Pre-load of new bearings

[duncan webster 1Participant @duncanwebster1]

and here's the relevant section from Timken on fits for angular contact bearings. Note 'mean tight fit', not sliding fit

**LINK**

hopefully the link will work this time, you want page 130

Edited By duncan webster on 22/09/2017 10:07:37

[Neil WyattModerator @neilwyatt]

Posted by duncan webster on 22/09/2017 10:06:37:

and here's the relevant section from Timken on fits for angular contact bearings. Note 'mean tight fit', not sliding fit

**LINK**

hopefully the link will work this time, you want page 130

Edited By duncan webster on 22/09/2017 10:07:37

Lots of interesting stuff. Counter intuitively two widely spaced taper roller bearings are less sensitive to them effects that a close spaces pair with a standard one at the tail.

Earlier it says "loose fits" on the inner ring are OK for taper rollers if experience shows this does not cause problems.

The mini-lathe design uses an interference fit at the front of the spindle where the loads are highest and I suspect this is why what I would call a 'firm push fit' is satisfactory at the much more lightly loaded tail end of the spindle.

Interestingly the SC4 uses a big taper roller at the front and a ball race at the back.

Neil

[Neil WyattModerator @neilwyatt]

[Colin LLoydParticipant @colinlloyd53450]

You guys are brilliant – everything you say makes sense to this novice – even if it doesn't necessarily apply to my situation. And it's true that you learn most through getting things wrong.

Nick – yes everybody suggest that if you replace the bearing – use taper replacements – and I would have done if there weren't so many adjustments and alterations that need to be made to fit these – especially for someone who didn't even know what the inside of a lathe headstock looked like – or whether this was going to be a descending path into chaos. I know better now and should I need to replace the bearings again – I will go for taper bearings.

Anyway, I took the back-gear off the lathe, again, to the point where I could look at the movement of the spindle and the individual bearings (I can nearly do this part in my sleep now). And at that point I could move the spindle, by hand back and forth by a few 10ths of a mm. And this movement was all on the inner ring of the angular bearing at the drive end – which tends to say that despite my efforts I had failed to apply any preload originally. May have been a similar movement at the Mandrel end but difficult to ascertain what was happening there as the mandrel shoulder covers the inner race and part of the sealing rubber skirt – but luckily (or most probably by design) not the outer ring – otherwise you couldn't apply pre-load to the mandrel end bearing.

But I think the initial problem was using one of the old bearings as a "pusher" item so that the bearing went in straight and to protect the new bearing from the steel pipe that was used with the leather hammer to drift the new bearing on (unfortunately several Youtube videos showed this was how you did it). Once the outer race was seated fully in its recess, no more pressure would move the inner race further along the spindle. And no amount of what I deemed to be reasonable tightening on the spindle through the C-nuts would cause the inner ring to move on the spindle to create pre-load.

While I appreciate that there should have been sufficient "give" on the spindle to allow the C-nuts to adjust the Pre-load – I'm loath to take the bearings off the spindle and start again – as I have found that these angular bearings, through their design, are prone to come apart in one direction if not the other. Again I realise that this is due to the very tight fit on the spindle – but again as a novice I thought this tightness was normal – will do better next time.

I made up a press along the lines that Neil suggested (thank you Neil) and gradually tightened the inner ring along the spindle – all the time making certain that the spindle would still turn easily – and I have got rid of any apparent play in the spindle. I will get a clock on the mandrel face and see what movement there is still left through hand action – not very scientific but at least a start.

I realise I'm actually applying preload – that will not come off when I release the pressure (because of the tight fit again) – and that this is therefore possibly a one-way action. Neil – you mentioned somewhere else applying pre-load by tightening up the C-nut by hand and then by another amount defined by an angle. Is there a torque load that I could apply that essentially does the same thing? Anybody?

[Clive Brown 1Participant @clivebrown1]

Is there a torque load that I could apply that essentially does the same thing? Anybody?

