High speed CNC spindle

[mu38&Bg#]

After disassembling the spindle, I now have more trouble. The shaft that's been damaged the has very little runout on Vee blocks. This is somewhat expects because I don't have the Vee blocks where I had bearing mounted. The second new shaft I have actually shows greater runout on Vee blocks. On top of that, the balls in the duplex bearing pair show bands. I don't know what this is caused by and am trying to find some information.

What I'm going to do is put this back together with the new shaft and give up. If the time comes for a better spindle, I'll look at this again.

This just shows that a simple programming error can destroy a CNC machine instantly.

Greg

 

[steamer]

Sorry to hear about that bud! I hope you didn't wipe the bearings out....are they smooth?

Dave

 

[mu38&Bg#]

They feel smooth, but spinning it I can ever so slightly hear a high pitch buzz. I do have more bearings and at the twenty some odd dollars a pair I found them for, it doesn't hurt, but I don't want to burn up another set. I'll continue to run this set until problems become serious. The only reference I found to ball banding was reverse loading of angular contact bearings. These are definitely installed correctly. if I knew how to get these apart I'd have a look at the races.

 

[Florian]

Hey Greg

What do those "bands" look like? Maybe there are induced currents going through the bearings?? (This would harm bearings and make "bands" on the races.

Maybe you can cut one of the bearings races if they are crap and see what the raceway looks like...?

I mean, what may has happened is This.

Cheers Florian
btw: If those shafts are all crap; just make your own and use it unhardened! If you use some high-tensile steel, the collet cone will last quite long. Then you could also make it shorter so the bearing is as near to the collet as possible...

 

[mu38&Bg#]

I don't think it's current. The upper bearing looks like new. The bands are just hazy. They don't catch on an Exacto blade. It's quite possible it's just the very early stages of wear due to the crash. I think the spindle squealed for a second the first time I started it afterwards. It ran for three minutes or so.

These bearings have one piece composite retainers. I'll have to figure out how to get them apart to look at the races. I'm thinking they'll just fall apart if I give the inner race a push.

I don't have the means to grind the spindle or I'd think about making one. Quite honestly I just don't have the time to work a new spindle unless the replacement shaft turns out to be unusable. I have a project planned that needs a .050" endmill with a .5" reach. It won't handle much runout.

If anyone is planning a similar project, I'd stick with 1/8" tooling.

 

[Florian]

Hey Greg

With the new Shaft, if it is not running true and you own a dremel, you could try this:

Clamp your Spindle in the 3-Jaw Chuck of your lathe.
Support the spindle in the front with a fixed steady.
Then adjust the compound slide to the collet cone (using a dial gauge with lever).
Now you clamp the dremel on the compound slide and set its center exactly to center height of your lathe.
You can use a Aluminum oxide grinding stone for grinding the cone.

You should also cover all the ways of your lathe and if possible use the shop vac when grinding.

Finally you can run the spindle slowly, in opposite direction to the dremel and use the dremel on the compound slide to grind the collets cone.

When grinding you should only take very light cuts and always move forward an backwards a few times until there are no more or only very few sparks when moving.

I think this can be done in 1 to 2 hours and you will have very small runout because you have ground the cone IN the spindle while it was running by itself.


So, if required and if you own a dremel, I would say give it a try!

Cheers Florian

 

[Mosey]

My old old Dremel is loose in it's bearings. I have a cross-slide mount that I made, but I assume that since it is wobbly, I shouldn't try grinding on the lathe with it, right?

 

[Florian]

but I assume that since it is wobbly, I shouldn't try grinding on the lathe with it, right?

No, not exactly!
Of Course you won't get the best possible surfaces this way but the grinding stone will place itself to where there is the least forces on the (dremel) shaft. And If you work carefully, you will get quite good results with that.

Of course you will get better surfaces with a grinding spindle but the surface you get with the dremel should be good enough for the collets. It is also important that you set the grinding stone that it touches the cone only on a small area and not on the full lentgh.

Id say try what happens.

By the way if the surface of the cone is not as good as you want it to be, you can make some aluminum arbors with the collets shape and then lap the cone with some diamond paste.

Cheers Florian

 

[mu38&Bg#]

I just got the new shaft into the spindle. The bearings don't make the whirring noise so maybe there will still be some life in them.

However! Taper runout is worse than the bent spindle, almost .002" TIR. That first collet extension must have been a fluke at virtually perfect.

I was hoping not to rework the taper, but it seems I have no choice now. I really doubt the Dremel. I used a much more stout grinding spindle, though with a little runout on the taper, on the lathe and I wouldn't use it for this. I might mount the spindle in the mill with a boring bar on the table and write a program to true it.

Greg

 

[steamer]

Either that or mount the assembled spindle housing into a steady rest and turn it. If it's not that hard.

Otherwise....the mill approach is a decent one......I think.

Dave

 

[Florian]

the mill approach is a decent one......I think.

I thought about this when i wrote my last post but left it away because i excepted the z-clearance to be too small for doing this...

But if you can manage to, this would be the best way I think.

Cheers Florian

 

[mu38&Bg#]

Well, it's back together. I had a heck of time balancing it again. I gave up and only got it close, but it runs smooth installed in the mill. I mounted a carbide tipped boring bar in the vise and wrote a program to true the taper. It turned out about .0002"TIR. A rod in a cheap collet was about .0006" TIR. I'll live with it for now, a good collet should do better anyhow.

Note to self, on the next (yes, I'll punish myself) spindle design balancing into the components with easy access.

Greg

 

[steamer]

Hi Greg,

The design of spindles is really tough. There are so many conflicting trade offs.
I think you did great!
:bow:
Dave

 

[steamer]

....and they're ALL hard to build! Now you know why they cost so much!
:big:
Dave

 

[mu38&Bg#]

Yes of course, you're right. It does work well. it just isn't strong enough for mistakes. I forgot to check the alignment of the Z axis after the wreck, and ended up fixing that yesterday. I just have to be very careful about not crashing this until I can decide about a studier machine.

Greg