Angular Contact Bearings - any experts out there?

[slow-poke]

0.0006 to 0.0022” width variance difference between the inner ring and outer ring is really high, that should be much tighter. What is the full part number of the bearings? Are the bearings mounted face to face, or back to back? If these are universal ground bearings, there should be no preload or end play, it seems like there might be two different grinds getting mixed together.

I purchased a package of five (non matched), 7001AC bearings,

The inner races are wider than the outer races by between 0.0006” and 0.0022” :

#1 0.0012”
#2 0.0006”
#3 0.0022”
#4 0.0017”
#5 0.0020”

I have installed them DF, (Duplex Front to Front), with a 0.005" shim between them that only touches the outer races. With that shim installed, the outer races should be in full contact, and there should be a small gap about 0.002" between the inner races with no tension on the retaining nut. As I torque the retaining nut, I would expect that 0.002" gap to approach zero. I don't have a sense of how much torque would be required to close that 0.002" gap and provide a bit of preload?

 

[Ischgl99]

I’m not accustomed to such a wide variation in ring widths in the same batch, the width specs are +0/-0.0047” on each ring, so they appear to be within specs.

The torque necessary will depend on how much preload you want on the bearings for that application. Since you are not using precision bearings, your preload will be a bit less to take runout of the components into account. This document from SKF will help you calculate a starting point for the locknut torque, you’ll need to adjust that based on your installation conditions.

https://cdn.skfmediahub.skf.com/api/public/0901d1968065f1f4/pdf_preview_medium/0901d1968065f1f4_pdf_preview_medium.pdf

 

[slow-poke]

I want to make sure I understand this correctly.....

My understanding is that these bearings will either have slightly wider inner races, or slightly wider outer races, and when properly installed the preload force on the bearing will move from the (unloaded distorted shape) to the preloaded non-distorted shape where the inner and outer races of the matched pair will be in full contact and some predetermined preload will exist. If my understanding is correct, that would mean to get the bearings to the preloaded non distorted shape, the clamping force would have to be applied to:
- the outer races if they were narrower than the inner races,
- the inner races if they were narrower than the outer races.

So in a application where the ballscrew retaining nut is applying the clamping force only to the inner race, the bearings would have to have slightly narrower inner races style. If the inner races were actually wider it would not be possible to preload them with the ballscrew retaining nut, because it would simply clamp the already touching inner races while leaving the outer races loose.

Is my understanding correct, or am I missing something?

 

[dfloen]

This may be what you want:

SKF

www.skf.com

 

[Ischgl99]

You don’t have a matched pair, so there is not a predetermine preload with the bearings you bought. The bearings you purchased have a width tolerance of each ring of +0, -0.12mm, there is not an assembled tolerance, so it is the luck of the draw on whether the inner or outer ring is wider. When assembling these bearings, you need to have some sort of adjustment to take up the end play, or add preload. These typically are not used as adjacent pairs since you need to make adjustments to get them set right. For an adjacent pair, universally ground, or back to back, or face to face with ground in preload is preferable. You can also get ball screw bearings that have higher axial load ratings than angular contact bearings.

You need a clamping force on both the inner and outer rings. The housing would provide that force on the outer rings by means of a shoulder and bearing cap that presses against the outer ring, and a shoulder on the leadscrew and the locknut does that for the inner rings. The only difference is where the gap is between the rings, that determines whether the housing or the locknut is providing the adjustment. Since you have these arranged face to face, a shim on the shaft would be necessary and the preload force gets applied to the outer rings. If you mount them back to back, the shim would go on the outer rings and the locknut would provide the preload force.

 

[slow-poke]

You don’t have a matched pair, so there is not a predetermine preload with the bearings you bought. The bearings you purchased have a width tolerance of each ring of +0, -0.12mm, there is not an assembled tolerance, so it is the luck of the draw on whether the inner or outer ring is wider. When assembling these bearings, you need to have some sort of adjustment to take up the end play, or add preload. These typically are not used as adjacent pairs since you need to make adjustments to get them set right. For an adjacent pair, universally ground, or back to back, or face to face with ground in preload is preferable. You can also get ball screw bearings that have higher axial load ratings than angular contact bearings.