If it's any help, this topic has prompted me to look at my Boxford booklet. The headstock has 2 taper-roller bearings and Boxford recommend setting the pre-load, by adjustment of the end-nut to give a specified torque necessary to slowly rotate the spindle. This is measured with a length of string wrapped around a face-plate and a spring-balance. A torque of between 1 and 2 lb. inches is recommended, with the lower figure for higher speeds I've never actually tried this, so can't say how easy it is to accomplish.

The fit of the inner bearing rings on the spindle is described as a "gentle drive fit" and as said, the ring-nut provides the pre-load adjustment. If overtightened, Boxford recommend a "gentle tap with a rubber mallet" after slackening the nut.

I realise that your machine differs, but you might get some indication of what's required.

Pre-load from a defined angle of turn on the nut might be good, but suffers from the problem of knowing what that angle should be.

Pre-load from torque-tightening the nut would suffer the same unknown as well as, IMHO, be very imprecise.

May I make one further comment, from your description the bearing fit on your spindle seems tighter than I would have expected.

Hope that is helpful.

Clive

Edited By Clive Brown 1 on 22/09/2017 13:48:50

[Neil WyattModerator @neilwyatt]

The best way to set the preload is to make sure that, after 20-30 minutes of running, the headstock is warm to the touch but still quite comfortable to leave your hand on.

A non-contact IR thermometer is used in industry.

be warned hot bearings die quickly, and may be dead before you notice the heat, so take it gently.

Correct preload also gives a no-shake fit once warmed up, a symptom of insufficient preload is slop when warm as the spindle warms up and expands more than the headstock. We are talking tiny distances here.

Neil

[Ketan SwaliParticipant @ketanswali79440]

Posted by Neil Wyatt on 22/09/2017 14:34:18:

The best way to set the preload is to make sure that, after 20-30 minutes of running, the headstock is warm to the touch but still quite comfortable to leave your hand on.

A non-contact IR thermometer is used in industry.

be warned hot bearings die quickly, and may be dead before you notice the heat, so take it gently.

Correct preload also gives a no-shake fit once warmed up, a symptom of insufficient preload is slop when warm as the spindle warms up and expands more than the headstock. We are talking tiny distances here.

Neil

Neil,

You are now entering the voodoo grey area with this one. This heat business is based on a combination of grease content, pre-load pre and post run-in over time. The subject is too long to get into detail. In Colins case it will be down to 'feel' more than anything else, and trying to keep things simple.

Grease: Factory greased for correct amount, and sealed, in case of AC bearings. If TR bearings, fill in internal content with 20 to 25% internal volume with GP grease.

Hand tighten, put on about 5 minutes turn (as unable to visualise degrees) on inner locking ring. Lock in place. hand turn chuck and feel the resistance. If all okay, turn on the machine on slow speed, increase to max over say 30 mins. Feel the temperature on head – hottest part you can find with some kind of temperature probe. Record the temperature. let it cool down. Repeat as necessary, tighten very little (here less is more), but please make sure not to over tighten. hand turn the chuck and feel the resistance every time. If it feels okay, then it is okay. Never over pre-load.

Note: there is a direct relationship between pre-load, speed and heat. During the running-in period, expect the temperature to rise sharply if you have put on high pre-load, and slowly if low pre-load. Very roughly speaking, the temperature at the hottest point at max speed continuous could rise (very roughly) to between 20 to 30 deg.C above room temperature. After a while of running in, the temperature should slowly drop and stabilise at some level (but I cannot remember around where).

Based on use, the bearings will bed in over time, and you will need to tighten the locking nut 'a little more'.

After you are happy with the initial installation, and once the machine is warm – suitable for regular use, put in a test piece of each material you are likely to turn. if you are happy with the result, put the test piece away with the date you turned it, with what. Three months later (or based on use), turn a new piece of the same material from the same reference bar-stock, compare it with the original turned piece. If you see ripples or a finish which is not as good as the original (under same conditions), that would be the time to add more pre-load, as hopefully this will suggest that the original bearings have bedded in. Then turn another piece, and hopefully this should be as good as the original.

Above is the simple method. there are more technical methods with the right equipment, but not necessary or critical in Colins case.

Ketan at ARC.

