Spindle Bearing preferences

[AussiePhill]

G'Day Gents,

I am about to fit some new spindle bearing to my turrett mill. I can get a matched pair of precission angular contact ball bearings, or I can get get the plain GP variety as individuals for much less. I am going to fit the matched pair as they are to spec on the required precision, regardless of the fact that there is a spacer between the outer races (bearing sales persons claim that this is not what they are designed for, which is correct as they are designed for back to back setup, but it's the only way to get precission bearings over here).

How do most of you set the preload? The spec' preload for those bearings is 21 kg. They are loaded with a nut. The only ways I can think of is A/, to preload them of the spindle and measure the compression...... Or B/, I could calculate the torque required on the nut for that pitch, to exert the required force.

Your thoughts gentlemen? There must be an easier way......

If nothing comes to mind, then I will sit down and work it out. It's not hard, Just anoying....

Cheers,

Phill.

 

[machtool]

regardless of the fact that there is a spacer between the outer races

Are you saying there’s no inner ring spacer, just one between the outer race’s?

bearing sales persons claim that this is not what they are designed for, which is correct as they are designed for back to back setup

Let me guess, you went to CBC?. If you install them back to back, there still back to back, just with a spacer in between them. Lots of reasons to use spacers. Lubrication reservoirs, limits cross heating, widens the reaction forces on the spindle etc.

My best guess is that the bearings you found are in a light preload – 21kg’s. That’s about all you will find in stock in this country. That should be specified on the bearing box. The part number will end in DUL P4. If you had an inner spacer that was the same thickness as the outer spacer, you just clamp them on the spindle and you will get the preload as specified on the box.

If you wanted to cough up the numbers, I’d have engineering tables that will show you the relative shift of the inner to outer ring to come up with that preload. I’d need the size, and contact angle.

Where did you pull that figure of 21kg’s from? From the new bearings, or an original spec.

Just make sure they aren’t short changing you on the contact angle also. There’s a propensity to only sell 15 degrees in this country.

Regards Phil.

Adding on edit. Checking the preload off the spindle just doesn’t work. I’d take it you intended to place a load of 21kg’s onto the inner ring, and measure the flashness or step between the inner and outer ring heights.

If the bearing outer ring isn’t constrained in it’s housing the outer just expands, giving you a fake reading. Remember the contact angle and how that is applied to the outer ring. An analogy say for a 15 degree bearing would be akin to placing 21 kg’s on a chisel with a 15 degree bevel. The forces to expand the outer are considerable.

 

[RC99]

Hi Phill, Are they 7207 bearings?

My el-cheapo bridgeport clone turret mill has SKF 7207BEP 102Y bearings in it....I am suspicious they are not precision ones and just generic angular contact ones...


I think at some stage the designers of these clone heads cheapened the design a bit.. If you go to page 10 on this .pdf http://igor.chudov.com/manuals/Bridgeport/Series-I-Parts-Lists.pdf this is what a genuine bridgeport looks like...

If your is like mine you will find there is only one outside spacer between the main angular contacts with an adjusting nut on top of that....Above that there is a plain deep groove ball bearing, which I can not remember the number of it but it is a common 6XXX type..

If we compare that to a genuine bridgeport as shown in the .pdf we see is has an inner and outer spacer between the main spindle angular contacts and then it has a third spacer which goes up to the top bearing (which I guess is floating) and the preload adjusting nut for the spindle bearings is above it...

My mill is giving me the shits with chronic chatter when boring chronic chatter even on small 3/8 endmills... I swear I can see the spindle moving under load...

 

[Forrest Addy]

There's several factors in rolling element precision spindle bearings often overlooked even by the so-called experts selling them. The preload setting is made by grinding the race faces individually to a given total offset. When flushed by mounting the offet produces an axial preload that greatly reduces radial and axial deflections from loading. The manufacturer has tables giving the total offset for bearing size and preload grade. The preload is in a range and the amount of inner race interferance on the spindle has an effect. Naturally when the nner race is dialated the offset (and thus the preload) increases as a trigonometric function of the cone angle.

