Torsional Vibration and Resonance - Basic Theory and Issues

Discussion in 'General Auto Conversion Discussion' started by wsimpso1, Nov 20, 2012.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Apr 19, 2019 #41

    Vigilant1

    Vigilant1

    Vigilant1

    Well-Known Member HBA Supporter

    Joined:
    Jan 24, 2011
    Messages:
    3,898
    Likes Received:
    1,717
    Location:
    US
    Thanks for finding that. So, his contention is that the slack-side idler provides damping for 1/2 of the belt system, and that we can't have resonance unless it exists on both sides. I dunno. But, if his setup has worked in practice, that says a lot (for that hardware and use, anyway).
     
  2. Apr 19, 2019 #42

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    I don't know either. But I think they went through stuff that didn't work as well.
    I think Great Planes suspended the Kohler redrive after problems.
    The opposed 1/2VW doesn't seem to work with a redrive, which implies the V-twin has an advantage. My interest is strictly direct drive.
    If I want more power, a four cylinder is my choice instead of redrive.
     
    Last edited: Apr 19, 2019
  3. Apr 19, 2019 #43

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,786
    Likes Received:
    3,009
    Location:
    Saline Michigan
    Same questions for everyone who says they are proven. How proven? How many flying, how many hours on the fleet, and how many failures in use? If they can talk MTBF, even better.
     
  4. Apr 19, 2019 #44

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    6,712
    Likes Received:
    1,766
    Location:
    North Carolina
    Slip on only half the cycle will damp resonance.
     
  5. Apr 19, 2019 #45

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    I respected the Smith's and miss the flying at Oshkosh. No more to be said by me about their former business....
     
  6. Apr 19, 2019 #46

    Vigilant1

    Vigilant1

    Vigilant1

    Well-Known Member HBA Supporter

    Joined:
    Jan 24, 2011
    Messages:
    3,898
    Likes Received:
    1,717
    Location:
    US
    Perhaps the short overall length of the V-twin crank plays a part in all of this.
    Going direct drive does make the TV/resonance analysis less complicated (fewer elements). Of course, it doesn't necessarily make the problem go away.
     
  7. Apr 19, 2019 #47

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    The longer crank does have more crank twisty things happening. My view is the even firing order is more likely to compound and resonate. The Vtwin has an odd fireing order that may cancel resonance. All just my guesses.
    It seems over the top to cut a VW in half then try to get more power again with a redrive anyway. Reports from the half VW yahoo forum say a prsu doesn't work.
     
  8. Apr 20, 2019 #48

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,786
    Likes Received:
    3,009
    Location:
    Saline Michigan
    The rudder pedal frames have the last tab brazed on, so I can take the rudder pedal system over to the powder coated and clean the shop. And now I can spend a little time on this.

    First stiff systems - Everything from the crankshaft to the tip of the prop blade has to accelerate and decelerate together and that takes a BUNCH of stiffness in every part and the reduction drive can not stand any significant lash, as it will bang back and forth through the lash.Two things happen: the first is just impact which will further amplify the firing pulses going both ways through the power path, and; Impact tends to excite ALL of the resonant modes in the system, which will also further amplify the pulses in the system. We have stiff systems in direct drive engines - just look at a Lycoming crankshaft when you get a chance. Beefy huh? The only stiff geared system I ever saw that worked was Ev Hatch's Powersport PSRU. One external spur gear beefily supported on both ends; a beefy internal spur gear in one big piece with the prop shaft; Lapped gears to take all of the lash out. It worked, but after Ev Hatch passed, the folks that took over did not seem to understand the Mazda rotary engine half of the business and folded. Ev Hatch understood Mazda rotaries and building high perf systems, and he still went through a bunch of iterations and a number of in-flight failures to get to his stiff system right. Alan Tolle flew no less than seven maydays with these things. Some folks are working on reviving this drive. Could be fun, but for a one-off, no do not even think about it.