[Ketan SwaliParticipant @ketanswali79440]

Just some further points of note for Colin:

– traditionally AC bearings do not have seals. They are open in the same way as TR bearings, because to the 'near' separable nature of the inner and outer ring/design of AC bearings, similar to TR bearings.ARC introduced the AC bearings with rubber seals (2RS), to reduce the possibility of contamination over time. However, it does mean that when you try to replace these AC bearings/remove them, the removal process can separate the inner ring from the outer ring more easily, as you have found, and this is due to the design of AC bearings.

– in terms of accuracy – specifically on a mini-lathe, it will be very difficult to see a difference between AC and TR bearings. Secondly, there is lower surface contact of 'balls' in AC bearings with the inner surfaces of inner and outer bearing ring, (two point contact). the 'roller' elements of a TR bearing are in full contact with the inner surfaces of inner and outer bearing ring. Result is more friction with TR, less friction with AC, under pre-load conditions, allowing the spindle fitted with AC bearings to run faster than the TR bearings, giving better finish at high speed.

AC bearings with pre-load will run a little lower maximum speed in comparison to deep grove ball bearings-to which very little pre-load can be applied. This results in deep grove ball bearings providing variable lower accuracy in comparison to AC and TR bearings, but still suitable for the job.

AC is a good compromise between deep groove ball bearings and TR bearings, and one gets the best of both worlds in this particular application.

Hope above is as clear as mud.

Ketan at ARC.

[Neil WyattModerator @neilwyatt]

Ironically, I found the main benefit of the TR bearings to be the extra 5mm between the flange and the headstock

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 22/09/2017 18:41:48:

Ironically, I found the main benefit of the TR bearings to be the extra 5mm between the flange and the headstock

.

Yes … When turning Iron, the extra distance helps stop dust contaminating the front bearing.

MichaelG.

[Neil WyattModerator @neilwyatt]

Posted by Michael Gilligan on 22/09/2017 20:22:09:

Posted by Neil Wyatt on 22/09/2017 18:41:48:

Ironically, I found the main benefit of the TR bearings to be the extra 5mm between the flange and the headstock

.

Yes … When turning Iron, the extra distance helps stop dust contaminating the front bearing.

MichaelG.

Actually it means it's MUCH easier to fit chuck securing nuts in the gap…

[Michael GilliganParticipant @michaelgilligan61133]

Sorry Neil … it was meant as a pun.

[Colin LLoydParticipant @colinlloyd53450]

Being an ex-scientist, I thought I'd do a series of experiments to ensure I don't "burn-out" my new bearings. The initial test was for a pre-load that seemed to take out all the free-play in the longitudinal and radial axes of my lathe spindle. The amount of preload was such that a mass of 125g acting vertically downward from a chord wrapped around my 100mm chuck was sufficient to just begin to rotate the spindle – not a lot of preload I guess. A 2.5mm hole was drilled into the top surface of the headstock box over the centre of the rear bearing to a depth of 7mm – so not right through to the outer bearing surface. Computer CPU thermal paste provided the thermal contact between a K-type thermocouple and the metal surface at the bottom of the hole.

Over a period of 40 minutes, the lathe was run at increasing speeds at 5 minute intervals and the temperature of the thermocouple recorded every minute – the result is shown in the graph below. The initial temperature was 18 deg C and initial rotational speed was 240 rpm. The arrowed figures show the increasing rpms (black is low range, red is high range). At the end of the 40 minutes, the headstock recorded 37 deg C while a contact measurement on the inner ring of the rear bearing recorded 45 deg C.

I shall now increase the preload by a fixed amount as judged by my rotational mass method and repeat the experiment. I need to find some figures for maximum operational bearing temperature and also some torque figures for preload that I can convert to my simple rotational mass method.

Eventually this may be of use to other mini-lathe owners.

[Neil WyattModerator @neilwyatt]

Posted by Michael Gilligan on 22/09/2017 21:27:15:

Sorry Neil … it was meant as a pun.

there isn't a 'scratching my head' emojii…

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Michael Gilligan on 22/09/2017 20:22:09:

Posted by Neil Wyatt on 22/09/2017 18:41:48:

Ironically, I found the main benefit of the TR bearings to be the extra 5mm between the flange and the headstock

.

