Hunting tooth life extension over an exact 2:1 ratio

[tensor47]

Hi all,

I do condition monitoring and over the years 3 gearboxes with exact 2:1 reduction spiral bevel gears have failed with spalling of the pinion teeth at the tips only.

The loads are smooth and about 50% of the design loads for this gear pair.

The following 2 helical gear sets have lower design loads but have never failed.

The manufacturer claims to have load compensated profiles for high area contact at full load. He says "AGMA 2003" but when I look there are only 3 hits and the most relevent is about plastic gears.

The lube is good.

The things that I wonder about are :-
Position adjustment of the crown wheel and pinion. (most polishing at the addendum)
Hardness of the core material and surface material.
Do AGMA standards calculate life expectancy with a knowledge of hunting tooth or exact ratio's ?

I don't have acccess to AGMA standards so if you know the page where this calculation is done, or if you know it's not done can you advise me please.


Cheers

John

 

[gearcutter]

Help us so as we can help you.
Better quality images would be a good start.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[mfgenggear]

Tensor

If you have more specifics of the gear design, material, Heat treat then a better analysis can be given by the great people here.

Mfgenggear

 

[mfgenggear]

wanted to add, if you do not know the above then send the gears to be analyzed by a gear shop.
verify all of the gear attributes, then send the gears out for a full met lab analysis, Material, Heat Treat, if case present. Case Depth & Hardness of Case & of Core.
then ask the lab to do a failure analysis,
These are are all crucial questions.
Do you have access to the designs, are loads acceptable?
what differences between the Helical gears vs The Spiral Bevel gears.
The setting of the backlash of bevel gears is very crucial.

HTH
Mfgenggear

 

[tensor47]

Hi Chaps,

Thanks for your input to date and response to your questions.

The pic sent is the only one for the first failure but it does show the tips are completely gone, not just the loaded side.
There was a single tooth fault at a low level, which then increased and decreased several times in a yearly cycle, and finally many teeth failed.

The most recent fault was acted on after several months with what looked like a single tooth fault, but again at the end became a multi tooth fault within weeks.
To me the failure modes look identical with the tooth tips missing.
Attached are 6 pic's of the latest failure.

Do you have access to the designs, are loads acceptable?
They did divulge the design loads for each gear pair, and we have measured the real loads which are smooth and about 50% of the design loads for this gear pair.
The following 2 helical gear sets have lower design loads but have never failed.
They claim the cause is overload, but the evidence does not support this claim.

What differences between the (parallel shaft) Helical gears vs The Spiral Bevel gears
Not sure what you mean.

The setting of the backlash of bevel gears is very crucial.
Yes I agree they are a factor to be considered. Does the tip loading suggest poor meshing geometry as the failure cause ?

Manufacturers keep material and hardening details a secret so I don't know the details.
I have recomended that a metalurgical and failure analysis be done.
I think a cross section of a good tooth and bad tooth with microstructure and crack pictures taken.
Also hardness over the whole area to find the case thickness etc.
These can then be compared to the spec's (if they make them available) at that stage.

I think the following is part of the cause :-
When new, there is a slight "click" at crown wheel speed which gets stonger over the years.
In my view the exact ratio imprints the fault on a particular pinion tooth.
I think a change to a non exact ratio will "average" the variations rather than reinforcing them, ie hunting tooth.

Do AGMA standards calculate life expectancy with a knowledge of hunting tooth or exact ratio's ?

I don't have acccess to AGMA standards so if you know the page where this calculation is done, or if you know it's not done can you advise me please.

Cheers

John

 

[tensor47]

Looks like it over writes so here is another one.

I've asked that the parts not be thrown away.

Cheers

John

 

[mfgenggear]

tensor

The gear teeth appear to have a very rough finish, the contact of the gear teeth appear to be incorrect & not central.
gears that are mis aligned will self destruct., if the backlash & alignment is off this can cause such a failure.
because it will mesh & act like the incorrect mod or Diametral pitch & pressure angle. so it will self destruct.
to verify this a contact pattern procedure should be completed(if possible do in the box with new gears) & then analyze the data. & then inspect the backlash.
this is the first step should be taken.

In reference to the Helical gear vs the Spiral Bevel Gears.
is it the same material?, the same heat treat, the helical gears are on center, & is backlash is correct.

has the oil been analyzed for contamination

If original gear supplier does not want to release the gear data then a new gear vendor must be contacted, & the gears must be
reverse engineered. without the data it is difficult to run the numbers.

I am not good with analyzing for pitting or fretting I will defer to the experts.
but I suspect that the tooth failure is several factors & is not easy to decipher.
I suspect there may be incorrect heat treat, bad alignment, rough finish & poor quality gears & lube issues.

as far as AGMA criteria, some of the other Senior Design Engineers will pop in.
I not in the office.


HTH
Mfgenggear

 

[gearcutter]

It would be helpful if you removed all of the grease and oil residue from the teeth surfaces before taking any photos.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[Strong]

"The loads are smooth and about 50% of the design loads for this gear pair."
Is this statement based on actual measurements of static and dynamic torque? A torsional resonance could cause/contribute to this failure mode.

