Technical Homemade Roller Cam for Flathead Six

Yes I have, but I didn't know who worked on it, thanks for that info. From what I understand the problem has always been getting the lifter to land on the cam softly throughout the RPM range.

To illustrate I've attached a couple of files. The first, ballistic 6500 rpm, is a cam lobe that is designed to throw the vlv open. On the Lifter lift graph, the black line is the motion the lifter would have if the spring keeps the lifter on the cam. The red line is the motion that the lifter actually goes through. The lifter leaves the cam because the spring is not strong enough to keep the lifter on the cam for this particular combination of mass, spring and RPM. At 6500 RPM, the cam lobe is designed to "catch" the lifter as it lands on the cam.

All looks well, the lifter will gently land on the cam lobe. The lift of the vlv has been increased from a low RPM lift of .504" to a dynamic, high RPM lift of .611"

However, as the second file shows, when the RPM is slowed to 5700 RPM, the lifter slams into the top of the cam lobe. This will quickly end the life of the vlv train and cam.

This is a simple model using a simple spring and no damping. Perhaps a variable rate spring (one that is designed to progressively cause coils to contact each other therefore changing the spring rate as it is compressed) or a specialized damping system could be used.

Although not illustrated here, instead of using this idea to get higher lift, the same concept can be used to change the duration drastically by having the vlv close at some point after the lobe event. But, as Ned Ludd points out, the problem then becomes how to get the lifter and vlv to gently close as the lifter lands on the base circle...specialized hydraulic lifters maybe?

With today's computers and the vast improvement in dynamic modeling over the past 20 years, it may be worth re-investigating.





Thanks
Kipp

 

I have been playing with your program. I find that it is a great visual tool for watching the affects to the graphs as the shape of the profile is changed. When it comes time to actually make the cam profile, a lift table in 1-degree increments showing the lifter height, will be necessary. I do not see this table in a form that shows that. The radius of curvature and pressure angle are also important pieces of data that I did not see. I am looking at this from a point of view that a proper and reliable cam profile can be made with the program. You are well on your way to achieving that.

 

Apparently the trick is dual valve springs in series - one very stiff and one very soft - with some sort of damping mechanism in parallel to them. The bit in Norbye is substantial; forgive me that I don't type out all of it.

 

Thanks! I really appreciate your time and comments.

The program is not finished yet, and the more I read, the more features I want to add...

The beginnings of a lift table is there. Click on either the "Intake" or "Exhaust" tab. Within those tabs are three more tabs, "Graphs", which is the default tab, "Lobe Profile", and "Lifter Movement Table". I believe what your asking for is on the "Lifter Movement Table". Note: The data displayed is the raw data the computer uses, because the mathematical model I used (specifically, two 7th order spline functions with x and y coordinates in the Cartesian graph of the cam lobe profile, as functions of the Cartesian degrees in the cam lobe profile, were 0 degrees is set to be the lobe centerline) is a linear representation of a round object, the computer calculates all points from 3/4 of a turn before lobe centerline to 3/4 turn after lobe centerline, therefore providing a 1/2 turn, or 180 degrees, overlap at the "center" of the base circle. This was necessary to produce a smooth cam lobe where the cam goes from +180 degrees to -180 degrees (in math terms, at the first branch cut for a circular object, where we go from 2 Pi radians back to zero) So from a practical stand point only look at the lifter movement table from -180 to +180 degrees.


Features that I plan to add:
1) allow the designer to manipulate the other graphs and have the computer calculate what the cam lobe would have to be shaped like to attain those manipulations. For example, the designer would be able to move the lifter lift graph and the cam lobe would change in response to the designer's input. I want to allow this feature for the following graphs: lifter lift, velocity, acceleration and jerk, vlv lift, Fourier series harmonics of the acceleration/Force
2) graphs of pressure angle and surface stress (Hertz stress) as a function of cam degrees
3) Harmonic analysis, looking at both the natural harmonics of the valve train and the harmonics of the acceleration graph (in math terms: the Fourier series that makes up the forcing function on the spring-mass system)
4) Regrind feature: if we wanted to regrind a cam, this would allow us to input the profile of the existing lobe into the program and this lobe would appear on the cam profile graph in a different color. Then when designing the new profile, as long as the new profile was inside the existing profile, we would know that it would be physically possible to grind the new profile into the existing lobe.
5) Grinding machine output. Since the grinder will have a different wheel radius than the follower and the geometry of the machine will probably differ from the geometry of the engine (ie, rocker table vs a lifter moving in a bore), the output to the grinder will have to be different than the lifter table. I intend to write code so the designer can input all the necessary dimensions of his grinding machine and the computer will calculate how the grinder needs to move in order to produce the desired cam lobe. This feature would allow for both a masterless CNC type grinder and allow for a master to be made that would then be used on a master type cam grinder.
6) allow the user to select which data they want output from the program in a universal tab deliminated ASCII format that can be imported into other programs or so others can write and build their own CNC machine interface software / hardware.