Yes … When turning Iron, the extra distance helps stop dust contaminating the front bearing.

MichaelG.

.

Neil,

I can't find a highlighter on the formatting, but I have emboldened and italicised the relevant words.

It leaped out at me … but I'm really sorry I bothered !!

MichaelG.

[Colin LLoydParticipant @colinlloyd53450]

Experimental Update: See above for 1st test (result labelled Temp(125g) on graph below. Increased the pre-load on the bearings – it now takes 163g mass to rotate the 100mm chuck. Did the same test again – increasing the rpm at 5 minute intervals and recording the temperature every minute. At the end of the 40 minutes, the end bearing inner ring recorded 47 deg C. Results shown as Temp (163g) on graph. Then did a complete run at max revs (nominally 2500) and recorded temperature until it plateaued at 42 deg C when the bearing inner ring recorded 48 deg C. This is labelled Temp (2500rpm) on graph. During this run it was noticed that the revs increased from 2455 at start to 2559 at finish (labelled RPM on graph- right hand scale). Not sure if this was due to bedding in of the bearings or lower friction through heating. Difference between Bearing inner ring temperature at start and end = 30 deg C which seems to be at the top end of Ketan's range. From this, I'm guessing I'm about there with pre-load but prepared to be told otherwise. Will reassemble rest of lathe and do some test cuts.

[Michael GilliganParticipant @michaelgilligan61133]

Nice work, Colin

MichaelG.

[Ketan SwaliParticipant @ketanswali79440]

Posted by Colin LLoyd on 24/09/2017 15:38:28:

Experimental Update: See above for 1st test (result labelled Temp(125g) on graph below. Increased the pre-load on the bearings – it now takes 163g mass to rotate the 100mm chuck. Did the same test again – increasing the rpm at 5 minute intervals and recording the temperature every minute. At the end of the 40 minutes, the end bearing inner ring recorded 47 deg C. Results shown as Temp (163g) on graph. Then did a complete run at max revs (nominally 2500) and recorded temperature until it plateaued at 42 deg C when the bearing inner ring recorded 48 deg C. This is labelled Temp (2500rpm) on graph. During this run it was noticed that the revs increased from 2455 at start to 2559 at finish (labelled RPM on graph- right hand scale). Not sure if this was due to bedding in of the bearings or lower friction through heating. Difference between Bearing inner ring temperature at start and end = 30 deg C which seems to be at the top end of Ketan's range. From this, I'm guessing I'm about there with pre-load but prepared to be told otherwise. Will reassemble rest of lathe and do some test cuts.

Great experiments, and a great log Colin. Your guess is probably right for the pre-load. Do post your results for the test cuts.

I know that others in future will refer to this thread for guidance. So, I would just like to mentions some words of caution for them:

1. Always avoid over-tightening. When one over-tightens the nuts, it means that the motor has to work harder, drawing in more power at times through the control board. In turn, this could have adverse (negative) effects on the motor, as well as the control board.So, be careful.

2. Brushed DC motors work a little differently from Brushless motors. the principal is the same when it comes to pre-load, but the behaviour of the motor and moreover the control boards will be different. So again, don't just depend on one factor/variable to make judgements and conclusions.

3. load on the spindle with an 80mm chuck will be different from the load on the spindle fitted with a 100mm chuck. So, again, keep this in mind.

4. Temperature on the front end (working end) – where the chuck is mounted is likely to be greater than the temperature at the rear end, due to the load of the chuck at this end. It is my understanding that the temperatures observed on Colins chart are at the rear-end. So, avoid prolonged 'running' at top speed at the top end temperatures. If it looks like the temperature is likely to continue to increase, stop the machine, let is cool down, start again. if same happens again, loosen off on the pre-load.

5. Colins process of bedding-in should be regarding as the starting point. It will take further time for the bearings to really bed in, based on use.

