[bluejets]

Unlike the crankshaft grinder, the cam grinder is largly “mechanical” in operation. The silver steel (drill rod to some) blank for the cam is first turned up in the lathe with bearing surfaces to size and the cam sections sized to a few thou over the maximum lift and centres drilled each end. Usually allow a little extra in length to drill a small indent to locate the drive dog.

The bed for the grinder is much like a lathe except that the bed and the spring loaded main tower are pivoted at a point “x” distance below the drive spindle centre. This can be seen in one of the photos. A cam profile is made from 1.5 – 2.0mm thick steel or aluminium and the tower carries this profile on a boss that is driven in sync with the spindle. This top shaft is 4 times “x” distance above the pivot point previously mentioned which allows the cam profile to be 4 times larger than the ground cam will be giving greater accuracy in the finished cam. During grinding, the tower pivots in and out following the curve of the profile.
The boss carrying the profile is indexed to allow the profile to be moved to the next cam timing location and it is changeable for a different “number of cylinders” cam to be ground.
When the profile is fitted to the boss on the tower, it aligns with a feed-in rod. This 8mm rod has a 20mm flat that the profile rubs against while it turns. The rod is driven back in (when grinding) at an adjustable rate from a small stepper motor that gets a feed-in pulse on each revolution of the spindle. The rate is from 0.5 thou to 3 thou and with the grinding motor and other factors, I usually set this towards to lower end. The speed of the main grind motor is also adjustable although I removed the original internal control (spring govenor) and fitted a panel mounted thyristor type as the original caused an amount of “hunting” and caused the drive belt to flex like crazy.
So to grind, set the Auto/Man switch to Manual. The cam profile is fitted to the top shaft in the first cylinder position and whether it is exhaust or inlet. The cam blank is fitted between centres on the lower spindle, attaching the drive dog in the previously drilled indent. The grind wheel carriage is run up to align with the first cam and the spindle shaft is turned so that the high lift part of the profile is sitting on the feed-in rod. The grind wheel (not running at this stage) is fed inwards until it just clears the cam blank by a couple of thou. A shutdown optical switch is then run up to this point and the measuring dro is set to zero. The feed motor is then run back out again and the profile again rotated until the back of the profile now rests on the feed-in rod.
All is now set to grind this, and, with the exception of a position change of the profile and re-alignment of the grind wheel, all remaining cams. If different profiles are used for exhaust and inlet, these are simply changed at the appropriate time. Normally I would grind all the exhausts first then change the profile and do the inlets.
Switch to Auto, press Start and go do another job while the grinder does this one. In the above I mentioned setting the initial grind a couple of thou short. So once this first grind has shutdown, a measurement can be taken and if needed, final cuts can be done and the shutdown point reset to this new value. A quick polish with some wet or dry paper and the cam is ready for hardening.
A mate of mine has built a similar machine using different types of recycled materials. The most talked about is a small fine-feed gearbox (out of an F1-11 no less) as the feed-in arrangement. It feeds continuously via a speed control and shutdown is via a micro switch. I like the optical switches myself as they are more accurate. He has also built a gear and rack unit to which he can mount his profiles during manufacture. It traces the shape of this profile onto a sheet of paper where it is possible to see the exact number of degrees the cam will be open or closed and the shape of the curve i.e. either moderate or “snappy”.
…………… Jeff

[One of Many]

Very impressive Jeff.

Can we hope to see a full write up and use in Model Engine Builder? I'm sure Mike would love to receive articles like this.

Is the PC controlling the steppers, or is there another board for I/O and motor control?

I have often wondered about something similar for radial cams. Maybe something for 4 stroke twin cylinders. Something like the small cams in Saito's, etc.

Thanks for posting the pic's too!

