Spindle design

You may wish to think about adding a flywheel to the pulley so you have some stored energy to even out the cutting speed when milling.

regards Martin

To amplify Kiwi Bloke's comments, you need a reasonable preload to resist forward loading, the actual spring rate doesn't matter. Disc springs are a good alternative, but I have the impression they are stiffer than wave washers, so you need to be more accurate in machining axial lengths. Again as someone says you can stack them up \/\/ which reduces the overall rate and gives wider tolerance

Wavy washers and belville wshers wil give you pre load but not control over end float .

You will see them used in electric motors where endfloat is not critical and to some extent desirable , look at things like lathes and mills where end float is critical they use either taper roller , angular contact or thrust bearings to keep endfloat in check .

I would be looking at either single row roller bearings with inboard thust bearings or angular contact bearings provided you can get them to handle the speed range you want. Polishing the spindle to get a nice firm push fit is reasonably easy so set up the housing as the press fit or loctite retained . Only the inner of the bearing on the adjustment nut end needs to be a little looser , A light to moderate tap fit is what you want and on assembly i would nip everything up lightly and run it until it comes up to a stable operating temperature then adjust the endfloat using a dti and the retaining nut ( should have a lock nut also )

Edited By XD 351 on 30/10/2018 12:08:18

Having had a quick look for belleville/wavy info I've realised just what a spindly bearing the OP is proposing, 6mm ID, 10mm OD, 3mm thick. He won't find a suitable spring I think. In fact for a 6mm shaft the hole in the spring needs to be say 8mm minimum, the only one I can find is 15.5 OD, 10.5 ID (EPL4 or 5) which means a 625 bearing perhaps, 5*16*5 or next step up 16.5OD, 11.5 ID and a 606ZZ.

This is all making it a lot bigger, so perhaps not feasible after all

For a unidirectional load the outer (stationary) member does not need to be a tight fit, but it won't do any harm as long as it's not too tight. The inner (rotating) does need to be tightish, so as I said before if you are adjusting via the inner you need to get it just right, too slack and you get movement between inner and shaft which causes it to get slacker over time, too tight and you can't adjust it. If you don't believe me read the manufacturer's bumph sheets.

As I suggested earlier, the spindle would benefit from being as large a diameter as the OP's application can accommodate. However, Belville washers are, IIRC, available in small diameters.

XD 351's comment about spring preload washers not controlling end-float is not strictly correct. They will not eliminate end-float, but can control it, to an extent that allows the design to be practicable. The force required to move the spindle axially can be designed to be higher than anticipated forces applied to the spindle by the cutter, etc. That this works in practice is borne out by the thousands of (fairly) satisfied Unimat users and its very widespread use in precision spindles.

The whole point of spring pre-loading in this application is because trying to set preload by screw-adjustment is difficult - you can't tell when it's correct and, in small sizes, the axial displacement required for reliably setting the correct adjustment is too small to be applied easily by a practicable-pitch thread, especially when a lock-nut disrupts the adjustment. If the adjustment of such a rigid set-up is such that the axial float is just eliminated, there will be minimal bearing preload - less than the 'correct' amount. However, setting the 'correct' amount is difficult - how do you measure it?

In much larger applications, opposed taper roller bearing-supported shafts are often set to a desired torque required to overcome the drag imposed by the preloaded bearing. Great fun, swapping shims, rebuilding, measuring torque with a spring balance pulling a string wound around a gear, taking it all apart and doing it all over again. And again. When, years ago, I rebuilt a Land Rover transmission, (LR service manuals detailed this method), I was surprised to learn that the local main dealer never opened transmissions - they just bunged in 'recon units'. The labour (and skill?) required was too much. Luckily, they carried boxes of shims for us idiots working for free, in our own time...

OK, enough of this counsel of perfection - don't be put off. You will probably get away with a screw-adjusted arrangement; make the thread as fine as you can and adjust it until there's just a little rotational drag, and there's no rattling on shaking the assembly. Spring pre-load makes life easier...

(Stupid emoji added to my previous post, apparently by 'auto-correcting' software built into this forum. Can't you just wait for self-correcting autonomous cars?)

Spring pre load is also variabale over time as the spring can loose tension .

With a mill , lathe or surface grinder even a thou or two of end float can cause issues or dig ins . If the OP uses angular contact bearing or a normal bearing with thrust bearings behind them and sets the spindle up with some pre load he will be ok , i set them up using a dti to just eliminate any end float then turn the adjustment nut 1/16 of a turn then lock it and have never had an issue with any of the lathes , mills and grinders i have rebuilt - i have never seen a wavy washer in these machines except in the motor and if i ever find a machine with these fitted they will be coming out ! If the OP uses a thread on the spindle with 40tpi he can essentially set the retaining nut up as a micrometer adjustable type by making a nut with a collar scribed with 25 equally divided lines so endfloat adjustment can be achieved with a resolution of 0.001 “ or better .