gear cracks are a separate issue from the gearbox drive coupler failures.

Possibly so.

The flexplate is a stock Honda part with a starter ring gear welded to it.

Does the 110 manual list a part number?

The cracks appear to be limited to the 110 engines and not the 3-gear setup on the 130/150/190 engine which use the same part.

Basis? I'm guessing very few are being dye checked, and the crack location appears to be hidden behind the drive pin disk.

If the cause is pitch/yaw induced bending it would affect every installation of this part.

To be precise, it would affect those who actually apply high pitch and yaw rates. Cherokee vs Pitts, so to speak.

Seriously John, look up the equation for gyro moment and run some numbers. It's been a while, but I seem to recall something like pitch/yaw rate x rotation rate x mass moment of inertia. The MMOI is easily obtained by bifilar suspension. I'm pretty sure the result will exceed any imaginable air load.
 
John, I agree with the other commenter who said that the smaller holes were added to the OEM part. So its no longer a stock part. In addition, the material gauge was never sized for carrying the ring gear, that has always been attached to the torque converter. So thats the second modification and a big one. The torque converter is a robust 20lb part.

I do think it is a seperate issue to the drive flange failure, but makes me glad I planned on going to a "proper" flywheel instead of continuing with the flex plate. If the failure progression is so bad after 350 hours its going to have to be replaced with a better part, meaning a flywheel.

I believe the flexplate/ring gear cracks are a sperate issue from the gearbox drive coupler failures. The flexplate is a stock Honda part with a starter ring gear welded to it. The welds are as good as any I have seen, they are uniform, and the plate shows no sign of warpage. There are no additional holes drilled.

The cracks appear to be limited to the 110 engines and not the 3-gear setup on the 130/150/190 engine which use the same part. Charlie Rosenzweig has the 3-gear version with the same flexplate/ring gear on his turbo 1.8L Honda, so it would be interesting to see if he has the same problem. If the cause is pitch/yaw induced bending it would affect every installation of this part.

So what is different about the 110 setup that could cause a bending vibration? The big difference is the bottom half of the flexplate is in the radiator airstream where it is subject to the prop pressure pulses. It may be the same phenomenon at work as cracked exhaust pipe-muffler joints because the pipe sticks out a bit too far from the cowling. The other engine installations use a horizontal radiator mounted under the footwells, vs the 45 deg angled 110's which is about 10" behind the prop. It could also be turbulent flow through all those holes are causing some resonance to occur.

Just one more thing to check during the CI.

John Salak
RV-12 N896HS
 
The flex plate is a Honda OEM part, right down to the "5TO" stamped on it. There are no modifications beyond the welded ring gear. See the attached photo of a used OEM part on ebay, same hole pattern and listed as a 2015-2020 Honda FIT part. The 110 Engine Manual was never updated past 2013 and still has the aluminum flywheel.

Dan, I have no doubt about the gyroscopic forces and the potential to be a causal effect for the cracks. I enjoy making steep turns in the 1.6 to 2.0G region on a regular basis. The gyroscopic effects are one of my major concerns with KeithO wanting to go to a "heavy" flywheel design, not to mention a forward C.G. concern (the reason my batteries and fuel pumps are in the tail cone). Sadly, my data points are limited to my aircraft and two others that I know of that have had cracked flex plates (all 110 engines). The cracks will eventually result in a complete separation. If there are 130/150 engines with the same crack patterns, I would go back to looking for a common cause vs looking at what is different. I believe there are a couple of RV-12s running the 130 engine, however I know nothing about them or how they fly.

The only way you find this is to wait for a complete separation or if you remove the gearbox and crankshaft drive flange to look for cracks.

John Salak
RV-12 N896HS
 

Attachments

  • Ebay Honda FIT Flex Plate 2015-2020.jpg
    Ebay Honda FIT Flex Plate 2015-2020.jpg
    235.4 KB · Views: 85
John, whats the background to the photo you posted of the flex plate on a scale with the core cut with a plasma cutter ? Im curious how old that photo is ? Seems someone knew of this cracking issue a loooong time ago ?
The bigger engines have an 8 bolt pattern, so they will not be using the identical part number, but one modified in a similar way.
 
Last edited:
Back
Top