Redrives for most commercial engines.

[Lucky Dog]

The rubber ring is not truly a damper. It is a soft torsional stiffness...a torsional spring, if you wish.

Damping means removing energy from the system, typically in the form of heat. A torsional spring merely stores energy and returns it to the system sometime later in the cycle. The typical rubber spring does have a very small damping coefficient, evident when worked, as it gets warm. However, it's nothing like the damping necessary to deal with a resonant period on a continuous basis. If it was, the rubber would melt. Centaflex, a donut supplier with good published design data, actually lists what they call Pkv, maximum permissible power loss, in watts. A Centaflex A rated for 0.1 kNm nominal torque (about 74 lbs-ft, or 50 HP if applied at 3500 RPM) is only allowed 25 watts, or 0.0335 HP.

Sorry if this all seems pendant, but the use of incorrect terminology does no service to the community....and doesn't foster much confidence in the vendor among those who may possess some small understanding.

So, according to your post, there is a degree of correctness in both member's statements. No harm done. The Rotax C drive supports the effectiveness of incorporating the Guibo ring as an harmonic insulator. Empirical data is often more useful than calculated projections. In this case, I'll side with the C drive and applaud this new design as a plausible improvement to low-power belt redrives.

 

[rv7charlie]

I didn't read it that way at all. *Every* device (even a gear train) has at least some damping; it will convert some of the energy going in to heat. The terminology problem for us shadetree guys is that we don't have a convenient shorthand for 'torsional spring' that's easily understood by all. Unfortunately, a lot of us get lost on 'torsional spring' but we kinda grasp the purpose of a 'damper' even if we misunderstand the principle by which it works.

At least we've got this forum, to begin to wave the fog away.

 

[DanH]

So, according to your post, there is a degree of correctness in both member's statements.

If that was your impression, I did a poor job. I'll try to be more clear. Referring to a rubber coupler as a damper is wrong as a soup sandwich.

 

[challenger_II]

Ok, so the guibo utilized in the Rotax C-Box application is an isolator. Isolates tortional loading and unloading of the engine input, and the propeller reflection, during operation. Must be fairly functional, as the C-box has a better track record on the 582 than the B-Box, with the spring washer-loaded dog clutch.

I believe Lucky Dog has a point.

 

[Hot Wings]

I believe Lucky Dog has a point.

Or at least a soggy sandwich?
There could be a minimal amount of damping due to hysteresis in the guibo. One of Great Plains early flywheel drive (prototype?) for the VW used rubber doughnuts around the perimeter of the drive plate. They did such a good job of damping that they melted, and if my memory is correct, fairly quickly .................. I thought I had a pic in my files, but I can't find it anymore.

 

[DanH]

Ok, so the guibo utilized in the Rotax C-Box application is an isolator. Isolates tortional loading and unloading of the engine input, and the propeller reflection, during operation.

Assume a simple two-element model, fairly realistic when discussing a very low connecting stiffness. When a forcing frequency nears or matches the natural frequency of the system (ratio near 1:1), it will resonate; torsional oscillation will amplify. When forcing frequency is not close to the natural frequency (forcing/natural ballpark greater than 1.2 or less than 0.8), the system can be considered as isolated from the forcing frequency.

Put another way, isolated is the opposite of resonant. It is a result, not the reason.

Here's the important part. How does a designer adjust the ratio to get the system into isolation? Using the 582 from your example, combustion events provide powerful forcing frequencies ranging from 66 hz (2000 RPM idle) to 226 hz (6800 RPM). Only real choice is to change the other side of the ratio, the natural frequency of the system. Push natural frequency down below 50 hz or so (i.e. less than 0.8 x 66), and the result is isolation. The reason is a low connecting stiffness.

I have no idea what you mean by "propeller reflection". The propeller is basically a very large inertia as compared to any other inertia in the typical system. As a large inertia, it oscillates with less angular displacement as compared to the smaller ones. Think of it as the relatively immovable object around which the other inertias oscillate, and you're roughly on the right track. For a better understanding, quantify the relative oscillations by plugging the stiffness and inertia values into a Holzer code and look at the resulting mode shape.