DC

[bluejets]

DC,
Always great to receive recognition for effort... thanks.... Not familiar with Model Engine Builder. Is it a US mag? We have an Aussie mag here that takes in all aspects of IC engines and had thought of running a couple of contributions past them so could possibly split it two ways.
PC is only used in this instance as a measuring device with a simple mouse dro program. It was one of my first attempts and it works fine.
The stepper motor is used as a feed-in for the grind wheel. It came out of a really old hard drive and so I used the board from there as well. All I had to do was give it a pulse and direction signal (from a 555 and a switch if I remember correctly) one for in and one for out direction. I did make a couple of boards for the rest though.
Not quite sure of your reference to radial cams....... Jeff

[NC Cams]

Not too long ago, Home Shop Machinist had a feature article, including a build how-to, on the creation of a mechanical cam actuated profile copying cam grinder. The method is a table top variant of the exact same procedure used by our Berco and nunerous other "rocking table" grinders.

There have been a number of attempts at using steppers to create "real" automotive cam grinders. However, the "step" function of the motors tends to leave visible and measureable facets in the lobes which literally create a "buzz" in the follower. Moreover, they have had problems with slowness as the steppers are usually adapted to heavy, existing mechanical, "rocking table" cam grinders. Somewhat akin to putting lipstick on a pig.

For small cams operating light valvetrains, this may be a non issue. But for full size/high speed applications, high speed/high power servos do a much nicer/better job especailly with regard to finish and profile dynamics.

This is NOT to disparage the phenominal efforts of our Australian member. I am very impressed with the project. This note is only a caveat to those enterprising souls out there who might get the idea that "all I have to do is scale this up".

It has been tried and it has its limitations.

When you want a true CNC cam grinder, the unequivocal #1 source of CNC automotive cam grinders is the Landis 3L masterless cam grinder. It can perfectly grind a V-8 Chevy cam with 16 lobes, made of hard face welded materail and/or tool steel in about 8 minutes. Only about $1.5million plus installation and mandatory maintenance contract.....

[One of Many]

And tremendous effort that is. I must have just missed the original crank grinder post some how. That too was great to see first in the Zone.

http://www.modelenginebuilder.com/

MEB is the quazi replacment to Strictly IC. They even have Robert Washburn(SIC owner) as the Editor Emeritous/contributor. I have known Robert since I was around 12-13 from his Hobby Shop about 30+ years ago. We still cross paths once in a while. I have been a subscriber to both from their beginnings.

To clarify, I should have stated Radial Engine cams. Disks with lobes. The stone would need to be smaller to reduce the ramp at lift, but that should not be much of a change/addition to your design.

DC

[bluejets]

NC,
Stepper motor drives a gear train reduction to a threaded feed screw.... no "steps" here. DC motor and fine feed gearbox would have worked just as well but decided to try something a bit different. Finished cam at 1/4 of original profile is a copy of the larger profile. Would have made no difference if I had used a crankhandle. (First one I made 10 years ago did) All up cost was less than $50. (Aus)
Jeff

[NC Cams]

In our case, we measure cams with a device that can read to within 0.000001" - we can detect even the slightest surface anomaly such as "pick off" where the wheel is lifted off the ground cam lobe. When you can measure values like that, you see pretty much ANY deviations that may develop in the cam profile.

Grinding oversize masters for a tracing style grinder at 4:1 or 5:1 normal size is a well known trick. IF there is an error/deviation in the profile, it will be diminished by about the same proportion when you "shrink" it back to size.

In our case, our masters start out as a 6" dia blank and then you cut in to the part to create the base circle and therefore generate lift. On small masters such as what you're using, you can probably get away with 1 deg cutting intervals when you cut your masters.

In our case, we found that you need to use a MUCH finer interval in order to generate a smoothly grindable/dynamically stable profile.

Again, for model engine use, your system is surely at an unparalled level of sophistication and/or development - I am fascinated by the device as I own and run a "real" cam grinder. However, I can assure you from our experiences in the business that when you scale things up to run full size engines, especially those at racing engine speeds, the requirements go up about as fast as the level of complexity.

