Camshaft design

[Keith Morganstein]

A website with interesting articles about cam design and history.

http://tildentechnologies.com/index.html

http://tildentechnologies.com/Cams/CamHistory.html

[Geoff2]

Very interesting read. Ford were using asymmetric lobes in 1949. Hmm. So much for them being 'modern' designs.

[DaveMcLain]

Very interesting read. Ford were using asymmetric lobes in 1949. Hmm. So much for them being 'modern' designs.

That article is interesting. A year or two ago I posted some stuff where I had plotted some cams from flat head engines like a Ford 8N tractor. The designs seem to be surprisingly radical until Chase Knight pointed out in the thread that you have to remember that since the engine has no rocker arm the camshaft lobe itself has to do all of the work to produce the valve motion. Another interesting thing about the 8N cam is that it does have a constant velocity area right where it comes off of the base circle. One time I had one of those engines going together on the engine stand. The valves hadn't been adjusted yet and were all set very loose. I could take a wrench and turn the cam by hand and when I'd do that the valvetrain was astonishingly noisy. It sounded like a teletype machine. Once I set the valves properly it was quiet.

[Stan Weiss]

This is a page from "Valve Gear Design" by Michael C. Turkish 1946

Stan

Turkish_Cam.gif

[hoffman900]

Larry has a nice page with a lot of good math. He's also a nice guy.

We've had some very good discussions on here about the math behind camshaft design involving Jon Schmidt, the late Harold Brookshire, Mike Jones, and Mike Shoe (spl?) (dacaman12). Also lots of good discussion on cam selection for different applications.

See here:

Camshaft Lobe Design
viewtopic.php?t=9026

Cam Design Basics:
viewtopic.php?f=1&t=6807

Velocity, Acceleration, and Jerk Graphing
https://www.speed-talk.com/forum/viewtopic.php?f=15&t=6615

Cam Acceleration
https://www.speed-talk.com/forum/viewtopic.php?f=15&t=8148

Cams with Dwell at Maximum Velocity (with acceleration curves of some of Harold's cams)
https://www.speed-talk.com/forum/viewtopic.php?f=15&t=11713

Which Has a More Aggressive Cam Profile? Tappet or Roller?
viewtopic.php?f=15&t=16426

Understanding Design of Motion Graphs Like Cams (Jon put together an excellent exercise in his first post. Download and give it a try! As he points out, Harold did this in the '70s by hand and with punch cards - no visualization)
https://www.speed-talk.com/forum/viewtopic.php?f=1&t=46446

Take-up Ramps and Valve Lash (with graph)
viewtopic.php?f=1&t=12662

Camshaft Guy's
http://www.speedtalk.com/forum/viewtopi ... =1&t=16691

Jerk at Lift Peak
http://www.speedtalk.com/forum/viewtopi ... =15&t=8284

? On Intake to Exh Ratios in Relation to Cam Lobe Design
viewtopic.php?f=15&t=6949

[panic]

Very good remarks everywhere.
What I find relatively ignored: there are 2 very different factors involved in cam design which are frequently confused and/or conflated:
1. the dynamics of controlling a reciprocating mass (especially the valve) at speed, and making it be where you want, and
2. where do you want it, when, and why

1. determines what's possible before you break something, spring load, compliance, etc. but may differ significantly from what develops power
2. is what the engine wants, regardless of how it's done - frequently not possible

[hoffman900]

Very good remarks everywhere.
What I find relatively ignored: there are 2 very different factors involved in cam design which are frequently confused and/or conflated:
1. the dynamics of controlling a reciprocating mass (especially the valve) at speed, and making it be where you want, and
2. where do you want it, when, and why

1. determines what's possible before you break something, spring load, compliance, etc. but may differ significantly from what develops power
2. is what the engine wants, regardless of how it's done - frequently not possible

There are a lot of people out there that can design camshafts, but they're mostly in industry designing production machines or designing curves for robotic motion. Thinking in terms of a robotic arm - they typically are working in 3D . Very few want to work in the automotive world and fewer still understand how to design the curves AND pick out the events. Guys like Harold, Mike, etc. were/are essentially one man operations. So they design the lobe for an application, put it out on the street, and get feedback from the builders. Over time they get an understanding of how this goes down. On the other side, there area lot of 'custom camshaft' people who are just picking out lobes from a camshaft company's catalog, but couldn't tell you where to start with designing a lobe.

Computers have made this so much easier.

Check this out: http://fourstrokedesign.com/content/cas ... traints-2/. Four Stroke Design is Brian Kurn's (Roush & Yates, formerly ECR and Hendrick) former consulting business and some work he did for Arrington Performance.

As a background, both Harold Brookshire and Mike Shoe (spl) designed cams for Arrington under 'Custom Camshaft Company' (for NASCAR Dodge engines, NHRA Pro Stock, and other applications) up until all that was sold to Howard's Camshafts in 2010-2011 timeframe.

Brian shows how computers have made cam design so much easier. You can build in all these physical constraints (valve spring velocity / acceleration, piston-to-valve clearance based on the ideal simulated events, etc.) then you can work with these constraints in a cam design program to create the lift curve to work within this box. With visualization, you can just match up the curves to fit into this box, then plug and play in your dynamic valvetrain simulation model, and this was in 2012. As we head into 2018, machine learning is growing exponentially. We're approaching the point (or are there) where we could build in all these constraints (acceleration, velocity, jerk, etc.) pick the starting and ending duration and lifts, and the computer can maximize the area of the curve (using several curve fitting techniques and using the best) while staying within the box. There is a reason all your Masters and PhD physics and math graduates are moving into the machine learning field. They come up with the algorithms, and the software engineers simplify it a bit and convert to useable code for applications.

