Spindle project on a budget, any hope?

[funkstar]

Hi all
Yet another thread on a homemade mill spindle. But I thought that the project might be of interest for some.
I try to improve my mill spindle as the current is a bit slow, week and runs really hot. I can have about 2000 rpm with continous load without the spindle running superhot. Have not dared to test max temp but, I cant put my hand on it without buring my fingers off (that is rpm;s over 2000 with load).
So as I am anyways modernising to Mach3 i tought that I might give a new spindle a go.
The project is based on the X3 replacement spindle on Little Machine Shop. I was promised that the runout would be better than 0.01 mm, and the price almost nothing I thought I took a chance.
I did a bargain on ebay on precision spindle bearings and the only thing I had to do was to make a new shoulder for the bearings to fit on. The X3 spindle had a 30 mm and a 35 mm shoulder for the bearings, but I needed a 40 mm. The spindle had some material that I could machine away and make a new 40 mm shoulder for the bearings. This way the bearings will come closer to the tooling and thus make the spindle more rigid (right?).
To reduce the runout as much as possible on the new bearing shoulder I had it set up in the lathe with a ball bearing in the support-dock (see pic).
The new 40 mm shoulder came out some 0.005 mm too thin, that is 39.990 due to that temperature fell after I have sanded the last 0.01 of.. Grr. But I went on anyways.
I have now made the parts except for the housing. When I measured the run out I first got it within 0.004 mm. Great Success!! But I soon realised that further out on the ground axis on which I made measurements it was more like 0.02 mm.. not so good anymore. And worse, I could not repeat the 0.004 when redid the setup. Its closer to 0.02 - 0.025 mm. Crap!
Now, as a goal of the project I aimed at around 5000 rmp and perhaps higher. Do you think I might be able to remedy the problem somehow? Is here any hope or is it just the bin..

Lots of more pictures and details on my webpage:
The Spindle Project

Cheers

Per

[funkstar]

Not much response.. but anyways:
Did some more measuring on the spindle and it turns out that I when I put a good endmill in a collet and turn the spindle I get the same runout as the conical surface in the spindle which is less than 0.01, more like 0.008 or even less depending on where on the conical surface I put the dialindicators tip.
If I put the collet in a certain position I can get a runout of about 0.004mm.
The runout varies when i turn the collet but maximum is about 0.008mm
And as I measure furter away from the collet, I put a 100 mm long endmill in the spindle with a shaft of about 80 mm, I get a runout of about 0.02 to 0.03 mm which shows that the mill is quite paralell to the rotational axis.
The runout-tops co-line when the runout is at its maximum and is separated 180 when at its minimum.

Its better than before, but still I had hoped for better.

As I intend to use the Tormach tooling system I am curious on how paralell I can get the mills. As the Tormach tool-holders get their reference from the planar surface of the spindle nose rather than the spindlebore it self (if I have understood it right), perhaps this will make a better and useful mill spindle.. I need to order a TTS-collet and a toolholder to experiment..

I will continue with the build at least for a while..


P.-

[vishnu]

Hi Per,

that's good work man, with a run out of .004mm is OK. You just remained me of my yet to take off BT 30 spindle project with power draw bar. I think once i complete my plasma build i will start on the next project. A complete mini machining center for both hobby & business purpose.

[Drools]

I always find these DIY spindle posts intriguing. I wish I had a metal lathe to try making some of the spindles I have seen others making.

[funkstar]

So I finaly had the existing spindle on my machine disassembeled and measured the same way as my newly built one.
The measurements was done with a good quality 20 mm thick, 130 mm long (not including cutting part) end mill. This was set up in a cheap 20 mm R8 collet. The same set up for both spindles.
First measure is the radial runout of the millshaft close to the collet.
Old spindle +- 0.003 mm
New spindle +-0.0025 mm
Second measure is the radial runout of the millshaft some 130 mm from the collet
Old spindle +- 0.0275 mm
New spindle +- 0.0275 mm
Both spindles showed the "high values" at the same angluar position indicating that the offset is pointing in the same angular direction as the angular deviation of the mill
Third measure is the radial runout at a point in the conical surface of the spindle
Old spindle: +- 0.0035 mm
New Spindle: +- 0.002 mm
And the last measure is the radial runout of the cylindrical surface of the spindle, this was only accesible from the backside of the old spindle. The new one is not accessible from the backside.
Old spindle +- 0.007 mm

I draw the conclusion that my new spindle is about the same accuracy as the old one. Good thing. I will proceed with some more measuring before actual tearing out the spindle from its hosing (not and easy job).
I have not yet managed to get hold of a collet and TTS tool holder to make some measures with that construction but hopefully I will soon get it.
Anyways, I feel confident enough to go forth with making the housing for the new spindle.