    Now for getting the resonance to low rpm, ideally we want the first resonance order signficantly below idle. That is what we do in cars. And we do this by putting a soft spring between the flywheel mass and the transmission input, so the geartrain is isolated from the firing pulses. Put resonance at least a half-octave below idle and you will have largely isolated the engine from everything downstream of the spring. The spring must be able to carry all of the engine torque plus the oscillation of the crank times the spring rate, which is the torque ripple that is left after the spring. And the spring rate is chosen to drive the resonance safely below idle. If the resonance is up in the operating range (even at the low end), any lash in the geartrain, splines, etc, will be exercised and be noisy, even if not destructive, and you have to go through the resonance when you shove the throttle up - some systems can get hung at the resonance point. Put the resonance down below idle and you go through it once as the engine comes up to idle after cranking/firing, and again when you turn it off. You already see this in the engine mount system going through resonance on start and on stop. The rotating parts are nicely handled this way too.

    So, the system has to be designed to have a suitably soft element to get the resonance below idle. If the soft element is not soft enough, the resonance only gets moved around in the operating range... Steel springs, rubber springs, etc will all work if you can get low enough spring rate and high enough torque capacity. A number of PSRU out there have rubber donuts or other elements that seem to work pretty well. Some rubber spring designs actually become rising rate springs: At low torque (and low rpm), the spring is soft and keeps the resonance below idle; as torque comes up also rpm goes up, and the spring gets stiffer, but the resonance is still safely below operating speed. This obviously has to be DESIGNED or LABORIOUSLY DEVELOPED. Tracey Crook did it. If you want a 160 hp Mazda rotary scheme see if Pat Panzera will work with you a little. Size things improperly and you will either destroy the rubber elements or have resonance in the operating range or both. Oh, and then there is that pesky thing that can happen on a fast descent - while the power is pulled back, the torque is high (but in the opposite direction to when power is being made), but the windmilling prop is running along at half to two-thirds of cruise speed - you can get into resonance if the spring rates are a tad too high.

    None of this sounds like territory for the non-technical guy.

    Belt slip - this will make the belt and sheaves hot in a hurry - they are difficult to take heat out of - and quickly cause failure of most belts.

    One-way-clutches - These will reduce or eliminate resonant inputs, but that is another part that has to be carefully selected, installed solidly, and if it misbehaves, you have big problems. Automatic transmissions used to have several in each, and there were full time engineers both in the OEM and at the suppliers of the OWC's dedicated to these guys. How do they misbehave? Well, some will centrifugally lift off if disengaged and then run at high speed, so you could back off power, windmill the prop and disengage the OWC (prop overspeed anyone?), and when you go to bring power back in the engine revs freely - the airspeed has to come back down so the prop slows in the re-engage speed before power comes on, and that might be quite low if you swipe the wrong part from the car dealer's part bench. All OWC can fail in a couple more ways: The OWC can lock - no slip either way - no longer isolating vibration and result in powertain failure from resonance, or roll over - slips both ways - and cease to connect the engine to the prop. They can work and are used in things like helo powertrains, but they are not for the amateur...

    Belt slip in coast only by using a spring loaded tensioner - Hmm, it had better have only a narrow resonance band and the tach yellow lined for that speed band. If you stay there for long, well, see the belt slip note above. Even if you stay out of the resonance zone, on a long descent with the power back, the prop could be trying to drive the engine, the belt is slipping and being cooked, with either short belt life at best and belt failures/total loss of power at worst.

    What can the amateur do? Perfectly, exactly copy something else that works. Engine, belts, pulleys, structure and bearings and shafts supporting the pulleys, prop - everything. Change anything and you are in unknown territory. And things are done differently in different sized engines, but almost everyone does things one way in big engines and almost everyone does things another way in the UL world. For instance UL belt drives have the sheaves cantilevered and the engine side sheave hung right on the crankshaft, while the V-8 guys isolate the crankshaft from overhung loads by hanging the sheave between two bearings, while the big sheave is usually hung between two bearing sets as well.

