Grinding Cams - Will this Work?

[dsage]

Here's the picture of the straight edge setup emulating passing a milling cutter across the blank. Note that the point of contact is (properly) a couple of inches off the center line. This produces the correct profile.

Sage

 

[mu38&Bg#]

This is what we were expecting. The offset is curious. A problem with the math would be obvious in the coordinates supplied by Camcalc. The numbers from Camcalc are symmetric so I don't think that's the issue.

 

[dsage]

After checking, Yes, it appears the data is symmetric so I guess it comes down to eccentricity in the rotary table chuck. It was just odd to see it show up exactly where I've seen it in some of my test cams when I make no special effort to put parts in the chuck in a particular place. I just zero it wherever it lands. That error is minor. For now this method of grinding cams is on the back burner due to the apparent errors in the flank. Even though, like I said, you'd be hard pressed to see (visually) as much error as shows up in the test piece. As well, the duration, lift and as best I can measure, the radii come out correct. It may not be fair to magnify the errors like this. Sort of like using an optical comparator. Minor issues can be made to look excessive But I guess errors are errors and I'd like to make them the best they can be.
I'll have to see what my programmer friend comes up with for a new program.
It's all been an interesting experiment.
I just thought it best to update this thread in case someone thought this method of grinding cams was a done deal and worked properly.
At the moment I'd say it's a far cry better than using cam calc to rough them out and then hand file them. I'll leave it at that for now.

" I'll be back "


Sage

 

[dman]

After checking, Yes, it appears the data is symmetric so I guess it comes down to eccentricity in the rotary table chuck. It was just odd to see it show up exactly where I've seen it in some of my test cams when I make no special effort to put parts in the chuck in a particular place. I just zero it wherever it lands. That error is minor. For now this method of grinding cams is on the back burner due to the apparent errors in the flank. Even though, like I said, you'd be hard pressed to see (visually) as much error as shows up in the test piece. As well, the duration, lift and as best I can measure, the radii come out correct. It may not be fair to magnify the errors like this. Sort of like using an optical comparator. Minor issues can be made to look excessive But I guess errors are errors and I'd like to make them the best they can be.
I'll have to see what my programmer friend comes up with for a new program.
It's all been an interesting experiment.
I just thought it best to update this thread in case someone thought this method of grinding cams was a done deal and worked properly.
At the moment I'd say it's a far cry better than using cam calc to rough them out and then hand file them. I'll leave it at that for now.

" I'll be back "


Sage

if the radius of the grinding wheel is not in the same proportion to the cam that you used while grinding then the error will be exagerated and not scaled to the cams you made. i an be wrong but that wheel in the mockup looks less than scale.

if you want to be 100% precise you can move the wheel up and down as well as the table in and out as it turns and then wheel radius wont matter.

if i have time tomorrow i can try to draft something up in draftsight to show how to find all the points as a visual aid and maybe come up with a spreadsheet to crunch the numbers.

 

[dsage]

Most of you guys are so smart. (compared to me).
Why didn't I think of that. Yes the size of the grinding wheel is far from scale. It's 7" diameter and it's what I used to grind the cams which are about .430 dia as a blank. For this test setup the grinding wheel would need to be 16 times the cardboard size or about 100" in dia. to be in the proper scale. (the cardboard cam is 6.25 base to tip)
I'm glad you caught that because I was dismayed at the amount of error showing up. It certainly was not indicative of what my steel cams were coming out like but I couldn't pinpoint the source of the error.
Certainly a properly scaled wheel would appear more flat (like the program is expecting and works fine with) and a lot of what shows as error would be removed.
Now I'm curious as to what effect that will have. I should be able to do that by drawing a 100" circle and use a portion of it in place of the 7" wheel. I'll try that tomorrow.

Thanks for that insight.

I had rejected the idea of moving the grinding wheel up and down early on because it would then need to be mounted on the quill of the machine (unless you had some other idea I've missed). There is too much "stuff" on the head and quill to get a setup rigid enough (I expect). If I were going to use the quill I would generate a cam program specifying a 7" diameter tool and effectively "side mill" the cams with a grinder.
Side milling is how I did my first set of cams and it worked just alright but your cutter needs to be factory sharp and holding cam blanks vertically is problematic. I had a lot of rejects with chatter etc. from flexure.

I had thought to model the whole thing in 3D CAD so it will be easier to visualize the geometry but it's probably quicker and easier to just keep making cardboard models. But if you know of a way to accomplish that I'd be interested in your results.Thanks for the offer on the analysis.