P.-

[Herbertkabi]

Very nice work!
But these 0.002 mm ... even 0.005 mm tolerances ....
Hmm ... you are keeping this spindle by lathe three jaw chuck from another end.
I mean lathe chuck jaws (centre) runout. You can do what you do but three jaw chuck wants never keep 0.00 tolarance. OK, it acts very little yo another end because length of shaft but nevertheless. Or what?
At least with my lathe nearest precision is possible only when chuck removed (ohh dear yay boring) and using direct collet. Or often I do use my own lazy man trick - "special one slot collet" between jaws, therebiforn precisely sized for every spindle part individually (slot location need to be well marked).
Regards,
Herbert

[funkstar]

thanks
Your right about the three jaw chuck, its not the best thing. Collets it much better (though I dont have any for this lathe), and a specially made collet as you suggest is probably the best here. I should have tought about that actually.
I also regret not to have taken a radial runout test close to the chuck as I did the machining.. next time :-)
However the radial runout of the bearing I did use closest to the maching surface was way better than the chuck and that provided me at least some remedy to the "poor" chucking. But of course, with proper machines the result would have been more accurate and reliable.
And in the end, this is a "as good as I can get it"-project and I am quite happy with the result so far of the final spindle and what I can measure from it.

P.-

[funkstar]

Ok, some update in this thread.
It has taken some time but now I have completed the spindle. I had a company manufacturing the housing as it was too big of a project for my lathe. The results was good. Nice finish and accuracy to some pretty tight tolerances that I got from the bearing data sheet. Surpricing these tolerances was on the +side making the bearing seats bigger than the bearings -0 + 0.008.
Anyways I anodized the housing and the bearings just slide in with a firm push. I like stuff that is so accurate :-)
I also ordered a TTS toolholder to see how the final result was to bee.
First results was bad. +-0.07mm runout just 50 mm from the front of the spindle. But as I pointed out in an earlier post, was that the TTS holder guide on the "front surface" of the spindle axle, so I put the thing in the lathe and had it run in the bearings installed in the housing and machined the front surface there. The operation improved the spindle tenfold.
Final result from the spindle with a TTS set screw holder and a 10 mm end mill mounted 50 mm out from the front of the spindle was +-0.0035 mm radial runout.
I also checked the runout close to the front of the spindle and it too was +-0.0035 mm. Moreover it also lined up with the runout 50 mm out, indicating just a offset and no (that is "almost no" as nothing is absolutely perfect) angular misalignment.
Not perfect but I am happy and confident that it will be a good spindle.
Cant wait to get it hooked up with the mill and tried out for real. Will get back with details.
Lots of details about the build and the mill i general at my site:
TRIAC Mill Project

[funkstar]

I also have changed my strategy for lubrication. I started out with grease in mind, but I now have a plan to extend the central lubing pump for the guides to also put some oil into the top bearing pair of the spindle. This oil will then slowly work its way through the top bearings and drip down to the bottom pair of bearings, then finally drip out and onto the milling table (quantities will be so low that I can live with this, just be careful not to work the mill with best shirt on..).
Now question, will the oil for the guides be able to work with bearings around 3000 to 5000 rpm... my guess is that it will but far from optimal.
Anyone have any thoughts on this?

P.-

[handlewanker]

Hi, lubrication will make or break a spindle bearing set-up.

Why are you lubing the top set of bearings?

In a spindle design I have, and have made for a toolpost grinder which is used horizontally, and as such the front bearings being angular contact were lubed with high melting point grease with a labrynth type seal at the front.

The other drive end bearings were two sealed axially floating radial type bearings to take the belt sideways pull, and as such were sealed for life.

The spindle was designed to run at 5,000 rpm.

In the vertical mode the top bearings are also sealed radial bearings spaced apart and floating axially, with the bottom bearings being two angular contact bearings mounted back to back, and with a double seal at the bottom.

If you don't seal the bottom bearings you will wear the lube as it spins out and also breathe the mist in too.
Ian.