Sun gear from a 3 pin planetary set 6

[gearcutter]

This was an interesting investigation.
As a manufacturer; I had a pretty good idea as to what had caused this as soon as I saw the damage. Proving what it was I suspected was occurring was therefore very quick and easy resulting in a happy customer/end user.
The part is a sun gear from a 3 pin planetary set.
Carburize case hardened and gear ground with a crowned profile on a wet gear grinder.
Surface damage is of a consistent nature on all teeth with 2 or 3 of the 18 teeth only showing damage on one end as can be seen on tooth #8 in the image.
Damage is on the loaded flanks only.
These parts normally show very little signs of wear or damage during service life.
Acid etching revealed that a too aggressive grinding operation resulted in the formation of untempered martensite on the ends of the teeth where the grinding wheel approached full depth of the crowned profile.
After further testing; a batch of over 200 parts was subsequently rejected.
The gear grinding machine that had been used to grind these parts is a modern state of the art machine that is designed to lessen the risk of grinding burn by detecting & monitoring changes in grinding wheel spindle speeds & loads.
The manufacturer was relying on this feature in their new machines and consequently had stopped using the acid etching test.



Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[rmw]

Looks like lubrication failure to me. I've seen the same failure on a 3-pin planetary set that went away when the gear lube was changed from alphaltic based LO to parafinnic base LO's.

Nice test of the new site too.

rmw

 

[gearcutter]

rmw;
I’m interested to hear your reasons behind suggesting that this is a lubrication failure.
How else would you explain the acid etch results that were obtained from the samples?
If it was a lubrication failure; I would suggest that the surface damage would be evident over the entire tooth length - not concentrated on the ends as in this example.
There is no evidence of scoring or scuffing or heat discolouration that one would normally associate with lubrication failure.


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[mfgenggear]

Gearcutter

Was there any re hardening. how severe was the untempered areas.
it would be interesting to see the results of a failure analysis.
I am curious to view the hardness values of the untempered areas & how deep it is.
This is very Interesting picture views & discussion.
while this is A break down of standard procedure. by not temper etching.

Mfggeareng

 

[tbuelna]

gearcutter,

I would agree with mfgenggear, surface de-tempering due to abusive grinding is easy enough to check for using nital etch. As I'm sure you're now well aware, it's always a good idea to process and inspect a set-up part with any new machine set-up. Scrapping parts at finish operations, like final grinding of gear teeth, can be especially costly due to the amount already invested in the part up to that point. Grinding burn of gear flanks and bearing journals is especially troublesome because in my experience most customers won't permit the gear to be re-tempered, so it must be scrapped.

Another thing you may wish to watch for with heavily loaded case hardened & ground gears is stock removal during finish grind. If too much case material is removed during finish grind operations, the remaining case may not have sufficient hardness/thickness to accommodate the operating tooth flank contact loads.

Best regards,
Terry

 

[CoryPad]

Great image and discussion.

 

[gearcutter]

mfgenggear - Yes there was re-hardening...........the reason why un-tempered martensite was detected.
Equivalent HRC values measured in these areas was around 64 - 66.
Here in Australia; unless the customer specifies it; I've found that it's generally not a standard procedure in the gear manufacturing industry. I recently conducted a quality audit of a large factory in Italy that produces gearboxes and, even though they had a Barkhausen Noise Analysis unit, no testing after grinding was performed unless it was specifically required.

tbuelna - I'm not really sure what you mean by "de-temper". I know that some people use the term when referring to a loss in hardness. If a loss in hardness is what you mean then, in this case, this is not what happened.
An excellent point about stock removal as this is often overlooked when designing gears that are to be finish ground after case hardening. There is usually a fairly large difference between the 'desired' case depth and the 'actual' case depth.




Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[tbuelna]

gearcutter,

Yes, what I meant by "de-temper" is a loss of hardness. Abusive grinding (grinding burn) can result in localized heating and quenching at the grinding wheel contact.

As you note, nital etch will also reveal areas with un-tempered martensite and also areas with decarburization. However, it takes a trained eye to tell the differences based on the color/shade of the etched surface.