My friend reported back that he has corrected his program to take care of the tip geometry but he has to do other things for a day or so before he can get back to figuring out the flank.

This is all very interesting. I hope it works out in the end.

Sage

 

[dsage]

OK, so here's the latest.
Yes, in fact using a properly scaled up grinding wheel - in this case 108" in diameter does get rid of most of the error. See the two pictures below.
I also figured out what the gap I mentioned before was on the one side. The card with the template on it was rotated slightly i.e a line from tip to base circle was not precisely level so obviously the outline scribed by the program was not going to follow the drawing. I decided to eliminate that distraction by tracing the path using a blank card and overlay the template after to fit inside the shaded area to observe the differences.
The first picture is the tracing on a blank card. The second picture is with the template overlaid. It can be seen that there is a slight error on both sides of the tip. The rest of the trace is right on.

I'm not sure all of the error is real though. Drawing a cam in AutoCad is difficult. I drew the cam using all the same radii used to program camcalc. But I measure the duration of the cardboard cam and it is 125deg whereas the data was for 138deg. I'm sure that comes down to my ability to line up all the curves in CAD and print and cut out the template. It appears the program is scribing something with more duration so it's probably closer to correct.
But for sure there will be some error using the large radius instead of a flat because a surprising amount of the grinding takes place off the centerline and the curve instead of a flat will affect that. No grinding is going to take place where there is a curve instead of a flat and that's what the template shows - a lack of material removal at the tip.
Considering I'm using about the largest wheel I can swing and the largest practical diameter wheel at 7" I'm going to have to live with the error.
The error as far as I can see is only about .050 on each side of the cardboard lobe. Scaled down that's only .003". Some of it might disappear if I could perfect drawing a cam with the 138deg of duration.

So, I'm back to thinking this method is pretty darn good considering the simplicity and accuracy of grinding over hand filing the output of CamCalc.
I'm glad I was able to prove it.

Have I missed anything (else) ??

Sage

 

[dsage]

Oh. BTW
Dieselpilot. You win the prize. I went back through the posts and you have long maintained that there would be a problem exactly where the test shows. Had you posted the cad drawing you said you created to show the problem I might have understood what you were getting at earlier. Sorry I was so slow.
Dman was also onto important considerations to do with the point of grinding etc..
Sorry I didn't listen earlier but seeing is believing I guess. A picture is worth a thousand words as they say.

Thanks to both of you for your help. And to everyone else for for that matter for keeping the discussion going.

Unless someone can come up with where I've gone wrong on this I'm going to go ahead and grind some cams. I understand where they are wrong now and I'm happy to live with the slight errors in favour of making superior cams in an easy fashion.

(I reserve the right to change my mind on all if this - again)


Thanks

Sage

 

[dman]

Most of you guys are so smart. (compared to me).
Why didn't I think of that. Yes the size of the grinding wheel is far from scale. It's 7" diameter and it's what I used to grind the cams which are about .430 dia as a blank. For this test setup the grinding wheel would need to be 16 times the cardboard size or about 100" in dia. to be in the proper scale. (the cardboard cam is 6.25 base to tip)
I'm glad you caught that because I was dismayed at the amount of error showing up. It certainly was not indicative of what my steel cams were coming out like but I couldn't pinpoint the source of the error.
Certainly a properly scaled wheel would appear more flat (like the program is expecting and works fine with) and a lot of what shows as error would be removed.
Now I'm curious as to what effect that will have. I should be able to do that by drawing a 100" circle and use a portion of it in place of the 7" wheel. I'll try that tomorrow.

Thanks for that insight.

I had rejected the idea of moving the grinding wheel up and down early on because it would then need to be mounted on the quill of the machine (unless you had some other idea I've missed). There is too much "stuff" on the head and quill to get a setup rigid enough (I expect). If I were going to use the quill I would generate a cam program specifying a 7" diameter tool and effectively "side mill" the cams with a grinder.
Side milling is how I did my first set of cams and it worked just alright but your cutter needs to be factory sharp and holding cam blanks vertically is problematic. I had a lot of rejects with chatter etc. from flexure.

I had thought to model the whole thing in 3D CAD so it will be easier to visualize the geometry but it's probably quicker and easier to just keep making cardboard models. But if you know of a way to accomplish that I'd be interested in your results.Thanks for the offer on the analysis.

My friend reported back that he has corrected his program to take care of the tip geometry but he has to do other things for a day or so before he can get back to figuring out the flank.