It is my understanding that Radius of curvature is a useful tool to not exceed the allowable surface stress. So instead of displaying the radius of curvature, I was thinking about having the computer complete the calculation and display the surface stress directly. What do you think? Is there another reason to calculate / display the radius of curvature that I am missing?

Again thanks for the input and suggestions, the more input the better the final product.
Kipp

 

The minimum negative radius is important to know on a roller tappet profile. The nose radius is important on a flat tappet profile. Obviously, the stress in these areas will be high when these values are too small, but I would prefer to know what the actual number is. Also the pressure angle number.

 

For this instalment, harmonic analysis (at least what I know thus far)

We're all pretty familiar with the concept that driving a system at its resonance frequency causes large oscillations (even though we may not have described it that way) For example, a wheel out of balance that shakes the hell out of our car at 55 mph, but below and above this speed the car rides smooth. At that certain speed the car shakes because the spring-mass system, consisting of the spring and the mass of the suspension, tires and car are being driven at their resonant frequency by the imbalance of the tire. Thus driving the system into resonance and large amplitudes.

Another analogy is pushing a child on a swing. If we push at just the right frequency, ie. once every time the child comes back to us, the result is a large amplitude for a relatively small push. But if we tried to push at some other random frequency, we would get slammed into by the child swinging, and the amplitude of the swing would be greatly diminished.

So there are two parts to causing large amplitudes of movement in an oscillating system, one is the natural resonant frequency of the system and the other is the frequency of the driving force. If the driving force frequency equals the resonant frequency, then we get a lot of motion for very little force.

First part, natural resonant frequency:
In a typical engine, the natural frequency of the vlv train, ie the spring-mass system consisting of the vlv spring and the mass of all the vlv train components, is around 24,000 cpm, cycles per minute, which equates to 48,000 engine RPM! Well obviously we will never drive the vlv train at its resonant frequency, so no problem...right?

Well, no. Thinking about the child on a swing analogy again, we could push the child every time they swing back to us, that would be the fundamental frequency, but we could also push them every other time, or every third time, or, as in an engine, every 7th, 8th, etc time. This means that we can drive a system into resonance, and cause large oscillations, by driving it at one of its fractional frequencies or fractional harmonics, of 1/2, 1/3, 1/4, 1/5, etc. of its fundamental frequency. (Note on harmonics: normally we think of harmonics as having a greater frequency than the fundamental frequency. For example a plucked guitar string may have a fundamental frequency of 440Hz. But it will also have harmonics which will have various amplitudes depending on where we strike the string, of 880, 1320, 1760 Hz, etc. for the 2nd, 3rd, and 4th harmonics, respectively. But no matter what, it cannot have a frequency of 220 or 110 Hz, just not physically possible. This is different than the "fractional" harmonics were talking about here. Here we're talking about driving a mechanical system at a harmonic that is lower than the fundamental frequency. So the "fractional" harmonics are at 1/2, 1/3,1/4, etc of the fundamental frequency)

Now the analogy of the child on a swing to explain vlv train harmonics breaks down and we have to dig deeper. The child is moving at their resonant frequency, but a vlv spring that is excited by driving it at some sub-multiple of its resonant frequency doesn't actually move. I mean, the vlv is NOT going to come off it's seat 24,000 times per minute as the spring-mass system is driven into resonance. What does happen, however, is called spring surge. This is like a slinky toy, where we can make a wave appear to travel back an forth along the slinky. The wave that we see is just an area of the slinky in which the coils are compressed while the rest of the coils are farther apart, and this compressed area travels back and forth. The same thing happens in a vlv spring. There is a wave, actually several at each increasing harmonic, that travels back and forth. When the wave gets to the retainer end it exerts a force on the retainer that is greater than the static spring force. This extra force is exerted on the vlv when the vlv is closed and, more importantly, is exerted on the entire vlv train and cam when the vlv is open. If we drive the spring into resonance this extra force becomes excessive and things break!

