Grinding Cams - Will this Work?

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
OK guys I'm no math genius. my highest math was trig. but I use it a lot. I'm very good with geometry.... the whell diameter does matter but not by much. on a model engine you can overlook it and get good results. biggest factor being that it doesn't change too much from start to Finnish on a multi cylinder engine. even then you are not likely to know the difference.

if you want absolute repeatability then worry about the wheel size but the relationship between the cam diameter and the wheel diameter makes the wheel look like a nearly flat surface. if you have a 1/2" can diameter and a ~4" wheel the worst case scenario puts the contact from the lobe to the wheel in a range that could never deviate from the centerline more than 1/4" drawing a 1/2" chord on a 4inch diameter gives a height of about .008" that's a good amount but its not the significant value we are looking for... if you swamped in a 50% larger wheel the chord height would change but it only drops to .005 a .003 difference so a massive change in wheel diameter made a noticeable but still small difference in the chord. remember this is a worst case scenario. the chord you are worried about in reality would be much much less than the cam diameter, in fact to avoid the lifter from scraping the lobe off the cam lift needs to be slow enough that the contact point always lands in an area smaller than the lifter diameter which on a pushrod american engine is often smaller than the camshaft. as the wheel gets bigger and bigger the change to the chord height gets smaller and smaller and thus the effects on geometry. remember the rocker ratio is not constant or linear, the lifter radius can vary, and the thing will wear, not to mentioned lash. a .003 change in lash will change timing more than a 50% change in wheel diameter if the cam is small enough.

the important part is that the valve accelerates open decelerates to the max lift and does the same closing so it doesn't release energy on direction changes and bounce or jump. the lift curve on a well engineered cam should resemble a section of a sine wave. the cam may look different from a sine wave wrapped around a radius because of how the contact point of the lifter moves across the lifter surface but if you have flat lifters and a large grinding wheel and you plot a single sine wave in g-code across a section of rotation equal to the advertised duration you will be in good shape. if you use roller lifters you can't ignore the difference in wheel radius to lifter radius when you design the lobe and a lobe design becomes mathematically complex to translate into g-code but if you are going off a blueprint that describes the lobes dimensions you can treat the wheel like a flat surface as long as the wheel is large enough.

even if you don't exactly match the ideal cam as the blueprint or your own design sees it, it's not the end of the world. I mean people successfully do this with a file and never plot the valve lift curve on the assembled engine or check that the lifters aren't binding.. its only a model after all.
 
Remember that the contact line between the finished cam and the follower sweeps across the face of the follower as the cam
rotates. It is the same when machining the cam. The contact point between cam and wheel only lies on the centre-line joining the cam and wheel axes at points where the cam slope is zero - at maximum lift and round the clearance arc. At all other points you are off the centreline. Unless the wheel diameter is infinite, it has a curvature. That is why the wheel diameter does make a difference.

Charles, that explains the issue in a way that I understand and can now see that the wheel dia does come into it.

Thanks, jason
 
OK. here's the answer to this delema:rolleyes:
use your micrometer:)

the diameter of the grinding wheel is important IF USING CNC PROGRAM
BUT NOT IF YOUR USING A FOLLOWER

a very simple explanation your "zero setting for your Z axes using an end mill" is set you run your program. When your done let's assume that your program says that you remove 250 thousand you take your micrometer and man your only at 200 thousand what is the next step:confused:
reset your "zero" to compensate for your end mill that change height or in our case the grinding wheel worn out a bit and that's it no rocket science there.
Now if you listen to the next video Joe Martin founder of Sherline does mention at 5:53 minutes in the video that the dimension of the grinding disk is important and then at 6:57 he takes mesurement and has to give a few more thousands to compensate for the disk warring out.
I' in the process of building a follower and the only reason I'm going to use a 5inch disk it will take longer to ware out:fan:




[ame]http://www.youtube.com/watch?v=PlZDg0rpIi8[/ame]
 
Charles and Nick:
(and everyone else)

I'm starting to see the issue with machining the cam this way, especially with the explanation of the changing axis produced by the changing lifter contact point during rotation. I've seen this before in a diagram and it's certainly easy enough to see it in operation on a real cam and lifter.
But I didn't really relate it to the process I'm planning to follow making the cams - which is a modified facet operation.
How much of an error it's goint to create I guess remains to be seen.
I guess the proof will be in the results. I hope to grind something soon and I'll let you know if it works out.
I can measure the lift, duration etc easily by putting a dial indicator with a dummy lifter pad on it in the (unused) Z-axis quill. Since I have the rotory axis I can measure the timing as well.

I'll let you know in a few days.

Thanks

Sage
 
Nobody is disputing that the method described by Sage will produce a cam.