Precision class angular contact bearings has burnsh marks for points of max eccentricity and lengthy codes marked on their races. Among the confusion of code characters is "DU" "DB" "DT" or "DF". Most all over-the-counter precision angular contact bearings come as a packaged pair in "DU" config. "DU" means duplex universal. Thus the ground-in offset is not only set for the particular preload but the offset is the same for both bearings. If the bearings have to be separated by spacers then identical length spacers are all that is required of the assembler.

Another is precision grade. The tracking error of the spindle axis is a product of the sum of all the bearing errors - which can be many: the roundness and concentriciy of the raceways to the mounting features, the individual ball errors and the way the balls are distributed in the races, the errors of the mounting fits are but the most influential. There are more. Bearings are tested for tracking error on equipment resolving into sub-microinches and tens of nanometers and traced on a circular graph where the error is represented in radial exaggerations of 10,000, 100,000 etc. The trace looks often looks like a potato to me.

Many think this quest for ultimate accuracy is far too refined for their simple needs but not so. The total error of a spindle is the sum of the many possible errors present in the bearing system. It's like a household budget: so much for rent, so much for car payment, bar stock, groceres, carbide inserts, utilties, cutting oil, heat, dentist, etc. A machine tool spindle has to have a total roundness error no more than 1/10 the finest tolerance the spindle is expected to hold. If the 0.0001" total error is to be held and a half dozen manufacturing errors contribute to this 0.0001" then it follows that no one error can be greater (over-simplfying here) than 16 millionths of an inch (metric cognoscenti get out your abaci).

The object of this excruciating quest is the roundness of the finished product. The most accurate features commonly seen in the general run of machined parts is bearing fits. They have to be to size within 0.0005" of size often less and roundness error introduced by the spindle forms part of this figure.

This is background for the assertion widely held in the machine tool community: spindles have to be supported by precision grade bearings. Using general purpose ABEC Grade 1 results in roundness errors of 0.0003" or more times the square root of the spindle bearing ID. Milling and lathe spindles are generally fitted with ABEC Grade 7 angular contact ball bearings with a 15 degree cone angle. Those equipped with taper roller bearings are fitted with comparable grade bearings. I'd list the Timkin grade but my book on preciion class taper roller bearing is misplaced so I can't set down the spec in confidence.

Another point: There has been considerable discusion in some quarters about light Vs heavy preload with some factions supporting "more is better". I submit that light preload is to be preferred for most machine tool applications. Heavier preload neans lower bearing deflections under a given load but also more heat generated in proportion to spindle RPM. Since finish cuts are generally light the forces are low and so are the bearing deflections. If the spindle has warmed up that portion of the spindle ahead of the spindle bearing and the spindle housing expands taking with it the trace of the cutting edge. The trace of the cutting edge goes deeper into the work in proportion to the net heat gain and the distance of the cutting edge from the spndle bearings. Since there is less heat build-up in a light preload than a heavy preload more accurate work is likely in the absense of dynamic thermal comp, recirculatng cooed spindle oil, or deliberate operator calculation in the case of a manual machine.

This is background for my assertion. Unless your application requres greater than usual stiffness AND the spindle speeds are lower, your spindle will serve you better if the preload is set on the light of the range thus allowing it ro run cooler.

I'm going to add mote to this later.

Folks in Oz have a specal problem in that they are far from the sources fo precision bearings suited for machine tool spindles. They represent a market too small for the world bearng interests to serve properly so those in need in NSW or WA do have to scratch to get their spindle bearing needs fulfilled.

 

[HuFlungDung]

AussiePhill,
I would not preload the bearings with a nut across an air gap. That method gives an indefinite preload which varies on differential heating. Also, a nut cannot be relied upon to maintain critical parallelism between the two races, which can cause extra bearing heating and not giving you the real stiffness that the bearings are capable of giving to the spindle.