This is all very interesting. I hope it works out in the end.

Sage

ok, if you are using cam then you can do that. if you did it by hand it could be easier to offset the wheel as it goes round. i haven't come up with a way to analyze it without actually offsetting, i have some concept of what would have to be done but it would be complex for sure, would take more time and energy than i have. maybe if i knew some computer programming or was better with spreadsheets and had no job i could come up with some way eventually but why when the cam programmers already have ways built into the software?.

i thought your fixture used the quill. but since it doesn't and especially since you have cam available then that's the way to go. but just to throw it out there you could turn the grinding wheel horizontal and the cam vertical and use the x and y. but enough being silly. sounds like you got the picture.

if there is a problem with symmetry or timing make certain that the wheel and cam are exactly on center height of each other as measured perpendicular to the y. if you project the motion of the cam it should line up with the center of the wheel exactly.

i haven't made a cam yet but i have put a lot of thought into the subject. i used to want to try this on full scale but it's really not worth it. if the oem's didn't know what they were doing the aftermarket would have certainly proved something weird works better by now. cam choice isn't as much of a mystery as they want you to think. david vizard has it pretty much to a science. while you can do as well or better with other methods or rigorous dyno testing the difference often isn't enough to matter and if you want to do better it's atleast a good starting point. his rules are simple enough to be understood by anyone.

 

[dsage]

dman:

Just to clarify a few thing because I know a few of them have been buried with all the discussion on the profile:

One of the unique things about the way I'm doing this is that I have the A-axis and Y-axis move commands on the same line of G-code. This way it describes an arc (tiny) between the two facets. The facets are 1 degree apart and the cut depths are accurate to tenths of a thou. No additional work is required on the cam after it's ground. I've improved considerably on the finish posted in the early pictures by using tenths. It takes a keen eye and very good light at the right angle to see the facets. In fact if I didn't tell you they were there you wouldn't notice them.

It's all automated so I just go watch TV or do other things while it's busy. It pauses about 10 thou short of the final base circle diameter. Then I measure what's left to remove and adjust the final cuts to bring it exactly to size. This is necessary because the wheel is wearing slightly as you go.

I don't use the quill for the reasons mentioned. You'll need to go back to the earlier posts to the first picture I posted. The grinder is fixed. Although I'm not using the mini-grinder any more. I made a proper grinding spindle and I'm using a proper 7" wheel.

I don't understand your proposed setup using X and Y. They are both on the table. I guess you could put the grinder on the table to move the wheel to the right position but then where would the cam go? It also needs to move slowly around it's axis.

As mentioned I could arrange the grinder like a big milling cutter and just side "mill" them. I use Mach 3. Apparently cutter compensation doesn't work properly and that would be necessary to use that method with a grinder since the wheel wears. I side milled my first set of cams with a fixed program of cuts and a carbide end mill. A good number of them had chatter marks. Holding the cam vertical is problematic since it's so small. Might be worth another try though since the tool pressures are almost nill with a grinder.

My grinding wheel is right on center. This is one of the reasons I'm limited to a 7" wheel. My rotory table center is about 4" above the table. Holding everything solid is key to success and mounting to the mill column helps considerably in that respect.

Thanks

Sage

 

[dman]

dman:

Just to clarify a few thing because I know a few of them have been buried with all the discussion on the profile:

One of the unique things about the way I'm doing this is that I have the A-axis and Y-axis move commands on the same line of G-code. This way it describes an arc (tiny) between the two facets. The facets are 1 degree apart and the cut depths are accurate to tenths of a thou. No additional work is required on the cam after it's ground. I've improved considerably on the finish posted in the early pictures by using tenths. It takes a keen eye and very good light at the right angle to see the facets. In fact if I didn't tell you they were there you wouldn't notice them.

It's all automated so I just go watch TV or do other things while it's busy. It pauses about 10 thou short of the final base circle diameter. Then I measure what's left to remove and adjust the final cuts to bring it exactly to size. This is necessary because the wheel is wearing slightly as you go.

I don't use the quill for the reasons mentioned. You'll need to go back to the earlier posts to the first picture I posted. The grinder is fixed. Although I'm not using the mini-grinder any more. I made a proper grinding spindle and I'm using a proper 7" wheel.

I don't understand your proposed setup using X and Y. They are both on the table. I guess you could put the grinder on the table to move the wheel to the right position but then where would the cam go? It also needs to move slowly around it's axis.

