Planetary gear noise problem 6

[spigor]

I have a planetary gear 32DP 20PA, ring dia. 100 mm. It runs at 150 RPM. The requirement is it must run very quiet, but a low humming sound is audible. I was thinking about using 64DP and/or 17PA to address that. Does anybody have experience / information on how much would it change the picture, or any other idea on what to do to silence that gear? If we doubled the number of teeth to use 64DP - should we expect the gear be much weaker? Thanks in advance for your input.

 

[dinjin]

What materials? I assume the gears repectively are
62 32 126 teeth combination. I wonder if you use 63 31 126 combo if that would help. I am thinking not to have a hunting tooth arrangement so the that the sun and outer ring is in sync. Also maybe a 62 32 124 teeth with 50 percent long sun, 25 percent long planet, and a 25 percent short addendum on the ring gear. What are you allowed to change? Do you have to hold the individual ring blanks?

 

[Tmoose]

I'd use PC soundcard and freeware FFT to determine the frequency of the hum

 

[spigor]

The material is nitriding steel, teeth no.: sun - 62, planet 1 - 20, planet 2 - 39, ring - 121. We can change pretty much everything, thought about having the numbers of teeth doubled and switching to 64DP, and using smaller pressure angle, like 17 deg., but looking for somebody with experience who could advise, such changes are costly though.
Tmoose, you mean determine the frequency and then find the member that has number of teeth matching that frequency? That we can do.

 

[MikeyP]

You will probably find that is happens at the "tooth passing frequency" which is the same for all members in the gear set.

Can you do a contact pattern test under load to see if the gear contact is doing what you expect. This is usually the first port of call in diagnosing these problems. Also you need to know if the parts are made to print - get them measured.

As with any problem like this if you have one that is noisy and one that is quiet then the answer is "simple" - Find the difference between them.

M

--
Dr Michael F Platten

 

[Tmoose]

The frequency of the hum is probably related to some gear mesh frequency, but it might be something else. That's why I start with the frequency.

If some panel of the gearbox happens to have a resonance close to gearmesh frequency even "good" gears will make it sing. It would be a shame to make SUPER gears just to quiet it down.
It may be more appropriate to selectively stiffen the yodeling panel, or if truly resonant, add constrained damping layer

 

[tbuelna]

spigor,

Gear noise is mostly due to the release of strain energy (as the gear teeth unload leaving the mesh) causing the gear bodies to vibrate. Anything you can do to promote uniform dynamic load sharing among all of the gear teeth meshing at any given instant will reduce perceived noise. Things such as tooth index errors or overconstrained conditions with multiple mesh points (such as your epicyclic has) can cause the loaded teeth to "fight" each other if not perfectly located. Ideally, the tooth load should transfer smoothly from the tooth exiting the mesh to the one trailing it. If there are geometry or index errors, the strain energy in the tooth exiting the mesh is released abruptly, causing the gear structure to vibrate. The reverse is also true for teeth entering the mesh.

The most logical corrective actions would be to increase contact ratios (via higher tooth counts or lower pressure angles), improve the precision of the gear profiles/indexing/mounting, or redesign the system's degrees-of-freedom or constraints such that it is not overconstrained (float the sun gear or carrier, reduce the number of planets to 3, attach the ring gear using a flexure).

You can also look at using a "non-factorizing" epicyclic, which uses tooth count combinations and planet gear positions that vary the mesh actions at each planet at any given instant.

Hope that helps.

 

[MikeyP]

I would disagree with tbuelna's description of the gear whine process. There is no pinging of teeth as they enter/leave mesh (it simply that the stiffness of the mesh varies depending where you are in the engagement cycle), but yes, "promote uniform dynamic load sharing among all of the gear teeth" is exactly what should be done.

