Troubleshooting help

I've got a Sainsmart Genmitsu that I extended the Y rails on using Openbuilds parts. 1m C-beams and 8mm leadscrews to go with them, including ant-backlash nut blocks.

I calibrated them to where they seemed dead on for the Y axis.

However when I recently tried boring a cylinder pocket into maple to put a threaded insert in, I see that I'm an oval. (see picture)

It was supposed to be a .395" diameter, and it varies from .385" to .405" (roughly, a little hard to get good numbers using a digital caliper, the lockscrew on it kept being in the way).

I thought it might be the CAM settings I was using to generate the toolpath (I was using ramping options in Carveco Maker), or possibly the feed being excessive. But after generating the toolpath with different ramping, and no ramping and about 1/3 the feed, and 1/2 the stepdown, I'm still getting a consistently larger X dimension than the Y dimension.

I checked the calibration for the stock X axis, and seems to be almost dead on over about 380mm. Less than ±0.2mm over that distance I'm sure.

So, while I'm sure of the effect, I'm not sure of the cause. Would it make sense to switch from the 10mm 5 lead acme leadscrew that is the default on the X to the 8mm acme screw from Openbuilds (along with gantry plate, anti-backlash nut block, etc..) to match my Y axis?

I'm fairly confident given the lower feed rates that it isn't due to excessive load for the router/machine.

(Speaking of which, I'm down until Friday due to the brushes wearing out on my Dewalt 611 after only about 20 hours)

I did try reducing the acceleration to 180 mm/sec2 from 300 to see if that would make a difference. It didn't appear to and I was about to put it back when the router failed to start.

 

Thank you. Given its behavior, I think your suggestion of loose coupler between a motor and one (or both) of the Y axis leadscrews is the most likely cause. I'll check it out as soon as I get home tonight.

 

Found it. Lock collars on the ends of the Y-axis leadscrews were spaced a bit far out. Enough you could move the gantry just a bit before one or the other lock collars would stop linear movement of the leadscrew.

To properly secure them, I need to use a clamp to hold the one lock collar against the bushing while I hold the other one in position and tighten the set screw. I've gotten 1 done so far and it looks good. Once I get the other one done, I'll test cut the bore again and see if it measures good.

 

Cut done. Hole is round.

Whoo-hoo! Time to get to work.

 

Ok, I'm getting out of round (slightly about ±0.010" variation) again. And my location of cut is off, and the size of my circles are off a bit too (about 0.020" undersize for a boss, haven't done a hole yet). This is not the same problem as before. No movement on the Y-axis leadscrews without power, and the lock-collars are where they should be with no play.

I've noticed that both of the retention nuts for the anti-backlash set-screws have fallen off. I'm going to be taking off the spoilboard tomorrow to reset the set-screws and put the retention nuts back on. Maybe a bit of blue Loctite to help there.

I've also notice that the X-axis gantry can be moved slightly. Pivoting around the gantry base-plates that are on either side of the leadscrew. I don't know if this is due to deterioration of the roller wheels, or something else. Figure about 80-100 hours running this machine.

I've got 2 weeks to finish the project I'm working on, and I'm very frustrated with this happening now. I'd buy a whole new (better) machine if I had the time, but I don't.

Any suggestions for a quick fix, other than what I'm planning to do with the anti-backlash nut blocks?

Yesterday, and earlier today I did some engraving with a 30° bit, and it looked nice an clean. No sign of locations not repeating, or the like. So, this might be due to only the retention nuts falling off and getting backlash. I'm hoping. It is still going to take me a few hours to get it off, fixed, and back on and my fixtures indicated in again.

Thanks for putting up with my frustrated rant as well.

 

Since this based off of a Sainsmart machine, it still has the original plate on one side of the gantry base, and it isn't adjustable that I can see.

Thinking a bit more on it, given the change is sudden, and I found the retention nuts off right after (I vacuum the c-beam channel regularly), I'm pretty sure that is my cause. And if the gantry was tipping, it should show up like a tramming issue, and I'm not seeing lines at step-down levels, so I don't this it is a current problem.

However when I do have time, I'm thinking seriously of changing over to ball screws and swapping out the rollers for precision rails.

 

Ok, retention nuts back on the setscrews on the anti-backlash nut blocks. Setscrews adjusted to give some resistance on the leadscrew, but not enough to present a motor problem. Retention nuts loctited and snugged up enough they shouldn't move on their own.

Re-checked the calibration on the Y and X axes, including moving back and forth, large and short amounts. Everything is as dead-on as I can tell by eye and mm scale.

Re-ran the program that was off. The boss is now centered on the part, which it wasn't before.

YAAYY!!

Small spoiler in the mix, the boss is slightly undersize. Given I've got about 1 - 1/4" stick-out on the end mill, and it is an non-precision collet on a router, I expect it is due to runout. I don't currently own a stand to be able to easily check runout, so I'll comp with program adjustments after trying to rotate the collet around the spindle part way, and cut a circle pocket to see if that gets me closer to programmed size.

However once this project is done and I've got some time, I am seriously going to be working on an upgrade, or maybe a new machine. I'm tired of fighting this one so much.

I'm limited to about 30" x 48" workspace though.

 

Just to follow up, and for your amusement. Actual spindle runout is 0.015".

I'm picking up a new router today since I've got time-pressure limits. And when I do have the time, I'm thinking of switching to a 2.2KW water-cooled spindle for a proper replacement.

 

Oh absolutely. I was getting a couple of suggestions for getting the 2.2, but since I'm mainly going to be working with wood, and maybe some very light soft metal machining, a 1.5 may be the right choice. I'll be doing more research before actually pulling the trigger.

The machine I have is 80x40 C-beam for most of its construction, so it is relatively rigid.

 

I know it has the ability to do so, but as evidenced by my experience, even with only about 60 hours of exclusive use on wood, I've got a Dewalt DWP611 router that went from OK to 0.015" runout. I want a spindle that is designed for long term usage that will have minimal runout and good bearings. It looks like I'm going to be using this machine a lot, and I don't want to have to spend anywhere near the time I've been spending on diagnosing and fixing problems.

To be clear, I know whatever I get will need regular maintenance to preform well, and I'm happy to do that.

Also, for additional information, I picked up a replacement DWP611 yesterday since I have only 9 days to finish a big job, and the new one cuts on size with the same Harvey Tool end mill I've been using up to this point.

And I do really appreciate the support I've been getting here!