Spindle Runout bigger than expected

One of main reasons I got my Carvey was to mill PCB’s, although from the day one I had problem with getting thin cuts and they were wider than what they should be.

After messing around with it I found that my spindle has a significant amount of runout, I am not sure if that would be everyone’s spindle, or just mine particularly, but all I know is, that it is keeping me from milling precisely using precision end mills, I went ahead and spent $150 on a set of t-tech end mills which are supposed to allow me milling beautiful clean tracks, but with this much runout I will break the 0.004 tool instantaneously.
I attempted to mill 0.5mm pitch using the V carving bits, but they only wobble in circles what causes the entire cut being 5 times wider than expected (roughly 5 times).

I ran the spindle at low RPM and used my microscope to look at how the spindle behaves when spinning and I came up with this short clip…

I am sorry for how shaky it came out, but I had to hold the microscope in hand.

I also took a picture of the tool’s tip as well as how the tracks are coming out, I know they aren’t deep enough to cut the copper, but driving it deeper would only bend or break the tip off.

Does anyone have a suggestion to how get rid of the runout?
Would retightening the set screw help? (One that holds spindle on to the motor shaft)

1 Like

so first thing I would do is get a dial indicator on there and see exactly what the runout is

with no collet or nut
with collet and nut but at the tool

probably a tenths indicator and then I would call inventables and see if you can get a spec sheet on the spindle

and find out what the runout tolerance is from the factory

it just might be that the spindle is not precision enough for the work that you want to do and you might have to buy a new different spindle that is more precision

also if your running .004" tools dont you need to run them at 60,000rpm and above to get any speed out of them?

Contact us at help at inventables.com when we measured the repeatability it was .0005"

1 Like

would you happen to have the spindle runout spec though?

Contact us at help at inventables.com when we measured the repeatability it was .0005"

1 Like

okay message sent thanks

Here’s a little clip of me checking the runout on the spindle and the tool, I think that I could adjust the tool’s runout a little to get it better, but with this much runout on the spindle alone, I doubt I would get it centered enough.

I made this little rig to hold the dial indicator:

and the indicator I bought on amazon for under $40 :

It’s unfortunate that my spindle was off by so much, but it’s the spindle manufacturer’s fault. I can’t wait to get it replaced!

yeah it looks like your .0035" runout on the shaft and .0075" runout on the tool

I emailed tech support and they are not gotten back to me yet on the manufactures spec

but I would defentily want to see those

but it looks like the collet and collet nut are making things alot worse you may have to get a more precision grade collet and nut

for running a .004" endmill I would image that you would need a runout less than .0005" for even less

It looks like the units are .000X on your dial indicator. Can you confirm that?

If so the measurements would suggest the run out is .0005".

That would be in tolerance. That’s 1/2 a thousandth of an inch or 1/6 the thickness of a piece of paper.

If I’m wrong please contact our Customer Success department and we can get it switched out and sent back to the manufacturer.

I have contacted the Customer Success department and it was a pleasure dealing with them. A replacement is on it’s way.

My dial indicator has the same scale as yours, every tick is half a thou or, 1/2 a thousandth of an inch, five ten thousandths of an inch or 0.0005

Just like the one you used in your YouTube clip, except that mine jumps about 8 ticks (0.0040") as opposed to yours which runs out 1 tick (0.0005")


You are absolutely right, my biggest worry is the shaft’s runout, I might be able to tap the tool straight and tighten in the collet, but if the whole thing wobbles, it won’t be possible to get the tool straight (I guess).

Inventables has posted a video 3 years ago of how they test their repeatability and they got it down to half a thou, which is definitely to my likings:

but it is not what mine looks like.

Here is another clip of what this spindle looks like taken apart, I am not sure if this is exact one they equip Carvey’s with, but if it is, then I found a bunch of these on internet and their specs say that the runout should be between 0.01mm - 0.03mm (note that these are in mm) 0.0004" - 0.001" , but I haven’t found out if this is exactly what they use.
They just look identical.

1 Like


ahh okay yeah i have not seen that video and I am glad that they are replacing your spindle

personally I think runout needs to be checked it quite a few places even on the shank and cutting edge of the tool not just on the shaft of the spindle as the collet will introduce runout if not correct

also how are your pre-loading the collet nut? are you torquing it down to a specified value? and are you seated correctly in the collet and or using a collet plug?

well yeah thats basically what I am saying i believe just in a longer version right?

do think that the spindle on the carvey is capable of .0005" or less runout all the way down to the tool tip?

with the factory collet and collet nut?

oh okay I gotcha and yes thats true

I think thats why when using small tools like that the run-out will have to be checked each time and the collet nut torque needs to be monitored

I guess its a little easier on atc machines as you do all your tool setting on the bench and have precision iso tapers in the spindle of course I image it might have to be checked on that each time as well

hah! well by far I was only inserting the tool and tightening using included wrenches to what “felt right”, not too tight to strip the thread nor too loose to let it unscrew, but I had it once let loose, so I go a little harder on it now.
I don’t even know of any tool that would be able to torque the collet. Maybe a torque wrench with a deep socket would work just fine, but torquing it never crossed my mind. I will take that into consideration.

I totally agree with @AngusMcleod , that is why I definitely want a “somewhat” straight spindle before I start messing with straightening out tool inside the collet. I need the good start before I mess around with other runout.

I know I am posting a lot of YouTube clips, but this one has taught me something too. It is pretty much what @AngusMcleod is saying

I know it is definitely not comparing apples to apples, but there are just as many points of possible runout.

What is a collet plug? is it the piece under the nut?

A world of difference with a new spindle!
This one is milled using cheap V bit without auto leveling the PCB, but it makes me very happy.
My T-Tech bits arrived today so I will try one of them out carefully

This is 0.5mm pitch so the machine has to be very precise and I need to figure out how to use auto leveling on UGS as I can see a lot of variation over the span of board.