So it is a given that you'll need an inner and an outer spacer. These should most likely begin as equal thicknesses and be very carefully machined so that they are parallel sided. This might involve something like making a special expanding mandrel for each spacer so that you can support it from the inside, and take a light facing cut on each end face without disturbing the setup.

Then you could use a press or make a press to preload the bearing/spacer/bearing stack. The press could be as simple as a nicely faced off base plate with a threaded hole (carefully threaded perpendicular to the base), a suitable length of bolt and a heavy washer. Use the bolt and washer to pull the bearing stack together.

You can apply light pressure to the stack and test the fit of the outer spacer, since it is the only one which is readily accessible. By 'test the fit', I mean you can attempt to slide it around (just a little) with a soft punch and a small hammer. It should still move when you've got the right preload, but as if restrained by thick molasses. The bearing literally cannot live at high speed if the preload is so tight that the outer spacer cannot be moved.

Normally, one would lap the inner or outer spacer to modify the fit and achieve the right feel to the spacer. I'd use a granite surface plate with self stick adhesive disk for the lap. Body shop suppliers typically have these types of abrasive in 6" diameter. Select about a 240 or 300 grit. Hand lapping is tricky because most people do not appreciate the significance of the direction and pressure angle of their hands on the part. Even though the part is laying flat on the lap, your fingertips always produce localized pressure. So you need to take a couple of strokes across the lap, rotate the ring to a new position, take a couple of strokes, making every attempt to equalize the rotation and the number of strokes you use on the lap. Mark the ring so you know when you have made a full rotation. Then stop and measure the results. Try to use the entire surface of the lap so as to keep its cutting action degrading uniformly over its entire surface.

Hopefully, you should not be removing enormous amounts by lapping: removing a full .001" is way too much work, you should be machining it closer than that You can check for gross errors that you introduce by lapping by using a micrometer as a gauge. Local pressure on any thick spots in the spacer can be worked while lapping, to bring them down. Measuring the thickness of the spacer requires a repeatable technique in using your micrometer: a slow spindle advance, meticulous cleaning of the part and the micrometer anvils for every measurement. I like to bring the micrometer up to a tension where I can still wiggle the spacer out from between the anvils if I pull on it.

Absolutely wash the ring off and clean your hands meticulously before handling your precision bearings. You might even wear rubber gloves while doing the dirty part and new clean gloves for handling the bearing.

Yeah, I know this is just a seat of the pants method that the average guy might actually have stuff on hand, to do a slightly less than expert job. If you are patient, and can measure the thickness of the spacers accurately (to determine true parallelism without fudging your readings ), you should get a far better spindle than any air gap preload will ever provide.

 

[AussiePhill]

G'Day Gents,

I'll answer you each in turn. It'll seem more logical to my tired brain that way....

Machtool

You are correct, there is no spacer for the inner bearings...... I can machine one up, so that's no problem. I just didn't think of it. Well that's not true I did think of it, but for some reason the idea formed and then went away without any real connection. <GRIN> We all know what that's like.
No I didn't go to CBC, for 2 reasons; 1/ they are rip-off merchants, 2/ trade discount given is very small.
The bearings that match the original sizes are 7207C. The spec on the box in stock is 7207CTRDULP4. Of the top of my head, they are 15° PA bearings. My NSK book spec's P5 min for mill/machining centre spindle bearings with a light preload (21 Kg for this size bearing).
My mill is an Elliott Millmore (toolroom model). Everything else is almost perfect. Some small marks in the table from the previous owner are the only defects.
You are correct in that I cannot take a pre assembled load measurement, as it is out due the the strech on the inner races. And thinking now, I cannot use a torque on the nut to set them, as it is not posible to accurately account for the force required to overcome the fit of the inner race on the spindle...... Hmmmm. that removes that easy solution.


.RC.

