Question on generated bevel

[[Ow...]]

First off, I know very little about bevel gearing and I’m provided very little information about the bevel when I’m asked to create a program using the best gear we have to compare others to.
I’ve generated nominal data on a master bevel gear and get decent <±0.005 variation on four flanks.
The problem I’m having is that when I check another gear that is also known to be of very good quality, the tooth topography shows like the tooth geometry would be close (see pics) to being the same “if” I was able to best fit the actuals in the Z axis and I see nowhere in bevel basic where this can be done. I’ve read that the bevel-advanced license does have this option?
The reason I believe the difference I’m seeing is in Z is because the shoulder area where the gears mounting distance would be measured (or located on in a fixture) is inaccessible on the rotary table and I’m using a different shoulder to use as zero when generating the nominal data and that shoulder is allowed variation greater than the mounting distance shoulder that is inaccessible. Make sense?
In basic bevel when in the topography tab under mounting distance, there is a button to reset the mounting distance correction value before measurement starts but, the mounting distance is not measured when generating nominal data from a master gear (that’s exactly what the manual states) so without that, I’m not sure there is any other way to make the best-fit needed?
This is difficult to elaborate on in written form but, does anybody have any idea on how I might get bevel to best fit the topography data?
Would the option in Bevel advance allow this best fit?

[[Ro...]]

Yes, GEAR PRO bevel advanced has the possibility to calculate axial-best fitted results to determine a shifted mounting distance.
If you send me an .act I can do that on your part if you don´t have the license.

[[Ow...]]

Please sign in to view this quote.

I've reached out to the Zeiss help-line to help answer a few questions regarding Bevel , maybe find a work-around and to get a quote on the advanced edition if needed. If it's just not in the budget, I may send you an .act file to help show how the advance edition could help and that would be greatly appreciated.
I can't even find any literature on Bevel Advance that would show what extra capabilities it would offer.

[[Ro...]]

The US colleagues should know the differences between base and advanced 😉
However, just have a look into the release information which you can find in every GEAR PRO software installation (e.g. in the menu bar => ? Info => Release info)

[[Ma...]]

Owen,

To be clear, the mounting distance optimization in the advanced version operates without re-measurement of the part. With the basic software you can change the mounting distance or the Z reference freely and accomplish the same task by trial and error, obviously not as efficiently. In your case where you are using a different axial reference than specified, the best solution may be to measure the distance between the two axial surfaces on each part and compensate the mounting distance for each part measured by the difference in this axial distance relative to that of the original reference part.

[[Ow...]]

Please sign in to view this quote.

Thanks for the suggestion Mark.
Measuring the difference between the two shoulders is good idea, I just don't know where to change it at once the difference is known.
As stated, I haven't worked with the Bevel much at all and never had any training specifically on it.
When generating the bevel nominals, the instructions specifically say to set the mounting distance to zero in the geometry window and when I've tried putting MD in, it doesn't work when trying to generate the data.

After I've generated the teeth and created the gear-pro model, I've tried adding an offset in the geometry window (as shown below) but, it wants to generate a new model (that's fine as I don't save it) but, when I try to re-run the actuals, it throws up an error saying it has trouble accessing some parts of the tooth?
You mention changing the "Z reference", where in the gear-software would the Z reference be changed?

Because we don't deal with Bevel much, even though it would add un-needed time, If I could make some trial & error changes, without affecting the integrity of the measurement to make the best fit, that would be great as the approximate 10K cost of the Advance license is a bit to swallow for no more than we're going to use it.

Trying to get this to work off-and-on while tackling many other jobs doesn't make it any easier but, I appreciate everyone's efforts in trying to help me learn/figure this out.

[[Ma...]]

Please sign in to view this quote.

Owen,
I don't have a lot of experience with "nominal generation" in Gear Pro but I used this function quite a bit with UMESS/GearBevel and the general functions are similar.
A typical procedure for generating nominals from a master gear would go like this:
-Define the part coordinates (and link to rotary table axis) in Calypso
-In Gear Pro generate nominal data for one tooth of master and save nominals.
-Measure 4 equally spaced teeth on the master
-Convert this actual measured data to new nominal data.
-Measure 4 teeth on the master again and flank form and tooth thickness deviations should be very close to zero. If not convert actuals to nominals again and repeat until results are very close to zero.

