Camshaft lobe design

[SchmidtMotorWorks]

BTW, the port velocites and waves are caused by the cam and its design, not the other way around.

I think it is most accurate to say that they effect each other in a continuous cycle, as each has to deal with the consequences of the other. An engine with little duct length has little inertia no matter what the cam timing is and long ducts can't benefit from inertia unless the cam timing capitalizes on it.

Simulation code mimics this by using trial values that are predicted for results and adjust them in a loop(s) until the trial and results converge. Most simulation code operates on this principle.

[RW TECH]

Harold,

Your posting is a very interesting read, and the way you express this information makes me want to comment on the fact that you are a true gentleman in all of your posts.

Sorry if I'm highjacking the thread, but this stands out to me for a good reason. One of my best friends was killed back in August (Mike Browne a.k.a. Wires & Pliers).

I was thinking about him a lot today, and one thing that Harold's post reminds me of is an expression I heard from Mike's uncle (Dick Chrysler from Cars & Concepts, known for the Hurst Olds, plus stints in Congress, etc.). Dick said "You Will Attract More Bees With Honey Than Vinegar".

I have no opinions whether you cams or anyone else's are the best thing going, but you absolutely have a very intelligent method of communicating with people and that's definitely something to appreciate.

[UDHarold]

Jon,

I started seeing this in 1977, before I knew of such things as simulation software. I do not use simulation software now, yet even my hydraulic cams from the early 1980s show inertia ram, and I rarely look at duct length.
I view the engine as a differential machine, and design my cams to maximize the differential. This consists of commonly optimizing the cam at different points, and optimizing might mean maximizing, and it might mean minimizing.
Pressures that exist in the intake port and the exhaust port prior to valve opening affect the velocities in a major way.
The pressure differentials affect how the air is going to flow, and what sort of velocities will be present.

UDHarold


RW Tech,

Thank you, and I am sorry to hear of the loss of your friend. As time goes on, we all lose someone we hate to see go.
Again, thank you for your kind words....

[David Redszus]

I had a concept of using a blend of Constant Acceleration, Constant Jerk, and polynomial curves into what I called a "Multi-Segmented Polynomial" equation, or 'MSP'. Constant Acceleration and Constant Jerk curves are forms of simplified polynomials, so polynomial describes them all.

Harold, I have to confess that I am a little confused. Acceleration and jerk are derivatives and a polynomial curve defines postion. How can they be combined?

When you say simplified polynomials, to what degree polynomial do you refer?

CamKing, I am really confused by your design method. How is it possible to measure port velocity (or valve curtain velocity) in a running engine? And won't the velocities vary greatly with valve size, crank angle and piston speed?

The original question refered to the proper matching of lobe shape to valve train design. Do either of you gentlemen use the same cam lobe for hydraulic lifters as for mechanical lifters? Or bucket followers or finger followers? Since in each case we are trying to produce a certain valve lift curve, would not the lobe shape have to be modified to accomodate the valve train mechanics?

[CamKing]

CamKing, I am really confused by your design method. How is it possible to measure port velocity (or valve curtain velocity) in a running engine? And won't the velocities vary greatly with valve size, crank angle and piston speed?

I'm not measuring it, I'm calculating it, every 1/4 degree of piston movement. When I design a profile for an application, I input Bore, Stroke, Rod Length, Comp Ratio, Valve size, Choke diameter, Desired max HP RPM, and a few other things.

The original question referred to the proper matching of lobe shape to valve train design. Do either of you gentlemen use the same cam lobe for hydraulic lifters as for mechanical lifters? Or bucket followers or finger followers? Since in each case we are trying to produce a certain valve lift curve, would not the lobe shape have to be modified to accommodate the valve train mechanics?