As noted above, I have only one spacer. My top bearing is a twin row ball race that is a press fit on the shaft. Well.... It is to get it off, but my bearing heater will relieve that problem for putting it back on.
The original angular contact ball bearings were made in itally, and were marked 4APM with 3RIV on the outer race and 1RIV on the inner race. The cage is machined from that fennoll resin stuff. Sorry Gents, but my brain is failing fast now...... These are high precission bearings. If they have a cage that is anything else, it is NOT a precision bearing.

My spindle didn't chatter in any opperation. The noise from the bearing when taking a reasonable cut, is what alerted me to the problem. The heavier the cut, the worse the noise.

For those who may be interested, what caused the bearing failure was the previous owner using a compressed air line to clean the machine. I found small pieces of different metals in the bearings...... As there is a steel facial laberenth ring to prevent dust and dirt from entering the bearings, the only way in is to be blasted in. Oh well, I suppose "some mothers do 'ave 'em....." Anyway, the noise (not very loud) has slowly gotten worse over the last year. I assume that being flushed with oil (from my liberal oiling) the dents have started to make themselves heard.


Forrest Addy

I was after P7 bearings, and they claimed that they had some, but when the box was inspected, they turned out to be P4's...... The number quoted as the P7 was - 7207CTDUEP7.
You are absolutely correct about the precision of the bearings that are available down here. I personally blame the accountants for the problem, as we never used to have the problem.....



HuFlungDung

You're correct. over the air gap, it is imposible to ensure that the inner races are concentric. Even with the interferance fit of the race on the spindle, the interferance is not great enough to ensure correct alignment. Therefore, machining up a new inner spacer is the only viable coarse of action to set the preload and the concentricity.


Gent's it's 00:40 here, and I need to go to bed. I have to be up soon to earn my daily bread.


Cheers,

Phill.

 

[donie]

I ran into a custom preload problem on a Moore jig borer.
The 62mm OD bearings were Barden 9s. The spindle nose pair were spec to be 250lbs, twice factory heavy preload.
Barden does not recommend such preload. Barden has no formula for spacer offset to effect custom preload.
A few tries with Moore, I was able to find a way to do it. Each bearing is loaded to 250lbs in a fixture using a hydraulic scale, and the race deflection measured, and then applied to the spacer offset.
In the case of this jigborer, the heavy preload is needed for its intended function.
Its works in this situation due to the high accuracy of the spindle components and the high accuracy of the bearings. However, bearing life is shortened.

Hydraulic scale
Hydraulic scale picture by donsmonarch10ee - Photobucket

 

[RC99]

Genuine bridgeports use 7207 angular contacts from Fafnir

part # 2MM207WI DUL It may be cheaper to buy a set from overseas and get them posted...

Lewis International Bearings Lewis International Bearings Pty Ltd (in Melbourne) show a set of Fafnir 2MM207WICRDUL [PR] as available as well as a heap of NSK 7207A1WIDBCA22 [PR] which may be similar specs....

 

[9100]

In some of the older instructions for setting up automobile rear end pinions they have you adjust the preload for a torque required to turn the pinion without the ring gear or seal in place, apparently based on the rolling friction of the rollers increasing predictably as preload increases. That approach allows you to set the preload and leave it together. I don't know how the original specs were generated or whether a bearing manufacturer would have them.

Bill

 

[machtool]

I was after P7 bearings, and they claimed that they had some, but when the box was inspected, they turned out to be P4's...... The number quoted as the P7 was - 7207CTDUEP7.

Your getting ABEC and ISO numbers mixed up there. EP7 is Abec7, closely equivalent to P4 in ISO. You would be hard pressed to find any thing here in a less than P4 grade, they just don’t bring the lower grades in, once you start asking for precision.

Remember the lower the number in ISO, the better the quality. Abec is higher, 9 being the best.

 

[AussiePhill]

G'Day Gents,

RC

Thanks for the bearing details. I really appreciate your efforts. It shows me that I have already sourced the correct bearings here in Perth. At least I know I got that right. Bridgeports only use a 15° pressure angle, and a P4 precission rating.

MachTool

Yep. I was switching them over for some reason. I'll chalk it up to tiredness. I have it right now.... Maybe....



