F1technical.net

Hi everyone,

I couldn't think of a better forum on the planet to ask than this one when it comes to this engine I'm putting together, I'm at an impass regarding the strength of springs to use, so I actually require some professional guidance as I do not know the formula used to calculate necessary spring pressures! The manufacturer for these cams really has no idea, they just make wonderful cams and expect Ferrari people to know this sort of thing. I built an engine similar to this a while ago, but I cannot recall the springs or spec I used for them.

I'm building a 2 valve 308 Ferrari rallye engine stroked to 79mm stroke with 83mm bore, displacement is just a little over 3.4L.

For all Ferrari engines, the camshaft directly actuates the tappet. STOCK springs employ about 63lbs pressure at the seat and 190lbs at .350" lift (8.89mm), these are good to 8k rpm with the stock cam profile, which doesn't use too radical of a ramp to open the valves. Duration for the stock cams are about 20 degrees less than the figures below but the ramp is not as radical nor is the lift nearly as high.

The new billet cams are of the following specification:

duration at .1mm = intake 291° exhaust 283°
duration at 1mm = intake 254° exhaust 246° -
intake lift = 10.75mm exhaust lift = 10.00mm

operating range up to 9000rpm, expected power output is in the 380bhp range. stock springs will not be considered for this as I'm sure this cam ramp will loft the valve and at max lift there is not enough spring clearance for me to be comfortable with.

TOTAL weight the spring will have to control (valve, tappet, retainer, collets, spring) is about 215-230 grams, nothing titanium being used, the tappets are standard 37mm Ferrari shim over bucket tappets identical to those used on countless VW, Volvo, Fiat, the valve is a touch lighter than stock and is 2mm oversize.


anyone who is able to offer their studied opinion would be greatly appreciated

The basic equation to look at is: spring force > moving mass * acceleration

So, you should try to measure the exact profile of the cam as a function of the camshaft angle, as this defines the displacement of the valve. Then replace the angle by the coresponding time (=time for one revolution of the camshaft at 9000rpm of the crankshaft). The maximum of the second derivation of this function is your acceleration.

exact cam profile below for those who might be curious

explain measuring the profile of the cam as a function of cam angle? I'd like to know how to do this, but I'm not quite sure... I'm doing some research this evening on it though for sure. Also going to try and find how to measure cam acceleration, or at least typical acceleration for a cam with specs about like this one.

.0133 second for 1 cam revolution at 9k rpm seems right. 75 cam revolutions per second at 9k rpm

moving mass in grams I assume? so in this case we'll say 220g as this is the total mass the spring will control

and I'm guessing this will calculate what I'm looking for in force or NM?

sorta related. Has a 308 and talks about cams

guy is making a 12 cly 4valve for his 308

the madness is fascinating

http://www.ferrarilife.com/forums/proje ... n-v12.html

flynfrog wrote:sorta related. Has a 308 and talks about cams

guy is making a 12 cly 4valve for his 308

the madness is fascinating

http://www.ferrarilife.com/forums/proje ... n-v12.html

I know Mark very well, we chat about the ports for various Ferrari heads and flow/cams and that sort of thing often. Brilliant guy, I should probably bug him regarding this too.

The time required for this intake lobe to accelerate from 1.2mm to 10.75mm at 9k rpm is .004703, if my math is correct. That is calculating 127 degree duration from 1.2mm-10.75mm required per the lobe design. Working on the rest, I still don't quite understand some of your post Blanchimont, specifically cam angle measurement and replacing that with time and the second derivation of that function.

ask the cam manufacturer

strad wrote:ask the cam manufacturer

That would be the easiest way to get the correct spring!

If you want to measure the profile by yourself, you need to be able to place a clock gauge at the camshaft ( http://www.peachparts.com/shopforum/2619802-post18.html ). Make sure it is positioned properly, which means it has to point at the camshaft turning axis and be perpendicular to the camshaft turning axis. Then measure the lift of the cam for every 5 or 10 degrees (or as in the link at every tooth) for one revolution. Now you have the data points (X = angle of camshaft, Y = lift).

Replace the angle by the time, where 360 degree is the same as 60s/9000 x 2 = 0,0133 s in your case. Now you have data points related to time. Try to get a function from these data points with the help of Excel/Matlab/any suitable program.

Once you have access to this function, you need a little math to get the second derivation, which is the same as the acceleration!
If your function would be (just an example): 0,005m x sin(t/0,0133s),
then the first derivation would be: 0,005m x cos (t/0,0133s) x (1/0,0133s),
and the second derivation: 0,005m x sin (t/0,0133s) x (1/0,0133s)² x (-1)

The maximum of the second derivation between t=0s and t=0,0133s is the acceleration you want to know.

Multiply the acceleration [m/s²] and the moving mass [kg] of the valve/tappet/... and you'll get the maximum force in Newton. Add a bit to the result to account for elasticity in the system and then chose an appropriate spring.