The mounting distance must be set to zero when generating nominal data and should be left at zero when measuring the master part. When measuring additional workpieces it should be possible to change the mounting distance in the program or apply a temporary offset to it at measuring time.

Alternatively, you could shift the Z coordinate in Calypso to account for the differences between the master and the workpiece.

Also, while it is possible to optimize the mounting distance by trial and error, you must consider what use if any that information will be. Do the parts have adjustable mountings? Have the mounting distances already been established by roll testing, etc.?

Good luck

[[Ch...]]

Gentlemen,

I agree the instructions say to put mounting distance at 0, however I would not recommend it. I've done it both ways in the past. 0 might work, but will never show anything meaningful on the reports.

I suggest after running alignment in Calypso, toggle over to gear pro and map in the master at appropriate M.D.

as Mark suggested, after the original file has been measured, there are a couple ways to store it as the nominal 'master'. After you do so, you should rerun to make sure results are close to zero (<.0001" basically inspected to itself, etc.)

If you are generating nominal data path without CAD model, I feel your pain, it can be a bit challenging. Having a CAD model helps a great deal to select grid points on the flank (make sure tooth is aligned to +X).

I am lucky because our Eng Dept builds our Gear Pro programs from G-AGE. I only have to make the Calypso program and minor settings in Gear Pro.

The M.D. Offset you are referring to is for parts cut at a different M.D., the only time I see this available is when program is started from Gear Pro or GAGE :
I believe this 'difference' input field only like metric values so be careful.

Good luck.
Welcome to bevel gears.

[[Ow...]]

As always, I appreciate everybody's suggestions.
I'm hoping to get back on it next week sometime and I'll be sure to update.

[[Ow...]]

I’ve finally got a chance to get back on these and while cumbersome and frustrating would be putting it mildly, I have one question.
After transferring the actuals to nominal and re-running, the global report that shows the results of all four teeth on one page, shows almost perfect with nothing more than .001mm error on the grid.
However, when you look at the individual report for each tooth, they are not perfect, with as much as .006 flank error on some parts of the grid like in the picture showing tooth33.
Why is this?
Is the global all four teeth on one page report the only report that shows the best fitting?

Tooth 33.jpg

[[Ro...]]

The Global Format shows always the mean value calculated from all measured teeth.
If you digitize e.g. 4 teeth and set them to nominal the Global Format should show (almost) 0 after a second measurement run.
That does not mean, that all individual teeth have to be 0.
E.g. if you have on two different teeth exactly on the same position one time +20µm and on the other tooth -20µm the mean value is 0.
However, the master gear method is a very sensitive thing.
I recommend always not to make the measurement plane (corner points) too big.
Especially in root and tip area roundings can dramatically influence and make someone go crazy.
If you have defined a useful area digitize in a first step 4 teeth and set the mean of them to 0.
After that rerun and measure e.g. 8 or all teeth and set the mean of them to 0.
After this procedure you should have pretty good master gear data.

[[Ow...]]

7/26/22 Up-date:
As can be seen in the plot pictures, I have not been able to get very much change in the topography plots regardless of offsetting with the Mounting distance or changing the Z with a secondary alignment on the Calypso side or trying tooth thickness modification with looping. I did make a probe where I could access the actual shoulder that the mounting distance is taken from but, it didn’t make any difference. I also tried regenerating the nominals with a mounting distance and got nowhere so went back to having the mounting distance set to zero and regenerated (digitized) the nominals for probably the 5th time. Taking Romens comments into consideration regarding going crazy, I’ve paid close attention to avoided the root and tip areas. Repeatability is good as I’ve checked each gear at least 3 times and they check the same each time.
I’ve talked with tech support a couple times (teamViewer computer log-in) and had zero success in getting any different results than I already have (see pics). Tech Support has downloaded the actual files and tried making adjustments in the advance version of bevel using axial best-fits and wasn’t able to get any better tooth fitting than I got using several different methods as descried above.
Tech support has been good, and I appreciate their time helping get this bevel-newbie guy going in the right direction and when the time is available, Tech support has agreed to take it even further and try making some modifications on an actual CMM with an actual gear and get back with me.
Questions:
Looking at the pictures of the global report for each gear (all new gears, never-used gears) would this kind of variance be normal?
Maybe the addendum and dedendum isn’t that big on this unknown angled bevel gear and I’m checking parts of the flank I shouldn’t?
Another thought, the area the tooth pitch is picked up (grid 3) always comes in close to zero on all 3 gears so I'm not sure if it's actually moving when I try to make the adjustments or if the grid is just set from that location and is always going to be zero?.

