Let The Fun Begain, New build on the Table.

[Jonathan]

Moving the ballscrew, whilst not disastrous, hardly gains anything. All you're gaining is some protection for the ballscrew from swarf and there are plenty of other ways to do that - for example adding seals or bellows. So even if the difference to the overall machine due to the loss in rigidity is small, you're doing this to correct a problem which is also very small since it's clear that swarf on the ballscrew is also only a small problem. I recall Jazz saying he'd worn out his ballscrews from letting the swarf pile up on them over a few years, but I've not come across anyone else who has had a ballscrew (or nut) fail for that reason.

For this type of build it will work absolutely fine. Have a look around the forum with such a design and ask how they are getting on ;-)

Having the ballscrew on the back on my machine has crippled the stiffness of my Y-axis and although the distance is greater, that shows how the effect can be significant.

[Bush Flyer]

When I am finished with a tools I am probably OCD about keeping tools clean and put back ready for the next time, It's the same when building a new RC model before finishing for the day I will tidy up and put all tools back where they come ready for the next time, and I have always been like that so keeping the ball screw clean will not be a problem.

[Swarfing]

Having the ballscrew on the back on my machine has crippled the stiffness of my Y-axis and although the distance is greater, that shows how the effect can be significant.

Correct Jonathan yours was a huge distance from centre hence your issue. Switching would save a lot of hassle having to sort extras like seals and way covers etc. My suggestion was a sound one but the choice is yours of course. BF your design is a good one i would like to point out and will work great without the changes so well done. Building these things is a costly venture and you spent the time listening and looking at what works already. Mine is no way near as good as what you have so you should be in a good place.

[Ross77]

Sorry Ross in this instance i would have to disagree

No need to apologise, the more opinions Mike gets the better.

Swapping it over as i said would result in the ballnut being roughly in the same position,

Yes but it would need a link arm or bar to connect back to the bearings hence lever arm. It just seems odd to me that people use ball screws to rid of backlash to 0.01mm and then mount them in systems that deflect more than that.

I still agree with Jonathan and don't see why you would risk compromising machine ridigity just to keep a ball screw clean, that's what guards are for

Any way i think this point has been laboured enough and mikes happy to keep his tool clean......

[Swarfing]

Yes but it would need a link arm or bar to connect back to the bearings hence lever arm.

Mmmm! not sure what you mean here? If you are making a box to encompass the supporting rail then you are actually increasing rigidity. The bearings will be supported both sides of the fence so to speak so can not for the life of me see how it is worse? the ballnut will still be roughly in centre and mounted on the backplate. I really don't see what there is not to understand here.

[Ross77]

Mmmm! not sure what you mean here?

Ok better explain myself in case I'm wrong.

Just to make it clear thought that its only because open bearings and supported rail are being used with a small bearing centre to spindle nose distance ratio. Ideally the bearing centres (on plan) should be increased but given Mike has already redesigned it a few times I didn't want to offer any more changes without justifying them with calcs.

If you are making a box to encompass the supporting rail then you are actually increasing rigidity.

Yes but its not so much the bearing mounting but the bearings themselves, open bearings can have quite a large radial clearance (10 to 50 micron), with the current set up and using a conservative 10 micron clearance (could be to 50), 100mm bearing spacing and 200mm spindle distance that results in 39 micron of free movement at the spindle. moving the ball screw the other side increases this to 51 micron.

This is free movement just to take up play, if you add any deflection due to load then this will obviously increase further. for cutting Ali the open bearings can run close to the max load capacity and as rigidity is a function of imposed load and load capacity then deflection will be high.

On the face of it 10 micron is a very small amount but remember this is only One element and doesn't include deflection in the part. if you lose 10 micron at every junction then it all adds up to a considerable amount.

Any way that's my take on it!

[Swarfing]

Ross i did say that the plate could be moved further back. In this design there is no reason why the ballnut could be more central and the spindle moved closer to centre. There is no need to have the nut pushed further away from centre.

[Jonathan]

Ok better explain myself in case I'm wrong.

I thought your first explanation was good, but the latest one is even clearer. I agree with the 39-51um figure and as you say it's likely to be more. Any compromise like this will reduce the stiffness of each element. You can analyse the stiffness of the overall machine as springs in parallel, so if you reduce one a very small amount it's not likely to make a tangible difference, however if one is reduced substantially such that the stiffness is lower than other elements, then that will dominate the system, causing the overall stiffness to be low. This means compromising one part can waste money you've spent on other quality (stiffer) parts. An extreme example would be using profile rails on two axes and unsupported rails on the other. In that instance changing the profile rails to supported rails isn't likely to make a difference, since the weak point would still be the unsupported rails. Similarly, 'making a box to encompass the support rails' wont make much difference, since that part of the assembly is already at least an order of magnitude stiffer than other parts.

You can get the ballnut in a more optimal position with the current design - it doesn't have to go behind to do that. Similarly the current design could easily have a 'box to encompass the supporting rail' added.

Switching would save a lot of hassle having to sort extras like seals and way covers etc.

Just stick a 15mm shaft seals, like these, either side of the ballnut and the swarf is wiped off easily. They're working well for me and not showing signs of wear, although admittedly I've not been using them for long.