Problem- Milwaukee Router has excessive axial play in the spindle

[millnturn]

I have a large CNC knee mill in my home shop. Recently, I made a mount to attach a Milwaukee 2.25 HP router to the Z-axis. It is a model 5616-20.

I was very happy with the router when I first bought it, as the runout and endplay was very good. I have put very little time on it, and all of it has been light cuts in wood and acrylic. The router is still essentially new. I just noticed it has developed .005" of axial (up and down) end play in the spindle. Radial is still tight.

I am considering replacing the bearings in the router with something of a higher quality. I did a little searching, and did come across this: http://www.vxb.com/page/bearings/PROD/Kit7550 It is a ceramic hybrid design. I am a bit confused by bearing specs, and wonder, would an ABEC 3 or 7 bearing be significantly tighter in terms of axial and radial play?

Has anyone else experienced this, and if so, had any luck improving it? I work mostly in metal, so this router is an accessory. I would love to have a nice, water cooled spindle running on a VFD, but it just isn't in the budget right now.

If anyone could share some insight on this, I would be very grateful.


Jon

[Fatboy]

I have the same MW router on the Joe's 2006 that I built in Nov. Can't say that I have any run out in the shaft but maybe I'll check it now.
I got an email today from vxb that a $20 order will get you a free $8 caliper.
My MW router is only about 6 months old but I did use it on my old machine to cut all the 2006 parts (that's how many it felt like). Also, I've used it to cut a few designs in aluminum with satisfactory results.
If you decide to change out the bearing please update this post.

Thanks,
Fatboy

[millnturn]

I would be very curious to hear what you find in terms of runout in your spindle.

I posed this same question to several senior mechanical engineers where I work, and their answer was unanimous; better bearings will likely yield a similar result. It is a fundamental design flaw. To get near zero runout, you need to use pre-loaded, angular contact bearings. This would of course, require changing the housing and rotor to accommodate a setup like this.

I guess I'll just have to start saving for that water cooled spindle I have been lusting after!

Jon

[wildwestpat]

Hi Jon

Since the radial run out is still ok but the axial (up and down) play has increased this may indicate that it is the plunge feed that has caused a problem. If the bearings still exhibit little or no movement when you tug in the X and Y on the spindle then the problem is in the way the bearings are mounted. Check for broken spring washers used to apply a pre load to the bearings and that the end castings supporting the bearings have not moved. I don't know how this particular machine is constructed but heavy spring or wavy shim washers are common. Also long hours of machining may have caused movement in the moror laminations as the varnish may have gone soft with the heat build up and alter the clamping force holding the parts together.

I have had problems with plunge cutting due to using router cutters that did not respond well to starting with a rapid plunge. I modified the CNC to cut in with a slow Z feed making a further pass to clear out to the required depth. I now use end mills.


Hope this helps

Regards

Pat

[millnturn]

Hi Pat,
Solid carbide, two flute end cutting end mills, and vee bits are the only thing I run. This thing has about 45 minutes of run time on it. I did measure radial runout when I first got the router, and found it to be very good. My guess is the axial runout was always there, and I only now noticed it.

I think I will just live with it until I can upgrade to a better spindle. After all, sign carving, and making plastic parts were my motivation for mounting this router to my mill in the first place. Neither of which calls for extreme precision. If I do need to be really accurate, I do have the mill spindle 5 inches away!

Thanks for the replies.

Jon

[wildwestpat]

Hi Jon

Cutting tools are obviously not the problem. The end float might have been there all along as you now suspect but I would monitor it from time to time to see if it is getting worse. I would still suspect the bearing mounting and or spring/shim arrangements. However too much axial pre load on the bearings is not good given the high spindle speed so it is just possible that the end float you are detecting is taken up when the spindle and motor frame are hot after a bit of heavy use.

Hope you continue to make good parts with the router as that support arm looks very good.

[Fatboy]

Ok, I just checked mine and yes there is a small amount of axial play but no detectable radial movement.
Couldn't say that this is a design issue or not. After all it wasn't designed to be mounted and operated as a CNC spindle. For use as a handheld unit it may be totally acceptable.

Fatboy

[FandZ]

Ah crap, after reading this thread I checked my Hitachi router. I've checked it recently and it was solid. Now I can feel a little up and down movement on the bit. I can hear it too. Just a few hours ago I accidentally ran it through .01 inch deep cut across an aluminum extrusion at 70ipm by accident. Makes me wonder if that caused it.

[Fatboy]

FandZ,
I have a feeling that this end play is there from the factory. Most all of the electric motors that I've ever taken apart had shim washers on the shaft to control end play. This part of the assembly proccess probably has a range of acceptable end play so some units may have more than others yet still pass QC.

Pre-loading the bearings would require a higher quality bearing and more stringent manufacturing/assembly process. This would increase the cost.
Keep in mind that the routers weren't designed for machine use but are accurate enough for hand routing.

When I need to replace the brushes in mine I think that I'll check to see if the end play can be shimed.

We should start a registry for people to post the end play and shim count/thickness on different routers.

Fatboy

[wildwestpat]

Hi folks

The shim pack will vary from router to router within the same series due to manufacturing tolerances. The motor field winding stack being possibly the major contributor to the variation.

Since routers use a universal comutator type motor the stator windings (rotating part) will heat up the shaft and the differential expansion between the frame and the shaft has to be accomodated. Thus applying any appreciable preload to the bearings when cold might if the shaft expands more in length than the frame cause failure of the bearings or much shortened life due to the increased load due to thermal expansion.

Any wavy shims / spring washers fitted also help compensate for expansion.

I have checked the four routers I own and find that they all have just perceptable end float that clocks in at less than 0.001 inch with the exception of a real cheap one (£25 with a free set of 5 bits from the local DIY emporium!) which clocks at 0.012 inch cold and rather less than 0.003 when hot.

If you use a flute router bit so that it drags the spindle down as it cuts this might be a solution on the rare ocassions where more depth of cut control is required.


Hope this helps.

Regards

Pat

[FandZ]

Originally Posted by Fatboy FandZ, I have a feeling that this end play is there from the factory. Most all of the electric motors that I've ever taken apart had shim washers on the shaft to control end play. This part of the assembly proccess probably has a range of acceptable end play so some units may have more than others yet still pass QC. Pre-loading the bearings would require a higher quality bearing and more stringent manufacturing/assembly process. This would increase the cost. Keep in mind that the routers weren't designed for machine use but are accurate enough for hand routing. When I need to replace the brushes in mine I think that I'll check to see if the end play can be shimed. We should start a registry for people to post the end play and shim count/thickness on different routers. Fatboy

You may be right Fatboy, but when I last checked I could have sworn there was no end play. I'd say the total up and down travel is about .005.

[FandZ]

Well after reading Pat's post, maybe I don't have anything to worry about. I did just cut out a part it and it came out perfect.