You need a clamping force on both the inner and outer rings. The housing would provide that force on the outer rings by means of a shoulder and bearing cap that presses against the outer ring, and a shoulder on the leadscrew and the locknut does that for the inner rings. The only difference is where the gap is between the rings, that determines whether the housing or the locknut is providing the adjustment. Since you have these arranged face to face, a shim on the shaft would be necessary and the preload force gets applied to the outer rings. If you mount them back to back, the shim would go on the outer rings and the locknut would provide the preload force.

Thanks that makes sense. I found a set of matched NSK 7001C bearings on eBay and ordered them.

With the new set being matched can I simply snug up and lock the ballscrew retaining nut without fear of over preloading them?

One of my buddies suggested, tighten a little and then measure torque required to turn the ballscrew, if less than 5-7 in-lbs, tighten a bit more until 5-7in-lbs is required and then lock it down. Does this seem like a viable approach?

The static load rating of these bearings is 2800N, and according to engineering toolbox, you need only 4Nm (3ft-lbs) to apply 2800N of force so I better go easy.

 

[Ischgl99]

Is that the full bearing part number? If so, they are not matched, and they are not appropriate for a ball screw. The C indicates they have a contact angle of 15 degrees, you want 40 or more. Low contact angles are good for spindles with radial loads, ball screws have very little, if any, radial load, so the higher contact angle is needed for the axial load capacity. Ball screw support bearings have a contact angle of 60 degrees as a comparison. Matched pairs will have a suffix to indicate how they were ground, and need to be installed, if that is lacking, then you need to assume they are not matched. Those bearings might work in your application, but not the best choice.

Download the NSK industrial bearings catalog at this link, page C70 starts the section on angular contact bearings.

Catalogs & CAD Drawings

Explore our complete product lineup of bearings and linear motion products. See our latest publications and technical figures, including CAD data.

www.nsk.com

Keep in mind bearing part numbers are not necessarily the same across manufacturers, so what NSK specifies for a particular bearing can be different for FAG or SKF.

Your friends recommendation is correct, it doesn’t take much torque on the nut to take up the end play and add a bit of preload. Finger torque is usually enough. Static load rating is the maximum safe load before you damage the bearings, you don’t want to be anywhere near that in this application. These are small bearings, so take it slow and only do as much as is needed, and no more.

Instead of randomly buying off of Ebay, you really need to call a local bearing distributor and let them select the bearings you need.

 

[slow-poke]

The full part # is NSK 7001CTYDULP4 (2 matched bearings in the same box)

Correct these are 15 degrees so not ideal but likely a lot better than what's in there at the moment. In the longer term I might make or purchase larger bearing blocks so I can accommodate larger bearings. The static rating of these bearings is 2900N

I don't have a sense for how many N of force these bearings would be subjected to in actual practice, this is a smaller 8x30knee mill (table weighs perhaps 100lbs) with ac servos directly driving the smallish 1605 ballscrews. The servo's can move the table rapidly, however I have limited speed to 100IPM, with acceleration of 4IPS for now, so that only works out to about 5N based on F=ma. That's assuming no friction.

I imagine the cutting forces far exceed what is required to just move the table, I did find a MIT paper that discusses this topic they tested a MUCH larger mill (800kg table) with 3kW servos and maximum force was about 3000N.

 

[Ischgl99]

I haven’t done the torque calculations on a ball screw, so I can’t help with that for now. The static load rating of the bearings is important to make sure you don’t destroy the bearings, but for calculating the life of them, you would use the dynamic rating. That will give you the number of revolutions of anticipated life so you can see if these will last a long time under the loading conditions, or have a short life. It’s important to get a reasonably accurate load on the bearings so that your life calculation is as accurate as possible. Just thinking out loud, if you had a scale to measure how much force it takes to push the table, that would give you the inertial and frictional loads, then add what the maximum cutting force would be and I think you would be reasonably close.

Thank you for the full bearing number, that helps a lot. Those are universal ground, so you can install them back to back or face to face by clamping them together, no shims or spacing rings would be necessary. Face to face gives you better misalignment leeway, and back to back is more rigid. I’m not sure if that will matter here, but for a spindle, back to back is usually preferred. You don’t need to worry about adjusting them, the preload is designed in. The L indicates they are light preload. Tighten up the housing bearing cap on the outer ring and the lock nut on the shaft and you should be good to go. Maybe do that in stages just to make sure nothing is binding.