accelerated life test on the bearing

[bernardg]

Hello all,

We make dc motors... and as you know, bearing is a part of the motor system. We use 608 ball bearings. Mostly, in the field... the bearings would take the radial (side) load... and a little axial load, depending on the application in which they are used. Right now, we are trying to make a fixture to do the accelerated life test on the bearing.

Lemme illustrate our test fixture, as best as i can...
5 shafts (6 inches long) with 2 bearings are interference fitted to the both ends of the shaft... and the bearings are loctited onto the endbells (with the machined bearing bore). The endbells are fixed in place using bolts.
Exactly at the center of the shaft (position could be adjusted to change the location of the point load), the shaft is coupled to independent loading fixtures which applies 100 lbs-force ea. on all the 5 shafts.

All the shafts are connected by 1 single belt drive, driven by a motor.
The belt drive is on one end of the shaft (between the center loading point and the bearing). The tension of the belt is say 10 lbs-force.

The way the fixture is constructed, this is how 1 belt could drive 5 shafts....
(i) the 3rd shaft has the belt below the shaft... such that the tensional load due to the belt drive would act in the vertically upward direction.
(ii) the 2nd & 4th shafts has the belt on top of the shaft... such that the tensional load due to the belt drive would act in the vertically downward direction.
(iii) and the shafts 1 & 5 has the belt such that it is gonna put the tensional load at an angle.

Based on the scenario that I gave you, could anyone guide me in calculating the forces that each bearing is gonna encounter.

Our inital thought was to test 10 bearings with the same radial (side) load. But over the time, as we started to build this fixture... it seems that itz difficult to put the same load on all the 10 bearings. Is there something I am missing here, i mean.. is it achievable to have equal load on all 10 brgs with some slight modification to my fixture? If not, what should I do different to make it happen?

Thanks for your valuable inputs.

Bernie

 

[Metalguy]

Have you considered using direct-drive motors-one per shaft? You can use inexpensive couplers that have rubber "spiders"-these do not put any end loads on the shaft. You MAKE motors, so you can test them at the same time! Have you considered using max.-fill bearings (filling-notch) instead of Conrads? Your loads sound ideal for them.

 

[MikeHalloran]

I don't understand why you're doing accelerated life tests on the bearings, when ball bearings have already been analyzed and tested to death. The bearing manufacturers can guide you to the right equations to let you predict the life of a bearing under a given loading condition, to any confidence level that you desire.

If you _must_ test, ten bearings is not a large enough population to give statistically significant results anyway.

-Mike-

 

[MikeHalloran]

It also occurs that a radial load in the middle is not a good analog of the loading that the bearings really see in service. The bearing nearest the output shaft suffers a radial load nearly as large as the radial force generated by whatever gear pass or belt that the motor drives. The bearing at the blind end experiences a much smaller radial load.

Which led me to another thought. Using Loctite on the bearing o.d.s is a good idea for dealing with bore misalignment in the end bells, but it makes the bearings susceptible to uncontrolled thrust loads associated with differential thermal expansion of the shaft and the motor stack.

For an example of a better bearing system, look at a typical alternator; one large radial bearing adjacent the pulley takes most of the radial load, and all the thrust load. A needle bearing in the blind shell supports the lightly loaded end, and suffers no thrust loading.

I suggest you do a careful design review of the motor bearings and their supports, considered as a system.

-Mike-

 

[bernardg]

To answer your questions...

As you know, in a dc brush motor, the primary failure mode is usually the brush wearing out... and NOT bearing reaching the end of its life.

We are currently buying the bearings from NTN. Because of the cost implications, we are trying to find a cheaper source...(a source, where the testing facilities are not comparable with NTN... but still would give us ample bearing life).

We intend to test 10 bearings of NTN and use that as the baseline data for another set of 10 local bearings. The testing is done at room temperature. There is no vibration, so no thrust load.... so we are not using the loctite for bearing retention.

I wanna underline this: "We are not trying to duplicate the load conditions in the real time operating environment in the field". We are trying to give a maximum side load... so that we can accelerate the testing, to get the results asap.

This accelerated brg test fixture doesn't accomodate the complete motor. All it has is 5 shafts with two ball brgs on both sides which are housed into a casting with a bearing bore.

Thanks in advance for all your inputs!

Bernie G.