valve spring selection

[AdamVR4]

After doing some searching, it appears that our community is pretty damn ignorant of the springs we put in our heads.

I currently think that I want springs that meet the following constraints:
1) small enough wire diameter to avoid coil bind with aftermarket cams
2) as close to stock stiffness as possible to avoid timing belt problems (overwhelmed tensioner, eccentric pulley loosening)

Determining #2 is difficult by itself because stiffness (k) data is so limited. You're lucky to find it presented like this:


the stock valve seat pressure is 52.9 lbs at an installed height of 1.492", free height of 1.83 for '93+ cars

This drive me nuts because that's a force, not a pressure, and the free height is often not provided, so you can't calculate a stiffness.

But in this case, stiffness, k = 52.9 lbs / (1.83"-1.492") = 156 lbs/in

The only other reliable spring data I can find is from 3SX!

k, 3sx = (they actually give us 2 data points! (http://www.3sx.com/store/comersus_viewItem.asp?idProduct=12128)) = 280 lbs/in

From here, all you would need is the spring mass, retainer mass, and valve mass to calculate the first natural frequency of the valve spring assembly. Does anyone know them?

efunda (http://www.efunda.com/DesignStandards/springs/calc_comp_resonance.cfm) claims you want the first natural frequency of the valve spring assy to be 15 to 20 times greater than the cam frequency to avoid a reduction in spring stiffness as RPM's increase. Based on the following, I guess that sort of stiffness is excessive in automobiles.

If the engine is running at 8000 rpm's, the cams are at 4000 rpm's ... divide by 60 ... 66.67 cycles per second ... ~67 Hz

Without data, I'm assuming a valve assembly weight of 100 grams (If I recall from dynamic systems, you only include a portion of the spring mass, but it's been a while...?). The natural frequency is then calculated using:

http://www.cs.wright.edu/~jslater/SDTCOutreachWebsite/images/gif/tac12.gif
(source (http://www.cs.wright.edu/~jslater/SDTCOutreachWebsite/nat_frequency.htm))

If you're doing the calculations in USCS, use lb/ft for stiffness and slugs for mass. In SI, use N/m and kg.

natural frequency given an assumed assy mass of 100 g's:
fn, oem = ~83 Hz
fn, 3sx = ~111 Hz

Can anyone dig up more data for me? Crower springs? Brian Crower springs? IPS? Dynamic? Masses of ANYTHING? haha ... Similar analysis done for DSM's or Supras?

Thanks,
Adam

[Hans@GZP]

Brian Crower Valve Spring Kit | Ground Zero Performance (http://www.groundzeroperformance.com/product/valvetrain/brian-crower-valve-spring-kit)

Supertech Valve Spring Kits | Ground Zero Performance (http://www.groundzeroperformance.com/product/valvetrain/supertech-valve-spring-kits)

[AdamVR4]

natural frequencies still based on 100g valve/spring/retainer assembly assumption...

OEM - k: 156 lbs/in... fn: 83 Hz
3SX - k: 280 lbs/in... fn: 111 Hz
BC - k: 275 lbs/in... fn: 110 Hz
STstiff - k: 267 lbs/in... fn: 109 Hz
STsoft - k: 254 lbs/in... fn: 106 Hz
STdual - k: 323 lbs/in... fn: 120 Hz

If any of this sounds wrong, please, speak up.

[Boostaddict]

I comprssion tested a handful of springs (stock, crower, web dual, beehive) and posted the compiled data at 3stech. Search beehive.

[green-lantern]

This isn't really relevant to the info you're looking for but I thought it was a pretty good read.

The Spring that Revolutionized Nascar - 2012 Daytona 500 - Popular Mechanics (http://www.popularmechanics.com/cars/news/vintage-speed/the-spring-that-revolutionized-nascar-6643778)

[AdamVR4]

I comprssion tested a handful of springs (stock, crower, web dual, beehive) and posted the compiled data at 3stech. Search beehive.

I found that but you said the x axis (deflection) was misleading and technically incorrect. If you can clarify that I can use your data.

[DocWalt]

In for info.

[Boostaddict]

I found that but you said the x axis (deflection) was misleading and technically incorrect. If you can clarify that I can use your data.

I have all the data still. I believe it was just the X axis label that was misleading. It should have simply been called deflection not deflection from installation height or something. I'll look at the excel data in the next day or so and let you know. Shoot me an email address if you want a copy. Not sure I still have your cell # or not.

W.

[AdamVR4]

I would have updated this thread last night, but as I was going through boostaddict’s data, I noticed that the springs almost universally exhibit some non-linearity. Specifically, our valve springs are less stiff between their free height and installed height than they are between the installed height and (approaching) bind.

If it weren’t for the OEM springs, I probably wouldn’t have noticed this. Based on the OEM engine manual data for ‘93+ springs, you can determine the stiffness in the free to installed height range. Over the same deflection range, boostaddict’s data nearly matched this. His old springs are actually a hair stiffer than ‘93+ OEM (perhaps he tested the ‘91-‘92 springs? They have a different free height than 93+; I need to review this later). However, over the entire deflection range, his old OEM springs were considerably stiffer than the 93+ (in the free to installed height range).

As expected, the conical kiggly springs are very non-linear, so I won’t be including them.

Bottom line, I need to make sure the stiffness data is all taken from similar deflection ranges. Given the relative availability of stiffness data in the free to installed height range, that’s probably what I’ll use.