Regards,
Terry

 

[gearcutter]

tbuelna,

I’ve always wondered why people use the term de-temper when referring to steel which has undergone a softening process as it seems, to me, to mean the opposite of what it should actually mean.
To ‘temper’ steel effectively means to soften so if you are ‘de-tempering’ wouldn’t that suggest that you are doing the reverse............to in effect re-harden?
I realise that it probably might just boil down to being a question of semantics but it’s always baffled me.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[tbuelna]

gearcutter,

The reason I use the term "de-temper" is because that is the term used by the guys I learned from. However, I can see your point about literal interpretation of the term meaning something else. I checked the definition of the prefix de- and apparently it can mean either "away from" or "reversal".

Abusive grinding creates severe localized heating on the part surface, which is then subsequently quenched by coolant or air. This accidental, uncontrolled heating/quenching alters the intended temper condition of the steel, thus it is "de-tempered".

If you have a suggestion for a more correct term I'd be happy to adopt it. I'm still not too old to change my bad habits.

Terry

 

[gearcutter]

tbuelna,

I've always used the term 'overtempered' when refering to areas which have experienced localized overheating and have lost the initial surface hardness. I've found that the drop in hardness is usually around 2-5 rockwell points.

For areas which have experienced localized overheating but are then rapidly quenched by either a grinding fluid or compressed air, I use the term 're-hardened' to help describe the retained austenite to untempered martensite transformation.

I'm not sure if this is the correct terminology that a Metallurgist would use but they are terms I recorded in notes I took at courses that I've attended that were run by either Ray Drago or Robert Errichello.

Best regards,
Ron.


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[gearcutter]

Here is a short but excellent report on the effects of grinding burn on aircraft engine crankshafts which have been nitrided.

http://products.asminternational.org/fach/data/fullDisplay.do?database=faco&record=1841&trim=false

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[mfgenggear]

Gearcutter

Thank you for the info.
it is very detail report.

Mfgenggear

 

[Frederick]

I know this is slightly off subject, but what no one has mentioned is the huge tensile residual stresses that improper grinding can create. My studies indicate that even if you do not hit the austenizing temperature under the wheel to create the formation of untempered martensite, you can induce stresses high enough to form micro cracks in the surface. As the surface is heated by the wheel, it expands. At high heats it can undergo plastic deformation because the area around the heated zone is not expanding. When the heated and deformed area cools, it is now in tension. Tensile residual stresses are a real problem on things like gears and dies. The lead to early failure.

My point here is that you do not need to get to tempering or austenizing temperatures to ruin a gear or other surface in grinding. The localized heating and subsequent cooling under the wheel is enough to create problems. Also noted during my studies is that it does not take much of a deviation among any of the grinding parameters to create the stress. Grinding wheel composition, abrasive grain type, abrasive grain size, abrasive grain concentration, wheel bonding agent, wheel hardness, wheel speed, work speed, depth of cut, wheel dressing and lubrication all play a significant role in the grinding equation. All those factors (and more, I am a bit time challenged and can't go into comprehensive detail here)have to be controlled to get good grinding results.

One thing that I found where I was working is that there was a tendency to buy grinding wheels that would "last a long time." This lead to hard wheels with tight bonds. That led to the abrasive grains becoming dull but not pulling our of the wheel matrix, which led to high temperatures and enough residual stress to crack the surface. Switching to a softer wheel led to more wheel changes but dies that lasted significantly longer. The wheels and the labor of changing them was cheap in comparison to making new dies.

There is an excellent book out titled Low Stress Grinding by Guy Bellows. It is out of print, but is listed on Amazon. Read this book and it will explain a lot about what happened to the gears in this thread and how to avoid this in the future.

 

[gearcutter]

Frederick,

Thank you for your contribution to this thread..........it is most informative.
That book sounds like an important source of information so I’ll try to get hold of it.
A star for you.

Regards,


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[tbuelna]

Frederick,

Great post. With finish grinding of gear flanks, the grinding method used (form vs. generated) could also be a factor. I would imagine that all other things being equal, a generating grind process would be less likely to produce grinding burn than a form grind process.

I'll also keep an eye out for a second hand copy of that Bellows book. I'm sure I can outbid gearcutter if one becomes available.

I've attached some text from a gear temper etch inspection document.

Regards,
Terry