What's new
What's new

Cam wear - Caused by Hard Milling?

gregoryd

Hot Rolled
Joined
Jun 30, 2007
Location
Mass
Where I work Cams for internal use machines used to be made from Mehanite and hard ground on a tracer machine with a master.

Over the years material was changed to A6 and hard machined at 52-54 Rc.

We are now chasing down a problem that we are not sure is related or not to the manufacturing method. It is difficult to get the whole story on when this became a problem, but cams that used to last years are lasting months. This is happening to hundreds of cams in a couple of different factories. Some issues are from set-up on the machine, but other are not.

What is weird is the wear pattern seen on the cams, they are getting grooving along the whole profile that is like the rumble strip on the side of the highway.

As you can imagine there are plenty of theories and variables to the problem. I'm interested in other peoples experiences with trying to hard mill cams and if anybody experienced similar wear or failure problems.

These Cams are about 6" in diameter, turn in the 175-225 rpm range, and have a 1.25" unlubricated roller cam follower on it.


We cut the CAMS on a MAZAK FJV-250
Rough Cutting is done with a Z Carbide ½” end mill
Finish Cutting is done with a EXO Carbide 6 Flute end mill
Rough and Finish Pass is done at 2500 RPM moving at 10 IPM (Inches Per Minute). Does not give a "ground finish"

Pictures in next post.
 
Images of a couple of worn cams
 

Attachments

  • cam1.jpg
    cam1.jpg
    20.2 KB · Views: 682
  • cam2.jpg
    cam2.jpg
    16 KB · Views: 628
This might be a good six sigma project. It looks like it could be caused from the milling process. One way to find out is have a few of those cams ground or turned on a lathe. Shaperhaven
 
If you put an indicator on that surface, how wavy is it really? Appearance alone is deceptive.
Have you done hardness tests on the crests and valleys of a worn cam? How soon does the pattern begin to emerge after installation of a new cam?

The periodic spacing doesn't seem to really jive with mill cutter eccentricity, not at the feedrate and rpm mentioned. I'm thinking the cam roller itself likely as some kind of balls or rollers with approximately that spacing. Maybe they have cut costs and quality on the cam rollers and they've got 3 cornered rollers in them now :D
 
Is the wear pattern similar around the entire cam? Is the pattern the same on arc ( dwell ) segments as on the rise/fall segments?

Are the cam profiles generated mathematically in a CAD program or what is being used to generate the CNC cut path?
Have you made a new cam and inspected the profile for any indication of the pattern, maybe checking the hardness at very small intervals...

Interesting problem.

Chris
 
Does not give a "ground finish" = gravel road
hard ground on a tracer machine with a master= paved road no wheel hop
key word is ground
just my 2cents
 
These Cams are about 6" in diameter, turn in the 175-225 rpm range, and have a 1.25" unlubricated roller cam follower on it.

I assume you mean something like this? http://www.rbcbearings.com/camfollowers/index.htm

What is the unlubricated roller cam follower "face" material? Is it the same (or very close to) material as what you have made the cam out of?

Has the cam follower gotten harder (or are the end-users replacing it with a harder follower) ?

Are there any changes which caused significant deflection in the cam-follower stem?
 
Greg:

There are so many possible variables that are or may be involved, some of which you may not have been made aware of, such that it would be extremely difficult to pinpoint the actual cause of the problem described.

However, it seems the old processes used when the cams were made of cast iron are still utilized for a very different metal.

Another questionable area is the selection of A6 for the cams. A6 wear characteristics are about midway between O1 & D2. D2 is a superior choice of steel over A6 for those cams.

Back to the process, milling the tool steel cams then hardening and drawing would give far better results with less surface stress than milling after hardening. Hardening and then drawing back to RC 56 - 58 will also help.

The lack of lubrication is the obvious suspect, but assume there is a valid reason...Food or medical industry???

A new "set-up" man may be excessively pre-loading the cam followers etc. An example of another step in the trouble shooting process. Without being on site, I feel like try to help in this situation is like shooting in the dark.

Good luck,

Amigo
 
Been said allready, but is it milling from point to point or following a continuos path. I would assume a basic nc program using plot points at a close spacing would provide such a wear pattern. While pretty good, not quite perfect. I guess...
 