I've attached a simple sketch of the first and second mode oscillation of a three inertia model, and a more formal mode shape plot from denHartog, first and second modes for a seven inertia model, in which the angular displacements of the inertias have been quantified. BTW, seven inertias means it will have four more natural frequencies and modes of vibration not plotted, each at higher and higher frequencies. Natural frequencies = inertias less one.

 

[challenger_II]

Ok, so the guibo is an isolator, rather than a damper. Got it. But it still does a decent job of keeping the engine and gearbox in a relatively happy relationship. Which was the point of the conversation.

 

[DanH]

Ok, so the guibo is an isolator, rather than a damper.

Neither. It is a torsional soft element. Sometimes it contributes to isolation, sometimes it contributes to amplification.

I've attached a spec page from the current Lovejoy Centaflex catalog. Note the range of available dynamic torsional stiffness values. A skilled designer does not simply grab some random "guibo" from an aftermarket auto parts catalog. He chooses a soft element with a specific engineered stiffness which will result in a desired first natural frequency, preferably below the lowest forcing frequency (i.e. below idle RPM). If he can get natural frequency far enough below the least forcing frequency, the rubber coupler contributes to isolation thru the rest of RPM range. Even so, the system must pass through the resonant range as RPM ramps up from cranking speed to a stable idle speed. When it does, the rubber coupler choice contributes to amplification.

The above describes the classic Rotax approach, practical because the high idle means the lowest forcing frequency isn't all that low. In the C-box example, a rubber coupler puts the first natural frequency well below the 66 hz forcing frequency at a 2000 RPM idle, so it only amplifies during ramp up....you know, the giant rattle and shake when it starts.

Sometimes the application requires a low idle. In that case, the RPM may need to pass through the resonant range every time the throttle goes forward. Within that resonant range (forcing/natural ratio between 0.8~1.2) the coupler choice contributes to amplification. An example might be a 3-cyl four stroke system with a coupler choice resulting in a 25 hz first natural frequency. It would idle fine at 500 RPM (12.5 hz, isolated), beat itself hard at 1000 RPM (25 hz, resonant, amplified), and be smooth again at 1500 RPM (37.5 hz, again isolated).

 

[PMD]

If that was your impression, I did a poor job. I'll try to be more clear. Referring to a rubber coupler as a damper is wrong as a soup sandwich.

In case nobody noted this: it depends on the hysteresis of the rubber compound. I will read the rest of the thread to see if I just repeated some other observation(s).

 

[challenger_II]

You didn't miss it, and you have a hugely valid point.

 

[Lucky Dog]

Ok, so the guibo is an isolator, rather than a damper. Got it. But it still does a decent job of keeping the engine and gearbox in a relatively happy relationship. Which was the point of the conversation.

Bingo!

 

[Lucky Dog]

Theory, however correct, does not solve problems and achieving solutions always requires compromise. The guibo solved the problem for Rotax, and did so for BMW's shaft drive motorcycles. It would be more helpful if the scientists on this page could step forward with practical solutions we could employ rather than well supported warnings of doom. "Perfection is the enemy of good enough." Your predecessors solved harmonic issues because they had to - armed only with slide rules and driven by necessity. I live near the annex of San Diego's Air and Space Museum, where all of their engines and engine cutaways are stored. So, every piston engine used in WWII was geared (helical, planetary, bevel - all types were represented) and all use some sort of crude harmonic isolation device: springs, odd ratios, floating crankshaft counterweights, ect. - not perfect, but many are still chuffing through the sky 70 years later. let's see some sketches. Reduction drives are the missing link between modern engines and a flourishing homebuilt aviation community. We know how to build things. All we really need is a good set of plans.

 

[DanH]

In case nobody noted this: it (damping) depends on the hysteresis of the rubber compound.

True, but again, the limiting factor is heating. The above Centaflex table offers an example. Note the very low permissible power loss values, in watts.

