Bearing play – Rotary Table

[James Hall 3Participant @jameshall3]

My first venture into gear cutting (timing gears for Farm Boy) – all went surprisingly well on a trial run (using aluminium) and the two spur gears actually meshed at the correct centre distance so my calculations were confirmed as OK. But they bind at one point which surprised me as I had taken great care to ensure that the blanks were concentic on their axles.

Investigation reveals the fault to be in my rotary table: runout <= 0.015mm which I think is acceptable, but radial play of 0.07mm on the table. The play was in the 6906R5 bearing in which the table rotates.
Simple, I thought, replace the bearing, which I did only to find similar play in the new one once all was reassembled.

The bearing came from a motorcycle spares shop – sold as a wheel bearing – and was very cheap compared with the same from a bearing supplier. I was a bit suspicious of the low price but reasoned that it must be OK sold for that purpose – I certainly wouldn’t want to be riding a bike with that much play in its wheel bearings!

Now, the dilemma: should I invest in a better quality bearing, or can I expect it to exhibit a similar amount of play. I’d be well p’d off to spend more, go through the whole dismantlin/reassembly process just to find myself still no better off.

Any thoughts?

[HuubParticipant @huub]

A common ball bearing like the 6906 will always have a bit radial and axial play. That is by design, other wise small  axial forces on the bearing would cause a huge load on the balls.

On my cheap rotary table, I can adjust the axial play by tightening a nut at the bottom of the rotary table.

In general gears have a lot of play. On a module 1 gear set, I adjust the play by placing a piece of common printer paper between the gears. That sheet of paper (80 grams) is about 0.15 mm thick.

During the cutting of a gear, the cutting forces can become quit high. Without support, the shaft that holds the gear blank will bend a little. Supporting the shaft with a (live) center can reduce this bending.

During the cutting you have to lock the rotary table to avoid moving of the table due to the cutting forces. Because I often forget that, I have added a pneumatic locking system to my rotary table. My rotary table is CNC controlled.

[cogdobblerParticipant @cogdobbler]

Talk to a good quality bearing supplier and get a good quality precision  bearing with reduced internal clearance. These usually have C1 after the number on them to designate tighter clearance.

Motorcycle wheel bearings quite possibly use a bit of extra clearance such as C3, or are just cheap and nasty.

Or you could turn up a nice tight bronze bush. I can’t see why a slow moving RT needs a ball bearing necessarily. For a real budget solution, turn a steel spacer with a thin bronze bush pressed into the centre.

[noel shelleyParticipant @noelshelley55608]

6906R5 ?  Not 6906 RS ? indicating rubber seal often marked as 2RS. Noel.

[DiogenesParticipant @diogenes]

Just out of interest what table is it? ..superficially at least, a ball bearing seems an odd choice..

[parovozParticipant @parovoz]

And on assembly, look to have some pre-load on the bearings to take up any clearance slop.

[James Hall 3Participant @jameshall3]

Noel Shelley – you’re right of course, 6906RS – can’t read my own writing.

[James Hall 3Participant @jameshall3]

Diogenese: No branding on it. Bought from Chronos and is, or is identical to as far as I can see, to Sorba HV6.

[Diogenes]

On Diogenes Said:

Just out of interest what table is it? ..superficially at least, a ball bearing seems an odd choice..

I don’t want to over-state this, but it reminds me of the recent discussion about lathe headstocks

Sometimes ball-bearings are just an ‘expedient’ [cheap & nasty] option, when a plain bearing would do the job better.

MichaelG.

[Michael GilliganParticipant @michaelgilligan61133]

https://www.chronos.ltd.uk/wp-content/uploads/ROTARY-2017-2-1-1.pdf

 

[parovoz]

On parovoz Said:

And on assembly, look to have some pre-load on the bearings to take up any clearance slop.

Yes, good idea. If you were to shim the inner or outer race axially  to put a little axial preload on the ball bearing, it should eliminate radial play.

[parovozParticipant @parovoz]

If it’s as per that SOBA linked then the threaded collar should provide the pre-load? Or as per Cogdobbler, a bit of shimming. Does the threaded collar have a locking / clamping screw to maintain its position when tightened? That way pre-load can be quite accurately adjusted. Many machines have this facility on shafts that require high accuracy / stability etc.

But if as it looks on the drawing, there are two bearings, then pre-load solves it. Many higher speed applications have very precise bearing pre-load requirements for the very reason of smooth running and play free stability.