Now determining the natural surge frequency of a vlv spring, not to mention the mass the spring is acting on changes radically when the vlv is on its seat from when it is being moved by the cam and vlv train, is a bit of a mathematical problem and beyond the scope of this thread. If anyone wants to discuss the math behind this, please PM me and I would be more than happy to share what I know. That said, once we know the natural surge frequency of the spring-mass system, then we also know the fractional harmonics and these are the frequencies we need to stay away from with our cam pattern so we can have parts that live.

Coming soon....Part two: Driving frequency.

Thanks for reading. All Q's and recommendations are greatly appreciated and welcome.

Kipp

 

Part two: Driving Frequency

First some general background on a periodically repeating signal. So around 1800 this dude named Fourier (pronounced Foy-yea, French name) figured out that any signal that repeats can be produced by adding sine waves of different frequencies and amplitudes. I'll use a square wave to illustrate the idea. The attachment shows that a square wave can be approximated by adding its odd harmonics together. The first column shows the sine waves that are "in" the square wave. The second column shows what happens when the sine waves are added together and the third column shows the frequencies and the amplitude of each frequency required to produce the square wave. This column is typically what is meant when people say frequency analysis, Fourier analysis, Fourier Transform, harmonic analysis, etc.

The important part to recognize is that this is much more than just some random abstract mathematical novelty. The fact is that all of those sine waves that comprise the square wave are actually present in the square wave. And in fact, we can filter out any of frequencies present. For example, lets say the square wave is an electrical signal produced by a switch whose contacts are opened and closed at a fixed frequency, like points in a distributor. Lets say the switch is opened and closed 100 times per second, thus producing a square wave with a frequency of 100Hz. Now lets say we wanted to filter out everything except the 5th harmonic. If we run the square wave through an electrical filter that only allows frequencies of say 450 - 550 Hz to pass through, the output will be a perfect 500Hz sine wave even though the original signal looks nothing like a sine wave and certainly does not appear to contain a 500Hz component, yet there it is! Furthermore, if we were to run the signal through a filter that passed only 350-450 Hz, there would be zero output because the square wave does not contain even harmonics, ie the 4th harmonic corresponding to 400Hz. Thus we see that those sine waves are physically present in the square wave produced by opening and closing contacts in our switch.

Okay, back to cam shafts...this means that the motion of the vlv train, which is a periodically repeating signal, contains several sine waves of various frequencies and amplitudes and when these sine waves are added up, we get the vlv train motion. Using Fourier's mathematics we can readily calculate the frequencies and amplitudes that are contained in the vlv train motion.

If we think about pushing a child on a swing again, what's important is not the motion, but rather the rate at which we push the child. In other words we want to know the frequency of the force that is applied. (In math terms: the driving force on the linear differential equation for a driven spring-mass system) Therefore, instead of seeing what frequencies are contained in the vlv motion, its better to see what frequencies are contained in the force that the cam applies to the vlv train. (in math terms, we'll perform a Fourier transform on the acceleration graph, scaled by the mass of the system)

Once we find out the frequencies that are contained in the driving force produced by the cam lobe, we can compare these to the fractional harmonics of the vlv spring. If one of the driving force frequencies matches a fractional harmonic with enough force, we will drive the spring into resonance.

To correct this problem, we simply need to reduce the amplitude of the culprit frequency. One way is to fool around with the lobe design and see what happens, trial and error approach. I'm thinking the better way to solve the problem is to have the computer calculate all the driving frequencies and amplitudes, then it will reconstruct the cam lobe by adding up all the driving frequencies that it just calculated. So now we have a cam lobe that is based on a sum of sine waves. The advantage to this is now we can reduce the amplitude of just the culprit frequency and the computer will produce the new cam lobe with that frequency suppressed.

Q's I don't have answers to yet. Just how much spring surge is too much? In other words, when one of the frequencies contained in the driving force matches the harmonic of the spring, what is the maximum amplitude this frequency can have and still let the engine live? In his book, Don Hubbard states that the resonant factor times the harmonic number squared cannot exceed .12 for pushrod engines and .22 for L-head engines, but he doesn't say what the units on these numbers are so I cant apply them. I'm assuming he came up with these limits by years of R&D. I'd like to take a mathematical approach to figuring out the limits the vlv train can live with. And, of course push those limits...that's hot rodding, right?




