Custom Bearing - Calculate # of Ball Bearings 4

[BrittToolEngineer]

Greetings:

We are designing an inspection fixture where a round part will sit on a plate and rotate about its center. The size is fairly large, so we are looking at designing custom thrust type bearings. General information: Ball bearing diameter is .250". We are cutting (2) grooves where these balls will ride, one radius is 34.500 and one is 36.500. I know how to determine the circumference and then divide by .250, and this gives me a perfect fit number of ball bearings; however, is there a rule of thumb to reduce that number so that the ball bearings have sufficient space to perform their job?

Thanks in advance for your assistance,

Brent

 

[kingnero]

If there is something that is well-established and that has a vast variety of sizes, it must be bearings...

Please define exactly why you can't choose one out of a catalog?
Those will have cages if necessary, and are engineered according to years of experience and real-life testing, I doubt you'll improve on that process for a one-off.

http://www.fusionpoint.be

http://be.linkedin.com/in/fusionpoint

 

[BrittToolEngineer]

I'm looking for one out of a catalog, maybe looking in the wrong spot. Basic needs for one that is caged, the top and bottom need to be flat and parallel, and we need approximate sizes below, small cross section:

- 19.7 OD w/ a 17.7 ID
- 15.0 OD w/ a 13.0 ID

Can you point me in a direction where simple bearings like this can be found?

Thanks again for your time.

 

[btrueblood]

Dimensions given are mm or inches? Did you mean diameters of 34.5 and 36.5 in first post, and radii in the 2nd post?

 

[BrittToolEngineer]

btrueblood. I apologize, I put dimensions in for a 2nd application. Lets go with the 19.7 OD and the 15.0 OD numbers (inches)

 

[drawoh]

BrittToolEngineer,

I do not have access to my machine design textbook at the moment.

As I recall, the round bearing ball contacts the curved race surface with a contact area of zero, resulting in infinite stress. There are formulas for this. What happens in reality is that the ball and race flatten a bit, creating a finite contact area, and a stress a bit below infinite. This works fairly well because the balls and races are very accurate, and they are hardened.

I assume you are buying the bearing balls. How accurately can you machine the races?

If this were my problem, I would use as many balls as I could fit. The rule of thumb is that journal bearings wear out, and ball bearings fatigue. More balls, less stress and fatigue. I would want to know how I was going to install them, and retain them into position.

--
JHG

 

[imcjoek]

I suspect your grooves would need to be ground & hardened for this to work. So you're looking at substantially large chunk of 52100 or similar.

Can you use crowned profile cam followers or needle bearings arranged around the periphery?
How about curved linear rail?

 

[BrittToolEngineer]

We are planning to machine the grooves, and we can hold tolerances tighter than +/- .001 consistently. I do agree with the plate, or at least the groove, need to be hardened to prevent or slow wear rate and we will take that into consideration for material. The quantity of bearing balls I think I understand now. I think if I calculate the total and take one out, this would work fine for the application. The weight on the bearing is minimal. The importance of the table is to turn true with respect to the center point and remain flat and parallel while the diametrical profile is checked. In summary, it spins slowly and weight is not an issue. I plan to design the plate that rotates to hold the ball bearings in with a groove as well. and that plate will mount to the base plate using a radial bearing, so the ball bearings will be contained.

These are my thoughts. I appreciate everyone's feedback. I am continuing to research catalog bearings that are large enough for this application diametrically, but are also small cross section.

 

[hendersdc]

You should be able to find a slewing ring bearing off the shelf in something close to your sizes, just google that term and you will get a lot of results. They will not be cheap but will be much easier than designing your own.

 

[BrittToolEngineer]

Thanks hendersdc. I have looked at those. The problem I run into with those is space. I really need to use two thrust bearings for this application, and I'm having a challenge finding two of those that will fit within the real estate that we have.

 

[hendersdc]

What is the max thickness of a bearing can you fit?

 

[BrittToolEngineer]

Thickness wise, I could go up to 1.00, prefer .375 or .500. However, OD and ID, I need to be around 19.7" (500 mm) OD on one a 15.5" (392 mm) OD on a 2nd one that will support the top plate on the same plane. From what I've seen in the slewing rings, I cannot find a combination that does not interfere with one another on the diameters (ID on the larger bearing versus the OD on the smaller bearing.

 

[hendersdc]

That is pretty tight for two concentric bearings, closest I could find in a brief search are these from VXB though I wouldn't expect much accuracy from them and the runout may ruin any inspection measurements you take from an instrument mounted on them. They are very cheap though so might be worth trying out.

[URL unfurl="true"]https://www.vxb.com/500mm-Lazy-Susan-Aluminum-550-lbs-Turntable-p/kit11283.htm[/url]
[URL unfurl="true"]https://www.vxb.com/392mm-Lazy-Susan-Aluminum-420-lbs-Turntable-p/kit11281.htm[/url]

 

[Lnewqban]

....is there a rule of thumb to reduce that number so that the ball bearings have sufficient space to perform their job?