Also, I was able to find mass data for 1G DSM springs, retainers, valves, and keepers. The total was a little higher than my assumed 100 grams (~125 grams or so).

[Boostaddict]

I believe the stock springs I measured were out of my orginal '93 heads with 100k+ miles, but I've had so many heads laying around and so much intermixed OEM valvetrain hardware (boxes full) that they could easily be newer ones that I purchased used at some point over the past decade or more.

[speedy25]

Interesting display of mathematics.

Heres the things I have known to be why a spring is chosen.

Performance cams have a higher rate of acceleration than stock cams. This can be from lift or the ramps to open or close. The spring has to be strong enough to overcome those forces. The you add in that the heavier springs increase wear in the valvetrain, and rob you of some precious horsepower. Some grinds the stock springs are just fine. The cam grinder should know what you need to install to make it work reliably.

Maybe your math knows how to balance the two.

Carry on with calculations...!

-SP

[AdamVR4]

Good thoughts, speedy.

source: http://www.tech.plymouth.ac.uk/sme/desnotes/valvet1.htm


3. Analysis - Acceleration of Cam Follower
The analysis of anything other than a simple configuration can be quite complex. The analysis will depend upon the type of follower and the detailed geometry. Because of these difficulties with the analysis it was common for accelerations to be determined graphically.

In this section we will make some simplifying assumptions that a knife edge follower is being used. This will not be very accurate, but will give some idea of values.

i) The most simple assumption for analysis is to assume that the opening and closing is simple harmonic motion (SHM). Assume that the engine speed is 4000 revs/minute, this gives a cam shaft rotation speed of 2000 revs/minute (in a 4 stroke engine the cam turns at half the crankshaft speed) so the time taken for 1 revolution of the cam shaft is 0.03 seconds
Also assume that the cam is opening and closing the valve for 120o of its rotation.
Hence the complete valve cycle is completed in 1/3 camshaft revolution, or 0.01 seconds.
The equation describing SHM is: displacement = amplitude x cos(omega x t)

where t = time and
omega is the 'angular velocity' of the system in rad/s, and is equal to (2 x 3.14159)/(the time for 1 cycle)
So the omega value here is 2 x 3.14159/0.01 = 628.3 rad/sec.

Differentiating the expression for displacement gives:

velocity = - omega x amplitude x sin(omega t), then differentiating again:
acceleration = - (omega)2 x amplitude x cos(omega t)

The maximum acceleration occurs when the term cos(omega t) has the maximum value of 1, this occurs at the extremes of the motion.

If the cam has a lift of 20 mm, the amplitude of the motion is 10 mm, and the maximum acceleration is given by:
accelerationmax = (628.3)2 x 0.01 = 3948 m/s2
ii) An alternative assumption is to assume that that the peak of the cam is finished with specific constant radius with a centre at a distance 'e' from the centre of the cam. The term 'e' may be called the eccentricity. The peak radius is often blended to the the base circle by a large radius curve.
In this case the acceleration of a 'mushroom' - flat foot - follower in the vicinity of the maximum cam lift is given by: (cam shaft angular velocity)2 x eccentricity. If the cam base circle is blended to the nose radius by a large radius, the maximum acceleration occurs when the follower is on this larger radius and is equal to the large blend radius x ( cam shaft angular velocity)2

4. Analysis - Valve Train Forces.
If the cam shaft is a direct acting overhead unit then all the components in the (short) valve train will undergo the same acceleration. The exception to this is the spring, for which an allowance may be made by using 1/3 of its mass as an effective mass.

Maximum force will then be:

F = max. acceleration x (masses of tappet + retainer + collets + valve + 1/3 of spring)
At least this much force must be generated by the valve spring at its maximum compression. Use the spring design applet with measurements and / or data from the workshop manual to check this.


So it looks like all I would need is the maximum lift of various cams to determine the maximum acceleration & minimum required spring stiffness. If anyone knows these off the top of their head...



Also, I KNEW I'd seen an effective spring mass before...
The exception to this is the spring, for which an allowance may be made by using 1/3 of its mass as an effective mass

With that assumption, the effective mass of the spring assy (spring, valve, retainer, keepers) is ~90 grams.

[donniekak]

You need more than maximum lift. As mentioned above ramp rate is very important. Many cams with similar lift can have wildly different valspring needs.
A good way to judge the ramp rate from a cam card is to compare total duration, to duration at .050" lift.
For example, a cam with 272* total duration with 210* at .050", is a much milder cam than one with 272* total but 220* at .050". The second cam will break many aftermarket valvesprings, and the first will work with stock springs.
Total lift is important, but more important is how fast you go from totally closed, to totally open, and from totally open, to totally closed.

[AdamVR4]

You need more than maximum lift. As mentioned above ramp rate is very important. Many cams with similar lift can have wildly different valspring needs.
A good way to judge the ramp rate from a cam card is to compare total duration, to duration at .050" lift.
For example, a cam with 272* total duration with 210* at .050", is a much milder cam than one with 272* total but 220* at .050". The second cam will break many aftermarket valvesprings, and the first will work with stock springs.
Total lift is important, but more important is how fast you go from totally closed, to totally open, and from totally open, to totally closed.

Thanks. As you can probably tell, this is the first time I've put any thought into cams/springs/etc, so I'll gladly take anything you're willing to spoon-feed me. I was really hoping to keep the analysis simple and general, but this makes that sound like it's bordering on a waste of time. At least it won't take long :) When it's clear that the results are totally useless, I'll look into a more thorough analysis.