I design the valve lift curve for the engine application. Bucket, finger follower, or rocker arm, I design the valve lift curve the same. I then work backwards through the valvetrain geometry to get the lobe profile. The type of follower's weight, and physical limitations may cause me to rethink the aggressiveness of the cam, but I don't change the shape of the valve lift curve just because of the follower type.
As for the difference between mechanical and hydraulics, All I change is the ramp height.
On both, I design a valve lift curve, starting at zero valve lift. That point will become my hot lash point. If I want a .012" ramp, I add .012 to every point, and them mate a .012" ramp on both ends. With a hydraulic, I use a .006" ramp on both ends.

[UDHarold]

David,

I am not a very good teacher, my wife says I have a tendency to lecture, or talk down, to people, and I do not mean to be that way. Maybe it's just her that is so affected.......
Rothbart's book, which has only one chapter on automotive cams, cira 1955, has excellent chapters on the basic curves you mentioned, as well as one on the polydyne cam. This is 1955 information, so that most cams designed back then were designed by desk calculator, and computers are not mentioned in the book.
Here is how I derive a polynomial equation for a constant acceleration curve, the rest, constant velocity, constant jerk, constant snap, etc, are derived the same basic way.
Acceleration=A, whatever constant you wish to use. Some work, some don't.
Velocity=AX+V, the original velocity at the beginning point of this curve.
Displacement=((A/2) times X*2)+(V times X)+D, the original dispancement at the beginning point of this curve.
X is the degree number, 1,2,3, or maybe .5, 1.0, 1.5, 2.0, etc.
Although I do not use this curve coming off the base circle, if I did, V=0, and D=0.
As you see, this is a 2nd order polynomial. For over 20 years, I used a Txas Instruments pocket calculator to calculate all the points in this curve.
To answer your other questions:
No, unless some racer wants to cheat by running a tight-lash solid in a hydraulic cam only class, with about .004"-.006" valve lash. Rarely you get a case where someone orders a solid lifter cam, and decides setting valve lash is too much trouble, and wants to run a hydraulic lifter. When he does, he ends up with a REALLY BIG hydraulic cam.
Bucket followers are just a sub-species of solid lifter cams, designed with the same programs. I have done a number of them for Kawasakis, Cosworths, etc. Finger followers require a completely different cam program, and extremely accurate valve train measurements. I do not fool with them, as each engine requires its own cam design.
Roller cams are designed with the same program as solids, at least I have for 35 years. The difference is in the cutting of the models or masters. In cutting a roller cam profile, you cut the base circle radius PLUS the roller follower wheel radius(These two make a constant for cutting), then add the lobe lift curve at every degree. You are in effect cutting the path of the roller wheel axle going around the cam.
Solid and hydraulic flat tappet cams are governed by the maximum velocity of the cam design and the radius width of the lifter. Fords have higher peak velocities than Chevrolets, and Chryslers have higher than Fords.
Roller cams are governed by base circle radius and acceleration rates.

Whew!!! This is enough for one night. Think about what I wrote.

UDHarold

[tjs44]

Harold,you have a PM.Tom

[SchmidtMotorWorks]

Cam Design Handbook - H. Rothbart (2004)

This book has a good section by Dimitri Elgin Chapter 16.

If you know the shape you want and want to code it yourself I would suggest using b-splines if you are familiar with their characteristics. If not polynomials are much better explained in cam books.

David F. Rogers, An Introduction to NURBS is the best, he was one of the pioneers but has the rare ability to explain it for the rest of us.

There are some good papers on SAE of how to do it with polynomial's but the some of them require optimization routines or lots of human effort to work out.

There is a good thesis you can buy from a University (IIRC in Finland) by Jouni Loupern or something similar. It describes a way to model cams with b-splines using Genetic Optimization Routines. In the long run, when this is fully refined, it has the potential to find improvements that are outside of human imagination if they exist.

[Cammer]

Members needing reading material on camshafts:

Camshaft Reference Handbook by Don Hubbard
__________________________

For good vibrations!

