Large Scale Gear Fatigue Study Done by Dudley 9

[ExRanger]

I have heard that sometime after WW2, Darle Dudley conducted a large scale study of gear fatigue that may have been funded by the US government. He supposedly came up with S-N curves.

Does anyone know if the results of that study are still available anywhere? I can find no trace of it on the internet, but that may be because its too old to have been digitized yet.

Thanks in advance for any help or leads.

 

[geesamand]

I have a copy of the the Dudley book at work. I'd imagine the reference has to be in there. I will try to remember it on Monday (try).

David

 

[tbuelna]

ExRanger,

Use of S-N curves for gear fatigue will likely give inadequate results. When analyzing gears and/or shafts, you should consider the total combined loads on the part. This would include tooth root stress due to bending, torsional effects in the gear shaft, bending in the shaft between bearings due to separating and tangential forces at the tooth contact, hoop stresses in the gear rim due to CF at high speeds, and dynamic tooth loads due to to torsional vibration and/or tooth geometry errors. Something like a Goodman (?)diagram would be more appropriate.

Also, fatigue life can be significantly affected by things like material quality. Gear steels are much better quality now than they were when Dudley was doing his best work.

Regards,
Terry

 

[dinjin]

I took a course with the author and he handed out a supplement. He did not test gears but used rollers
against rollers to simulate the contact of a pinion against a gear. He was encouraging roll tests to be done for a good approximation and felt from the limited tests that they had done, 40 percent increase was typical. I left all of my gear stuff at work when I retired and it probaly made the scrap bin. I will check the next time I stop by the company if it still is available. His big push was definitely long and short addendum gear systems especially on internal gears. He was bought out by a group and his book is available from them and maybe the supplemental material as well.

 

[dinjin]

It was not Dudley that I took the course from, it was Buckingham.

 

[ExRanger]

tbuelna,

Using S/N curves developed from small, polished material specimens for gears would be inappropriate unless one actually factored in a variety of things (this should be worked on).

However, many times S/N curves are developed based on tests from actual gears and thus they include all of the effects you mention. These are the kind of curves I had in mind, and I thought I had understood that Dudley tested actual gears with a range of different parameters. This kind of testing is very expensive because you have to make gears as the test specimens.

 

[israelkk]

ExRanger

I do not really know what is your intention, is it for the historical value of this information or for practical use. If it is for practical use then you are missing the point. All the accumulated data is embeded in AGMA gear rating standards based on the accumulated empirical and theoretical knowledge and develpments from Dudley's work, Buckingham's work and the contribution of all others in the last ~100 years. Therefore trying to invent the wheel and use Dudley's test results to design new modern gears is wrong. Isolated tests of actual gear testing data is questionable because there are many variables beyond the actual material properties. For example: the casing rigidity, the shaft rigidity, the gear accuracy such as AGMA 12, the bearing accuracty such as ABEC 7, etc,. (I assume the bearings and gears Dudley used were less accurate than the standard bearings and gears you can cheaply buy or manufacture today).

 

[ExRanger]

First of all, I never stated I intended to use the data for anything, so please don't anyone jump to that conclusion.

Secondly, give me a little credit. I know gear materials and manufacturing have changed over the years. That's trivial.

My understanding is that the AGMA actually used Dudley's test data at one point in time to establish gear S-N curves and allowable stresses (AGMA 411 I think it is). My interest is to see the data and understand how the tests were done, what material(s) were used, etc.

Again I don't intend to use this data to directly design anything. I never stated that I did.

 

[tbuelna]

israelkk makes an excellent point about gear profile and index accuracy. Dynamic tooth loads due to index or profile errors can be 2 or 3 times what you would expect statically. Reliably producing class 13 or 14 gears back in Dudley's heyday was not really an option.

Also you underestimate the impact of modern metallurgy on the fatigue life of both gears and bearings. The recent improvements in contact fatigue properties, especially with rolling element bearings, is due almost entirely to improved metallurgical processes.

AGMA formulas are widely accepted for gear design and they are constantly being reviewed and updated. But since they are somewhat generic in nature they must also be very conservative. AGMA may well have used some test data produced by Dudley at one time, but Dudley's data has likely been superseded long ago.

Here's some good reading regarding the results of modern FEA techniques vs. a conventional S-N or Goodman interpretation of gear tooth root stresses:

http://130.15.85.212/proceedings/WorldCongress07/articles/sessions/papers/A488.pdf

Regards,
Terry

 

[ExRanger]

Tbuelna,

As I said, I am aware that modern materials and processing improve gears in many ways. And again, I do not intend to use Dudley's data to design anything. But trust me when I say that companies are still using it. If they use old stress "allowables" that underestimate the capabilities of modern gears, then what they are doing is in at least some ways conservative. That is good enough for some people and applications. I agree that its not the best we can do analytically.