The question is what will the actual machined profile of the cam look like? It will not be what is expected based on the output from Ron's Camcalc data. It will be something different. How different? I see about 0.04mm more material adjacent to the nose radius for a 2mm lift on a 5mm base circle if cut with a 4" wheel. If you've been conservative with follower size it should not run the edge to the follower into the cam.

Will it matter for a model engine? Probably not. If I was making a cam for a high performance engine that was actually intended to drive a load at high power and high speed, I'd be concerned about many things in addition to the actual cam geometry. I'm sure the Sherline unit does all the proper calculations to account for grinding wheel diameter. The dressing of the wheel has such a small impact, it's negligible for our purposes.

Don't worry about valve acceleration rates if you're making simple cams. Without ramps and smoothing, it's going to jerk. Unfortunately, high performance cam design is a black art and those in the know keep the secrets close. Software for engine cam design runs in the thousands of dollars.

Luc, if you are using a follower machine you can compensate with the proper follower shape. That is not in question here this is how cams have been made since the beginning of time, it's simple geometry. The mechanical interface of the cam pattern and follower will take care of the issues at hand. Gcode for a cam grinder is also simple geometry, but the Gcode needs to be calculated either before making gcode if using a simple machine control, or the machine control must be able to calculate what's called polar interpolation on a C-axis, to use relatively simple input data.
 
I'm sure the Sherline unit does all the proper calculations to account for grinding wheel diameter. The dressing of the wheel has such a small impact, it's negligible for our purposes.
Hi Greg, I agree with you 100% but if you look at the video at the 6:57 you see Joe mesuring the cam total hight
it didn't match the program and he says that he have to remove 3.5 thousand to be good. I think that this will only
work if someone can figure out hte % of ware on the stone itself and compensate with it in the program.
Joe says it took him nearly 3 years to come out with a good cam profile using his set up
 
I don't think the Sherline is fully automated. It's up to you to set the offsets manually to get the finished base circle size. It seems to perform the tool path once, and you have to feed the wheel for each pass manually. On a stepper motor pushing the motor through a few steps manually is easier and faster than changing the offsets. I've never investigated their cam grinder. I'll build one long before I put down that sort of cash on something appearing so crude.
 
this is my next project, It was originaly German and got it translate

should be fun and usefull for cam and cranks
what do you think
 
this is my next project, It was originaly German and got it translate

should be fun and usefull for cam and cranks
what do you think
oups forgot the picture:wall::wall:

cam follower.jpg
 
Well All I can say is it works !!
And it works REALLY Well.

I didn't have much time so I made the first cam out of a wood dowell. That way I could take big cuts and not have to worry (Except for a little smell of burnt wood). Also if I got mixed up in the feed directions I wouldn't prang up the works.

The surface finish is smooth albeit a bit blackened.
Measurements are a bit hindered by the metal lifter not sliding on the wood so well but using the proper diameter lifter on the end of an indicator here are the results:

The lift programmed was .074" I got .075"
Base circle diameter programmed was .274 I got .272
(This is under my control and I probably fed in too much. I think this is what the guy from Sherline was measuring to tell when he was done. I went in a couple of thou. too much It's also tough to measure and avoid including the the sloped part of the cam)

The duration was programmed as 135deg. I got 133.5deg
(I measured that duration at .002" of lift. The program does not specify a lift. I'm sure there would be at least a couple of degrees extra if measured as soon as the indicator came off zero instead of waiting for .002" .)

So I'm going to say that this method works. All of the important parameters are there. It is possible due to the axis issue that the peak of the cam may not be exactly at the 0deg point but if it's shifted it will make no difference I don't have to put a cam lobe on anything pre-existing. Also if I make another lobe next to this one it will be shifted by an equal amount eliminating any problems. The whole shaft will just need to be rotated one way of the other.
(I don't even know if this is a problem)


This is certainly a LOT easier than any other method I've ever tried. And it's certainly good enough for a model engine. I hope I'm as successful with a steel version.


Picture of the setup to follow

I'll try a steel one soon.

Sage
 
I'm sure you'd find that the nose radius is actually larger than designed. If you use a reasonable flank radius there will be no problems. A large flank radius (flat) will actually result in a negative flank radius (roughly equal to wheel radius) using this method.
 
I'm not sure what you mean by choosing a "reasonable" flank radius
The flank radius was certainly reasonable in the design. I'd have to try to measure it and see if it is significantly different than designed.
You've surmised that the nose radius would be wider. I can section one and use a radius gauge to check. If that's true then I guess the flank radius must be a bit steeper or flatter? The curve still has to come from the base circle and join up with the nose in a smooth fashion.

But you know something; it really doesn't matter. The results of using this method are so much better than hand filing from the facet method surely nobody can complain. The results are right on the design figures. It's not like it milled a rectangle or a circle instead of a teardrop shape. Almost anything will work in a model but we do strive for perfection moderated by simplicity and this sure meets those requirements.