I'm not saying there is no damping available with a rubber coupler. Quite the contrary; note relative damping, which Lovejoy defines as the ratio of damping work to elastic deformation work for a period of vibration. I am saying it can't be expected to contribute significant damping for very long, because it will destroy itself. It's pointless to think of it as a damper. The system must be designed so it does not continuously run in a resonant range, which is why the key parameter is the coupler's torsional stiffness.

 

[DanH]

Theory, however correct, does not solve problems and achieving solutions always requires compromise.

Good design is the art of intelligent compromise.

The guibo solved the problem for Rotax

Because they applied the science. Same for those old school slide rule guys. Nothing about what they did justifies random application or poor understanding of concepts. Heck, 1930's SAE papers are some of the best, clearest explanations you'll ever read.

would be more helpful if the scientists on this page could step forward with practical solutions we could employ rather than well supported warnings of doom.....Let's see some sketches.

No warnings of doom. The goal is to increase wider understanding of fundamental physical principles, which is very practical. Progress is slow, but it's happening. Twenty years ago some of us were trying to move folks away from sprung clutch centers and splines, as in "Hey troops, there are engineered torsional soft elements for this!". Now we're merely trying to teach folks the how and why of choosing them. It ain't pickin' random couplers from a discount auto parts catalog.

Better to teach a man how to fish...

 

[TFF]

The ones in the pictures of Air Trikes solve his problems but those are probably from the driveshaft of a 200+ hp car on something that’s 30 hp give or take. I know someone with a VK30 and for the Continental engine it’s coupler is V8 car flywheel size and about 6” thick. Size matters. Some sizes are not off the shelf like most of the ones we really want.

 

[DanH]

Size matters. Some sizes are not off the shelf like most of the ones we really want.

Choosing a coupler from an auto parts catalog usually means there is no stated nominal torque capacity or dynamic torsional stiffness information. The choice is pure TLAR, the million monkeys approach to drive design. Even if it seems to work, there is no guarantee the next one has the same values as the first one, as we live in an era of cheap look-alike auto parts from endless third world sources.

We have one clown out there, a popular alt engine vendor, who buys auto couplers, then saws through half the connecting members. He claims it solved his problems too.

 

[TFF]

I have blown enough of the gublos in my Alfas and helped change plenty in Mercedes. It’s not the 30 hp engines for me, I want 250 hp and a box that can handle snap rolls at will. That is a well engineered drive. That’s right, none of those parts are in an auto catalog.

 

[bmcj]

Referring to a rubber coupler as a damper is wrong as a soup sandwich.

Mmmmm…. Soup sandwich

 

[DanH]

Previously I wrote:

Choosing a coupler from an auto parts catalog usually means there is no stated nominal torque capacity or dynamic torsional stiffness information. The choice is pure TLAR, the million monkeys approach to drive design.

Just for fun, I went digging to see what technical information was available, focusing SGF, whose rubber couplers are found on a lot of US vehicles. Although I was not able to find information specifically identified as automotive, they do publish tech data for industrial applications. Fair enough.

I then directly compared an SGF rubber coupler to a Centaflex A rubber coupler with near identical nominal torque ratings. The table is below, and nicely illustrates the lunacy of picking a coupler without data, or application of fundamentals, because all "giubos" are not the same.

Both couplers ballpark at around 200 Nm nominal, roughly 150 lbs-ft. However, look at the dynamic torsional stiffness values. The SGF coupler is roughly 5 to 8 times stiffer than the Centaflex, which would push the resonant range up the RPM scale, a highly undesirable result.

For those persist in the notion of rubber couplers as dampers, take a look at the permissible power loss. The stiffer SGF allows dumping only 14 watts into the coupler, which is miniscule. While on that subject, I've also attached a paragraph from an SGF application guide.

 

[challenger_II]

Ok. So, that being said, we should for Heaven's Sake NOT use a guibo. Not even in the factory designed application on the Rotax 582/C-Box combination. Got it!