Thanks Gents for all your help. I will now make a spacer for the inner ring and fit the new bearings. I had a look in the parts list for the machine and it shows an inner spacer...... So, in summary, someone (I can think of some "choice" words here... ) has previously checked/cleaned the bearings, and left it out/lost it....... Oh well, let's see what's in the scrap bin first..... I have to make a mandrel first..... This is where super glue comes in REAL handy. It has a very low shear strength compared to it's tensile strength, and will secure items nicely for light cuts. But a simple impact in shear (sliding it off) and, hey-presto, it's released.



Cheers,

Phill.

 

[RC99]

What sort of $ are the spindle bearings worth??? Just wondering if it would be worth putting a set in my mill..

 

[SAG 180]

What sort of $ are the spindle bearings worth??? Just wondering if it would be worth putting a set in my mill..

It goes up exponentially with the precision of the bearing so it's worth going a few steps up in precision. Also precision balls are a lot cheaper than precision tapered rollers, just look at the price for a set of Gamet precision hollow roller tapered roller bearings.

If your situation is anything like mine, getting someone who can buy at trade in Brisbane, Sydney or Melbourne is your best bet. The four major Australian bearing sellers are owned by the one guy I understand (I may be wrong), giving the illusion of competition.

As an example buying bearings locally at a certain nameless shop that offers bearing related services is 30% cheaper in Brisbane than in Cairns. Keep in mind this is "someone off the street" retail price in Brisbane compared to "You are a valued customer" 60% off retail price in Cairns for quantities of 60 bearings or so.

I'm also seeing a brand named FBJ in a few shops that are Chinese but labelled "Japan" on the boxes from a company that has been warned by an international bearing association to correctly label their country of origin. Even though the bearing shop claims they are every bit as good as their name brand, they had less balls in the bearing than the original name brand.

[FONT=&quot] FBJ Bearings & Singapore Continue to Deceive Country of Origin[/FONT]

"The BICC is carefully monitoring bearing manufacturers and dealers for deceptive or wrongful Acts against the interests of a 'level playing field' or that of the Consumer & Competition act 1974.

More recently, "FBJ Bearing" brand false marketing by their Singapore operations owned and controlled by 'International Bearings Ltd', and its ever increasing dominance in the Far East Asian, Australian & Russian markets.

After constant reminders to FBJ Singapore operators that 'they are acting against the Trade Act', FBJ continue on there merry way to deceive & distort bearing dealers and end users by not labeling "Made in CHINA" on FBJ Bearing or packaging. Just view their internet site.

When questioned, FBJ are very quick to reply the real 'Country of Origin' is JAPAN which in percentage terms gains them approx 200% more in both Income and Profitability. And whilst FBJ total sales are extremely small compared to major players, manufacturers such as SKF do nothing to assist the BICC to eliminate this illegal behavior. "

 

[Lakeside53]

What sort of $ are the spindle bearings worth??? Just wondering if it would be worth putting a set in my mill..

For a BP, the 2MM207WIDUL list price is over $US600 per set. I get them on Ebay brand new for less than $100... You need to watch the Bardon and RHP bearing numbers also (many don't, so you get better deals)... both originally used by BP before they bought Fafnir.

Be careful.. there are a couple of players selling sealed 7207dul magneto bearings - Abec 3 for $299- these are not correct, and they advertise them as abec 7. They also sell a sealed top 6206 - wrong.. oil needs to flow down though these. I don't buy any "BP solutions" - just by OEM bearing number.

 

[AussiePhill]

G'Day Gents,

An update for those who are interested.

The bearings cost me AU$286 for the pair. That's trade, not retail. I didn't ask for the retail price. Swap them over RC, they aren't that expensive. Nice price for P4 bearings? I was expecting to spend more, but I also wanted P2 accuracy, even though P5 is considered the minimum.

