[yantra3d]

What sort of runout are you guys getting with your x3 mills? Has anyone tested to see how accurate the spindle is "out of the box"?

I recently started running my mill and after some minor cutting I visually noticed excessive runout so I tested it with a dti on the ID and was shocked to see it was out by .01!

After loosening the top bearing block assembly and tapping it around a little I was able to get it down to .002 but this still seems like too much to me. I am seriously considering pulling the spindle from the mill but before I do I would like to know what sort of accuracy to expect so I don't waste time trying to fix something that may not even be broken.

I'm also suprised at how much deflection there is in the spindle, head, and column. Just a slight push on any of these three will have a major effect on the indicator (~.008). I guess I expected more rigidity from this little mill. :|

[yantra3d]

Well I guess I'm off to pull the spindle... .002 is just too inaccurate to live with and I can only assume that the runout should be less. It would be great to know a target figure that I could strive for but I guess I'll have to go it alone.

I also wanted to mention that there is a lot of play between the spindle and the quill housing that can be detected by pushing back and forth on the top (spline) and bottom of the spindle. And yes, the quill is fully locked.

Does anyone else have this slop in the spindle? Is this to be expected on the x3?

[ihavenofish]

i dont have an x3, but i have a kx1 cnc made by sieg. my spindle runout on the taper is a bit under .001"

i dont know if .002" is bad for a stock x3, but from what ive seen in peoples conversions, its usually the roller bearings that cause much of the innacuracy. my mill for some reason has angular contact bearings instead.

[yantra3d]

Hey thanks for the response... I appreciate that!

I pulled the spindle, mounted it to my table and setup indicators on both the inner taper and outer housing so that I could adjust the preload on the top bearings. I noticed that the ground washer between the c-nut and the bearing was no where near flat so I carefully lapped it on my surface plate and then lightly lapped the c-nuts before attempting to adjust the preload.

I was able to get the runout down to within .0005 with a finger tightening of the nuts. I then mounted the spindle back in the head and tested the runout and it was down to within .0003". The play or slop I was experiencing has also disappeared so it looks like we're back in business.

He whole endeavor took less than 2 hours to complete. Needless to say I am now satisfied.

[ihavenofish]

Hey thanks for the response... I appreciate that!

I pulled the spindle, mounted it to my table and setup indicators on both the inner taper and outer housing so that I could adjust the preload on the top bearings. I noticed that the ground washer between the c-nut and the bearing was no where near flat so I carefully lapped it on my surface plate and then lightly lapped the c-nuts before attempting to adjust the preload.

I was able to get the runout down to within .0005 with a finger tightening of the nuts. I then mounted the spindle back in the head and tested the runout and it was down to within .0003". The play or slop I was experiencing has also disappeared so it looks like we're back in business.

He whole endeavor took less than 2 hours to complete. Needless to say I am now satisfied.

wow. thats not bad at all.

[yantra3d]

I say .0003" but it's more of a guess since my dti only divides by .0005 and it's bouncing back and forth between one division. It's probably safer to say that it's less than .0004" but greater than .0003".

Regardless I'm happy to have that problem corrected... now I can move ahead with my build. These delays are killer but I keep repeating the buddha's words like a mantra on this project, "Ambition is like love, impatient both of delays and rivals". (If only it helped!)

[neilw20]

I compensate for the tower bending.
1.2KG is about 3 microns.
http://www.cnczone.com/forums/showthread.php?t=94115

[yantra3d]

Hey, I was having that same problem where the return pass (left to right) was off from the full pass (right to left). I can't fully visualize your description but it sounds ingenious. What a brilliant idea... And it works no less!

I have to say that I was surprised to see the column flex under so little force. But I have no other machine experience to compare the x3 rigidity to so I guess I am learning it's limitations as I come across them.

I'm really glad that I planned to replace the stock motor on this x3 since the added weight of the motor mounted on the side of the head is probably more than enough to have a considerable effect on cutting accuracy. I also noticed that the vibration from the running motor is being transferred down through the head and is measurable by a dti on the spindle! Not to mention the killer noise that motor radiates - along w/ the noise the gearing is putting out. A recipe for insanity! After several hours in my little shop I can still hear it running in my head. If I didn't need a chip tray so bad I would make the AC spindle motor a top priority. Be happy you got the sx3!

[ihavenofish]

Hey, I was having that same problem where the return pass (left to right) was off from the full pass (right to left). I can't fully visualize your description but it sounds ingenious. What a brilliant idea... And it works no less!

I have to say that I was surprised to see the column flex under so little force. But I have no other machine experience to compare the x3 rigidity to so I guess I am learning it's limitations as I come across them.

