[nicad]

I am in a dilemma. I run mostly "fine step" contouring in aluminum, and I'm wanting to get faster feedrates. I have run a Taig spindle (Maxnc machine) around 15k rpm with my own 2+HP DC motor, and also a KaVo 500W 40krpm spindle. The surface difference between the KaVo and the Taig/Maxnc is like night and day. The KaVo leaves a beautiful surface, but it has little to no torque for some of the more heavy duty cutting I do on the same part. Swapping spindles for every part is out of the question.

So, I am looking to build my own "fix all" spindle. I am trying to go for the smallest (desktop mill sized) spindle yet still maintain precision... which leads me to the Sherline industrial line of spindles with an ER-16 cut directly into the spindle. However, after talking to a Sherline rep, he assured me I would have problems pushing the stock spindle with stock bearings up to 30krpm. So, I am looking at getting the sherline spindle, replacing the bearings with PN 35435023 from MSC (Nachi grade ASME-7 matched set 13* contact angle, rated 35krpm). These have the same dimensions and will drop right into the Sherline spindle, and would be driven, via pwergrip GT belt drive, by my 2hp DC motor geared up aprox 4:1 to get very close to 30krpm at the spindle

My main question is, AM I CRAZY? or should this work ok? with the angular contact bearings being held about 3" apart on each end of the spindle (in "sandwich" configuration), will thermal expansion over the length of the spindle be too much to maintain a percise enough runout? (stock bearings they run 0.0003" runout @ 10krpm).

please help!
Thanks-
Colin

[HuFlungDung]

Colin,

You should be made aware that angular contact bearings require a mechanical method of applying and maintaining a preload, usually with threaded retaining nuts. Even if you buy 2 matched, preloaded pairs, each pair must have either the inner races or the outer races locked together. That is, a simple snap retaining ring "sitting in a groove nearby" is not going to do the job. Depending on how the original spindle was constructed, you might have to modify something to achieve this type of construction.

That high speed is going to generate some heat, too. The bearings really need to be lubed and cooled with an oil mist or spray system. If they run in grease, I suspect the drag will be too heavy, and the grease will be mostly thrown off, but the heat has nowhere to go. If the bearings are submerged in oil, the constant stirring of the oil will make heat, too.

If you decide to go ahead anyways, I think that you will not want to preload the bearings against one another across a 3" distance. The heat of the spindle will cause significant growth in length, and the preload will be altered. The more length between bearings, the worse this effect will be. This is why typical spindles have the bearings mounted back to back in pairs. The preload is entirely between the two back to back bearings, and not across the top end and the bottom end of the spindle. In many cases, the top bearing just acts as a steady, and floats a little bit in the housing. It is a precision fit in the bore, but as the spindle heats up, it can push this bearing upwards without fouling against a shoulder.

[c-c-cncboy]

Hey Nicad,
I have seen discussion here on CNCzone, on allowing a few tenths more endplay on the Sherline spindle when upping it to 10,000rpm because of bearing frictional heating of the spindle. Clearly, the housing is in contact with the surrounding air so it enjoys a little convenction cooling and some radiative cooling, too! The poor lil spindle gets neither. For 30,000rpm I wonder if having one bearing an easier fit in the housing, and backed with wave washer/s to allow it to move axially while never being loose, could be worth playing at? Wave washers - wavy washers - wave springs - can be selected for a range of spring constants and travels. I use English EMO brand but there are beautiful examples from Smalley and others . Regards, Terrence

[c-c-cncboy]

Ooops - that's "convection" cooling

[c-c-cncboy]

Colombo, Perske, Precise spindle websites will show how much effort/money goes into good, fast, production spindles:
www.elettromeccanicagcolombo.com
www.pbsspindles.co.uk
Then later you pay to have them rebuilt ... and rebuilt ...
Quote of the day? "Better a short life and a merry one!"
Regards, Terrence

[steveg]

nicad an interesting project! are you by chance the the same
nicad that posts on AO?

the bearings that you list are normally used placed against
each other, or separated by an inner AND outer race spacer
of the exact same height .

the manufacturer applies the correct preload during manufacturer
so the end user just has to correctly lock the outer and inner races to housing and shaft.

most machine builders lubricate small bearings of this size with
grease.(yes including at high speeds)
Kluber Isoflex NBU-15 is the prefered brand and grade of grease
for most greased spindles.

A quantity of 15-25% of the internal volume of the braring is
used (the internal volume is listed in the bearing catalogues.)

[nicad]

Steve- yep thats me. I'm forever looking for ways to manufacture same(or better) quality faster to bring my end prices down.