All the best, and sure you’ll get it sorted OK.

 

[old martParticipant @oldmart]

I bought a Soba 6″ rotary table in almost new condition and found that the central bearing had a lot of radial play. Upon dismantling it I found that it used a needle roller bearing running on a sleeve fitted to the top of the table. The needle roller was marked “India”. The hardened sleeve was good to leave in place and I made an aluminium bush from the neck of a high pressure floatation bottle from a helicopter. Years later there is no measurable radial play. I didn’t have any bronze to hand at the time. The RT gets dismantled, washed and relubricated twice a year.

Why they used a needle roller in something that is slow moving makes no sense to me.

[James Hall 3Participant @jameshall3]

Thanks for all your helpful replies – I’ll try to respond as concisely as possible.

Preloading the bearing in question to eliminate radial play may be difficult – the rotating table has a machined (ground?) bottom wich rests on the similarly machined top of the base and axial play is eliminated by a round nut (with locking screw) underneath via a thrust bearing. The bearing in question therefore plays no part in the axial positioning of the rotating part and could not be preloaded by tightenng the nut unless some very delicate (and probably inpractical) shimming were used to impose a preload. The bearing could be preloaded by using spring shims, but I’m not convinced that this would provide a positive enough preload to deal with some of the quite high axial loads that might be expexted.

I agree that this is a situation not really calling for a ball or needle race and that a cheapo bearing has probably been used as the least expensive option in a budget table. Interesting to note that someone has found the same excessive play – I wonder if this is a common problem in budget tables and has remained unnoticed. I also wonder how better quality tables are constructed in this respect.

I think that I will proceed with the suggestion of a plain bearing but using a brass bush and steel collar – brass should be adequate and considerably cheaper than bronze (it has to fit over a 30mm diameter shaft so quite a large lump is required). I take the point that some minimal play might be required to allow for any slight misalignment but could hopefully make this good with some corrective machining.

[IanTParticipant @iant]

Well James, a 30mm ID Oliite bearing will cost you about £8. Not thought about it too deeply but it might be an alternative. Brass bar isn’t that cheap these days…

Regards,

 

IanT

[PeteParticipant @pete41194]

Whatever you could attach onto a 6″ diameter rotary table and then machine, any radial or axial loads are going to be tiny. For the slow rotation speeds and in reality, very low number of rotations over a lifetime of M.E. use. Ball or roller bearings and especially cheap one’s aren’t the only or even best solution. These 2 videos might give some ideas and I think very useful information.

If your rotary table is mechanically set up like mine, the threaded adjustment on the end of the spindle only adjusts the axial end play and nothing else. The bearing or if changing to the better cast iron bushing on the lower end of the tables spindle does the location of the spindles radial position. The mention about taking these RT’s apart for a good cleaning and fresh lubrication once in a while is a good one. Even more important after buying a new one and before even using it for the first time.

[cogdobblerParticipant @cogdobbler]

<p style=”text-align: left;”>Cast iron might be cheaper and more durable than brass?</p>

[James Hall 3Participant @jameshall3]

Ian T.: I’m a fan of oilite bearings and seriously considered this route – thanks for the suggestion. I decided against it in the end because I may have to machine the table spigot and would then be stuck with having to match an oilite inner diameter. I’m also not sure as to the precision of them – probably fine, but there’s not much there if they did need some machining – they’ve been fine in some less accurate applications and I’ve used them for in the past.
As it happens, a suitable piece of brass round (50dia.x50) only cost a couple of quid more and with less risk of the above problems so that decided me.

[James Hall 3Participant @jameshall3]

Pete: Those videos are useful, many thanks – I note the solid bearing in the Vertex table which rather confirms that the ball race in the budget tables is probably just a cheap and cheerful solution.
I’d already decided on ‘tuning’ my table in the ways shown in the video (subject to some careful measurement) so it’s reassuring to see that my thinking is probably along the right lines.
I had already fixed positioning lugs on my table base – but removed them again – too easy to dink the mill table when getting a heavy RT into position, and only a matter of a few minutes to align by running an indicator accross the RT (though the other way round, of course).

[James Hall 3Participant @jameshall3]

Cogdobbler: wish I’d thought of that – cheaper, yes indeed. Don’t think that in this application wear is likely to be a problem within my lifetime (I’m 76).

[Author]

Posts