[tjd10684]

I do NOT trust my z axis it has terrible back lash and repeatability issues. I measured it once at 0.023 backlash but even with compensation I still cant move the head up and then back down and hit the same zero. After reading some of the other threads on here I have come to believe that the ball nut might not be tight. I did a quick check and it "seams" to be but I think I could have been more through. So I was wondering if anyone has run into this problem and if so how did you fix it. I am not above buying another ball nut to fix the problem so if anyone know what kind of ball nut I should be looking for that info would also be appreciated. Heck I am not above replacing the entire assembly (stepper, ball screw and nut) if it will help.

FYI I did replace the motor with a MUCH heavier (lists @ 42# MTR-002-3BD36) 2HP motor. That is why I would consider replacing the stepper. But again I don't know a lot about the electronics inside the cabinet if someone could suggest a stepper that might be more powerful than the one that is on there that would be a direct replacement I would gladly take that under advisement.

[Novakon]

The sources for possible Z axis backlash errors come from several possible sources. They are as follows:

1) Thrust bearing loose (most often the case)
2) Motor coupler slipping (can cause erroneous results)
3) GIBs not tight (small error)
4) Ball Nut (very small error)

If the thrust bearings have been removed, it is VERY important to replace them in the correct direction. If they are not oriented correctly, the bearing balls will slip from the chase and they will fall apart. I have seen this problem many times when assisting customers replacing the thrust bearings. If they are installed correctly, the nut on the screw may be loose. Sometimes the fit seems tight from the factory, but will slip into place as it is used. This is just a matter of tightening the nut onto the thrust bearings.

The coupler can slip, depending on the load and torque applied. Make sure it is tight.

The Z GIBs will have a small effect on the backlash (as opposed to the X/Y GIBs). Check the tightness of the GIB by using the 2 thumb screws to ensure it is snug, but not binding. Remember it is a tapper and is designed to eliminate the lateral movement, but it can be overtightened. Also make sure the locking thumb screw is securing the GIB from longitudinal movement. This will have a very heavy influence on the available torque from the stepper motor, meaning the GIBs are too tight.

The weight of the motor has little influence on the movement of the Z-axis. The gas springs will support some of the weight of the head to not cause a problem. Check the gas springs to make sure they are working. If they are not, they may also add to the drag of the Z-axis slide. These can fail if you over run the gas spring stroke. If this is the case, it is better to remove them until you can replace them...it still will work. Also, if you program the speed of the Z-azis to be lower (about 50ipm) the stepper motor is in its better torque curve. In high speeds, stepper motors drop off in torque rapidly.

In review, the ball screw is rarely the problem. check the sources of drag and loose parts. the adjustments of the slides should be done periodically and should be checked at various times based on use.

If you should need more support, please contact us at [email protected].

Regards,
Novakon Team

[t300]

can you post a video or directions on how to check these items? I would like to check them on my machine as well.

[ihavenofish]

just to add, the gib being too TIGHT can also cause backlash, because everything will compress. however, this isnt likely to be more than a few thou.

such a massive error that isnt consistent strikes me as either a loose coupling, or mach3 is missing steps.

if it turns out mach3 is dropping steps, it could be the gibs too tight / binding, bad pulse timing, bad computer, or the gecko driver.

the first group generaly make a stalling noise, youd most likely notice, especially in something like a backlash test with an indicator.

the last one -the gecko- makes no noise, but will only happen if the axis is moving very slow, under 1ipm, or with crazy slow acceleration settings. the work around it to switch mach3 to sherline mode, which changes the timing to something the gecko seems to prefer. (it has no impact on performance).

[tjd10684]

OK so does anybody have an exploded view of this assembly? I have not disassembled anything since the day I got it so everything should be in the correct orientation. I cant verify this but I don't think it was doing this when I got it so it stands to reason over the year something like the thrust bearing may have worked its way loose. I have checked the gibs several times they seam to be good I get very little lateral deflection under heavy cuts and I did attempt my z test under what I thought to be both loose and tight conditions very little effect on the results so I put it back to what I think was just a little on the tight side.

So I guess these problems should be under the z stepper. So I guess I will be removing it tonight and checking all the screws for tightness. Thank you for responding here, I did contact Khai about the z issue evidently John had been copied. That was 8 days ago I have not heard back. Now I know your busy up there I just thought I would use the resources available to me, thats all.

[ihavenofish]

if you take the cover off the z motor mount, you will see the z axis motor coupling, and the lock nuts that tighten the ball screw support bearings.

if you move the axis up and down, youll be able to see if:

- the coupling is slipping or compressing the red spacer.
- the screw is moving up and down relative to the bearings.

i think that should be the first step here.

[AiR_GuNNeR]

I just got done re-assembling the whole Z-axis during my rebuild so I am very familiar with it.
The nut is bolted to a L shaped block by four bolts. Check that he bolts are not loose. NOTHING was loc-tited on my machine. You should be able to get to the bolts no problem. The L bracket is bolted to the head by four bolts and two tapered pins.