I found full specs on those bearings and they are only rated up to 28krpm. Other bearings with higher speed rating are getting a little too far out of my range... typically 3x the price. I couldnt imagine having to build a spindle for a "real" machine.

To get around the thermal expansion, I could "back-to-back" these two bearings on the nose end of the spindle (image here: http://www.micro-tools.com/Merchant2...ode=SHER-40230) and then a regular bearing on the top, as I believe someone mentioned. Any suggestions for use for a bearing on the top? 20mm ID, 42mm OD and 30k rpm rated. somewhere in the $80-$90 range..

OR

Go ahead with the one on top and one on bottom configuration as mentioned by c-c-cncboy with the spring washer(s) there ready to take up the thermal expansion. How much "preload" do these usually have on them? In otherwords, how many LB of spring pressure should I plan to place on the top bearing to complete the "sandwich"?
I really like this method more.. for a few reasons-

- "drop in" modification in the Sherline spindle assembly (bearing "sandwich").
- cheaper since its only 2 instead of 3 bearings.
- according to the bearing data sheet for these (http://www.nachi-fujikoshi.co.jp/eng...df/183-198.pdf part #7004CDB), when used in the duplex is when they are limited to 28krpm.. when in single they are rated to 35krpm!! I assume this is due to cumalative heat when they are next to each other??

So a cooling mechanism is called for as well... hmm. Will something as simple as pressurizing the housing between the bearings so that air escapes through them provide enough cooling? (this is the cooling method of the Kavo spindle). Or how about air blowing over cooling fins in the head/housing?

thanks so far all!
Colin

[c-c-cncboy]

nicad - Colin!
I'd humbly suggest a look at the cooler kits from ARTX at www.artxltd.com which use a vortex tube to provide below zero air from a regular compressed air supply. They are fanf**kingtastic for very little $ and you will achieve much better cooling for the same airflow. And yes I think internal housing cavity cooling is best, cuz that's as close to the source as you'll get. Smalley and others have online tables and calculators for the wave spring force. I know that several professional spindle makers use a closed loop cooling system, but if you're prepared to pay for a little air, the ARTX vortex cooler is s-o-o-o-o inexpensive. For casual use rather than 16 hours x 6 days a week, I think you're pointing in the right direction. SteveG and HuFlungDung have given you top tips on the bearings, too. Regards, Terrence

[ESjaavik]

Kavo use air to keep coolant and debris out of the bearings, not to cool them. At 50K RPM any normal packing will burn up, so there is a labyrinth seal instead. The cooling is by a mounting block with water running through it. Similar to the CPU-coolers.

Why not try it? The cost of 3 bearings is not that great. They will wear fast, but how fast? I would put in a thermistor very close to the bearing to monitor it's temperature. When it rises, you know you should back off. Also when it starts to heat up faster than it did before, you know it's time to order new bearings.

Don't use springs to set up preload. It needs to be rigid. The high precision beraings I've come across have the preloading built in. That's one reason they need to be a matched pair. You tighten both the inner and the outer races against each other.

[nicad]

Any good sources for a 3rd bearing (if I were to put these two angular back-to-back on the nose and needed a 3rd bearing in the top)? again specs need to be:

20mm ID
42mm OD
30K rpm+

somewhere in the $80- $90 range.

everything I have found that is "high precision, high speed" is an angular contact bearing... since this bearing would be "floating" (of sorts) in the top, i suppose the angular ones are out?

This would be very close to that spindle layout that I belive h3ndrix was designing in another thread.(http://www.cnczone.com/forums/showth...&threadid=3001, specifically this image: http://www.1turk.org/cnc/ss5.gif), but smaller and faster, and no ATC (yet).

Thanks-
Colin

[HuFlungDung]

Yes, I agree with what Esjaavik said. Although you might maintain a proper preload on a simple rotating shaft with the use of a spring washer, remember that a spindle has real mechanical force applied when it is drilling, or other wise reacting to the cutting forces. You definitely do not want any float whatsoever in the position of the tool tip.