That doesn't account for spring surge, or valve bounce. I'm a bit surprised that there isn't a commonly available cookbook approach, similar to what you propose.

Ferraripilot-

It is no simple matter to design a reliable set of valvesprings that are capable of operating at 9krpm for a high-performance engine that was originally designed to operate at a very high 8krpm. The most difficult aspect of your spring design is getting the spring(s) to fit within the existing diameter and height space available, while also providing the required installed force, linear rate, natural frequency, travel without binding, and fatigue life. Unless you have substantial experience designing valve springs the learning curve can be quite expensive, when you have to replace engine parts due to valve spring failure.

Since you are building a high-performance Ferrari engine, I can assume you have a fair amount of financial resources. It is actually quite quick and easy to get a custom designed cam and spring set from companies like Comp Cams. All you need to do is provide them with some basic information about your engine and how you intend to operate it, and they will provide you with a properly matched set of cams and springs. Usually for a very reasonable price.

riff_raff wrote:Ferraripilot-

It is no simple matter to design a reliable set of valvesprings that are capable of operating at 9krpm for a high-performance engine that was originally designed to operate at a very high 8krpm. The most difficult aspect of your spring design is getting the spring(s) to fit within the existing diameter and height space available, while also providing the required installed force, linear rate, natural frequency, travel without binding, and fatigue life. Unless you have substantial experience designing valve springs the learning curve can be quite expensive, when you have to replace engine parts due to valve spring failure.

Since you are building a high-performance Ferrari engine, I can assume you have a fair amount of financial resources. It is actually quite quick and easy to get a custom designed cam and spring set from companies like Comp Cams. All you need to do is provide them with some basic information about your engine and how you intend to operate it, and they will provide you with a properly matched set of cams and springs. Usually for a very reasonable price.

Thanks for the posts everyone. Regarding spring surge and lofting the valve and valve bounce as someone mentioned earlier, I believe the status quo is to go for springs 5-10% stronger than the required calculated value in NM might be, correct?

Comp Cams and Crane Cams are both companies I've worked with before for other projects, but they've made it known to me that they have no interest in making Ferrari cams as they are very difficult to make (the cams are hollow so they must be drilled, have lots of oil holes and various slots in the journals) and with all the custom race team work these companies already have they're just not interesting in the learning curve.

I use Cat Cams in Belgium, they have been making nice billet cams for Ferrari's for a while now, the only issue is they really don't know anything else about Ferrari valve trains other than how to make the cams and they're not really interested in doing the research required to come up with a spring recommendation. There's no easy way unfortunately other than crunching the numbers.

I've weighed the components exact:
44MM diameter intake valve 115G
stock lifter with shim 62G
stock retainer 15G
collett 5G
stock valve spring 45G
242G total


I've done quite a bit of research and digging and I *think* a spring with the following specs will work out:

weight is just a few grams lighter than stock
11.8lbs per mm
installed height, force at the seat: 90lbs
fully open force 210lbs


I probably could use a lighter valve spring and new lighter retainers and remove some 30G from the equation and use a slightly weaker valve spring to reduce parasitic drag loss.

Does this look about right?

Ferraripilot,

What you need to do for starters is a basic stress analysis of your proposed spring design. There are a couple free online tools that you can use to do a basic check of the stress levels in your spring. You will need to input basic variables like spring rate, installed height, free length, number of active coils, wire diameter, travel, solid height, allowable working stress limit of the spring material, elastic modulus of the spring material, and mean coil diameter.

http://www.efunda.com/designstandards/s ... signer.cfm

Thank you for that. I'm still bothering my cam manufacturer for more specs regarding the cams if they're not willing to run a proper analysis to determine what's needed. Very frustrating.

In the meantime, I've done some comparative analysis regarding other engines of similar valve configurations and their spring requirements given a similar camshaft specification as I provided earlier. It turns out the early Ford RS2000 8v head uses a very similar valve size arrangement and does not use a cam follower bucket as Ferrari uses, but rather a sort of rocker type follower which weighs about the same as a Ferrari bucket follower meaning the mass they needed to control with the valve spring is about identical. The mentioned Ford engine in a mild rallye configuration can get away with running 85lbs on the seat and 195lbs at 10mm lift. Seems reasonable enough, and I'm all for reducing any parasitic additional spring force if possible.

Ferraripilot-

When it comes to valvesprings, there is a huge difference in the profiles of the cams used for finger lifters versus buckets. A finger follower that uses a roller or convex pad allows a more aggressive lift profile to be used on the cam lobe than a flat bucket does. The finger follower also allows a larger outside diameter for the valve spring than a bucket usually does, which makes it easier to fit a valvespring that is not overstressed.

One thing to add in is that the valve spring itself has mass which needs to be moved. Remember as the valve moves in and out its mass includes a portion of the spring moving. Now the complicated bit comes in that each part of the spring moves at a different rate so therefore it adds some complication to the overall momentum calculations for the spring and valve.