Dudley's data has been replaced with new data in the AGMA specifications, as you say. I know this, and never made any argument to the contrary. You're not telling me anything new. Once again, I do not intend to use Dudley's data myself.

I have been involved with the development of dynamic models for gears meshes and I am well aware of the differences between the static and dynamic forces. However, this does not necessarily negate the concept of using an SN curve to design gears depending on what you want to accomplish.

There certainly are better ways to handle fatigue calculations than using Goodman diagrams, SN curves, etc. That's a good discussion - it has nothing to do with my original question, but its something that I'm very interested in. Thanks for the link.

 

[tbuelna]

ExRanger,

With regards to endurance limit in Ch.3 of "Handbook of Practical Gear Design", Dudley references his own work:

ref. 130 Dudley, Darle W.: Gear Wear, Wear Control Handbook, ASME, New York, 1980, p.755

Hope that answers your original question.
Regards,
Terry

 

[ExRanger]

In case anyone is interested I finally located a paper having results from this study:

Seabrook and Dudley, "Results of a Fifteen Year Program of Flexural Fatigue Testing of Gear Teeth", 1963.

Its AGMA paper No. 63-WA-199

They broke around 3000 gear teeth of varying materials. Not of any real practical use today, but it helps one to understand how gear design has evolved and what underlies some of the design parameters used.

 

[GKBUCKINGHAM]

Interesting Topic. Earle Buckingham was my GrandFather, I spent many a Hour changing the Rolls on these Roll test machines as I was growing up. This brings back memories.

FYI. I am working on taking all Gear Programs that my father used for 25 years and putting them on the web.

 

[mfgenggear]

Mr Buckingham

You honor us by your visit.

I can not wait to view your web site.
I have three of yours & your fathers books section 1 thru 3

thanks

 

[ExRanger]

GKBuckingham,

I have a well-worn copy of Analytical Mechanics of Gears that has been useful many times over the years.

Small world - interesting to hear that you were involved with those fatigue tests. I would be very interested to see the website you mention when its done. Thanks for dropping by!

 

[GKBUCKINGHAM]

Anther note. My father (Eliot) donated alot of my Grandfathers Papers to the American precision Museum located in Windsor Vermont. I do not know all what is there, but if you are ever passing through, you might want to check it out. The web site is

http://www.americanprecision.org/

Also the Orignial Project was done at MIT and was funded by several Large Manuafactoring companies. United Shoe being one of them. caterpillar was also involved. I think there was several more. This is how my Grandfather got to MIT, as the companies told MIT, they would fund the Study, but only if my Grandfather was the one doing it.


In the 70's we ended up with 4 of the Roll test machines from United Shoe. My father used these up till about 1997 when he retired. he was doing alot of testing on Plastic rolls and powdered Metal. He sold the 4 machines to a company who I believe is still using them.

Caterpillar redesinged the machines, and I belive is still doing testing with them.

The Gear Programs I am working on were originally written in 1979 and were used upto 1997. This acutually was my first job, as I was in college during this time, and did most of the coding, with my father checking out the results.

My father sold off all the rights to the Books in 2000 but I kept the rights to the programs.

The programs are written in quick basic, but I am converting them to Java. The first program I am working on is the Worm Gear. This includes the Basic Design, Contact Diagram, Profile, Buckingham Load Factors and the AGMA Rating.

It takes time, as I have to check all the Formulas, and adjust the numerical precision to make sure I get the same results as the old program.

My plan after I get the web site going, is to acutally make the program available on the Iphone and Blackberry as well

I then have Helical Gear, Spur Gear, Spiral Gear and a Internal Gear program to do.

I have not figured out how to charge for usage. But I plan on making it free for some, so they can evaluate it and give me feed back on it.

 

[gearguru]

Mr. Buckingham,
I always admired the work of your grandfather. I also have the "Analytical Mechanics of Gears". This book helped me to learn how to solve many of my gearing problems.
Thanks for keeping his a your father's valuable work alive.

 

[tbuelna]

GKBUCKINGHAM,

Your Grandpappy was the man!!

I have a copy of his book "A Manual of Gear Design". I am truly impressed by all of the detailed, meticulous, and extensive hand-written tables and references within it. It must have been a real labor of love, created by a mechanical engineering artist absolutely dedicated to his profession.

Your grandfather's masterpiece holds a place of distinction on my bookshelf.

Regards,
Terry