As always, your thoughts and recommendations are welcome

Kipp

 

I commend you and your work on yor program,lots of effort here.
Not to get off track and just a suggestion from my background is all I have to offer.
Most of my racing/performance background is in v-twin mc. Now we know s&s bought out crane and I have a long relationship with people who work in r&d there. Now they moved cranes r&d and spintron to viola.
I had a lengthy discussion with the guy who ran the mc portion and from what I got from it was this,I hope it sheds some light.
As he explained harmonics in a platform varies,some engines when tested had to have odd numbers of cylinders assembled in order to test,if not the frequencies caused severe float. The worst offenders were the engines with 45 degree valve arrangements.
Ie a bbc may have to be tested with 4 cylinders mocked up while. Sbc three. Harley needs both cylinders assembled.the other cylinders would cancel the frequency.
Even when a cam ,lifter,spring has evidence of float at 5 k and not at 5100 it could change with a retainer or pushrod weight change.
Then the discussion on beehive springs scared me to the point of never use them in anything except a all out titanium race valve train.
Lots of analysis and available data will get you in the game but actual spintron testing of a particular combination is the only other way to know if the valve train is happy other than test and inspect as you develop.
Andrews sells a cam design programthat seems to be at the cutting edge at a rather pricy sum. Andrews was a good friend of Harvey crane,who wrote the book on witch many cam designers still use today.
Keep us informed as you go,you offer information that is not easily obtained.
Well done.

 

Wow what a thread! Even more wow to Harvey for doing so much the long way around, time, time, and more time..I found an old engine years ago that I am pretty sure was an L head and had a roller lifter/rocker arm arrangement. Once the foliage drops some more I'll go see if I can find it...I know its been tried in the past, air cylinders replacing the valve springs and may still be used in F1..Benifit, no harmonics, takes less effort to open valve and air pressure closes it; per Pro Stock Engineering pertaining to a set up on a Ford FE engine he [Paul Benoit?] developed somewheres in the early 70's.... ...

 

The first attachment is a graph I found on Comp Cams web site. It illustrates the above paragraph.

The force exerted by the spring onto the retainer is comprised of both the static force and the dynamic force. The static force is the force due to compression of the spring. The dynamic force is the wave that travels up and down the spring. The second attachment is a screen shot of a lobe on my cam program. The blue line on the Lifter Acceleration/Force graph shows the static force the spring is exerting onto the retainer. It is based on vlv lift only. (After I do some more reading about how to mathematically model vlv springs, I will be adding the dynamic force to the program.)

The first attachment shows spring force at two different RPMs. The blue 3000 RPM line is fairly smooth. There is very little spring surge at this RPM. In other words, the spring is not being driving into resonance. Because there is very little spring surge there is also very little dynamic spring force and this line looks very much like the static force calculated in my program.

Now the red line, 9000 RPM, is a different story. Here the spring is displaying resonant behavior. One can imagine the small sine wave, which is the dynamic force, as being added to the static force. (Or a fancy way to say it: the dynamic force is superimposed onto the static force)

There's a couple things we can get from the 9000 RPM line. First, we can determine the resonant frequency of the vlv spring. Looking at the first 360 degrees of crank shaft rotation, there are 3 1/2 oscillations of the vlv spring, thus the resonant frequency of the vlv spring is 3.5 X 9000 = 31,500 cpm (cycles per minute)

Second, we can determine the actual force on the vlv train. The static spring force is the same for all speeds and can be read from the 3000 RPM line to be 350 lbf with the vlv fully open. Without the dynamic spring force, this would be true at 9000 RPM as well. However, with the dynamic force, the actual force on the vlv train is 400lbf. If our vlv train can only handle say 375lbf, by looking at static force only, we would think that everything would be fine, but in fact, it will break because the actual force would exceed the limits of the vlv train.

Third, increased chance of vlv float. Notice at the beginning of max vlv lift, the 9000 rpm line is lower than the 3000 rpm line, this means that the spring is exerting less than the expected static force in this area and may not be enough to keep the lifter on the cam, thus leading to vlv float.

Fourth, vlv bouncing on its seat. Earlier I said, "I mean, the vlv is NOT going to come off it's seat 24,000 times per minute." That is generally true. But looking at the 9000 rpm line, we see that the seat spring force drops from a static 125 lbf to a low of 75 lbf. If we drive the spring into more resonance, the seat force could go all the way to 0 lbf. If we drive it even harder the end of the spring will actually come away from the retainer, and there would be nothing to keep the vlv closed, except cylinder pressure!