I was checking the top bearing anddecided to replace it also. The only problem is the type and size...... It looks like a custom bearing. The OD is 62mm, the ID is 30mm and the width is 20mm. It is a twin row ball bearing. It is almost imposible to tell if it is deep groove or angular contact, due to the cages. The closest bearing I can replace it with is a twin row angular contact at 23.8mm wide. By making another spacer that the bearing buts up against, that is 3.8 mm thinner (which leaves a spacer 3.36mm thick) it will solve the width problem. The funny thing is that the spacer is there purely to locate the bearing due to a very small shoulder height, which is too small for a bearing to shoulder against. Any way...... That bearing cost me AU$54.....
The original bearing has no seals, but the new one does. I have to double check the point where the oil enters the quill, but I now wonder if pull the seals. If the oil entry is above the bearing then I'll pull them, and if below.... Perhaps it would be better to leave them in? I checked my books, and the max spindle speed (4800) is below the max speed for grease (7100 with oil being 9500 rpm). I feel that maybe the oil speed is a better choice, but I will need to check the oil entry point, which for some reason I thick is below the top bearing......


Cheers,

Phill.

 

[machtool]

The only problem is the type and size...... It looks like a custom bearing. The OD is 62mm, the ID is 30mm and the width is 20mm. It is a twin row ball bearing.

Wouldn’t that just be a 4206, double row, ball bearing. The only other available size is a 2206 self aligning, Cant see one of those being used there. I cant see it being an angular contact also. Its just there for the radial load of the pulley. Its axial capacity is redundant if you have those swish precision angular contacts taking care of that all ready.

Phil.

 

[RC99]

Its just there for the radial load of the pulley.

Phil.

Should there be any side load on the spindle up there? I am thinking it would be splined to fit in the female spline on the top half..

I am guessing it looks similar to mine..

 

[AussiePhill]

G'Day Gents,

MachTool,

You're absolutely correct. It is an 4206. I'm not happy. That size isn't listed in my NSK and Timken books. I cannot find my ********* book, so I just looked on line, and there it is..... I'll swap it over tomorrow.

RC,

Yes mine looks similar to yours. The top bearing is critcal to stop vibration, as the main bearings are too close together to stop minor movement. The top bearing, being so far away (relatively speaking) stops the very minute movements that constitute vibration. The top bearing takes radial, and, a combination of radial and axial loads. Under light side loads on the cutter, the load is purely radial. Under heavy side loads, the spindle will deflect (bend) between the bearings, and this bending action applies the axial loads that go with the radial loads. The radial loads are incresed greatly in the 2 rows by the deflecting of the spindle.
Under medium to heavy loads, a single row ball race can often struggle to curb vibration. The twin row deep goove (or a twin row angular contact) will stabalise the spindle far better as it will not let the spindle deflect as much before the balls take up the axial loads. If you have only a single row bearing, see what can be done to fit a double row. It will reduce the chatter dramatically. The other factor that affects the deflection of the spindle, is the grade of steel and hardness of the spindle. If it's too soft, or even soft cored, then it will deflect much easier than a tough and hard (right through) material. Cost cutting is the main reason for weak spindles and/or single row bearings.

Back to the bearing shop for me in the morning.... I don't know why they didn't offer me a twin deep groove bearing in the first place.


Cheers,

Phill.

PS Sorry for the above lecture on machine design.

 

[Phil Burman]

How does this work, they both have the same modulus of elasticity.

Phil

The other factor that affects the deflection of the spindle, is the grade of steel and hardness of the spindle. If it's too soft, or even soft cored, then it will deflect much easier than a tough and hard (right through) material.

 

[RC99]

G'Day Gents,

MachTool,

You're absolutely correct. It is an 4206. I'm not happy. That size isn't listed in my NSK and Timken books.

You need to download this file.. http://users.beagle.com.au/lathefan/1bearing.pdf it's 7MB and is the bearing catalogue section from Bearing Services... Very good for ID'ing bearings..

Mine is fitted with a single row 6XXX bearing up the top..Looks like I have some mods to do to fit a proper double row...

Thanks for the explanation..