Nonidealities in very large spur gear train 4

[craig0a]

Hello,

I am having some trouble with theory not matching practice in a spur gear train that I designed. The train consists of about 200 1" stainless steel spur gears held in place in a horizontal plane between two plates. They are driven from above by a vertically mounted motor connected via a flexible shaft coupler.

Problem 1: In theory, spur gears don't have any thrust. But when running this system, the motor/mounting system is actually being lifted upwards (over 20lbs.) I am not sure where this thrust force is coming from: small angles in my train being accumulated?

Problem 2: Slowing down and speeding up the system to/from rest is not happening uniformly over the gear train. So what happens is that either the motor is being torqued by the moment of inertia of the gears or the motor isn't putting out enough power to turn the gears at the slowest speed and it is torquing itself; in either case the net result is that my flexible shaft coupler is being very twisted, probably almost to the shear point, and my gear train locks itself (that is, before attaching the motor I can turn it by hand, but after the bad slow down, I can turn it by hand in the opposite direction of rotation, but it is locked against rotation in the direction the motor had been driving it). Is there a safe way to slowly introduce and remove rotational speed to the system so that components aren't being caught between resistance and drive?

I apologize if these are commonly known problems which standard solutions; I am still in training as a mechanical engineer.

Thank you for any suggestions and help!

Best regards,
A. Craig

 

[MikeHalloran]

How are the gears 'held in place'?
If there are no radial bearings, you will have problems.

Stainless steel generally galls when rubbed on stainless steel; therefore it's unusual to find stainless gears meshed with other stainless gears. Is that what you did?

Have you found any wear products within the assembly?


A good drawing, or a photo, would be helpful.




Mike Halloran
Pembroke Pines, FL, USA

 

[craig0a]

Hello,

Each gear is fitted with a ball bearing at the top and bottom of each shaft and each bearing is fit within a hole in the plate. That being said, the bearings are only 1/8" thick, so the contact area of support isn't very high.

I understand that stainless-stainless is not the best idea, but the system is under water and corrosion was definitely a concern. The system is being run at a relatively low maximum speed, however, only 300 RPM, so I was hoping I could get away with it. I'm afraid I do not know specifically what to look for as indicators of wear. Visually the gears do not appear to be unevenly worn, but if there is a way to quantify this I can certainly try.


Thank you!

Best regards,
A. Craig

 

[MikeHalloran]

I suspect that you did not have the holes in the plates jig-bored (or bored on a super precise CNC mill), or at the very least match-bored while stacked on an ordinary mill.

You also need an occasional spacer between the plates, with an alignment dowel going all the way through both plates and the spacer, so that each gear's bearings are aligned 'dead nuts' with each other. ... and so that the center distances are as close as possible to ideal dimensions.


Your thread title is a little misleading, as the subject object appears to be, not a train of very large spur gears, but a very large train of small spur gears.



Mike Halloran
Pembroke Pines, FL, USA

 

[3DDave]

If there is a slight lack of parallism in mounting the gear shafts there will be a side load. Given the number of gears, the load you are seeing is pretty small. As the gear teeth come into and out of mesh there is sliding; if the sliding is at an angle you'll get thrust. The limit on the thrust load is the tooth load times the coefficient of friction. Usually water is not used as a lubricant; the friction level may be almost as high as dry steel on steel.

If there is galling, one feature will be that material is pulled along the faces of the gears and form burrs on the ends of the teeth.


There is mention of 200 gears - is that 200 gears in series? Backlash and elastic compliance across large numbers of gears is also a good source of problems. The last gear in the series sees only its own inertia, but the first gear is subject to loads 200 times that.

Look for a video of Arthur Ganson's "Machine with Concrete" It's a worm reduction, but it makes the point about backlash nicely.

To know any better solutions, you'll have to tell what the original goal is.

 

[craig0a]

Hello,

The plates were not jig-bored, but I thought the CNC machining used was precise enough for the purpose. It seems that I might need to remake the plates, but before that I will certainly try adding spacers to see if that helps.

