Quantifying valve float 1

[CrankAngle]

Gentlemen,
I am rebuilding my 12.5:1 CR SCCA racecar engine, and I am installing relatively large lift and duration cams. The is too much subjective speculation with regards to valve float.
I am attempting to quantify valve float. The assumed governing equation is:
Resonant frequency = sqrt(k/m)

It is assumed that the driving frequency is the cam speed, which is half of the crank speed. One valve is treated as a spring-mass system, assuming zero deflection of the valve stem and cam follower and no damping. Also, the sprung mass is comprised of the keepers, retainer, and valve, and 90% of the valve spring contributes to the sprung mass. (The rotational polar moment of inertia of the finger followers is treated as a slender rod, pivoting about one end. Non floating valve motion amplitude is governed by undamped forced vibration. Since we all love equations:

mxDD + kx= Fsin(wt)
This is the governing equation of motion, however how to quantifying the conditions at which the follower leaves the cam is still not evident.

I measured a minimum piston to exhaust valve clearance of 0.025" at 8 degrees BTDC on the closing ramp. (which will be machined to allow 0.080" clearance).

Below are measured values of my race engine:

308-05-0300 Ti Retainer mass: 8.4 grams
Keeper mass: 1.0 grams
311-05-5350 Pro Series valve spring mass: 43.6 grams
(Assuming a linear spring rate of 300 lbf/in) or 52538.05 N/m
OEM B16 Exhaust valve mass: 41.8 grams
OEM B16 Intake valve mass: 46.2 grams
(I still need to determin the moment of inertia of the cam followers).
 
I calculated a natural frequency of 118.4 Hz for an intake valve, and 121.2 Hz for an exhaust valve. This assumes the camshaft speed is half of the crankspeed, thus the driving frequency is 79.16 Hz. Resulting in the intake valvetrain driven at 66.84% of its resonant frequency, and the exhaust valvetrain driven at 65.27% of its resonant frequency. (This is assuming 9.5k rpm crank speed, and 90% of the valvespring mass contributing to sqrt(k/m))


At the peak lift angle, the piston is miles away, and valve clearance is not of much concern.

My question is two-fold:
Is this a reasonable manner to quantify valve motion? Also, should we assume that engine speed, valve spring rate and valve train mass are the only factors to consider? Should the opening and closing ramp rates be explored? (Such that the opening ramp rates can be compared to the spring-mass EOM velocity rates?)

Thanks,
Andrew

 

[patprimmer]

Airflow or air pressure difference on each face of the valve can also have an influence, especially with forced induction.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

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

for site rules

 

[GregLocock]

"Is this a reasonable manner to quantify valve motion?"

No

" Also, should we assume that engine speed, valve spring rate and valve train mass are the only factors to consider?"

No


" Should the opening and closing ramp rates be explored? "

Among other things.

The motion of the valves doesn't resemble a sine wave, so using an equation that assumes a sinusoidal response seems a bit, um, counter-intuitive.

I suppose you could say that any attempt to run a valve at the engine rpm indicated by your equation is likely to float, I'd agree with that, but it is hopelessly optimistic for predicting the onset of float.

Have you looked at Heywood? Have you looked at SAE papers? Have you looked at the IMechE anthology of papers on enegine design? A google search for valve float prediction throws up a few sugggestions.





Cheers

Greg Locock


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

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

 

[pontiacjack]

Nowhere do I see you mention the key parameter of the lobe profile- the maximum negative acceleration. If the camshaft manufacturer won't give you this number, you've got three options:

1> Switch to another cam company.
2> Analyze a pair (INT & EXH) of your cam's lobes with a "cam doctor" setup.
3> Tediously emulate a "cam doctor" with a degree wheel and dial indicator (two degree cam intervals, four degrees crank); then manually (also tediously) calculate velocity and acceleration approximations for each interval (only needed over the nose of the lobes).

I've actually done this (using option 3> and all the inertia calculations as you did) and promptly destroyed the valve train, first time on the track. I've subsequently used spring values advised by the cam company; I suggest you do the same.

[I have to assume that I erred by ignoring forces (friction, lube viscosity, cylinder pressures, blower pressures, etc.) other than inertia.]

 

[dicer]

If there is a spring that is a huge variable. If you have ever seen high speed video's, its apparent.

http://www.youtube.com/watch?v=yfmb-tCo2yA

http://www.youtube.com/watch?v=nsa6kq-qqIE&feature=related

 

[CrankAngle]

Gentlemen,

I appreciate the feedback. I removed the pistons from the engine last weekend to increase the valve clearance. The piston valve pockets will then allow for 0.085" clearance. I am using the camshaft companys' recommended valve springs. The company informed me that the configuration was "tested" to 11,000 rpm. However they were unwilling to share any valve train dynamics data. I have mapped intake and exhaust valve clearance from -20 to 20 degrees ATDC using two dial indicators and a degree wheel.

However measuring valve displacement with dial indicators may be inaccurate as any misalignment with the indicator axis with respect to the valve axis will measure the cosine of the angle between the two axes times the valve displacement.

We have a Spintron Valvetrain Apparatus at my work, however it may be a while before the rig becomes available.

I will measure valve displacement vs crank angle. The magnitude of the measurements might be inaccurate, however the opening and closing ramp rates should be close.

With regards to transient cylinder pressures influence on valve float, I believe the force acting on the valve would be equal to pressure times the valve area. The force would need to be evaluated at the instantaneous cylinder pressure at the exhaust valve opening crank angle. A Kistler pressure transducer should do the trick.

Greg, I have Heywoods book in my office, however I may have overlooked the section where he quantifys valve float. I thought of his book as more of an IC overview from a thermodynamic perspecive. Are there any recommended SAE papers on this?

In terms of the Valvetrain system not acting as a sinusoidal motion, I agree that the displacement as a function of time does not apply, and perhaps a step-response would be more appropriate. However as a spring- mass system, a free body diagram of the system resembles a forced excitation occurring in a periodic manner.

 

[patprimmer]

You do not need the pressure of the air on the valve, you need the average pressure difference on each side of the valve which is a difficult figure to acquire as air is flowing over the valve at constantly changing speeds. You not only might have boost pressure in the manifold, but you might have some vacuum in the chamber.

Regards
Pat
See FAQ731-376 for tips on use of eng-tips by professional engineers &

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

for site rules