Remove the face plate from the Stepper mount via the four philips head screws. That will allow you to access the pinch coupling. A single allan bolt can be loosened on the screw end, and four bolts holding the stepper in place to remove the stepper. Be careful that the coupling doesn't come apart if you remove it, (easy to put back in place, but you don't want it hitting the ground). the coupling has a hard plastic liner between the two ends. Make sure it's not split or broken out. You should have NO relative movement between the two ends of the coupling.

Ok, what I would do is remove the stepper, leaving both ends of the coupling in place. This gives you something to grip. Rotate the screw by hand both directions hard, but not hard enough to move the head. You might want to tighten the gib so you can't move the head this way. The point is, you should NOT be able to rotate the screw at all. If it rotates, keep moving it and look down the line to see what's loose or moving that should be.
Hope this helps!


There is a spanner nut that and a retainer underneath it that loads the two angular contact bearing. The retainer has a tab pushed into a notch on the spanner nut that keeps it from loosening.

With that much backlash, you should be able to see movement from somewhere where it doesn't belong.

[Novakon]

OK so does anybody have an exploded view of this assembly? I have not disassembled anything since the day I got it so everything should be in the correct orientation. I cant verify this but I don't think it was doing this when I got it so it stands to reason over the year something like the thrust bearing may have worked its way loose. I have checked the gibs several times they seam to be good I get very little lateral deflection under heavy cuts and I did attempt my z test under what I thought to be both loose and tight conditions very little effect on the results so I put it back to what I think was just a little on the tight side.

So I guess these problems should be under the z stepper. So I guess I will be removing it tonight and checking all the screws for tightness. Thank you for responding here, I did contact Khai about the z issue evidently John had been copied. That was 8 days ago I have not heard back. Now I know your busy up there I just thought I would use the resources available to me, thats all.

The exploded views are available and I can create one for the selected area later tonight and send it. I have talked recently (a couple of months ago maybe) to a customer about the Z-axis on their machine and have provided exploded views to him also on this same subject. I have not seen any recent message about this though. I don't know your user name so I am not sure where to send it, but if you send me an inquiry through [email protected], I will send you a breakdown of the Z-azis assembly.

Should you have any technical questions, please use the support email.

Regards,
Novakon Team

[AiR_GuNNeR]

I don't know how much it will help as there really isn't much to it, but here is a shot of the end of the Z axis screw/stepper mount area.
I turned a piece of aluminum that would fit into the coupling so I could mount my drill on the screw to run the head up and down while I worked on the reassembly.
Directly underneath the coupling is the spanner nut that screws on to the end of the screw. Underneath the nut is retaining tab. One of the tabs is bent up into a notch on the spanner nut. The bearing cap is underneath that. Take that off and you have access to the two bearings.
Hope it helps,
Eric

[tjd10684]

well thanks everyone. I tore it apart tonight and started checking the all the bolts. turns out that the retaining bolts had worked there way loose. I took them out cleaned and put some locktite on them and tightened them back down. checked backlash 0.005 and it was repeatable so I just used the backlash setting in mach. now it goes where I tell it to. The whole thing was not nearly as hard as I thought it might end up being. guess I should have taken pictures while I was in there but oh well I hope others can find this thread now and be able to fix their problem.

Thanks again everyone who helped me out.

[ihavenofish]

5 thou is still far to much backlash. ideally it should have much more than .0005"

something may still be loose. ( i doubt the ball screw nut is that worn).

[AiR_GuNNeR]

5 thou is still far to much backlash. ideally it should have much more than .0005"

something may still be loose. ( i doubt the ball screw nut is that worn).

I agree, the nut/screw combinations on mine are a real nice backlash free fit. In rebuilding mine, I ran across several screws that were only lightly tightened, and not a hint of loc-tite on anything.
How tight do the two ends of the coupling feel to each other. They should be a stiff fit. I suspect if you ran the gibs too tight the plastic spacer between them may eventually get peaned thinner and perhaps give you some backlash.
With the head tightened down, can you manually rotate the screw back and forth at all? It should be nice a and stiff.
Eric

[tjd10684]

When I had everything off I was able to move the screw and feel the backlash get taken up. at the moment I am going to finish a little production run and then while I was taking all these measurements I found another problem so it looks like its going to be best to just strip it down and check everything.

My new issue is the table is not parallel to the movement. It is 0.008 over the x travel and 0.004 over the table width in y. I have been running this thing for quite a while without checking it so I can't be sure this is not a new development although I suspect its not. So my next step is to tear the whole thing down clean everything do some checking of individual components fix where necessary reassemble and check everything is behaving like it should. Its going to be a long process but I hope to have an even better machine after its done. So it looks like there is a rebuild thread for me in the near future.