I know there is a lot of speculation on causes, (and it is difficult to solve without knowing the entire system) and what I was hoping for was someone out there that has seen cam wear like this one and solved it. Or had gone through some pains when they switched over to hard machining cams over grinding.

We design our cams with DYNACAM and then export a minimum of every 1/4 degree into Unigraphics where we use a cubic spline function. NXCAM then generates the path and then a macro sends it over to G-code.

What is strange is that many differnt cams have the same number of "stripes". Even across the dwells. We are considering that there are a couple of factors coming together to cause this problem, like rough cam surface and stick-slip of the follower, or something like natural frequency.

Failed surface is 58-62 Rc. (I was wrong in my original post on the starting hardness - it is 58-62)

Groove depth depends on how long it has run but is around 0.010" deep.

Inspection right after machinig with a CMM does not show the waviness, but is in spec (+/- 0.002").


Amigo: you bring up good points. We think there may be some combination beteen the material properties of: toughness, resiliancy, hardness interacting with the design elements of loading and lubrication causing the problem.
 
Would it be possible rig up a ballbar test on a brand new cam follower, riding on a new cam in the machine? You might be able to check some kind of periodic error in its very early stages. Maybe a failing ball bearing on the shaft supporting the cam could be the culprit?
 
I don't know about the Spline function you are using. Anytime I've attempted to cut a "spline" path with our CAM system, it actually cuts little Facets on the part - Almost like what you have. You need Good, CLEAN, Trimmed Geometry in your CAD, or its Not going to be a smooth arc!

Can you give us a sample of your G-code that you are using?
 
It's not a bad idea to ask if anyone has ever had a similar problem and solved it, hoping to get lucky with a quick fix. But that's what we're talking about - luck. Luck is good, but if you don't get lucky, the solution will be found by thoroughly understanding every element of the system and of the problem. The system involves everything from start to finish, from the time you buy raw materials (and the materials themselves) to the cams and the manufacturing processes (and tools/tooling) you use to make 'em. You also have to become familiar with the cams' use, the machines they are installed on, the products they are making, etc etc etc. Every detail is fair game, since you don't know what is causing the problem. You don't even know if the problem is "us" (cam maker) or "them" (cam user). Once you have that knowledge, finding the solution will become much clearer.

Here are some of the things I would be asking: When did the problem begin? What was changed about that time, or about the time that the first defective cams were installed? Are you using the same material, from the same supplier, etc. You get the idea. List and eliminate every possible variable. I'd also look at the nature of the failure. The worn cams' appearance resembles "gears" (for want of a better analogy) - count the number of worn grooves in each cam. Is it always the same, or at least the same for each type of cam? Do cams from a particular machine always have the same number of grooves? (If not, you have real problems). Let's say that every cam (or every cam in a particular type of machine) has (pick a number) 40 worn grooves. Where does the number 40 occur in the process? For example, let's say the cam has a roller follower that makes 40 turns for every turn of the cam. (Or it could be 20 turns per revolution of the cam, with two "hits" per revolution, etc) Now you have a legitimate culprit to investigate more thoroughly. Or, let's say that the machining operation that produces the cam face that wears out is an interrupted cut that takes 40 cut segments to complete the circumference.

These methods don't give you a clear pointer to the root cause, unfortunately, but they make finding the root cause much easier, especially if the problem is really convoluted. I have personally seen situations where the solution to the problem was utterly unintuitive, and was found only by thorough research and often after much time and money had already been wasted on speculation and tinkering. The sooner you dig into the problem in earnest, the sooner you will solve it. I also recommend that one person be in charge of the investigation. That way, the person understands every assumption and ever piece of data.

Good luck to you!

Steve
 
You are all taking a great problem solving approach. This has been a back burner curiosity for many of us, but none of us have been able to dig into it fully. We are just starting to do it now because the problem has gotten much worse lately. We are starting that full investigation since the speculation is not getting us anywhere.

There is an incredible amount of hearsay on this. Nobody knows exactly when the problem started, and few who are around today were involved when certain things were changed (cam material from meehanite to A6, post case hardening and grinding to prehard and milling). The biggest problem about this problem is getting data on some of the very questions you guys are asking!

