Camshaft lobe design

[SchmidtMotorWorks]

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 Engine Pro book on this site (if you don't have it buy it!) has some good examples of formula to illustrate this idea. While it does assume a 2nd harmonic intake length in all cases it does consider for example the effectiveness of plenum resonance in different manifold configurations to partially determine the difference between the duration of the intake and exhaust.
If the program allowed for different intake lengths, I assume there would be an additional factor for that. I have experimented with adding a factor for ram-tuning effectiveness (determined by the length and wave used) to his formula and when you adjust it to match manifolds other than the 2nd wave optimum stuff in Engine Pro, the cam recommendations match very close with what we see in use for non professional classes. Without those adjustments it is much more like pure racing cam parameters.

[David Redszus]

Induction tuning relates to a rpm point, either torque or power, or somewhere between. Camshaft event timing can be selected to be compatible with induction tuning, but has little to do with camshaft lobe shape.

The original question related to lobe shape as a function of valve train mechanics. We still do not have really good answers. But it is a very complicated question.

[MadBill]

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.

A little bird told me that CK's design approach was also 'adopted' by another Brickyard supplier a few years back and enabled their F1 team to "leap" ahead to the tune of 30+ HP...

[CamKing]

The original question related to lobe shape as a function of valve train mechanics. We still do not have really good answers. But it is a very complicated question.

I think we may be on different pages here. Cam designers design Valve lift curves, and then generate the lobe shape to give you that curve.
The lobe shape is generated by taking the valve lift curve, and working backwards through the valvetrain.

After I design the valve lift curve, I work back through the rocker arm to get the lifter motion. With the lifter motion, you can generate the lobe profile. All you need is the base circle diameter, and the radius of the follower(Flat tappet's are 52"-58" radius. I use 999,999,999" radius for a flat over-head bucket follower). I don't care what shape the lobe takes, as long as it gives me the valve lift curve I started with.

[gazunk]

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

How were you able to supply camshafts to the Cosworth-Chevy engines without full knowledge of the engine and valvetrain design details. I heard that even the GM engineers had a difficult time getting this information at the beginning of the program.

[David Redszus]

I don't care what shape the lobe takes, as long as it gives me the valve lift curve I started with.

While I completely agree that the valve lift curve is a critical design objective, it is only one of several factors. The valve lift (actually, valve curtain area) defines air flow at various rpm, but does not define actual valve motion.

Derivatives define dynamics. However well a cam/valve may flow, it could be a problem if the dynamic behavior is wrong. My main point is that engine builders pay very little attention to cam derivatives, that is if they know anything about the effects produced. Then we are repairing or replacing valve train components which were quite adequately designed if the cam was correct.

Since any cam lobe can be mathematically defined (much like a fingerprint), and the derivatives examined, wouldn't it be nice to know much more about how cam will perform before it is installed?

Cam design is after all a science and not a black art.

[CamKing]

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

How were you able to supply camshafts to the Cosworth-Chevy engines without full knowledge of the engine and valvetrain design details. I heard that even the GM engineers had a difficult time getting this information at the beginning of the program.

Wasn't the Cosworth called the Gen-5??
The last one I worked on was the one right before that. I thought it was the Gen-4. It was the dog that caused GM to buy the cosworth.

[CamKing]

I don't care what shape the lobe takes, as long as it gives me the valve lift curve I started with.

While I completely agree that the valve lift curve is a critical design objective, it is only one of several factors. The valve lift (actually, valve curtain area) defines air flow at various rpm, but does not define actual valve motion.

Derivatives define dynamics. However well a cam/valve may flow, it could be a problem if the dynamic behavior is wrong. My main point is that engine builders pay very little attention to cam derivatives, that is if they know anything about the effects produced. Then we are repairing or replacing valve train components which were quite adequately designed if the cam was correct.

Since any cam lobe can be mathematically defined (much like a fingerprint), and the derivatives examined, wouldn't it be nice to know much more about how cam will perform before it is installed?

Cam design is after all a science and not a black art.

Besides the nose radius and inverse radius, I don't look at the lobe shape at all. Look at cam checking equipment like the old Cam Doctors, and Audie Cam Pro systems. They don't check the lobe, they check the lifter motion. That's why they use a radius to check rollers, and a flat surface to check flat tappets. The Acc, Vel, and Jerk #'s are measured at the lifter and valve.
I only look at the motion of the valvetrain. The rates at the lifter to the valve. That's what change dynamics. I design valve motion, and then calculate what the lobe shape needs to be to give me it.

[gazunk]

"Wasn't the Cosworth called the Gen-5??
The last one I worked on was the one right before that. I thought it was the Gen-4. It was the dog that caused GM to buy the cosworth."

Now I understand. I did not realize the flogged the "dog" so many times. I understand that the relationship between GM and Cosworth was a tedious one even though the results were fine and it was mostly invisible to the rest of the world. I did hear that the GM design team was far more aggressive than Cosworth were expecting based on their previous encounters wtih Ford.

[gazunk]

"Cam design is after all a science and not a black art"

With modern technology this is now very true. It is now possible to get a cam about 99% optimized with the very first piece of hardware. No more cut / paste / cut / paste/ etc. With high speed valvetrain test rigs and "in cylinder" pressure data you can do the design on the computer.

The old trial and error method works. The new technology just works better and faster in the right hands.

[CamKing]

The old trial and error method works.

I've never liked this method.
If the first cam I designed for a project didn't do what I designed it to do, I f'ed it up, or was given the wrong info to start with.
The only time this doesn't hold true, is when I'm trying to manipulate the cam profile to bandaid another problem on the engine(restrictor plate, compression limit, ect), or if I don't have all the parameters on the engine.

[CamKing]

I did not realize they flogged the "dog" so many times.

The "dog" was the first one GM had complete control over. The ones before it had independent engine builders free to do their own development.

[JCR]

Cam design is after all a science and not a black art.

UDHarold would beg to differ. His old posts are very much worth reading. viewtopic.php?t=6807

[OldSStroker]

"Cam design is after all a science and not a black art"

With modern technology this is now very true. It is now possible to get a cam about 99% optimized with the very first piece of hardware. No more cut / paste / cut / paste/ etc. With high speed valvetrain test rigs and "in cylinder" pressure data you can do the design on the computer.

The old trial and error method works. The new technology just works better and faster in the right hands.

You have evidently "been there, done that", gazunk, or at least are familiar with OEM practices. I welcome you here and enjoy your posts.

Jon

[gazunk]

Jon,

Both actually. And thank you.