It's going to be so easy now to index the intake and exhaust separation by simply rotating the shaft the desired amount using the rotory table, zeroing the angle reading and grind the other cam.

I'm going to have to make a steel one now so I can get a really good look at it. (In a couple of days).

If anyone can do the math to quantify the errors from the design values it might be nice to see where the problems lie. That's way beyond me. Perhaps there is some sort of distortion that is not visible.
Remember these values are entered into the CamCalc program to give depth of cut vs. degrees rotation for the facet milling method. The only change I made was to join the facets with a smooth move between facets.


Cam Open Angle (degrees): 135.0
Valve Lift (thous): 74
Flank Radius (inches): 0.542
Base Radius (inches): 0.137
Engine Speed (RPM): 10000
Nose Radius (inches): 0.056
Tang follower width min (inches): 0.235
Max acceleration (g): 285
Min acceleration (g): -108
Lift Area (inch degrees): 5.803
Nose Transition delay (degrees): 17.11




Sage
 
The amount to turn and the depth of cut can be calculated from a program called CamCalc which is free.
Sage

Can you elaborate on what website you are using/downloading Camcalc.
Once upon a time this worked for me, but not sure why not anymore. Its either broken or my PC isnt allowing something necessary to install.
http://modelenginenews.org/design/CamTable.html
 
Luc
You posted a photo of a cam grinding device in the tread as your next project. I failed in my attempt to look it up in google. Can you post more info on this grinder or send me a private message if this would be considered taking over the thread
Art
 
Luc
You posted a photo of a cam grinding device in the tread as your next project. I failed in my attempt to look it up in google. Can you post more info on this grinder or send me a private message if this would be considered taking over the thread
Art
Hi Art, I'm not completely done with scaling all the cad drawings and
their are some stuff that I will modified so I can use it for Crank and Cam'S

Can you post more info on this grinder or send me a private message if this would be considered taking over the thread

what does this quote nean I dont understand
cheers
 
Petertha:

You have the correct link to Model Engine News and the Cam Calculator. It should work right there on the web page if you enter the data. See the note below the entry form. He says it has only been tested on certain Browsers. Works for me on IE 8. (Which I don't think is on the list but works for me). I think it uses Javas script. Do you have Java installed?

Sage
 
I'm not sure what you mean by choosing a "reasonable" flank radius
The flank radius was certainly reasonable in the design. I'd have to try to measure it and see if it is significantly different than designed.
You've surmised that the nose radius would be wider. I can section one and use a radius gauge to check. If that's true then I guess the flank radius must be a bit steeper or flatter? The curve still has to come from the base circle and join up with the nose in a smooth fashion.

But you know something; it really doesn't matter. The results of using this method are so much better than hand filing from the facet method surely nobody can complain. The results are right on the design figures. It's not like it milled a rectangle or a circle instead of a teardrop shape. Almost anything will work in a model but we do strive for perfection moderated by simplicity and this sure meets those requirements.

It's going to be so easy now to index the intake and exhaust separation by simply rotating the shaft the desired amount using the rotory table, zeroing the angle reading and grind the other cam.

I'm going to have to make a steel one now so I can get a really good look at it. (In a couple of days).

If anyone can do the math to quantify the errors from the design values it might be nice to see where the problems lie. That's way beyond me. Perhaps there is some sort of distortion that is not visible.
Remember these values are entered into the CamCalc program to give depth of cut vs. degrees rotation for the facet milling method. The only change I made was to join the facets with a smooth move between facets.


Cam Open Angle (degrees): 135.0
Valve Lift (thous): 74
Flank Radius (inches): 0.542
Base Radius (inches): 0.137
Engine Speed (RPM): 10000
Nose Radius (inches): 0.056
Tang follower width min (inches): 0.235
Max acceleration (g): 285
Min acceleration (g): -108
Lift Area (inch degrees): 5.803
Nose Transition delay (degrees): 17.11




Sage

my assumptions are there will be distortion but too little to notice. it won't be visable and may be difficult to measure without a scaled up print out and a comparator. it will likely be more acurate than the facets and file

I may be able to make up a spread sheet to figure it all out but it will take me time since I'm not really that familiar with spread sheets.. depends also on how in depth I want to go with plotting the curves. but lash will affect the critical timing events more so than the wheel diameter. wheel diameter would be critical if it was closer in scale to the cam lobe like a 1" wheel would need much more compensation the get ideal results.

if you use cam software and make it aware of the grinder geometry then problem solved. I believe master cam 8 and maybe 9 have been opened up for educational use and no longer need dongles or keys to use. its old software but will more than serve the purpose. there is more modern free software available as well but Im not familiar with them and find them lacking in features I like. makes me adjust my drafting style too much.....
 

Latest posts

Back
Top