Walt

 

[tbuelna]

tensor47,

I'm not sure you have a gear geometry problem per se. Looking at the first picture you provided, it appears to me like you may have a housing structural stiffness problem instead. You have a cantilevered pinion, and the outboard pinion bearing does not look like it is well supported by the housing. As a first step, I would suggest an analysis of the structure around the gear-side pinion bearing to determine if there is adequate stiffness to prevent excessive mesh displacement. This is a very common problem with overhung pinions.

Good luck.
Terry

 

[tensor47]

Hi chaps,

Here are my answers in order of your comments.

The pics are from the manufacturer, and we need to leave them as they are for the forensic part later. So we are stuck with what we have to look at, sorry.

Yes the load has been measured via the VF drive outputs and current/power limits are known. It's also measured as current at a high sampling rate to look for short term peaks and there are none of significance.

Yes Terry, the bearing housing is held onto the upper and lower halves by a ring of 6 or 8 bolts to the end of the gear case and now that you point it out there is a big cantilever leverage there to allow deflection of the pinion rel the crownwheel.
Well spotted.

In terms of general design excellence, they talk about deflection of the crownwheel and pinion in bending and torsion and tooth bending, and even about housing deflection.

I wonder if they did all this design work prior to 1996 when these gearboxes were supplied ?

Here is another pic to help.

The AGMA design allowance for exact ratio's is something I'd really like to know about !

Cheers

John

 

[gearcutter]

"I have recommended that a metallurgical and failure analysis be done."

I think your own recommendation, in this case, is sound advice.


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[tensor47]

Hi Chaps,

I forgot to cover the torsional resonance possibility.

The VF drives ramp up and the gear noise/vibration increases with speed to full speed, so there is no resonance being passed through.
This does not prove there is not one just above running speed which has an amplifying effect, but the load measurement says the real dynamic loads are only 50% of the rating for this gear pair, so it's not overloading the gear pair.

Also there is a soft pin and bush coupling to the motor and given the inertia of the motor and the low speed agitator blades (mult by ratio squared) it's likely the torsional resonance is very low, say 5 to 10 Hz.
This is close to crownwheel speed of 11 to 13 Hz but there is nothing out of the ordinary in the vibration at this frequency.
It's also well away from the toot meshing frequency of around 500 Hz.

Regarding the overhung pinion design.
I also looked at a group of 4 gearboxes (same manuf) on the same site, driving large (like porridge mixing) agitators purchased in 1980, which have lower power levels, and are physically a bit larger. They have a larger overhang, but there has never been a C&P wear problem.
I will try to get the real KW data and pinion characteristic diameter to see how the tooth loads compare.

I also looked on the web at the latest design from the same manuf and the overhang seems to be less than the problem machines, so there is a clear design path to ever less overhang from 1980 to now.

I'm going to suggest mounting a horizonatal and vertical proximity probe from inside the outer gear case to measure the relative movement of the pinion bearing housing from no load to load.

Cheers

John

 

[mfgenggear]

I came across this about overhung load.
It's worth investigating.
Nu luck with the AGMA spec's I have in regards to Hunting Tooth.

Mfgenggear

 

[geesamand]

I see you now mention agitator.

If this is the low speed pinion set and there is no quill / flex coupling on the output shaft, then you may be experiencing excess deflection transmitted into the gearbox. Not all gearboxes used in agitator service are appropriate for the overhung loads.

I know AGMA suggests hunting ratios however according to other literature the hunting ratios are only strongly recommended with thru-hardened material that will lose material and develop significant transmission error. With carburized gearing (particularly lapped gearing) it's not a big deal. (That said, make sure the match marks on the pinion set align)

David

 

[tbuelna]

tensor47,

To be specific, it's not the overhung "loads" that are the problem, it's the displacements they create at the mesh. Thus, having a very stiff mounting arrangement is most critical. If you consider that the hydrodynamic oil film thickness at the mesh contact is measured in micro inches, it becomes apparent that even very tiny amounts of deflection in the bearing housing structure can cause problems.

To be honest, I don't think you have a problem with meshing frequencies and structural modes. These are not usually energetic enough to cause this type of damage. I would recommend taking a close look at your housing stiffness. It also appears that you have a fairly wide face width for your gear diameters. It might be worthwhile to look at applying some face geometry modification.

Good luck.
Terry

 

[tensor47]

Thanks very much for your thoughts,

To David, it's the input gear set which does have a pin and bush coupling, so this is not an issue.
(The following 2 helical gear sets have lower design loads but have never failed) from initial post and there is a pic.

You are right in general though, as an example the Beconsfield gold mine in Tassie had problems with the output shafts of their agitators being very long and the gear cases were not stiff enough to keep the last gear set aligned which lead to tooth failures. Major court case at the time and a coupling was introduced.

It looks like we are converging, which is good and I'll let you all know what happens.

Cheers

John

 

[gearcutter]

tensor47,
What part of Australia are you in?

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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