[c-c-cncboy]

Nicad, it seems my inputs on wave springs and aircool have been read out of context. Hu and Es are perfectly correct in their design advice. From Hu, I endorse the concept of a preloaded bearing pair at the bottom of the housing. Similarly from Es, I acknowledge that Kavo just uses an air purge to try and exclude detritus.
My suggestion for WAVE SPRINGS was to control the axial float of a top bearing, not to provide preload to the bottom pair. The bearing manufacturers have preload ideas under control for you.
My suggestion for VORTEX TUBE cooling, is based on very happy experience. Clearly it has the benefit of excluding detritus, but for middle-duty spindle design it reduces the cost substantially.
There's not a lot out there commercially, between mean little engraving spindles and the big expensive professional spindles (names mentioned already), that's why we're all interested. The vortex tube air cooler offloads the cooling investment to your air compressor, instead of requiring you to pay for an expensive closed circuit cooler (or trickle mainswater to drain - SMACK!) - you already own the compressor, so for casual spindle use it just works a little harder.
18 of my vortex-cooled spindles are in regular use out there, in rapid prototyping mills. They are lasting around 1000 hours at 20K to 40K rpm, between quite cheap trips to the bearing hospital. The coolers last forever. My first spindle really sucked and I redesigned totally. The last ten are sweet, with a low-air-pressure interlock to E-stop and an overtemp sensor that triggers a warning horn but takes no other action. Regards, Terrence

[Scott_bob]

This is what I like about the zone, brain storming! Experimenting! Let's take it apart, and make it go faster! Pretty cool.

I just wanted to suggest a little perspective though.
How much faster can you feed with 3-4x the spindle speed with your existing CNC control?

We also tried higher spindle speeds (30k) but found poor life, and frankly poor results due to the inaccuracy of the machine control not the spindle. Even at 10k RPM we got poor performance. So there was no point in increasing the speed of the cutter if the machine could not take advantage of it.

Mold and surface machining require smooth imperfection free motion 1st. Then "if you have that", increasing your speed by 3-4x will give you feeds that can take advantage of this.

What kind of CNC control do you have?

[ESjaavik]

Use 2 angular (15 degrees) bearings at the bottom and a regular bearing at the top. It can probably be smaller as it carries less load than the bottom bearings. X-mounted preloaded angular bearings can actually run at higher speeds than non-preloaded.
And try to avoid plunge cuts. Enter the cut at around 20 degrees. This because you don't have any thrust bearings. They will not like the speed.

[c-c-cncboy]

It's clear that Esjaavik has done this work before! Spiral lead-in or ramp lead-in will help the tool, the bearings and (why are we here?) the cut finish. Many low-cost CAM packages don't offer much choice on lead-in style. Even if you need to go back and alter the code manually it's essential. Regards, Terrence

[steveg]

back from 3-12hrs :tired:
Colin one of the things touched on by ESjaavik is the
need for a bearing seal. the bearing that you have chosen
have no sealing at all!
spindles normally use a labyrinth (non-contact) seal to keep all the nasties out.
Sherline uses dust-shielded bearing, and a thin steel plate to do the same.
This makes the conversion more than a simple drop in.


One quibble with ES however , angular contact bearing ARE
thrust bearings! (chair) so the consern about plunge
cutting is maybe overstated.

Scott_Bob raises an interesting question about your control
and machine. One aspect of "high speed" and/or hard machining is to
run the machine at high ipm but with a tiny chip load, in this case
you get the metal removal rate
you are looking for as well as the speed and surface finish
you need.(edit: the point of this comment is of course
that if you could improve your control/machine the
kavo spindle might do the job you want)

About lubrication, many new machines lubricate their
spindle bearing with a mist lubrication system , really just
an elaborate air lubricator. this has the added benefits of
both cooling the bearings and placing the spindle under
positive pressure, keeping everything out.
The down side of mist is that you have to remember to add oil.
I just recently removed and reinstalled, after being
sent away for rebuild, a BT40 12000 rpm spindle that failed
because the air mist feed line broke! (in fairness the
spindle allready had about 24000hrs on it and had been sounding
a bit tired for the last year)
Grease on the other hand is for life and requires no attention.

One last thing Colin, check with a few local bearing houses
I've always felt that those large catalogs were Way overpriced.

[IJ.]

I've looked at a "Richmill" high speed spindle and they are using an air line for positive pressure inside to keep dust out of the bearings my guess is it provides some cooling as well as this has to be a factor at 30000+ rpm (it's good to 50000).

[c-c-cncboy]

Found this in "Manufacturing Talk" today: http://www.manufacturingtalk.com/news/fid/fid111.html
It's a nice reminder that reducing spindle bearing wear by cooling them with water sometimes can be a problem in itself. I still smile each time I walk by one of my chilled air cooled lil beauties. BTW bearing makers distinguish between "seals" and "shields" and I chose metal "shields". Regards, Terrence

[steveg]

usefull design guide http://www.ca.nsk.com/webadmin/resou...oaded/c859.pdf

[IndHobby]

Sorry wrong thread