CRANKSHAFT TORSIONAL VIBRATION

[RREY]

Hello all. I am new in this forum though I´ve been following some posts for a long time.

We are developing a 4 cyl, N/A 2.0L endurance race engine and so got designed and machined custom billet crank, rods, pistons... Once engine is ready and on the dyno, we´ve found that there is a HUGE torsional vibration from the powertrain!. I´ve neve seen this before!. It happens only at 5200-5500rpm (about 90Hz) and it is so bad that even shears the intake manifld bolts. This is obvioulsy unnaceptable for an endurance engine, specially taking into account that the peak torque is at 5900rpm.


The point is: how is possible the billet crank has such bad harmonics??. How can we move these harmonics to a (much)higher or lower rpm?. How the reciprocating weight can affect this or even compression ratio?

Thank you in advance.

 

[DLite30]

Torsional vibration is like a silent killer, as it doesn't manifest itself normaly in vibratory manner that's easily identified. You can break a crankshaft due to torsional vibration and never realize there's a problem until it's destroyed. It's also difficult to measure, especially on an engine that does not have extremely precise speed control.

So, my guess is that what you're witnessing is a set up of unblanaced forces and a coincident natural frequency of engine components or the test stand itself.

I would recommend starting with weight balancing of the flywheel, crankshaft, etc.

 

[Tmoose]

Inline or opposed 4?
If inline, Did it originally have a couple of balance shafts?
Did you change the rod length/stroke ratio?

Sudden onset and narrow "critical speed" suggests resonance of system, often related to mounting. Might even be how the intake is mounted. Some inline 4s have heavily braced intakes.

A $40 device like this could help identify if the vibration is primarily 1X crank speed (unbalance mostly) or 2x crankspeed ( secondary vibration, inherent with inline 4)

http://www.manddsmallengine.com/briggs/tools/19200.html.

A real vibration analyzer instrument or consultant could provide better more detailed info.

 

[ivymike]

don't know what your budget is, but you might consider having someone analyze your dyno configuration as well as your intended installation. It's possible that what you're seeing is a "dyno-only" behavior related to the mass elastic system of your current installation. Do you have a different dyno coupling you can switch to?

 

[ivymike]

oh, is this a chassis dyno or an engine dyno you're on?

 

[ivymike]

actually, I'm scratching my head a bit after re-reading your post. having a bit of trouble picturing what your setup is. "powertrain" suggests to me that you've got a transmission and driveshafts in the setup... but "engine is ready and on the dyno" suggests you're testing the engine alone. If you're testing the engine alone, then I'm a bit surprised by the behavior you're describing - the whole engine is rolling back and forth by a large amount? Normally I'd have expected your crank thrust bearings to fail, or your dyno coupling, or the crankshaft itself before seeing a large displacement of the block & manifold. Is the engine mounted on compliant feet, then connected to a large-inertia dyno via a compliant coupling?

Is it possible for you to upload a sketch of your system?

 

[RREY]

Hello all and thank you for your replies

- This an engine dyno, no gearbox in it. Engine alone
- Yes, the engine is an inline 4.
- Crankshaft, flywheel and even clutch have been accurately dinamically balanced .
- Vibration only occurs ONLY at that rpm range (5200-5500), not before nor after. Therefore it seems the harmonics match that frequency.
- We never had such an issue with an engine before!!.
- This engine (as OEM) doesn´t have balance shafts for the typical second order vibration in these inline 4s.
- Rod ratio has been changed for a longer ratio. It shouldn´t affect.
- Crankshaft is custom billet 4340 steel. 29bs. Stock is cast, 38lbs

TMOOSE:
I have never seen that device before. I see it is a tachometer based on vibrations. How can it help??.
What vibrations analyzers can you recommend?

We will check this week the behaviour of the engine in the car in case this vibration occurs only when assembled to the dyno (I doubt it).


POSSIBLE SOLUTIONS:

1.- To move the harmonics up or down in the rpm range, we will need to alter the elastic response of the crank, thus its natural frequency. For that, my guess is lightening the crank would be the most straight solution.

2.-On the other side, fitting a torsional damper on the crank end would help as well but that means adding about 7lbs rotating mass!.

Thank you.

 

[rob768]

If the problem is caused by torsional resonance, the behavior will change if the engine is coupled to for instance a gear. This, due to the cahnge in mass-elastic system. How do you know it is a torsional vibration anyway? It may well be transeverse, due to resonance of the set excited by unbalance, excessive vibration from bearing clearance etc.

Measureing torsional vibrations is easy, provided you got the right equipment. We use a high resolution laser device for that.

Don't jump to wrong conclusions!

 

[Strong]

RREY,

I would not conclude/assume that the vibration is torsional without a proper vibration test. If vibration frequency is at 1xSS, it could be a lateral vibration with a resonant support structure. I would conduct a structural vibration test (Impulse-response method) with engine off and then measure lateral and torsional vibrations Vs speed. I measure torsional vibrations with a strain gage telemetry system and with an encoder-FM system.

Walt

 

[rob768]

I would love to see you measure torsional vibrations on a crankshaft of a small engine like that .
But i agree: first establish the source of the problem, not just assume it is torsion

 

[RREY]

Thanks for your replies.

I see you all agree that we might have gone to fast with the diagnostic, I understand.

It is not easy to measure torsional vibration in such small engines as this one (as rob768 points out). "Common" vibration in an engine depends on rpm if it comes from unbalanced masses. Even if the crank is not internally balanced according to the reciprocating /rotating masses you could notice that at certain rpm the "internal" balance is not good but... it will never shear the bolts.