The gears are not fully in series, but almost, so yes, backlash is certainly very much present in the system. Is there a safe way to handle it when accelerating the system or is the answer simply "smaller gear train, more motors"?

Thank you both very much for your help!

Best regards,
A. Craig

 

[gearcutter]

The gear flanks will show evidence of mis-aligned engagement..........show pictures and some here might be able to help you, rather than trying to guess what you have.

What is the gear data that you used?

Pitch
Pressure angle
No. of teeth
Major diameter
Minor diameter
Mounting distance

The printing industry uses very low pressure angles to help avoid similar issues that you are having in regards to 'smoothness' of engagement.

How can the axial thrust being generated cause the motor mounting assembly to move?
Aren't you using some sort of flexible coupling?

So as we can help you; please help us by posting pictures/drawings.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[craig0a]

Hello,

Thank you very much for your help!

I am using a flexible coupling which allows for parallel and angular misalignment but is axially stiff, so it is still allowing a thrust to transmit.

Here is the manufacturing data on the gears:

Pitch 32DP
Pressure angle 20
No. of teeth 32
Major diameter 1.0625"
Minor diameter (unspecified)
Mounting distance 1.000"

Attached is a photo of a gear pulled out of the array very close to the drive gear. Please let me know it is not showing what it is necessary to see.

 

[gearcutter]

You're not making it easy for us.
Unfortunately the image you've posted is of very poor quality.
For one thing; it's out-of focus......probably taken with your mobile phone.

From what I can see; there doesn't appear to be any evidence of severe misalignment.
'Blueing' the assembled gear mesh would be one way of finding out how even the load distribution is.

If you could post higher quality images; we might be able to help you find out what's causing the irregular surface appearance.

If you want to increase the 'smoothness' of engagement;
* Use higher quality gears, e.g. gear grinding.
* Use a lower pressure angle, like 14.5 degrees.
* Increase the number of teeth by using a smaller pitch.
* Add tooth tip modifications, like tip chamfering.
* Use helical gears. Remember that in a straight-line 'driver-driven-driver-driven' arrangement like yours, axial forces are balanced out on all the gears in-board of the end gears. Only the end gears will generate axial thrust.


Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[craig0a]

Hello,

I am terribly sorry the photo wasn't helpful! I took it with my camera, but I was using macro focus so that the between -gear surface details would be clear. I take it a large, overall view with less detail is more helpful in this case? I will take and post better photos shortly, but again, I am so sorry to have taken your time.

Thank you for your advice on gear choice. Unfortunately, I don't think it will be possible to replace the gears at this point. The current solution that I am working on is to split the gear train into multiple smaller parts so that the accumulation of backlash is somewhat less and the inertia seen by the driver gears is somewhat reduced. At present I am looking into connecting the smaller trains to the same motor via a belt drive system (with torque limiters), but if this is a horrible idea, I would greatly appreciate being pointed in a better direction.

Thank you so much for your help!

Best regards,
A. Craig

 

[gearcutter]

I still haven't a clear understanding of what it is that you're trying to do.
Post drawings/images and then maybe someone can help you.
Until then, your leaving us all guessing.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[craig0a]

Hello,

Yes, I know I am being frustratingly vague; this is part of a research project and at present I don't think it is possible to provide a detailed schematic or an exact description of purpose. I am terribly sorry about this; I do not mean to waste people's time by being unhelpful. I am certainly grateful for all the advice and time that people have already invested in helping me!

To describe the setup as far as possible in one place, the key features are:

1. I have multiple hundred spur gears which I need to have in a defined spatial location in a single horizontal plane (so all rotation axes vertical). So far, I have done this by sandwiching a gear/shaft/bearings ensemble between plates which have holes cut in them to hold the bearings on the gear shafts. I can then locate the plates as appropriate. The plates themselves are small enough that I have space to build around them, if I need to.​

2. The gears are not driving anything, so their absolutely rotational position is irrelevant (i.e. backlash in and of itself is not a problem). All that matters for me is a constant and uniform rotation across the gears (after starting up, transients when starting are not a problem) and being able to safely start and stop the motor relatively frequently.​

Previously, I was connecting all the gears into a single chain and driving one of them by connecting to the shaft of a vertically mounted motor. At present, my idea is to fit some of the shafts with both a gear and a pulley, make smaller gear chains within my plates with the pulley above the plates (supported by an additional structure), and use a belt drive to link each small chain to a central motor shaft. My thought is that by using only small chains, the angular misalignment and backlash will not accumulate sufficiently to create substantial thrust and the belt drive will allow slack during acceleration (but torque limiters will be in place to ensure safety.)