Knowing this info about this particular set up, we can reduce the dynamic force oscillations by analyzing the component frequencies in the Acceleration / Force graph, which is the force the cam is applying to the vlv train, and see which component frequency corresponds to a fractional harmonic of 31,500 cpm and reduce that particular frequency. That may lead to other unwanted stuff, less duration, less lift, etc. and there will have to be a compromise at some point, but we can have greater control over the design of our cam knowing the resonant behavior of the spring.






Kipp

 

I didn't know that, I simply assumed the best way to do it is in a completely assembled engine, ie all vlv train components, pistons, etc assembled and drive it externally so you don't get noise from combustion in your recorded data. Interesting...

If you get a chance please post some pics of this engine. I would love to see that set up. I'm sure some other guys on here would like so see it as well.

Thanks
Kipp

 

Wow. Just caught the head thread today. Just making sure all is well. Bumpity bump bump. Sounds like a pretty cool cam thread.

 

Pun intended?

 

I'll have to get back to this one, although I followed the cylinder head thread for a while, I lost it........Could someone share a link, Please?



Edit:

Disregard, I found it, and saw that he be busy moving to a new crib!

 

ok, I know its an old thread , but its an amazing thread, a great read if your into cams.
I have to ask, with the Op's obvious intelligence level, why are you designing a cam that has been done over and over and over, when in reality, you say you want instant opening and closing of the valve?
Why not use the luxury of todays technology and your massive brain and make a set up that opens and closes the valves with selenoids?
Am I being crazy, why wouldnt something like that work? very little wear and tear, instant programmable operation of valves. Does the motor need the delay caused by ramps to "wait" for the piston and engine cycle to get where it needs?
I guess I answered the question myself, I dont think the motor rotation would keep up with the valves. crap I dunno, just thinking

 

Last I heard with the forum Redux he was, I believe "discouraged" is the proper word, from posting anything more on the HAMB that involved CNC, Computers or other forms of modern electronics.

I tried to entice him over to Dogfight, as I would love to see more progress on his home built cam grinder, but I don't know that has taken.

 

Alright, Alright.....

I'm still alive and not "discouraged!" in the least bit, just been busy as all hell. Moved to VA, had to settle for a place without a garage. UGH!,

Updates at the T-Roadstrer house: House needed stuff in preparation for a proper garage, ie, I jacked up, build a floor for, and moved the existing 16X20 garage and a 12X18 finished building accross the back yard in order to have access to the acre of land behind the house, the buildings were in the way of having access to a "properly" sized garage. Had to put in a wider door on a brick house (fun, not really) and build a deck with ramps and build a fence around the pool. Been spending about 50% of the past year at sea aboard the Theodore Roosevelt, right now we are conducting training missions in the middle of the Atlantic in preps for our upcoming 9month deployment to the gulf. (it's pretty amazing that I can surf the net from the middle of the ocean, didn't have that in the '90s!)

All of that said, don't loose hope, the Cam Project lives on... I've been doing a shit ton of research on the electronics and the programing of a microcontroller that will actually run the "CNC" cam grinder to produce the cam. There were vast gaps in my knowledge of computer controlled mechanical devices and feed back loops that I've been filling in by reading everthing I can get my hands on in my "spare" time while at sea. (luckily it helps with my professional knowledge well, Cam grinding and controlling neutrons, one book at a time...)

I could post about the electronics for the controller, but this isn't the correct forum for that. However, if anyone is interensted in the progress I've made with the microcontroller, software and electronics please let me know and I'll gladly share all I know / have learned and I'm always interested in new and different ideas.

I've gotten my knowledge up to the point that I can start to get concrete traction in the fwd direction on the schematics and software for the "CNC" grinder and plan to work on them during my upcoming deployment. If I have time, I'd like to get back to the Cam design phase around September time frame, then I'll have more Hot Rod cam tech stuff to post.

Thanks for reading and I'll catch ya on the flip-flop.....
Kipp

 

Wow! I just caught this thread today. I'm blown away. I was an Autoshop Teacher in one of my prior lives. As I read through this I kept thinking, "What would this guy have looked like as a student?"
Then it came to me that he must be a greeter at Wal-mart. Just smiling and saying, "Good Morning" to everyone while simultaneously deep in thought about cam design, computer code, machine development, metallurgy and who knows what else.

Now, we find out he can move buildings and in his spare time serving our country aboard an aircraft carrier.
I SALUTE YOU, SIR!
When will the cam design program be available as an app for an iPhone?