[bluejets]

NC,
Understand the requirements you have for racing engines. I'm just over the moon to get this thing to work to the point where it is reliable and consistant and the final product performs better than expected at times.
Currently working on the beginnings of a V8 with the patterns and castings completed and machining of the block started so might have some questions for you at a later stage on cam designs. My Dad was involved with this type of thing back in the late 50's when he helped develope an engine that held the water speed record for hydros. Many things have changed since then and then again, many remain constant.
Jeff

[dynosor]

Assuming that one has access to a cam grinder, manual or CNC mill, the question is where does one get the co-ordinates to feed it that would result in a functional cam?

I found a handy calculator at Model Engine News that would enable you to lay out cams and examine acceleration profiles at a given RPM. You get to input lift, duration, flank, nose & base circle radius and RPM. The calculator also spits out a table of cam angle and lift that you can feed your cam grinder.

http://modelenginenews.org/design/CamTable.php

Obviously this calculator doesn’t know if any cam profile you generate is good for the intended engine, but it looks like a handy tool that would aid in making a number of experimental cams to try out.

Speaking as a complete novice, the most obvious constraint on any cam would be maximum lift. The next limit I would consider is maximum acceleration/deceleration in the context of part inertia, strength and valve spring rate & preload. I would start with typical duration values for the type of engine you are building, remembering that effective duration is reduced by valve lash. The nose and flank radii would be driven by satisfying the above.

The goal with the first few profiles is to create cams that work and that do not destroy the valve train or the engine. Cam profiles for best performance can be developed later.

Possible machining set-ups are described here:

http://modelenginenews.org/feeney/pg7.html

[bluejets]

The proof is in the pudding as they say and with my engine running to 10,000 in the wink of an eye I'd say the cam design is quiet sufficient thankyou. If you want to run around with calculators and the like, all the best to you. Just one request, show us some of your achievements in 3D.

[Mcgyver]

Jeff, very nice. is it the V8 that's ticking along at 10k?

[NC Cams]

Bluejets: Do race wins with real cars count as 3D proof?

If so, we had 12 wins in 16 races in NASCAR's elite Winston/Nextel cup division over a 4 year period. We supplied cams to a particularly well known team during that time frame.

We also had a bunch of Craftsman truck race engines running with "tool steel" cams for the past 2 seasns and currently have some off road "trophy trucks" running with DOHC tool steel V-8 cams - more on the way.

Currently, our work is mostly OEM prototyping that, due to confidentiality reasons, we can't discuss but the clients are well known MAJOR players in their industries.

I'll admit, these are definititely not 10K model engines but, rather, 9.5K RPM REAL engines, or there abouts anyway.

[dynosor]

The proof is in the pudding as they say and with my engine running to 10,000 in the wink of an eye I'd say the cam design is quiet sufficient thankyou. If you want to run around with calculators and the like, all the best to you. Just one request, show us some of your achievements in 3D.

No arguments here - it is what works that counts, not theory. However, if I dream up a profile that causes a calculated 10,000 G acceleration it is probably not going to work for very long, if at all.

How did you determine your cam profile?

I have not made any cams yet - still trying to figure out how I could do it.

The only thing I can show is a practise straight 6 cylinder crank machined from solid - I figured that is the most difficult significant component to make, with the cam shaft at a similar degree of difficulty.


http://www.bimmerfest.com/forums/att...1&d=1194461147

I am busy setting up to make a second crank with larger journals to be closer to scale.

[NC Cams]

The requirements needed to go from a turning to a finished cam, especially a full sized automotive cam are NOT inconsequential.

As mass goes up, forces and dynamics become MUCH more substantial. Basically, what you can get away with on a model engine can/will result in havoc and broken parts when one attempts to "scale things up" from model to "real car" levels.

Another real tough thing to deal with in "real" engines is material compatibilities. Cams and lifters are NOT no brainer parts - simply making them "hard" doesn't alays work due to the wear and load phenomenon that they encounter. There is a lot of science involved in selecting and prepping the parts properly.

As far as grinding, spec'ing mat'ls and/or heat treating, we do offer that service on a boutique basis to/for our clients. These are NOT necessarily a DIY viable process - especially, the H/T and material spec'ing part of the process..