As mentioned I could arrange the grinder like a big milling cutter and just side "mill" them. I use Mach 3. Apparently cutter compensation doesn't work properly and that would be necessary to use that method with a grinder since the wheel wears. I side milled my first set of cams with a fixed program of cuts and a carbide end mill. A good number of them had chatter marks. Holding the cam vertical is problematic since it's so small. Might be worth another try though since the tool pressures are almost nill with a grinder.

My grinding wheel is right on center. This is one of the reasons I'm limited to a 7" wheel. My rotory table center is about 4" above the table. Holding everything solid is key to success and mounting to the mill column helps considerably in that respect.

Thanks

Sage

sometimes i just say thing that are hypothetical. not actually suggesting them. but what i meant about the arrangment was to tip the rotary fixure on it's side so the axis was vertical and parallel to the z.. you could call that an auxiliary "c" axis though if you already have indexing on the spindle then c might not be available, you could still call it "a" for your own purposes. then you could mount the grinder so it's axis was vertical and the wheel would be horizontal. in that case you could use x and y. but it's all hypothetical. without a tailstock things might get difficult. i have ideas on how to deal with that too but i don't want to get carried away. you have a good setup, you just may want to tweak the program for the sake of accuracy. or you might be satisfied...

i'm not always critisizing nor suggesting, sometimes my mind just wanders into hypothetical land when i type.... don't take me too seriously.

 

[dsage]

I'm not sure I've seen a C-axis, I assume that's on a lathe? I'l have to look that up to fully understand what you're describing.

So, you're saying hang the grinder from above (i.e. the quill or head) with the wheel horizontal, rotate the cam which is standing vertical in a rotory table and move the table in one (or two?) axis' to correct for the off axis grinding?

Is that correct?

Interesting. I guess that does add an extra dimension of flexibility to move around and get rid of the off-axis grinding problem. But I'd have to think hard what the program would look like. I'm not a mathematician.

I think in that case it would be just as easy to set the grinder horizontal over the table (fixed) and consider it a really big tool bit and write a CNC program to generate a cam shaped outside profile and side mill it. The CNC program would take care of the position to correct the point of grinding. The program would be easy as well. Just draw the cam and send it through a CAM processor and specify a 7" tool bit. (already done it with a milling cutter). Yes a tail stock would not be possible in this arrangement and as I found out before it would help things considerably.

Sage

 

[dman]

I'm not sure I've seen a C-axis, I assume that's on a lathe? I'l have to look that up to fully understand what you're describing.

So, you're saying hang the grinder from above (i.e. the quill or head) with the wheel horizontal, rotate the cam which is standing vertical in a rotory table and move the table in one (or two?) axis' to correct for the off axis grinding?

Is that correct?

Interesting. I guess that does add an extra dimension of flexibility to move around and get rid of the off-axis grinding problem. But I'd have to think hard what the program would look like. I'm not a mathematician.

I think in that case it would be just as easy to set the grinder horizontal over the table (fixed) and consider it a really big tool bit and write a CNC program to generate a cam shaped outside profile and side mill it. The CNC program would take care of the position to correct the point of grinding. The program would be easy as well. Just draw the cam and send it through a CAM processor and specify a 7" tool bit. (already done it with a milling cutter). Yes a tail stock would not be possible in this arrangement and as I found out before it would help things considerably.

Sage

technically a is parallel to x, b is parallel to y though one may change based on the other and go in line with the z. and c would be parallel to the z but i think it sometimes used for actual spindle orrientation rather than used as an auxiliary axis but spindle orientation can also be set with m19. you never hear of it because there is little need for it. 5 axis is enough to hit any side of the part from any angle so c is never really used.

if you didn't need such a large wheel it would be possible to hold the part rigidly and move the part around the grinding wheel.

again i was being hypothetical. you are much better off using a tailstock. cam software will crunch the numbers for you, just program it as side milling as you suggested, it's a good idea.

 

[dsage]

>>> if you didn't need such a large wheel it would be possible to hold the part rigidly and move the part around the grinding wheel.

That's exactly what I've already described. It's a standard side milling setup. I described it several times, once in my last post. You could use any sized wheel with no errors. The rigidity of holding the cam vertical is problematic (at least it was with a standard milling cutter. May be a non issue with the low tool pressures exerted by a grinder. I also mentioned a few other reasons why I abandoned it.

Sage

 

[dsage]

We've finally come up with a program to grind cams properly. I started a new thread at:

http://www.homemodelenginemachinist.com/f26/grinding-cams-cnc-program-beta-test-21127/#post223130


Sage