 

Hey, Kipp;
As in the past, so now: I'm (still) very interested in following/learning what you do w/this. Thanks for posting what you have, & what you will. & if it gets posted on another location, please either let us know here, or at least PM me, or email me the link. TIA.
BTW: Stay safe, & well.
Marcus...

 

Just reading this as an intrigued observer, thanks for the update!

 

On the flip side of this, I saw this same fella drink a bunch of beer and fall into a campfire once... but he is kind of a genius.

-Abone.

 

"In theory, there's no difference between theory and practice. In practice there is."

NASCAR, back in the day, the best 427"s making 500 horsepower, the cam/lifter/springs would randomly fail in less than 500 miles at 7,000 RPMs. More cars dropped out from engine failure than from wrecks. Today's 358"s make 750 horsepower and turn 9,500 RPMs for 500 miles and seldom-to-never have a valvetrain failure. The Spintron was the only way to get there. Design it, break it, evaluate the failure, make it better, run it until it breaks. Repeat.

jack vines

 

Kipp,

I came across this only yesterday when I Googled Hubbard's excellent book to pass the info to a friend. You are doing a great job. Like you, on and off over the past couple of years I have developed some similar software. I started with software to record and analyze data from my flow bench and started to include a facility to input valve opening data so that I could plot flow against crank/cam angle instead of only valve opening. Then I added features to analyze the dynamics of the valve motion, and then animations of the whole valve train motion. It will also calculate cam profiles for flat, linear motion curved/roller followers and rocker type curved/roller from specified valve motion. Probably the most useful addition was a cam synthesis (cam design) feature. With this you just specify required lift and timing and max. values for jerk, acceleration, takeoff and landing velocities, click on the GO button and it will spew out the optimum profile for those parameters.
I have been thinking about building a cam grinder also, but I am aware of the precision needed to make high performance cams, so that would not be an easy task. Calculating a profile to small tolerances is one thing but reproducing that in metal is another. Anyway as we seem to be treading a similar path I wondered if you would be interested in any collaboration on the software. If so you can contact me directly at [email protected]

 

I hope you two can get together on this. I need some special cams ground. I could be your field rep!

 

I must protest, I didn't "fall into a campfire." You see, what happened was:

We were discussing heat transfer and the fact that it takes time for heat transfer to occur, like why if you remove the thermostat the engine will overheat and why people can walk on a hot bed of coals...which, when standing around a hot bead of coals from a healthy backyard bon-fire, naturaly led to the famous words:

"Dude, hold my beer, watch this!!"

Thanks for the laughs and keeping it real. The '36 is looking bad ass. BTW Saw a pic of a really trashed ex-circle track '40 Ford in a pile of crap for sale, thought, now there's a car for A-bone, reminded me of the first time I saw your '34 in pieces in the corner of your back yard, what a beaut' that turned out to be!

Latter,
Kipp

 

Update,

Thanks Tony for sharing your program, I emailed you directly, look fwd to hearing how you went about your cam program, etc.

Nothing Hot Rod specific to report, but I have been working diligently on the microcomputer, electronics and assembly-level program to use a Raspberry Pi, (I don't know why they call it that, it's made in the UK, gotta ask the Brits I suppose) a 700Mhz computer on a chip with digitial I/O ports for $35 (US). Can't beat that price. Anyways, lots of reading and code writing to learn. I did manage to get it to print words to the screen and light an LED! Doesn't sound like much, but I'm laying the ground work to be able to control a precision CNC Cam Grinder.

I've also been reading a book on the Mathematical Analysis of Feedback Loops and the 27th edition of the the Machinist Handbook (yes cover to cover, I'm on page ~300, only 2200 more to go!) lots of good design info on springs, steels, etc.

All for one goal, rollered and blown 201cuin Flathead, going blap-eded-blap-eded-blap at the traffic light, shaking the car and causing me to have to go around tightening screws every month or two. And maybe some high ocatane fuel, combinded with lots of vlv overlap for that sweet smell of raw half-burnt fuel coming out the exhaust. Bad Ass.

 

My brain just stopped.

 

My brain lurched when I saw that Tony Foale reads the HAMB.

 

So.....back in the states...SD, CA.....if you want, checkout Theodore Roosevelt, CVN 71. Thanks for all of your patience with the "Cam Project", I'm getting there...and anticipating grinding cams circa 2020....

Thanks,
Kipp

 

Are you still moving forward? I have a couple of 201's and really was moved by this topic.

Sent from my XT1585 using The H.A.M.B. mobile app