Avoiding the friction between two consecutive balls, especially for high rpm's, is the main reason of the separator in normal ball bearings.
The number of balls in your case should leave some room for the lubricated balls to naturally find their place in the tracks, IMHO.

I would recommend you to observe the following:
- Provide your bearing with a flat surface as a solid base, which eliminate any tendency to the twisting of the racks and the need for extra energy to induce rotation. Due to the dimensions, your tracks may not be very rigid by themselves.
- Keep the lubricated tracks and balls protected from contamination with sand, wind driven dirt or water.
- Rather than trying to copy the radius of the balls in the cross sections of your tracks, machine V-shape tracks. This will make both tracks self-centering under weight, eliminating any undesired lateral movement. Bonus: each ball will have two points of contact with each track and small debris will go to the bottom of the track, not interfering with smooth rolling for moderate contamination.

Best luck with your bearing!

"Engineering is achieving function while avoiding failure." - Henry Petroski

 

[israelkk]

Look for full complement bearings in bearing catalogs (bearings without cages only balls/needles). Some catalogs give technical information how to design your own bearings including the number of balls with reference to the grove diameter.

 

[Tmoose]

"it spins slowly and weight is not an issue."
Machining your own races within 0.001" met your goals.

Sounds like a much smaller bearing would suffice.

http://www.l-07d.com/images/plinth1.JPG

http://www.vxb.com/Hardware-Turntab...&Click=54773?gclid=CMDu94T43dICFdRMDQod7YwOUw

fancy -

http://www.kaydonbearings.com/RK_turntable_bearings.htm#NG

 

[Jboggs]

I will join the chorus strongly urging you to go with a commercially available unit! One BIG advantage of that is that it is designed by people that are in that business and no what land mines lie ahead. they aren't doing this for the first time. I would also urge you to seek your help from a good reputable local power transmission distributor, like Motion Industries or Allied Bearing. They have people who know a hundred times more sources than you do and are paid to help customers achieve success. Put their knowledge and experience on your side. You want an Application Engineer, not an order taker. You will be glad you did.

You might also have more luck if you do your search for "rotary stages" rather than just thrust bearings. Lots of companies make rotary stages.

Also, have you considered using cam followers to support your upper plate? Does it absolutely have to be supported 100% all the way around?

 

[mfgenggear]

sounds like jig grinding required for the V grooves.

 

[Nescius]

I am not a bearing engineer; I have no idea how to determine the tipping point where a cage helps in some way. I believe many bearings have cages only to allow assembly, and because it's cheaper, not because it's beneficial.

Filling a race with balls, no cage, is certainly not necessarily a problem. It's done all the time in bicycle crank and wheel bearings. In bikes, the rule of thumb is to fill the race, then remove one. It's easy to put in too many if you try to maximize the number in there.

Ok, now for the crazy idea...

Ground flat upper and lower plates. Upper plate rotates and is located by a smallish vertical shaft in bushing, ground hole, or similar. Fill the space between the ground plates with a sea of balls. Some sort of superficial fence out near the edge of the plates would keep the balls contained.

Small ball bearings are very inexpensive and quite uniform. The flat ground plates would also be inexpensive and very uniform, especially compared to big precision circular grooves.

I have seen a crude application of this on a drive-on car hoist. In that case, the purpose was to allow easy pivoting of the front wheels. The plates were on a bed of balls, plastic balls, if I remember correctly.

 

[tbuelna]

We are designing an inspection fixture....

Based on this requirement, you probably want to use a catalog bearing. Grinding large diameter ball bearing race surfaces having a level of precision acceptable for an inspection fixture is not a simple thing. Here are a couple catalogs for large diameter, thin section ball bearings. It looks like there are standard sizes that should meet your requirements.

http://www.kaydonbearings.com/downloads/catalog390/Kaydon_Catalog_390.pdf

http://www.rbcbearings.com/literature/pdfs/RBC-ts.pdf

http://www.jp.nsk.com/app01/en/ctrg/index.cgi?rm=pdfView&pno=e1153

To improve the axial/radial runout accuracy and rigidity of your inspection fixture bearing system I would recommend using some preload. If there is axial space available, you can use a duplex pair of back-to-back angular contact ball bearings (Kaydon Reali-Slim type A) manufactured with a controlled gap between the inner race faces. Clamping the inner races together at assembly will produce the required amount of preload. If axial space is limited in your fixture base, you can also use a single 4-point ball bearing assembled with preload (Kaydon Reali-Slim type X). You can order catalog bearings from most manufacturers with whatever amount of preload you require. Either of the bearing arrangements described will handle combined radial/axial/moment loads.

While these are standard catalog bearings, just remember that they are also fairly expensive. A 16" bore, .50" section, 4-point ball bearing will probably cost well over $1000 new. If you only require a single bearing, you might search e-bay for one that is new-in-box.