Delighted to see this report.  The important part for me is Dave experimentally confirming there’s an improvement.   One hidden thing to look for though is excessive wear, likely to take a few hundred running hours to appear.

My guess is the lathe designer chose a Belleville washer as a cheap simple way of managing bearing life.  I believe it’s a compromise: if the bearing is left loose, the lathe performs badly.  Therefore tight is better.   But, as tighter causes more wear and tear on the machine, it’s important not to overdo it.

Worst case, the lathe performs brilliantly for a few hours, then seizes in a cloud of smoke.   More likely, the extra wear and tear will be  slow, perhaps taking years to become obvious.

This kind of problem is often detected in critical machinery by periodically sampling the oil and having it analysed for metal content.   Gives a good insight into how fast wear is happening, and often which parts are failing fastest.   Used to improve safety by taking rapidly wearing engines out of service early, and to save money by deferring servicing of engines in good condition.      Analysing the oil feels over the top on a Unimat though!

Dave

With the U3 having a head that is 60mm wide we can bring that down to 0.0007″. Less if measuring between the inner bearing faces.

Then add in the fact that the whole head gets warm when running for some time not just the spindle in isolation so that also means the bearings will move to take up some of that 0.0007″ and you may be down to two tenths

This problem does not seem to get mentioned for mini mills and mini lathes that essentially have a similar setup of preloaded non taper bearings.

Interestingly when the CNC has been running non stop for a couple of hours at 5000rpm the spindle is noticably tighter than when cold so if spindle expeansion were the only factor then it should be looser not tighter. Ketan says this is fine and he knows a thing or two about bearings.

They are indeed Michael but you are thinking the same as me that it is not just the expansion of the spindle taken in isolation that matters it is the rate that other parts expand (or contract) as well that should be taken into consideration. As well as where the heat is coming from eg is it from cutting through the chuck and into the spindle, the bearings or the motor heating the whole assembly.

Also as you know preload is best set on a machine that has been warmed up not one that has stood in a cold shed during winter for a week without running so components are at their running temperature and therefore size.

The CI mill/lathe heads tend to put the bearings further apart than the little alloy U3 so the actual change of preload is probably similar eg longer distance x smaller differential change = shorter distance x larger differential change.

Just did some numbers.

Differential movement on the U3 between head and spindle at 50degF change would be 0.0009″ over 2.5″

On a Minilathe/mill with say 4.5″ over the bearings in a CI head then differential would be 0.0003″

I have not taken into account any angular position of the balls which may affect any percieved change in gap

But if preload is set on a warmed up machine it would not be worth worrying about as there won’t be any significant differential change. This is why commercial CNC machines are put through a warming up cycle at the start of the day before being put to work.