In addition to tbuelna's suggestions re contact ratio and tooth counts, letting one or more components float is OK as long as you are not in a situation where you get double flank contact. This will be very noisy. Radial tolerance stackup is very important

Also planetary gear sets have a tendency to suffer from "modulation" of the whine noise. Without going into too much detail, this is largely caused by run-out, pitch errors and piloting. It can make even a gearset with very smooth mesh stiffness sound awful because the warbling noise you get is perceived as much more annoying than a constant hum.

M

--
Dr Michael F Platten

 

[tbuelna]

Dr. MikeyP,

I don't have the benefit of a university education, let alone post graduate level work such as yours. Since you don't like my layman's explanation of what causes gear mesh noise, would you humor me with a brief explanation (2 or 3 paragraphs) of what does?

I am always interested in learning more about gears, since I am self-taught on the subject.

Thanks.
Terry

 

[MikeyP]

OK. Flattery will get you anywhere (although my education had nothing to do with gears).

Let's assume we have a pure involute spur gear with a transverse contact ratio bewteen 1 and 2 (i.e. depending where we are in the mesh cycle there are either 1 or 2 pairs of teeth contacting)
Assume there is a constant torque going through the gear pair.
This means there is a constant force across the gear mesh.

When we are at a point in the mesh cycle where there is only 1 tooth pair in contact then all the force goes through that tooth pair. The tooth bends a bit as it is not perfectly rigid.

When we are at a point in the mesh cycle where there are 2 pairs of teeth in contact then the force is going through 2 pairs of teeth. If you assume that this force is shared equally between the two tooth pairs then each tooth takes HALF the force that it did when there was only 1 tooth pair in conctact. Half the force equals half the tooth bending, so the gear mesh displaces half as much as it did when there was only 1 pair of teeth in conctact.

So the displacement and stiffness of the mesh varies depending on where you are in the mesh cycle. Varying displacement = vibration = noise.

The varying displacement is usually called transmission error (TE)

If you have higher contact ratio of say between 2 and 3 (e.g. by using a helical gear instead of a spur gear), then the stiffness variation is smaller an so the displacement variation (and hence noise) is smaller.

Of course anything that makes the gear less than a perfect involue (misalignement, manufacturing errors, lead/profile mods) will also induce a vibration whether or not there is any load through the gear.

OK not 2 or 3 paragraphs, but I hope it helps.

Ta

M



--
Dr Michael F Platten

 

[gearcutter]

Dr. MikeyP,

Thank you for the excellent and easy to understand explanation of what TE is.
May I ask you a question which follows on from your description; the deflecting teeth alter the base pitch at the point of contact. The change in base pitch causes tooth tip/root interference as a result of the non-involute contact.
My question is: does the tip interference generate vibration and therefore noise?


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[MikeyP]

If the gear is designed properly, then it will have tip relief to prevent tip contact. If tip contact does occur then it will probably be noisy.
Often the tip will wear smooth after a while, but it still a bad design.

--
Dr Michael F Platten

 

[tbuelna]

MikeyP,

Thanks for the explanation!

I am familiar with TE and its many potential causes. Most of my (limited) experience is with relatively high PLV gears, where perceived noise is not so much of an issue, but dynamic tooth loads are. Using helical/double helical gears with their better face contact ratios and using careful tip profile modifications are very effective at reducing dynamic tooth contact loads.

And of course, with regards to sources of vibration, one should also check meshing frequencies against the structural modes of the associated gears and shafts.

Regards,
Terry

 

[unclesyd]

No one has asked if you are seeing any heat buildup above 140F or rapid heat buildup from a cold start?
Do you have the noise with a cold gear box?