Diagnosing torsinal vibration in these engines is a complex taks, as a bad criminal, it always leaves clues for the smart eye:
- you can clearly see the vibration in the belts running with the crank nose.
- If you lock the igntion timing to , let´s say 15º and run the engine all the rpm range, you can notice that, at the natural frequency, the timing is not accurate and perfectly still as it should be.

Said that, my diagnostic is torsional vibration.

It could happen that these vibration come from the camshafts because we are using really aggresive profiles and the amplitude of the (positive-negative) torque to move the cams is over average indeed. I wonder if, if this is the case, the belt from cams to crank could transfer such alternative torque to the crank.

I´d be grateful if you please tell me some good equipment to measure this effect directly (laser or similar).

Thanks again.

 

[Strong]

RREY,

I measured torsional vibrations on a 2-cylinder in-line motorcycle engine with a magnetic sensor detecting timing gear teeth. An optical encoder might also be used. It is possible on small engines!

Walt

 

[btrueblood]

Agree with strong, I've done it with optical encoders on a 20-hp v-twin engine.

 

[ivymike]

the timing gear / flywheel ring gear methods are the ones we use most often. we also occasionally mount a high-res rotary encoder on the cranknose.

 

[Strong]

RREY,

Have you run engine disconnected from dyno? What type of coupling/belts between engine and dyno? I am guessing that torsional resonance, if present, is probaly not internal to the engine. Of course, we are all working with words without photos, drawings, or Vibration Data!

Walt

 

[Tmoose]

The adjustable reed tachometer would help identify frequency(s) of vibration, and to some extent the direction and amplitude.
Some frequency information is Better than no frequency information at all.

In the early 70s A local balance shop balanced the rotating assembly of an Opel inline 4 that a local speed shop had hot rodded via boring, stroking, and adding a brace of webers on special manifold. The engine reportedly had a serious vibration problem that included "shaking the manifold off the engine." They checked the component balance and found everything as intended, but offered some cash as a good faith compensation. I started working there a few years later, and a few years after that we started doing field balancing and eventually vibration analysis.

A few years later a local fork truck distributor came in with a new model of a respected brand that had been receiving big complaints from operators that the vibration thru the seat and controls was unbearable. We spent some time doing a vibration survey using the old swept filter BALMAC 222D. There was indeed very objectionable vibration when operated at some rpm, but the "Vibration profiles" showed whopping spikes at 2 X crankspeed. The new forktruck had an inline 4 cylinder almost rigidly mounted as before, but the new engine was much more powerful as a result of a several hundred cc displacement displacement increase to over 2 liters. I did some research into engine theory, and recommended to my boss that better engine balance would not help (because the high vibration was at 2X), and made up some experimental isolation mounts for the seat and controls that took out most of the sting. I think the manufacturer re-did the engine mount design.

At that point The Opel engine was long gone, but when I heard about it I strongly suspected that our vibration analyzer would have indicated the onjectionable vibration was at 2X crank rotating frequency. I'd bet the engine stroke was increased but the stock length rods were used, which increases the acceleration disparity at TDC and BDC and thus the secondary vibration would have increased by some significant proportion. Add in possible fatso cast Jahns hot rod pistons and the secondary forces would increase even more. Who knows what engine mounting resonances, cantilevered manifold resonances, or even bending modes of the engine/transmission assembly were waiting eagerly to be titillated by the frisky new engine with stronger secondaries and extended rev range.

4 cylinder Secondary vibration is approximately a force equal to one unbalanced piston yanking the crank upward twice per crank revolution along the cylinders' centerlines.

 

[SomptingGuy]

Since nobody else has mentioned it...

Do not confuse torsional vibration of the crank (i.e. twisting) with instantaneous speed variation of the whole crank through the cycle. I4 engines will always show significant 2E speed variation, caused by pressure forces or inertia forces, depending on engine speed. Most automotive I4 engines will have a point at around 2500 rev/min, where the two forces cancel to give very low 2E. Then as the speed increase, inertia forces dominate. Secondary balancers are required to eliminate 2E inertia forces in an I4.



- Steve

 

[GregLocock]

It seems unlikely to me that a crankshaft of conventional design for a 2 litre would have a primary torsional mode at 90 Hz, UNLESS you have coupled the entire driveline inertia in - usually that is decoupled via the clutch assembly.

Any competent engine lab will be able to measure TVs on an engine like that. Equally you should be able to get a fair stab at the analytical modal frequencies, using Holzer or FEA, or even an experimental modal analysis.







Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[jeyaselvan]

Try to narrow down on whether your excessive vibration is torsional or linear vibration related. You have a lot of bouncing force arising out of 4 cylinder inline engines, which do not have 2nd order balancers.

If torsional, I believe the first thing you need to do is to find whether your excessive torsional vibration is due to resonance or due to your torsional excitation by itself. a) If this is due to resonance, if possible, then you may shift this torsional natural frequency away from your operating speed (for constant speed applications.As Greg points out, we can predict torsional natural frequencies and modeshapes more closely.) If this is not possible, then you may need to do a full blown torsional analysis and check for torsional stresses (do not forget to include stress concentration factors) to ensure reliability. Excessive torsional vibration is by itself no damaging,if not related to drive comfort b) But if this is due your excitations, then there is no easy solution other than by altering the stiffness (playing with shaft diameters & lenegths) & inertias (may be a larger flywheel) in the drive train which is what is characterised as torsional design.

Either way, I suggest you to do full blown torsional response analysis with your torsional excitations to ensure reliable designs.

Regards
Jeyaselvan

 

[Tmoose]

Hi RREY,

any developments on the engine vibration?

Dan T