I’m lucky to have the time to get this figured out so, before I take more costly actions (paying and sending gears to Gleason) I thought posting here might bring something up I’m doing wrong and if nothing else maybe provide future users tips getting started. 😕

[[Ch...]]

Owen,

Congratulations. You're doing well. I think I've been in your situation a while back learning Gear Pro bevel, hang in there.

While the topographical results could be real, I think the runout seems a bit high. Are you using R/T Qualification with a feature on the part? (at RT 0° & RT 180° ,etc.) High runout will effect topographical results as well. If not, have you indicated the part in to run best condition ?

nice photo, I've inspected similar bevels with splines in the bore. I have to assume the datum journal is the OD on the bottom of the part. If so there's not a lot there left to access.


Good luck, let us know if you need further help.

[[Ow...]]

Please sign in to view this quote.

Thanks for the reply Chris.
I’ve always thought I’m very analytical when learning and figuring things out but, everybody will have that “can’t see the forest through the trees problem” and having people making suggestions will often help clear it up.
Yes, using RT method two and using the bottom cylinder as Datum at zero and 180. The ID spline is a max dia fit and I would consider using it as datum but, so far, the Spline has been concentric with the Ø55MM OD hub (less than .010 runout) and I like using the biggest surface (the cylinder) I can scan.
I've considered using the mounting distance shoulder for spatial which is underneath the gear but, It's not very big in dia. maybe Ø65mm, and I don't think I'd see much difference because the gear teeth seem to be perpendicular to the shaft OD.
Regarding the runout of the pitch, I agree it's excessive and the more excessive it is, the worse the global plot looks. In the pitch evaluation, I can turn eccentric correction on an make the pitch look better but, it DOES NOT change how the topography is evaluated, I wish it did but, it doesn’t . That said, while the global topography shown is excessive (all four teeth) each individual tooth plot shows the flanks getting smaller about the same amount so, if it was the runout, you’d think they would be bigger on one side and smaller on the other?

[[Ch...]]

Owen,

great job, this stuff takes a lot of patience.

As you can imagine with a lot of runout / eccentricity the pattern or /topographical should shift or move around the tooth typically following the runout.

Did you put an indicator on the OD to verify as well? Sometimes the RT Qualification compensation can only do so much, and when I'm checking a master bevel or master gear/spline gages. I always indicate in as well as RT Qual on part for optimal results.

For spatial datum : I would say you have to use whatever is specified on drawing. However even when the Primary Datum is not the mounting surface, I've had engineers, in fact our current ones as well, say they always want the mounting surface as primary for bevel inspection.

What does Production/Engineering say about the runout ? I suppose it could possibly be real.

[[Ma...]]

Owen,

It is typical that the individual teeth show some deviations even when the average tooth plot is very close to zero. We measure multiple equally spaced teeth to eliminate the effect of axial and radial mounting errors on the average flank form. With the amount of cumulative pitch deviation you are showing the variation between teeth is likely to be significant.

Another thing to consider with bevel gears is that they are normally sold in sets and gears from different sets may be significantly different but both may run correctly with their respective mates.

Generally speaking, the datum surfaces used in the CMM should be the same datum surfaces used in the final application.

[[Ow...]]

Appreciate the replies guys.
It will probably be next week sometime before I can get back on them but, it being a straight ring bevel, I'm going to remove the rotary from the equation and check it without rotation using the ID spline as spatial alignment and see if it changes any.
I'll probably also use 8 to 12 teeth (currently using just 4) to best fit the master to.

Thanks again, you've brought some good food for thought.

[[Ow...]]

UPDATE:
I finally had a chance to get back on this project (luckily it’s not high priority) and re-created the program to measure the part OFF the rotary table to help ensure it wasn’t adding any bogus alignment/rotation errors that may be affecting the results. Without the rotary table, it performed no different.