    Billski
     
    Last edited: Apr 20, 2019
    rv6ejguy likes this.
  9. Apr 20, 2019 #49

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    I didnt see any spring in the idler tensioner in the Valley Big Twin PSRU belt drive. (Edit, I don't yet visualize how a belt tensioner works as a one way clutch)
    Larry Smith did an assembly video:
     
    Last edited: Apr 20, 2019
    Vigilant1 likes this.
  10. Apr 20, 2019 #50

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    I didn't notice them on first viewing, on the second viewing of the Valley PSRU video there is six large balls inside the large pulley. I don't know what they do. Could be a type of one way clutch?
     
  11. Apr 20, 2019 #51

    pictsidhe

    pictsidhe

    pictsidhe

    Well-Known Member

    Joined:
    Jul 15, 2014
    Messages:
    6,712
    Likes Received:
    1,766
    Location:
    North Carolina
    A flat twin has a very strong fundamental torque amplitude. It is an even fire engine. So, one evenly spaced firing pulse per revolution. A big factor is that the pistons are in phase, so you are accelerating and decelerating them both at the same time. In a V-twin, the piston accelerations cancel. Well, nearly, the finite con rod length means that's not quite perfect. V twins are also uneven fire, so the fundamental from the firing pulses is a little lower. I suspect the problem with VW redrives is that they have been thrown together without considering TV. Ye olde suck it and see engineering. So far, nobody has got lucky. It is possible to make a 1/2 VW redrive, but it will need much more engineering than previous attempts. I looked into a harmonic absorber for piston acceleration cancelation and it needed to be big, and heavy. hence my preference for a V twin...
     
  12. Apr 20, 2019 #52

    poormansairforce

    poormansairforce

    poormansairforce

    Well-Known Member

    Joined:
    Mar 28, 2017
    Messages:
    602
    Likes Received:
    131
    Location:
    Just an Ohioan
    They are tensioning balls. When the shaft gets torqued the balls are pinched on the ramps formed by the pockets and act as a spring for the idler based on the torque applied by the engine. That's why he keeps mentioning the .030 shim on the shaft. It needs a specific clearance to work correctly.
     
  13. Apr 20, 2019 #53

    poormansairforce

    poormansairforce

    poormansairforce

    Well-Known Member

    Joined:
    Mar 28, 2017
    Messages:
    602
    Likes Received:
    131
    Location:
    Just an Ohioan
    In essence the idler releases some of it's pressure between power strokes so the prop can't turn the engine. If you watch the vid where he pulls up in the Backyard Flyer you can see the idler pulsing back and forth. So it's not responding to a spring and it's mechanical properties but instead it responds to the engine load or lack thereof. Brilliant!
    He calls it a one way clutch then corrects himself and calls it a one way clutch "action".
     
    wsimpso1 likes this.
  14. Apr 20, 2019 #54

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    Wow. It's like a farmer that invented the engine, the PSRU, the clutch, the wing swing fold, the triangle wing spar that doesn't need struts, the prop, the landing gear, the welded aluminum tube and on and on.
     
  15. Apr 20, 2019 #55

    wsimpso1

    wsimpso1

    wsimpso1

    Well-Known Member

    Joined:
    Oct 19, 2003
    Messages:
    5,786
    Likes Received:
    3,009
    Location:
    Saline Michigan
    Those are the springs for the tensioner. Clever design. I bet it took some development to get the size and durometer of the balls and the shape/size of the mating recesses right.

    The senior Smith called their earlier spring loaded tensioner a device that slipped the belts on coast side implulses, this looks to me more like it will do something more sophisticated. I said that slipping the belt will heat it up and is likely to result in short belt life. I stand by that comment, but I do not think they are actually slipping the belt much, if at all...