Ironically manufacturing thinks they are getting bad cams, machine shops thinks it is manufacturing setting up the equipment wrong and not maintaining things correctly. Engineering thinks it is a combination of factors aggravating each-other.
In school one of my professors wrote papers on the difference in vibration that a cam would produce based on how it was cut - linear vs. circular vs. cubic interpolation. I have also seen some references in other threads that other members that tried to mill cam surfaces has increased wear issues. That is primarily why I tossed this out - to see if anybody had seen a wear pattern like this that I could bring into the investigation, and why part of my speculation in on surface finish.

I have gotten some good info from you guys already about A6 maybe not being the best choice (which was probably make on hardness alone and not wear resistance), and surface stresses from milling maybe being more than grinding. (Thanks AMIGO & SA100 !)

We are setting up for a couple of tests, and are providing samples to one of our metallurgists to determine if we are getting the material we specified and to classify the failure mode. We found a similar image of fretting on a bearing failure site. (below).

We are also going to have a couple of cams made from different methods (grinding vs. milling & case hardening vs. full hard) and run them side by side on a test bed to see if one style has different life characteristics. We will also be able to put an accelerometer on it to see how much the acceleration curve varies due to the different methods.

To answer some specific questions not yet answered:

1) HUFLUNGDUNG: Not sure if a failing shaft bearing would be the cause since this problem is happening on over 100 different machines of 4 different types. What we are trying to determine if there is any commonality with the cams that wear out the fastest, like many cams close together loading up the shaft radially and causing deflection, or if there is torsional vibration whipping the shaft.
2)HUFLUNGDUNG: Also in the speculation column is if we are running the followers close to their natural frequency (around 3 Hz). Once we get some more information like the failure modes we will probably call in the follower manufacturer for support as well. I'm sure that they have seen more problems than we can think of.
3) RSTEWARD: Let me see about getting the code.

I'll post updates if people are interested.
 

Attachments

  • bearing fretting.jpg
    bearing fretting.jpg
    26.9 KB · Views: 1,174
(cam material from meehanite to A6,
Can you go back to getting the Cast Iron? seems that is a better place to start...

I may have missed the manufacturing steps, but when was the Heat Treat done?..
before milling the cam shape or after?

Cincinnati Mill made all their machine tools from Meehanite (those that were not weldments), some machines had the ways flame hardened then ground and I do remember some cams used in the machines were made from the same material and milled using hydraulic tracer controled Hydrotels, but it's too long ago to remember when the flame hardening was done, if I'd have to guess, it was after milling.
 
I work on machines that have all kinds and shapes of cams and never have seen a cam follower impart a synchroness pattern in a cam like your examples.I can't tell if your photos are of the heels.the ramps or the lobes.Cams that have not been over speed{inducing "float") regardless of material generally wear the opening ramp first{unless yours are continually reversing}.Cam followers like all rolling element bearings generally start to fail by the rolling elements becoming brinnelled or flatspotted causing the roll to stick & slide and tend to wipe highspots off the surface they are supposed to roll against. I just don't see cam followers putting that uniform of a pattern on all the different cams and applications.If your cams are unidirectional then the serations on the worn cams would look like mini cams with wear on one side more than the other.Not knowing the type of loading,speed and application I would have to say that the only thing in common that all the cams share that would cause that uniform pattern is the cam manufacturing process.The bearing photo can mean different things depending what it is.If it is a cam roll I dont see the same pattern inside where the rollers are so the pattern was induced by what it was rolling against;if it is an inner ring than most likely caused by vibratory loads while the bearing is stationary.The only thing I can say for sure is that I'm glad its not my problem!
 
Out of curiousity, if you divided the pattern spacing by the cam RPM would you come out with a similar frequency, and do you see the same frequency on different machines? (or a whole number multiple)

A little background on the machines and the plant where they are used would be nice too...
 
Does the problem map to a specific lot number of material?

Does your vendor supply lot numbers?

If you don't record lot numbers, you can't recall faulty work from the field.

You might also confirm the specific alloy. I hear tell there are laser thingys that will indicate the metals mix without having to destroy the part or wait for results.
 








 
Back
Top