I know this is difficult to discuss on a forum so, bear with me as I try to explain.

When re-creating the topography nominals, I used 12 teeth to create the nominal data via actuals-to-nominals and paid close attention not to get in the root or the tip area.

Checking the 2nd part to the newly produced Master part topography didn’t look any different than the 1st time with the rotary table program (shown earlier in this post) and because I had no luck using any kind of mounting distance inside the software (probably because the instructions say not to when using the master part method), I used space points on the tip angle to gauge how far the tip profile was off from the original master gear and used that amount as a reference for a Z-offset adjustments with a secondary alignment to potentially simulate a mounting distance offset that is not accessible because - I don’t have the advanced version of Bevel.
Adding the Z offset (0.100mm) with the secondary alignment did change the tooth thickness measurement to where it was correct but, it didn’t hardly do anything to the profile plot, a little but, not much. I believe this is because of the acute ratio between the Z-axis and the profile data which is angled approximately 65° to the Z axis and the Z offset moves the profile more to the left or right (uphill/downhill) of the angle and the nominals of the profile stays the same along the pitch angle, making it not a true 3D best fit that I think would actually move it in the vector angle of the tooth profile. So, I’m not sure what the axial best-fit available in the advanced software would do but, if it just compensates in one direction, I can’t see it doing anything different than what I did.

I did call Tech support at Zeiss and they grabbed the program data/actuals and couldn’t get the axial best fit to do anything on their end with the advanced option 🙄 .

That said, on the alignment/feed-in page of the bevel gear program, there is a grayed-out checkbox for a tooth thickness best-fit (advanced option I don’t have) and maybe that measures the gear tooth width and adjusts the profile and nominals to where the tooth thickness nominal should be?

I realize this is big learning curve with me not working with bevel gears much but, if the 10K advanced gear-pro license is needed to make the reports reflect the actual tooth profile condition along the actual pitch line, I’m not sure why anybody would use the bevel gear basic option to begin with 😡 and even it’s not cheap.

[[Ch...]]

Owen,

Hi, thanks for the update.

Are you entering a 0.0 Mounting distance?, this is what I interpreted the instructions as saying as well originally. However I would recommend using the nominal/basic mounting distance always. From my experience now, having worked with it for years, I think this 0.0 may be referring to any MD offsets.

Regarding tooth thickness correction : I would always recommend using this as well on bevels and on cylindrical gears.

I agree their prices are a bit high. I understand supporting the developers, but I think once upon a time it was one license fee for Gear Pro and you got basic and advanced everything, cylindrical, bevel, worm, hob, etc.

Can you share a summary page of the mapped master ? before measuring it to itself, and one or 2 charts of single tooth topography, so we can see any of your concerns, errors?
4 teeth is usually enough to map a master - the grinding/forming processes are usually pretty repeatable, etc.

Regarding tooth thickness - have you experimented with making custom bevel charts yet -it's also a learning curve, but you can get the results to show in proper inches if needed 0.00010" etc (looks like you may be in metric) , although it can get a bit cluttered. The reason I also mention this is because the tooth thickness shown, may be chordal to circular, and may differ from you print requirement.

Also if I didnt mention before it is good practice to mark the start tooth or all tooth mapped on the master for re-checking later.

Good luck.

[[Ow...]]

Please sign in to view this quote.

Thanks for the quick reply Chris.
When I enter a mounting distance on the geometry page, and then go to data generation to map the tooth, the probe tries to scan the tooth well above the part. I tried creating an offset alignment that would compensate for the added mounting distance , the dimension between calypso z-orign and MD, and make the software think the tooth was where it should be but, that didn't work either so, I just continued to use zero in the MD.

On the tooth thickness correction,I did have it turned on but it didn't make any difference, it still measured the tooth thickness being wider (meaning it needed to be shifted up or more correctly, up parallel to the 65° tooth pitch axis) but, the axial besfit-nominal tooth thickness option is only with the advanced version and that's where I think it would maybe like loop check until it found the tooth thickness nominal and reset the flank nominals to that location.

It might be a day or two, really busy today but, I'll get summary pages loaded and let you look at them when I get a chance.
Thanks again!