6. The process Colin carried out with pre-load applies to AC and TR bearings only. Please avoid using this guidance for deep groove ball bearings, as in real terms – deep groove ball bearings are not really designed to accept 'proper' pre-load. Although I am aware that some people ignore this comment, it really is not good for deep groove ball bearings.

Over the years I have seen a small amount of bearing/motor/control board failures resulting from 'overdoing' things. So, one should keep these points in mind, as whatever you do, experiment, you do so at your own risk. Comments made on here are for general guidance, and each machine behaves differently.

Ketan at ARC.

Edited By Ketan Swali on 25/09/2017 20:52:38

[Neil WyattModerator @neilwyatt]

That's really interesting, Colin.

Thanks for reporting back.

Neil

[Colin LLoydParticipant @colinlloyd53450]

Sorry for late reply – been Fell walking in the Lake District. But also wanted to improve the wiring system and circuit board protection of the lathe control box. So have routed the normal leads that hang off the control box to 30A rating banana sockets on the side of the control box. This enables the control box to be easily removed in future without straining the current connections which, imho, would eventually fail or strain the circuit boards. Also the original glass-fibre sheet that, sort of, covered and protected the circuit boards was again, imho, inadequate. So I have used some thermoplastic sheet (Worbla's Black Art) to mould a perfectly fitting cover for the electrics.

Finally got round to doing a simple test cut. Please remember that I am a novice lathe operator and the picture shows just a thin cut along the 25mm dia aluminium rod and a facing cut using just a left-hand turning tool – not a finishing tool. I was pleased with the long curls of metal that came off and with the whole solid feel of the process – compared to the awful gougeing that occurred before applying pre-load. So I think I'm there now and have also learnt a lot along the way.

[Ketan SwaliParticipant @ketanswali79440]

Looks reasonable Colin. As long as you are happy with the results, that is what really matters.

Ketan at ARC.

[Keith MoatParticipant @keithmoat36026]

Are the plastic spacers on the spindle crushable ?

I.e. supposing there was a slight difference in the distance along the spindle between the inner races of the original deep groove roller bearings and the distance along the spindle between the inner races of the angular contact bearings, when under preload (less distance ?), would one still be able to slightly crush the original spacers to give enough preload ?

Keith.

Edited By Keith Moat on 11/05/2018 13:13:50

[Ketan SwaliParticipant @ketanswali79440]

Posted by Keith Moat on 11/05/2018 12:56:34:

Are the plastic spacers on the spindle crushable ?

I.e. supposing there was a slight difference in the distance along the spindle between the inner races of the original deep groove roller bearings and the distance along the spindle between the inner races of the angular contact bearings, when under preload (less distance ?), would one still be able to slightly crush the original spacers to give enough preload ?

Keith.

Edited By Keith Moat on 11/05/2018 13:13:50

If you go with AC bearings, just put enough preload based on earlier guidance in this thread. This should not really 'crush' the spacers in the way you think. However, some like Mr.Rudd have preferred to make them out of alluminium to avoid leaving it to chance. Whatever you decide, don't over think it, and note that if you put on too much pre-load, it will result in your motor needing to work harder to turn the spindle, which in turn could damage it and/or your control board.

Ketan at ARC.

[Keith MoatParticipant @keithmoat36026]

Thank Ketan, I'm just about to embark on this project, is it just a case of tighten until the spindle is just a little tight to turn by hand then back off a touch to protect the control board, or would you keep it a little tight.

Does ambient temperature make a difference to how tight to set it initially ?

Keith

[Ketan SwaliParticipant @ketanswali79440]

Posted by Keith Moat on 11/05/2018 13:33:27:

Thank Ketan, I'm just about to embark on this project, is it just a case of tighten until the spindle is just a little tight to turn by hand then back off a touch to protect the control board, or would you keep it a little tight.

Does ambient temperature make a difference to how tight to set it initially ?

Keith

Yes and Yes and have a read through this whole thread before you make a start on pre-load. your point about how much to tighten, covered from my post on 22 Septemebr 2017 onwards. Regardless, I would strongly suggest you read from the beginning of this thread/post.

Ketan at ARC.

[Author]

Posts