If this is still too vague, please don't worry about trying to help! I greatly appreciate your time and help so far, but I do not wish to be frustrating or waste your time!

Best regards,
A. Craig

 

[gearcutter]

No one here is asking you for a "detailed schematic or an exact description of purpose".
Surely a simple sketch or photo of the assembly isn't too much to ask.

I've been able to enhance your image somewhat.

There is evidence of misalignment - shown by the uneven load distribution contact pattern that can be seen on opposing flanks. Also by the fact that contact on opposing flanks is at opposite ends.
As yet, there doesn't appear to be any plastic flow so this suggests that the load concentration is not severe.

The most worrying thing seen in the image is what looks to be contact interference (tooth tip interference) in the root area. As the contact appears to be below the active profile; this will result in what's known as non-involute contact. Motion of non-involute profiles will be irregular.
The interference is occurring due to a reduction in mounting distance (hard meshing), probably as a result of the misalignment.
Strangely; only one of the tooth gaps, in the image, is showing this mark.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[craig0a]

Hello,

Wow, it is impressive that you are able to understand so much about the system from the wear pattern, especially considering a poor photograph! Would additional photos be helpful, and if so, what should I be trying to show?

Thank you again so much for your help!

Best regards,
A. Craig

 

[3DDave]

Pitch 32DP
Pressure angle 20
No. of teeth 32
Major diameter 1.0625"
Minor diameter (unspecified)
Mounting distance 1.000"

Unless the gears are cut with backlash, that's a nominal tight fit. If possible, remake the plates by stacking them and match machining, and change the mounting distance to 1.005. That will increase the amount of backlash, but there needs to be some for the train to move smoothly in spite of irregularities in the mesh.

 

[craig0a]

Hello,

We actually are considering remaking the plates at this point, so I will certainly experiment with the spacing before finalizing the specifications. Thank you very much for your insight! The mounting distance was one of the few things that I hadn't been thinking could be wrong.

Best regards,
A. Craig

 

[tbuelna]

craig0a,

I think what you have is a mechanical loss problem. If I understand correctly, you have 200 identical spur gears working in series, with each gear supported by a pair of sealed, grease lubed radial ball bearings, and no lubrication of the spur gear contacts except for water.

Basically you have 2 gears and 198 idlers. Since the gear contacts have no lubrication other than water, the friction loss at each of the 199 gear meshes is likely quite high. Just from mesh contacts, each gear likely has a loss of over 1%. When you add the seal friction and breakaway torque of 200 ball bearing pairs in series, what you get is a self-locking gear drive. The input shaft will only rotate until the accumulated backlash has been taken out in the gear meshes to the point where total friction torque losses exceed the input torque.

 

[gearcutter]

Well put tbuelna.

Ron Volmershausen
Brunkerville Engineering
Newcastle Australia

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

 

[craig0a]

Hello,

Thank you very much! That was a very clear and helpful description which helps me to understand what is going on in my system. I greatly appreciate you taking the time to explain!

Best,

A. Craig

 

[mfgenggear]

May I also suggest to rework the gears by grinding or gear teeth honing. this will also reduce friction & increase backlash.
the the guys pretty much nailed it in the head. precision grinding the plates flat is is cheap & easy to do. then jig bore or jig grind the holes.
very busy gear box for sure. more gears more friction & power loss. I surprised it did not self destruct.

I suggest doing a center distance calculation by Earl Buckingham & Eliot Buckingham, Manual of Gear Design section two available at Ash Gear.
this will tell you what the major diameters of the gears need to be without causing interference. & what the TIF(True Involute Form) should be.
the gear class quality must be improved. the quality of the lead & involute & the total index error.