[SpeedyDad]

Why is this happening? I have been milling shock towers for R/C cars for a while now. With my latest batch, the work is shifting for some reason. The actual material isn't, but rather where the machine cuts is. I do a first pass with a 1/8" cutter to create a recess for a screw head. Then, I switch to a 2mm bit to create the actual screw hole. Until recently, the screw hole ends up dead center on the recess, Now, it shifts for some reason. If I go back and put the 1/8" cutter back in and rerun the recess tool path, it comes out the same as before. What has happened here?

Here's a pic. You can see the offset screw hole in the recess.

[SpeedyDad]

So I spent a lot of "quality time" with the machine tonight doing various tests. Looks like its not the machine at all but the bits. I was getting diamond coated diamond cut 3/32" router bits from DrillBits Unlimited. They cut carbon fiber so easy and cleanly. Problem is they have been out of them for a little while and mine were starting to show signs of dulling so I got some 2mm ZrN coated diamond cut bits from Think & Tinker. I am convinced that what is happening is bit deflection. When I cut the same profile with my old 3/32 bit, all is good and centered. When I plunge with a 1/8" cutter, all is good and centered. When I cut with the 2mm bit, some are ok, but some are off center. I think that the smaller 2mm bit is deflecting and then just following its own hole and deflecting more. This was all at 10ipm and .04" doc per pass. As a last test, I tried the 2mm cutter at 10ipm but only .02" doc per pass. What do you know, it came out centered. I'd have never thought that little difference in diameter would allow that much deflection on the initial cut. I'll cut out a test part tomorrow to test this more.

[vtxstar]

When drilling even the 3/4 plywood on my Momus build I can see deflection with even larger bits using a manual mini-mill.
This was most evident when drilling on the edges of the material and the different densities of wood layers played havoc with a straight path.

It is good to hear that you have a suspect in the problem.

-=Doug

[SpeedyDad]

Well, I thought I did. As you can see from the pics below, not so. I have more experimenting to do.

The recess is milled with a 1/8" bit and the hole is plunge drilled with the same router bit. Should have worked.


Look at the holes in the left recess vs the right. The right holes seem to be shifted up.


I'm at a loss

[vtxstar]

Hmm... is it always off by the same amount and direction?

Check the set/locking screws of the pulleys. It may be tight but allowing it to shift just enough in certain positions of the travel.

You could whip up some Gcode that pokes a hole, travels in one direction, pokes another, travels back to the first, pokes it again...repeat..
You would then see which axis is slipping.

I doubt very much you have a tooth slipping on the belt/pulley.

You could remove the belts and manually move X and Y to test for binding points. It could be the motor struggles cog past one.

These are just ideas to try.

-=Doug

[stelni]

Hey Speedydad, I had a very similar problem on my Momus last week. I started to notice some small errors here and there.
It did not happen all the time, but most noticeable just after a direction change.

As Doug said, I traced it to a loose pulley on the stepper. The setscrew that was up against the flat on the motor shaft had actually stripped he hole in the cheap aluminum pulley.

I have ordered a couple of steel ones and hope to be back up and running shortly.

The way I found this problem was to power up the steppers so they were holding tight, then grab the gantry and try to force it back and forth, with a fair bit of force applied, it moved a 1/16" either way!

...Nick.

[SpeedyDad]

I received my new cutting bits and am still having the same issues. It just won't behave. I checked the machine and it is tight and smooth in all axis. A couple bearings were a hair loose but not enough to cause this much deflection. I checked all my pullies as well and they are all tight

This is my latest attempt at cutting this part. As you can see, the holes are offset in a different direction than the part above


What could this be? Ghecko? Power supply? Stepper motor? I'm at wits end.

[vtxstar]

First, I would stop cutting your good material until you find a resolution. Use some mdf or plywood until resolved.

Without digging through your old posts, what software are you using to drive the G540? What power supply voltage for the motors are you using?

Did you try some G code (I mentioned earlier) with single axis repeated movement between two points to discover which axis is the problem?

Do you remember when this happened first? Could it have been happening all the time and just lately really gotten bad?

Yeah, it sucks to have a problem like this. My first machine looses steps often. But in the is case, it would have to be loosing the same steps all the time.

-=Doug

[SpeedyDad]

I use the good material because I want to test in the exact environment it is cutting under. Forces on the machine are different between wood and carbon fiber.

I had an "aha" moment yesterday. I noticed that the offset errors were not consistent. Sometimes the holes would be off and sometimes the recesses. I also noticed that the holes and recesses on the part were better the closer they got to the back of the machine. Once it got a little past center, everything seemed to clear up while things toward the front of the machine seemed to be all over the place. I asked myself what could be causing this. I just stared at the machine and realized that the telescoping hose I have for my vacuum system goes a little slack just past centerline. Up until I changed to the new router (went from a Ridgid to a Bosch), I hadn't had any problems. At that point I realized that I had changed the routing of the hose and it was now attached directly to the Z rack of the machine instead of the carriage. What I believe was happening is that the even small leverage of the hose pulling on the Z rack was pulling it around causing the odd shifting. Because it attached to the top of the rack, more leverage and because, as the Z rack went up and down, it pulled on the hose more at the bottom of the travel than the top compounding the problems. No machine (except the high dollar commercial units) is perfectly smooth and this was just pulling against any little slop. Since this machine is also belt driven and, even with proper tension, the belts can stretch a hair when force is applied to them causing inconsistent cutting. I played with the machine to remove the little slop and ran it using just my shop vac holding the hose. The parts came out much cleaner as you can see my the pic I attached.

[vtxstar]

Indeed an AHA moment. That is good to hear.
Post a pic of the new vac setup. Your .com site seems to have only the original build pics (as shown in your build log post).

I am considering a vac setup and want some ideas as well as 'don't do this' tips too.

Congrats!

-=Doug

[SpeedyDad]

Here is my dust collector thread. It shows the v1 version. I've modified it since then but it is the same concept still

http://www.cnczone.com/forums/momus_...collector.html

[jckstrthmghty]

Since this machine is also belt driven and, even with proper tension, the belts can stretch a hair when force is applied to them causing inconsistent cutting.

Glad you discovered the source of deflection. I've been skeptical about using belts where any type of cutting forces is required. Probably great where little force and high speeds are required like a lazer.

[vtxstar]

The pics are not showing in the vac post.

This new software is odd at times...even the response post has no pics but broken link X's.
Probably not anyone's fault.

Got an external link?

-=Doug

[jim_r53]

The pics are not showing in the vac post.

This new software is odd at times...even the response post has no pics but broken link X's.
Probably not anyone's fault.

Got an external link?

-=Doug

The pics show fine for me. Maybe a browser problem? I am using Firefox.

Jim

[vtxstar]

Just loaded Chrome and the thumbnails show as pictures now and not red 'X's.
Thanks for the tip Jim.

-=Doug

[SpeedyDad]

As a final update on this thread, here's the results. After tightening up the machine and rerouting how the hose ran, the router was back to almost perfect. Still a hair off though. I studied the action of the router as it worked and noticed when the router descended to the work, the bristles on the dust skirt did not displace like the old ones did when they contacted a solid surface. They resisted moving, thus causing feed issues. I pulled the skirt off and found the bristles were stiffer than on the old skirt. I immediately cut out a new skirt base and used some of the old skirt bristle material and installed that on the machine. The result is that my machine is back to its original great results. Its amazing the number of compounded errors caused by changing the dust system when I changed routers.

[vtxstar]

Thanks for the update. I was wondering how that came out for you.
A friend of mine is hot to start cutting heli parts once I get running -- told him to not hold his breath.

-=Doug