    When the positive side of a power pulse occurs, the belt will hold and be quite stiff - just the load vs elongation of the straight segment of belt between the sheaves - and it will have the resonant frequency of the engine and prop inertia with the belt and whatever other springiness the sheaves and shafts have in between. Then on the deceleration between firing pulses, the belt is pulling in the opposite direction, deflecting the tensioner, and resulting in a much lower effective spring rate than on the power side, which gives a much lower resonant frequency in this direction. Having two widely different resonances available greatly reduces the amplification that the system will see when firing pulses coincide with one or the other... Then the second part is involved - the decel motion on a V-twin is just a few degrees, my back of the calculator estimate is 7 degrees. Well, the tensioner deflecting will let more than that much motion happen without the belt slipping, the tensioner moves that much belt. The crank goes through the power pulse with one resonant frequency, then goes into coast and slows down, but the belt tensioner lets the engine decel pretty much by itself without driving the prop shaft side into the same decel while having a system resonance at a much lower frequency. Half of the vibe input is greatly reduced and half of the time the system is responding in one resonant pattern, while the other half of the time, it is responding in a much lower resonant frequency pattern. This can nicely tame any vibe issues due to primary firing pulses and probably the pulses that occur at twice firing as well.

    Clever stuff when you get into it. I still bet that they did quite a bit of development...

    Billski
     
    Last edited: Apr 20, 2019
  16. Apr 20, 2019 #56

    BBerson

    BBerson

    BBerson

    Well-Known Member HBA Supporter

    Joined:
    Dec 16, 2007
    Messages:
    11,855
    Likes Received:
    2,252
    Location:
    Port Townsend WA
    Are the balls rubber or steel or aluminum? They look grey color.
     
  17. Apr 20, 2019 #57

    Hot Wings

    Hot Wings

    Hot Wings

    Well-Known Member HBA Supporter

    Joined:
    Nov 14, 2009
    Messages:
    6,312
    Likes Received:
    2,269
    Location:
    Rocky Mountains
    Good probability they are one of these:

    https://www.mcmaster.com/rubber-balls

    Based on color I'd guess the silicone version.?
     
  18. Apr 20, 2019 #58

    poormansairforce

    poormansairforce

    poormansairforce

    Well-Known Member

    Joined:
    Mar 28, 2017
    Messages:
    602
    Likes Received:
    131
    Location:
    Just an Ohioan
    I agree, I think he basically created a variable spring rate to kill the resonance as you mentioned and no, the belt doesn't actually slip, just changes the tension on the belt slightly. But it's enough.
     
  19. Apr 20, 2019 #59

    poormansairforce

    poormansairforce

    poormansairforce

    Well-Known Member

    Joined:
    Mar 28, 2017
    Messages:
    602
    Likes Received:
    131
    Location:
    Just an Ohioan
    If that's what they are then the development may have been easier???
     
  20. Apr 20, 2019 #60

    Vigilant1

    Vigilant1

    Vigilant1

    Well-Known Member HBA Supporter

    Joined:
    Jan 24, 2011
    Messages:
    3,898
    Likes Received:
    1,717
    Location:
    US
    Yes, but there's still a lot of "magic" in the depth of those recesses, the offset of the idler "ear" on the plate, etc.

    FWIW, Gene and Larry used the same design on their VW PSRU. A video of the installation of that redrive is here:


    1) It's a shame that the work to develop this is not still being put to good use. I wonder if Larry would consider selling the machined parts as a kit, or the whole thing?
    2) Criticality of the Generac vs another (still available) big block twin? As billski mentioned previously, details are critical if attempting to replicate a system that s known to work. The Generac (993? cc) isn't available anymore, but other engines of similar displacement and 90 deg V layout are readily available.
    3) At the "cost" of needing to do all the testing/validating again, could a similar result be achieved with fewer custom parts?

    ETA: I'm assuming that the literature that comes with the drive explains checks/maintenance. Those balls are going to change with time due to compression, vibration, temperature, abrasion of the outer surface, etc. You'd need some way to know when they need to be changed. Getting at them requires some disassembly (incl prop removal), so it would probably be easiest to do it when the belts need to be changed.
     
    Last edited: Apr 20, 2019

Share This Page

arrow_white