I have solved a lot of gearbox noise and/or heat buildup using different Lubrication Engineers products. would contact them to make a recommendation.see their products are available in your area

I have used LE products on very high horsepower speed increasers to prevent heat buildup and the dreaded hum. We use LE in the fractional horsepower planetary gear seats driving metering pumps.

http://www.le-inc.com/productPages.jsp?productID=200

 

[tbuelna]

MikeyP,

I did a bit of reading up on TE and noise in epicyclics. Here's what AGMA had to say regarding "engagement impulse":

"Whenever a tooth engages, it picks up part of the
load carried by the previously engaged teeth. Reduction
in loading on the original driving teeth causes
them to deflect toward their unloaded position, thus
imparting a tangential acceleration to the gear body.
Also, because the previously engaged teeth are deflected
slightly due to loading, the newly engaging
tooth cannot make the smooth contact that it should. Instead, it impacts against its mate on the meshing
gear, sending an impulse through the bodies of both
gears"

Regards,
Terry

 

[MikeyP]

Assuming that the gear has proper tip relief design then the engagement will be smooth. Of course, there will still be a time when the tooth isn't in contact and then another where it suddenly is in contact. The closing velocities are quite small - Smith (Chapter 3, "Gear noise and vibration" - J Derek Smith) gives a typical example where the impact contact force is two orders of magnitue less than the TE forces.

The other thing that suggests these impulses are not an issue it that you never see them in test data. Vibration measurements typically show meshing order and its multiples plus maybe some sidebands. An impulse type excitation would show up as a response across a wide range of orders.

Ta

M

--
Dr Michael F Platten

 

[gearcutter]

AGMA to hold a Gear Noise webinar on 18 May.

http://learning.agma.org/store/provider/provider09.php#blank

Gear Microgeometry Based Load Distribution Analysis for Gear Noise, Dynamics and Design Assessments
This webinar will demonstrate techniques for using the gear microgeometry to minimize noise excitations and also to evaluate the gear contact stresses, bending stresses and tribological factors such as film thickness and flash temperature. The gear pair used in this study is the high speed ratio of the NREL-GRC wind turbine gearbox. In addition to the load distribution analysis, we shall explore ways of using load distribution analysis to come up with new macrogeometries that perform as well or better than the original design. Finally, a simple dynamic analysis of the gear set will be performed to observe the response of various natural frequencies to a transmission error excitation.

Following the presentation attendees will be able to:
• Have an appreciation for basic gear noise excitations
• Understand gear microgeometry on design parameters such as transmission error and contact stresses
• Have an appreciation of a method to evaluate many different gear designs using load distribution results.
• Understand how transmission error excites the simple dynamic system of a pair of gears and accompanying drive and load inertias.
• Appreciate the effects of microgeometry manufacturing variation on the performance of a given gear design.
Speaker: Donald R. Houser
Professor Emeritus
Ohio State University
Department of Mechanical Engineering
Professor Emeritus Donald Houser is retired from the Faculty of the Department of Mechanical Engineering at Ohio State where he spent 35 years on the faculty. Dr. Houser is the founder of the Gear Dynamics and Gear Noise Research Laboratory (GearLab), an industrial research consortium with 45 participating companies. He has supervised over 120 graduate student theses and has consulted on gearing problems with over 60 companies. Professor Houser has over 150 gearing publications and is author of chapters on gear noise in the Dudley’s Gear Handbook (McGraw Hill, 1991 – out of print) and the Noise and Vibration Handbook (Wiley, 2007). He is a fellow of the American Society of Mechanical Engineers (ASME) and recently received the Darle Dudley award from ASME. He currently is active on the following AGMA technical committees: Helical Gear Rating Committee, Bevel Gearing Committee, and the Sound and Vibration Committee.


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[MikeyP]

Prof. Houser is an excellent communicator and teacher. I can highly recommend his lectures.

M

--
Dr Michael F Platten

 

[unclesyd]

Here is a process we used on some large gear sets used to draw large tow Nylon fibers. i don't recall the exact AGMA classification but it was 12+ and quite pricey.
We also used the Rem Process on some very high speed gears in our draw texturizing machinery. We weren't comparing the finished gears against others as this was a new process.

http://www.remchem.com/page3-34/Material

 

[unclesyd]

Here is a link to several papers on gear whine. The first paper has a lot of good references attached.

http://www.gearsolutions.com/search/?search_txt=whine