calculating valve train spring requirements, help

[Ferraripilot]

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.

Yes, I've sort of quantified the valve spring weight about by placing a rough guess as to what an aftermarket spring will weigh, which is usually at least 10% lighter than a stock spring which is how I deduced the above figure.

To further complicate matters, I could take into account hysterisis. Rather than do all that, I just want a relatively simple kineto static analysis that doesn't consider the transient states and hysteresis but rather just a quick number and I think that will be close (F=ma).

A huge problem I'm having is I cannot get the exact cam lift curve for the valve so to take the second derivative. The lift curve is essentially position. The second derivative I want will give acceleration. This camshaft manufacturer does brilliant work, the best I've ever seen for Ferrari cams, but their communication is less than great.

[Ferraripilot]

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.

So the convex finger follower, if it weighs about the same as a flat bucket, you're saying it can run more aggressive profiles with equivalent springs and not concern itself with lofting at stages that might concern a bucket follower? That's very interesting and makes sense as I suppose a convex finger can remove some of the radical behavior an aggressive cam profile might yield.

[Ferraripilot]

I managed to stump someone from Eibach springs yesterday, I believe the only way to produce an actual analysis with my set of perameters is to either a) do it myself, or b) find an engineering firm who are capable and willing


I'm going to search the university research systems JSTOR and see if there are any published works that are free. Springer Verlag (Netherlands) seems to have a good amount of research on kineto static analysis and non-metriculated status type materials. I will do some more digging tonight, but I have a feeling Springer Verlag may want big $$$ for anything more than abstract examples.

[Tommy Cookers]

my 2cents worth is there is no problem
presumably the original spring is a long way from being coilbound ?
and has a margin of rpm capability and a big margin of (fatigue) life, and is anyway not made from the world's best spring material ?
a few percent increase of wire diameter will give a lot more stiffness and suitably more spring force and more resistance to surge
there is no better valve spring material in the world than the best US material
you will have a better rev limiter than the original and you won't expect the springs to last 10 years ?

[Ferraripilot]

my 2cents worth is there is no problem
presumably the original spring is a long way from being coilbound ?
and has a margin of rpm capability and a big margin of (fatigue) life, and is anyway not made from the world's best spring material ?
a few percent increase of wire diameter will give a lot more stiffness and suitably more spring force and more resistance to surge
there is no better valve spring material in the world than the best US material
you will have a better rev limiter than the original and you won't expect the springs to last 10 years ?

With these cams, the original springs will be within .010 of coilbind so I obviously cannot use them :/ Ferrari used a very short spring with a very high rate (15.5lbs/mm) in stock application as the stock cams had only 8.9mm lift and around 235 degrees duration @ .050 for the Daytona/Boxer/Dino/early 308. Specs are right around 65lbs seat and 200lbs at the nose. They are stout little springs.

Ferrea springs/valves in FL told me today to run 250lbs at the nose and around 80lbs at the seat, with no analysis of the information whatsoever which sort of bothers me.

I'd say if the entire engine lasted 10 years I would have a winner. This engine will be run hard, often, so I expect 5-10 years would be a good lifespan.

[Tommy Cookers]

sounds like you have to try a spring design with 1 less coils than the original (and super material)
that's what is usually done
how many coils does it have ?
or some unusual design, maybe a volute/taper/'beehive' ?, or ovoidal wire ?
if you can find someone to make this

EDIT
profblairandssociates.com

has mountains of free expertise

[riff_raff]

So the convex finger follower, if it weighs about the same as a flat bucket, you're saying it can run more aggressive profiles with equivalent springs and not concern itself with lofting at stages that might concern a bucket follower? That's very interesting and makes sense as I suppose a convex finger can remove some of the radical behavior an aggressive cam profile might yield.

It's not the relative weight between the bucket and finger follower that matters, it's their inertias. The entire mass of a bucket follower moves the entire distance the valve travels. A finger follower does not do the same, since one end rotates about a fixed axis.

As for cam profiles, the cam's radius of curvature is far more limited when using a flat follower than when using a convex or roller follower.

[Ferraripilot]

It's not the relative weight between the bucket and finger follower that matters, it's their inertias. The entire mass of a bucket follower moves the entire distance the valve travels. A finger follower does not do the same, since one end rotates about a fixed axis.

As for cam profiles, the cam's radius of curvature is far more limited when using a flat follower than when using a convex or roller follower.

That's interesting, makes me wonder why Ferrari have remained with that paradigm for so long. Their old V12s obviously used hair springs and a different type of follower.

[riff_raff]

Here's a picture that illustrates the difference between the contact pressure angle of a flat follower versus a roller or convex follower:

[Ferraripilot]

Thank you! Very interesting



Update here, a tuner in Germany I know of has run this kineto static analysis in the past, although they would not provide me with an abstract of the work, they did tell me that a spring with 11.6lbs/mm is required for the cams I will be using and a seat pressure of 82lbs should be employed meaning at 13mm lift I will see 233lbs at the nose. The previous spring I mentioned should work fine then, although it is a bit more stiff at 12.2lbs/mm. Reducing to these other spec springs could probably free up a tiny bit more power....

[riff_raff]

Your proposed valve spring design sounds a bit optimistic. If you'll provide some additional basic dimensions such as max spring OD and installed spring height, I'll do a quick stress check on it for you.

[Greenlight]

You have most of the information that you need to calculate the spring force requirements.

The remaining information has to do with the cam lobe acceleration curve.

I need to know the camshaft acceleration value at maximum lift and the acceleration value when the acceleration curve changes toward a positive value (acceleration reversal point). Additionally, I need the valve lift at the acceleration reversal point.

The sketch below may help to clarify what I need.

[Ferraripilot]

You have most of the information that you need to calculate the spring force requirements.

The remaining information has to do with the cam lobe acceleration curve.

I need to know the camshaft acceleration value at maximum lift and the acceleration value when the acceleration curve changes toward a positive value (acceleration reversal point). Additionally, I need the valve lift at the acceleration reversal point.

The sketch below may help to clarify what I need.

Kind thanks for your post. I requested from the camshaft manufacturer these figures a few occasions now and there has been no response. This company makes a wonderful product but they really fail in offering support!

What I can add however is that the cam features a somewhat aggressive lobe which opens the valves from 1.2mm to 13mm (max lift) in a period of 119 crankshaft degrees which I believe at 9000rpm requires .0133 of a second (I think that's correct), all to control 242grams (it could be 220g if I replace these lifters with lighter units actually, which I will probably do).

I need flexibility for the occasional missed shift up to around 9k rpm. Normal shift point will be probably around 8200rpm.

I've learned more about valve springs these past couple weeks than I thought possible!

[Ferraripilot]

Your proposed valve spring design sounds a bit optimistic. If you'll provide some additional basic dimensions such as max spring OD and installed spring height, I'll do a quick stress check on it for you.

the particular spring I'm sort of being pushed to is as follows:

F1 : 76in (339mm)
F2 : 245in (1090mm)
ID : 0.71in (18mm)
L1 : 1.54in (39.2mm)
L2 : 0.95in (24.2mm)
Lc : 0.83in (21mm)
Max Lift : 0.59in (15mm)
OD : 1.26in (32mm)

I believe this calculates to a rate of 11.26lbs/mm, sounds like it might be a bit low for the application but I could be wrong!

[Greenlight]

You can get the required cam lobe curves from anyone who has a cam checking device such as Cam Doctor, Cam Pro, Dr. Doctor, etc.