I'm really glad that I planned to replace the stock motor on this x3 since the added weight of the motor mounted on the side of the head is probably more than enough to have a considerable effect on cutting accuracy. I also noticed that the vibration from the running motor is being transferred down through the head and is measurable by a dti on the spindle! Not to mention the killer noise that motor radiates - along w/ the noise the gearing is putting out. A recipe for insanity! After several hours in my little shop I can still hear it running in my head. If I didn't need a chip tray so bad I would make the AC spindle motor a top priority. Be happy you got the sx3!

all machines of this size are going to flex in the column. the x3's "fault" seems to mostly be how the column is mounted to the base. its prone to twisting at the bottom, rather than bending along its length.

youll have play in the spindle bearings, play in the gibs, play in the lead screws nuts and their support bearings. you also have the moment arms from having the z axis screw behind the column which will make the head bob up and down when you switch directions. how much depends on how tight the gibs are and how little the side plates goung around the column flex.

the thing neils talking about is something that will be in nearly every meachine - even bigger ones. hes less concened with absolute precision but rather repeatable passes for good surface finish as i understand. his way of compensating seems to be pretty clever, especially if you have a long production run that has predictable error on each pass.

id not let this put you off too much. understanding and dealing with flex is part of the job. the trick is to eliminate any source of random movement.

[neilw20]

Previously, I had observed similar Z problems (like tower bending and head bobbing) on machines even weighing in a 10 tons.
I just hadn't figured out how to fix it until I had my wife pushing down on the head, then adding a weight and removing it at the correct moment.
I don't think it is the tower attachment on the X3. It just bends. It is only a small section. Having extra (permanent) mass on the head won't help either, but will help to absorb high frequency vibrations.
SX3 is still good value for money/performance/quality.

[ihavenofish]

I don't think it is the tower attachment on the X3. It just bends.

ive done an FEA test on the x3 and x1 columns before. the result was 80% of the deflection came from the column connection at the bottom.

but yes, at ~$1000 you cant be all that choosy. the BF20 machines have a closed columb which is better, but its smaller, so its just a different version of the same problem.

[neilw20]

I can easily believe the base attachment point is flexing.
Does that mean I can just add a big stiffening structure at the bottom attached to the main base?
Do you have any of the FEA details. That would be most helpful.
Much better than my QUICK fix which only took 4 hours total.

[ihavenofish]

I can easily believe the base attachment point is flexing.
Does that mean I can just add a big stiffening structure at the bottom attached to the main base?
Do you have any of the FEA details. That would be most helpful.
Much better than my QUICK fix which only took 4 hours total.

i dont have the files anymore sadly, but it woudnt be too hard to redo if i had the measurements again.

the thing that seemed simplest at the time to me was just putting a plate bridging the back of the column at the bottom. 6x6x0.5" would probably do it. the drawback was that on the machine i was looking at (the cnc kx3) the stepper motors are tucked in there and made it more challenging.

your little fix is probably better in some ways, because even if this plate reduced the flex by half, you still would have some left.

[yantra3d]

But still half the flex means double the rigidity. I'd be willing to try this out on my x3. I have a piece of mic6 tooling plate left over that is just that size that might do the trick. It's actually 5/8 thick.

If I could get the dimensions or file to you would that help to give an idea of where to mount the plate? Or do you remember off-hand where about it was located? Thanks for the info and tips.... very valuable.

[ihavenofish]

But still half the flex means double the rigidity. I'd be willing to try this out on my x3. I have a piece of mic6 tooling plate left over that is just that size that might do the trick. It's actually 5/8 thick.

If I could get the dimensions or file to you would that help to give an idea of where to mount the plate? Or do you remember off-hand where about it was located? Thanks for the info and tips.... very valuable.

my tests only work on a single body. the things i did were to close the back, and to add ribbs to the flanges where the bolts are. those both helped.

i think any plat bridging the back of the column with a few bolts on each side would help a bit. how much is hard to say cause the fea doesnt take into account separate pieces and bolts.

someone else ive seen on this forum basically closed the entire back of the column with a massive square tube.

note that ive never actually done these mods, only did some fea on the basic design to find out where the weak points were.

[neilw20]

But still half the flex means double the rigidity. I'd be willing to try this out on my x3. I have a piece of mic6 tooling plate left over that is just that size that might do the trick. It's actually 5/8 thick.

If I could get the dimensions or file to you would that help to give an idea of where to mount the plate? Or do you remember off-hand where about it was located? Thanks for the info and tips.... very valuable.

yantra3d..

I will live with my balance weight thanks.
I can tune every move perfectly, and I don't care if it bends now.
I always 'tune' production programs anyway, and it is all hand written stuff.

[yantra3d]

Unfortunately I've already tied up all 6 axis on my controller (running both mill and lathe using lathe spindle as closed-loop for threading) otherwise I would try the motor counter balance mod you devised. The idea that you can fine-tune it for any given offset is brilliant. Who knows, I could end up giving it a try at some point if it continues to bother me. I will likely try adding the support plate first so I'll post my results if it shows any improvement. Thanks again for sharing your x3 experience... Again, very valuable info.