[dynosor]

The requirements needed to go from a turning to a finished cam, especially a full sized automotive cam are NOT inconsequential....

Thanks NC, I have gained significant insight by reading all your informative posts on the subject. Each one of us wants to produce the best cam/engine we can with the tools and expertise at hand. Most of us have a lot to learn and will probably never produce a commercial quality valve train. Fortunately, this lack of ability is of lesser consequence than for someone who wants to set themselves up producing cams for sale to others.

This is a hobby forum and while I appreciate being told what are the correct methods, designs and materials, I think protecting others from trying anything that would not be commercially viable on a hobby engine is not your responsibility.

That said; I recognize that I seem unable to follow my own advice – most of my post to CNC Zone involve trying to persuade others to see things “correctly” with regard to global warming. In the GW context I mostly just want everyone to think for themselves so that they actively develop their own opinion based on facts and logic and don’t just adopt that of anyone who speaks charismatically or with authority.

For instance:

http://www.cnczone.com/forums/showthread.php?t=45545


http://www.cnczone.com/forums/showthread.php?t=42162

[NC Cams]

The charming thing about designing cams is that they are locically responding devices that pretty much DO follow the F=MA axioms. It is not a question if you can/will follow the "rules" but whether or not you don't violate or sodomize them entirely.

Regardless of the engine, smooth valve operation is the goal. Yes, you can pound things open and close but that breaks parts, wastes power and ultimately time and effort. So, why bother???

Learning how to do/make things traditional and technically PROPER ways is akin to crawling before, walking and walking before running. I didn't even think of doing a race cam until I found that I could duplicate a known production cam from scratch. When I know how to duplicate it, I knew the vagaries and WHY the design programs did what they did and how you could bend/mold the cam curves into the shapes you ultimately wanted/needed.

The "beauty" of a well designed cam is in the smoothness of motion. The lack of jerk. I show guys cams running on the cam grinder. Essentially, you can literally feel the motion that the cam will apply to the valvetrain when it moves the rocking table on our Berco. As well it should as the same force signature is being generated by the cam in/to the engine.

For grins, I show guys two cams - a stocker and a "violent" race cam. Amazingly, engineers who should know better typically guess them WRONG. In my example, the race cam is actually LESS violent than the stocker. Part of it is due to design philosophy, some due to simple duration and lift.

Yet, the guys often wonder, "How the hell does he get such violent valve action to be so smooth????" Lots of time and effort spent studying the math and practice in designing things via the computer. WHen you figure it out and understand the mess, you can literally LOOK at the graph of the profile'd dynamics and tell if it will work or not.

But when you know WHY this is the case, you can do a better job of designing and making either hard core race or banal stock stuff. Spend the time to learn the basics and how to do things as properly as you can and with as much verve and character as you can muster. Learning the basics enables you to find even more exotic solutions because you know more about where and HOW to look.

OVerlook the details and you're missing out on some really neat stuff - trust me, its there if you take the time to look and find it....

[bluejets]

Thanks mcgyver.. V8 is still in construction ..... it is the 4 cylinder that cranks out the tune.... sweet as .....

NC ..... try it with 25cc's ........

[bluejets]

Mcgyver ..... 4 pot is over here ......
http://www.cnczone.com/forums/showthread.php?t=35241 ...... take a peek...... presently being shown at engineering conv. in Capital ...... also fitting it to a Rivierra 1020mm timber hull (home grown)

[NC Cams]

Re: 25cc vs 2500 cc

It really doesn't matter - ALL engines have to follow the same laws of physics and thermodynamics.

We went thru the same sort of "hobby vs real car" arguement with R/C cars. The R/C guys contended that my "big car" ideas wouldn't work with the hobby cars. They continued to do that as I was trouncing them with under powered/superiour handling cars.

The point is, the engineering works - regardless of the scale. It is purely up to the user to develop and recognize the proper corelation coefficients however.

[bluejets]

NC ..... I say again ...... try it with 25cc's ... and then show us your 3D.