Surface Durability Confusion 2

[ZCHSC]

I'm designing a light duty gearbox based loosely on an existing unit purchased from an outside vendor. It exclusively uses PM spur and bevel gears, made of a relatively strong alloy (FL-4405 or 5305 with heat treating.)I'm using MITCalc, as well as Dudley's Gear Handbook, in the gear design process.

I am confused, however, about surface stress and durability in the gears. According to the software calculations, HP throughputs commonly encountered in normal operation are enough to create massive surface pressures (in excess of 250 KSI in the differential pinion). MITCalc predicts a surface durability saftey factor of less than 1 for these conditions, despite that this transmission is intended for a sub-500 hour life. According to my MPIF handbook, 250KSI is the extreme limit of rolling fatigue even for the highest rated alloys.

On one hand, I know these gears are sufficient based on thousands of purchased units that have operated well beyond their predicted life. On the other hand, I cannot justify the current gearing dimensions with any calculations. Bending strength is never an issue.

Any thoughts on how to treat these questionable surface pressure predictions?

 

[gearcutter]

I too have MITCalc but do not use it for gear calculations as it is always overly conservative.
I don’t know why this is and have just put it down to “you get what you pay for”.
The surface pressures you have calculated; did you use equations from Dudley’s book or are the results a MITCalc output?


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

 

[ZCHSC]

MITCalc is very conservative for sure. I agree with "you get what you pay for"...for $300 it was worth trying. It's almost worth $300 to have something that can crank out accurate CAD models as quickly as MITCalc can. But you're totally right.

I might lobby for the KISSSOFT package if it looks like MITCalc is going to be more of a headache that it's worth.

But the next step will be to sit down with the Dudley book and grind out the stress; although I gotta say I seem to have more difficulty then usual following that book. It seems to lack the logical progression and design procedures that less detailed design texts are good for.

 

[Matt51]

If you use Dudley's book, you can do very preliminary sizing using unit load and K factor. With a simple hand calculation, you can see if you are in the right ballpark. Estimating Bevel Gear Size, page 247 of Handbook of Practical Gear Design, 1984 edition.

 

[Matt51]

Fairfield Gear used to have free software you could download, and I thought it had bevel gear ratings. This was ten years ago, I do not remember if they still offer it. I thought it worked very well, and was entirely accurate.

 

[gearcutter]

I have a PDF copy of a load calculation example according to Gleason straight coniflex bevels I can send you. It is straight forward and well laid out.
You can contact me via this link; don't post your email on this forum as the thread will be deleted.

http://www.aussieweb.com.au/email.aspx?id=1194181

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

 

[zafiro301]

I encountered a similar problem using KISSsoft. I don't think it matters which software package you use, the calculation methods for rating are basicaly the same in both software packages (ISO, DIN, AGMA, etc).
I have the latest edition of prof. Dudley's handbook, but my problem is that the customer demands calculation to be done according to ISO 6336 or DIN 3990. First I tried with KISS soft and the calculated safety factor for contact stress was too low. Then I decided to make a simple MATHCAD calculation based on ISO 6336 and I got almost the same results as KISS soft.

Ivan Zafirovski

 

[israelkk]

zafiro301

If your customer demands calculation to be done according to ISO 6336 or DIN 3990 then this is what you should do. ISO 6336 or DIN 3990 are more conservaive than AGMA. There are few comperative articles you can found on the web that shows examples for the differences.

Your customer should be willing to pay for larger and heavier system if he demands that the calculation should be done according to ISO 6336 or DIN 3990.

 

[zafiro301]

Thanks, israelkk

The company I work for has signed an awful contract for the production of the reducer, so the customer is in a position to demand larger and heavier system without paying for the "extra rating" of the gears.
I've studied the ISO 6336 carefuly. University professor told me that its an english translation od DIN 3990, but more conservative as a result of modification of the S-N curve.
Best regards
Ivan Zafirovski

 

[dinjin]

I would think that the static calculations would be more important than the durability calculations for sub 500 hour
operation. Also, how much wear is permitted before the gear is inoperable? We used to use a factor of 5 for service life verses B10 life of bearings. I would think something like this would be applicable to gears as well.

 

[Tmoose]

at 1000 rpm 500 hours would engage a tooth on the pinion 30 million times

 

[tbuelna]

ZCHSC,

You need to provide lots more info. Contact stress is only 1 of many things that you need to evaluate in a gear mesh design. You should also check for tooth bending and scoring. Even a very high quality carburized VIM-VAR alloy steel gear would have a relatively short number of acceptable load cycles at 250 ksi Hz stress.

If your contact stress is truly 250 ksi, I'd guess that you're probably not operating in EHL conditions anyway, assuming your PL velocities and transmitted powers are average. So your bigger problem will likely be a scoring failure.