The process is MUCH more simplified. Harold started out by fitting the ends of polynomial equations by hand and calculating them out with punchcards at 2am on Sunday mornings. These guys then had a rudimentary understanding of valvetrain dynamics as well and were inventing as they went. As a millennial who grew up in the age of Matlab and beyond and equipment like Spintrons to validate your work, that's just insane, and why I have a lot of respect for the guy. The cam designers of old really worked for it and made too little for what they were capable of intellectually.

If anyone has any siblings, nieces/nephews, or grand kids doing this kind of stuff, get them involved! You see a lot of bemoaning how the youth isn't into cars, but give them a reason. They may see flowbenchs and the stuff as rudimentary (and they are), but get them involved in writing some machine learning algorithms and they may see the joy in this as a hobby.

[hoffman900]

This is a good read as well

Development of Valvetrain for Formula One Engine (circa 2009)
www.f1-forecast.com/pdf/F1-Files/Honda/F1-SP2_09e.pdf

Check out section 3.3: Optimization of Cam Profile, with the use of a Bezier curve.

I've taken it out of the above for ease of reading:

[pcnsd]

I have read the laments on the lack of substance at ST these days. I will disagree. There have been a number of high end discussions in my short time as a member and this is certainly one. Don once issued a "Best of Speed talk" compendium for a given year. I nominate this thread for best of 2017 consideration. Herein already is a course of study that will keep your mind busy for at least a few months of spare time. Longer for such as myself.

[hoffman900]

Some more:

A Good Book on Automotive Cam Design:
https://www.speed-talk.com/forum/viewtopic.php?f=15&t=36080

Camshaft Design
viewtopic.php?f=15&t=16183

Vizard on 106* LSA vs 112* LSA in a BBC
viewtopic.php?f=1&t=16186&sid=35999bcfd ... 5d12b11bda

Camshaft Design Question (UDHarold?)
https://www.speed-talk.com/forum/viewtopic.php?f=15&t=7151

Valve Timing Event??
http://www.speedtalk.com/forum/viewtopic.php?p=73892

Tight Lobe Separation Angles
viewtopic.php?t=938

What Makes a Cam a Hi RPM Cam?
http://www.speedtalk.com/forum/viewtopi ... =1&t=13975

Flat Tappet Cam vs. Roller Cam, Early Ramp Rate Question?
viewtopic.php?f=1&t=4967

More Ex Duration
http://www.speedtalk.com/forum/viewtopic.php?f=1&t=9445

Exhaust Lobe: Fast, Slow, Symmetrical, or Asymmetrical.
http://www.speedtalk.com/forum/viewtopi ... =1&t=15431

Camshafts and Valve Springs
viewtopic.php?p=115582

Degreeing an Asymmetrical Cam
viewtopic.php?f=1&t=9555

Cam Duration and Lift Larger on Intake Side?
viewtopic.php?f=1&t=6529&start=15

This should keep everyone busy for a while. I forgot to included Buddy Rawls in the list of names in my above post. He's posted some great stuff as well.

[GARY C]

I have read the laments on the lack of substance at ST these days. I will disagree. There have been a number of high end discussions in my short time as a member and this is certainly one. Don once issued a "Best of Speed talk" compendium for a given year. I nominate this thread for best of 2017 consideration. Herein already is a course of study that will keep your mind busy for at least a few months of spare time. Longer for such as myself.

I would have to agree, out of all the forums I have seen I still think this is the best one out there and thanks to guys like Hoffman that can bring up the older relevant posts you realize that anything that can be covered has been covered here and still is.
On Speed Talks Face Book page they always have links to the most in-depth threads that have been done here.

[CamKing]

Very good remarks everywhere.
What I find relatively ignored: there are 2 very different factors involved in cam design which are frequently confused and/or conflated:
1. the dynamics of controlling a reciprocating mass (especially the valve) at speed, and making it be where you want, and
2. where do you want it, when, and why

That's backwards from the way I do it.
I design the optimum valve lift curve for the power application, then make sure I can control it.

[kirkwoodken]

Just want to reiterate something I've posted before. IMO, making a solid roller cam is a very labor intensive process which has no shortcuts. Beside grinding, there is plating, straightening, checking hardness, and checking accuracy of the finished product. I'm not asking the grinders/designers to raise their prices, but don't ever complain about what a special cam costs. If it were not for the few gifted who have chosen to do this work, the rest of us would have some pretty crappy rides.

I am not related to any cam grinders, but have worked in grinding shops for a large part of my life. Holding + or - a "tenth" ain't an easy way to make a living, but it's done every day by someone.

[panic]

That's backwards from the way I do it.
Ooppss - my bad, those numbers not given in priority order, of course the engine's needs are paramount, thanks for clarifying.

If something new comes along to replace oil as we know it (or a non-friction surface treatment) so we can use bigger springs, the dynamic part will become really interesting...

[hoffman900]

That's backwards from the way I do it.
Ooppss - my bad, those numbers not given in priority order, of course the engine's needs are paramount, thanks for clarifying.

If something new comes along to replace oil as we know it (or a non-friction surface treatment) so we can use bigger springs, the dynamic part will become really interesting...

The pneumatic buckets like have been used in Formula One for almost two decades now would be a giant leap.