[AiR_GuNNeR]

When I had everything off I was able to move the screw and feel the backlash get taken up. at the moment I am going to finish a little production run and then while I was taking all these measurements I found another problem so it looks like its going to be best to just strip it down and check everything.

My new issue is the table is not parallel to the movement. It is 0.008 over the x travel and 0.004 over the table width in y. I have been running this thing for quite a while without checking it so I can't be sure this is not a new development although I suspect its not. So my next step is to tear the whole thing down clean everything do some checking of individual components fix where necessary reassemble and check everything is behaving like it should. Its going to be a long process but I hope to have an even better machine after its done. So it looks like there is a rebuild thread for me in the near future.

When you rotate out the backlash, if the screw moves up and down, then the problem play in the angular contact bearings in the stepper motor mount. Tightening the nut should fix that. If the screw rotates, but does not move up and down, then the play is in the ball screw nut.

Regarding the parallelism, more than likely it is caused by the saddle. If you look at my rebuild thread, I have pictures of mine with measurements on what I had to do to get the ways scrapped in parallel and on axis.

[tjd10684]

nope the screw does not move I stuck an indicator on the head before trying to move it and I could rotate the screw slightly before I got any movement of the head. It was smoother feeling than my x2 or any of the bridgeports here at work but it was undoubtedly the nut taking up the backlash.

I hope your right since its shorter it should be easier to correct but I dont have the scraping skills you have so I will most likely try to bring it to a surface grinding shop and just flake it once its ground. I dont know how I am going to deal with the inside gibs but if I can get a shop to take it I guess its not really my problem.

A question specifically to you AirGunner would you care to take a guess at the weights of the head table saddle? I am wondering if its something I can man handle or if I am going to have to get a buddy to help.

[AiR_GuNNeR]

A question specifically to you AirGunner would you care to take a guess at the weights of the head table saddle? I am wondering if its something I can man handle or if I am going to have to get a buddy to help.

The table saddle I would guess at 20-30 pounds. It is easily handled by one person. The table is a bit heavy. I can lift and place it myself with a bit of strain. I'm no body builder, (6-1, 230lb), so someone of moderate strength should be able to handle it. No problem at all with two people.

If you were referring to the head, I'd guess it to be 40-50 lbs without the motor. I took the motor out to make it easier to manipulate while scraping. Again, with the motor in place, I could lift it off the column, but it wasn't fun. You should be able to remove the nut from the head without taking off the head. Simply lower the head down. First loosen the bearing nut. Next, loosen the stepper mount screws, then remove the four bolts holding the nut block onto the head. There are two pins behind the screw. You should be able to pull the block loose from the head by letting the screw angle away from the mill. Now fully removed the loosened bearing nut and the screw can be pulled out from the bearings. DON'T LET THE BALL SCREW NUT UNSCREW FROM THE SCREW! You'll be chasing down itty bitty bearings otherwise.

By the way, it looks like these ball screws are available on eBay for around $60 free shipping. You would need to make sure you got the correct size, but here is a quick example that looks just like the ones on this mill.

http://www.ebay.com/itm/1set-SFU1605...item20c59d14a3

[Novakon]

I would be very careful about the replacement ball screw you buy on E-bay. The 1605 ball screws are rolled and of lower quality...not bad for a router or some other home project. If you want to maintain the better backlash specs you are working toward, you need to use a precision ground ball screw. The machining needed to make the ball screw fit the bearing blocks is an additional expense. I also sell the 1605's and don't recommend them because the ball nuts are different and will not fit the mounting blocks. The headache you will create you can't imagine. Been there, done that.

We have precision ground spare parts available to replace the ball screw. Take a look at the NM-145 Spare Parts available and consider this as an option. We are working toward maintaining an even larger source of original spare parts as needed for the future.

Very nice job you are doing exploring your machine.

Regards,
Novakon Team

[AiR_GuNNeR]

Good catch. I didn't look closely enough. Rolled screws have a seam like ridge between the threads. Ground screws will be smooth.

[ihavenofish]

Good catch. I didn't look closely enough. Rolled screws have a seam like ridge between the threads. Ground screws will be smooth.

no. the ridge is for a seal. some screws have, some dont. its got nothing to do with the quality or grade.

a good c5 rolled screw would be adequate, but they arent cheap. the main thing for me is the end machining. these chinese places tend to do an absolutely aweful job. not concentric at all, so the screw binds up at the ends of travel. rolled screws are harder to get right, because the outside surface of the screw may or may not have anything in common with the ball track.

almost anything youll find "better" than what novakon will provide cost a fortune unfortunately, so its easiest to just buy from them.

[tjd10684]

At the moment I am not concerned with the screw it is repeating +/-.0005 and that is good enough for anything I am doing. If it came down to it though and I NEEDED a screw I would most likely just buy a screw and nut from McMaster-Carr because I know that I would be getting a good part and as long as its less than 500 it would be one of the cheaper things I have "laying around".