“Vibration of Cam Mechanisms” by M P Koster, Macmillan Publishers, Ltd., London, 1974

“Vibration Analysis of Cams” by C N Necklutin, Trans. 2nd Conference, Mechanisms. Penton Publishing Co. Cleveland, OH, page 6, 1954

“Vibrations in Valve Mechanisms” by T Warming, ASME paper No. 49-opg-2, 1949 Oil and Gas Power Conference.

[CamKing]

As you all know, Cam King and I have completely different views on cam design, as do almost all other cam designers that I know.

The funny thing is, we often end up at about the same place.

My design method is completely different from Harold's. I'd say they're worlds apart. I don't pay attention to .020"#'s, .200"#'s or "Major Intensity" when I'm designing a profile, but the results come out very close.

I remember back when the hot 2bbl cams were Harold's NF73 and my M74359.
NF73: 279@.020", 251@.050", 162@.200", .360" lift, .016" lash
M74359: 281@.020", 252@.050", 163@.200", .359" lift, .018" lash
The difference in lash, makes these cams almost identical in the engine.

I also remember when the 360 sprinters were running Harold's NR71 and my R76434 on the intake.
NR71: 284@.020", 255@.050", 181@.200", .4347" lift, .014" lash
R76434: 286@.020", 256@.050", 182@.200", .434" lift, .018" lash
Again, the difference in lash would make these almost identical in the engine.

Harold and I don't agree on how to design a cam, and we don't do anything the same in the design proccess. I couldn't design a cam the way harold does, and I have nothing but respect for the results his cams have produced.

[gazunk]

"Maybe I'm just really lucky, but I've watched teams of engineers from GM with all the latest software fail to come within 20HP of my Indy car designs"

Cam design is not this simple. It may work well but it is not likely to be optimized. I doubt that GM has "teams" of cam designers working on Indy engines. What Indy car engines are you referring to?

[MadBill]

Historically, what later proves to be a completely wrong theory ("The Sun orbits the Earth."), neatly explains observable phenomenon and accurately predicts outcomes, e.g. eclipses.
More recently, the best brains of the twentieth century tirelessly sought more data to determine if the expanding universe had enough momentum to overcome gravity and expand forever, or if the expansion would slow, halt and reverse, eventually producing a Big Implosion.
What was discovered instead is that many observable astronomical phenomenon ("Not only is the universe expanding, it is accelerating!") are explainable only by the presence of an overwhelming percentage in the universe of Dark Matter, undetectable by direct observation and hence also only theoretical at present.

What does this mean as far as cam design? Bleeped if I know. Guess we'll have to check back in another 20 years and see how it's being done then!

[gazunk]

"Do you account for the change in port pressure from inertia and waves?

I don't see how this would be possible without a full-fledged simulation program."

This is in fact mostly correct. Many people can use a simplified equation to develop good robust camshafts that can yield very good results. However, once you get into highly tuned engines, like those currently used at Indy, the game changes. A cam designer must now be able to look at cylinder, inlet manifold, inlet port and exhaust port pressures (hugely affected by exhaust system geometry) as well as the PV curve for each cylinder to be able to move the ball up the field. The last 2 - 3 % in power is the toughest to come by.

[CamKing]

What Indy car engines are you referring to?

Well there was the Buick V6 turbos, then GM's IRL Gen-1, then Gen-2, then Gen-3, then Gen-4 engines.

[OldSStroker]

"Do you account for the change in port pressure from inertia and waves?

I don't see how this would be possible without a full-fledged simulation program."

This is in fact mostly correct. Many people can use a simplified equation to develop good robust camshafts that can yield very good results. However, once you get into highly tuned engines, like those currently used at Indy, the game changes. A cam designer must now be able to look at cylinder, inlet manifold, inlet port and exhaust port pressures (hugely affected by exhaust system geometry) as well as the PV curve for each cylinder to be able to move the ball up the field. The last 2 - 3 % in power is the toughest to come by.

Well said.

Mother Nature is not a "one equation" kind of gal, at least not in my experience with Her.