[skillakali]

Hello, I have been reading up on the forums and couldn't find my problem so I decided to ask here.

I have just built my machine and got everything working great but for one thing. All my parts have a slight wave to them when doing curves. I have noticed a problem when I was trying to set the backlash on EMC when I saw that my scale is off. I don't know if that is the right term to use but it's what comes to mind when I'm trying to describe this.

Everything is good when I jog by .1 in. but when I jog by .01, the scale goes off by a +/- 2 degrees when I get closer to the .05 mark. Then when I jog it another .05 it returns to normal. I had tried different control software and its the same. So I decided to take the motor off and check the steps with no power.

I put a degree wheel on it with a gauge and stared turning by hand. The gauge I used is the old x axis crank wheel's gauge marked 0-100. Now the motor I have is a 200 step per rev. I started to turn/count and to my surprise when I did 100 steps (1/2 way or 180 deg) it lands on 51 and not 50 as would have been expected. I spin it another 100 steps and it lands perfect on 0. I also clocked the gauge differently but get similar yet different results, in other words it's still off by the 1/2 way point but at a different location +/- 2 deg. I have done this test numerous times and came up with the same result.

My diagnosis is that the motors teeth is off but I have not read up on steppers other than that its 2 big wheels with lots of teeth. I would think 200? Now is it possible that when manufacturing these teeth, that one set could be off a little and cause these problems? I am pulling out my noob hair trying to figure this out because I just want to cut a perfect curve with out small waves.


Thanks for any help.

[wehrmeyer]

This is an interesting problem in a technical sense. When cutting a circle consider the x axis. At the starting point the cutter is in the correct position. As the table moves to the left the cutter is further out of position ( to far from the axis of the circle) until it is 1/2 step out of position. Since steppers move 1 step per pulse when the x axis corrects it is now 1/2 step too close to the axis. As the travel condinues the cutter moves closer to the ideal position then past it until the cutter + 1/2 step. And the process repeats. So the lenght of your little waves should be 3.6 degrees if you take 100 steps to make a circle.

Since there is no perfect circles how close to the ideal do you need to be? If you change the steppers to a model with fewer degrees per step you come closer to the ideal. Or you could place your work on a rotary table, cut you circles a 1/2 step too large
index the table 1/2 the number of degrees of the little waves cut it again and repeat until the circle close enough to the ideal.

[skillakali]

Thanks for your reply.

Well, I have these motors on a lathe right now so I would like to cut a nice curve. After watching GMT's video of turning a rook and cutting the EMC's pawn program, I thought I'd try to make a rook of my own for practice.

Now I thought these motors would be accurate on this machine to at least .001 in. With the problem of the motor slows down when the motor reaches 180 deg then catches back up at 360 or vise versa depending on where I am measuring at, it shows up on my work more so on curves.

As you can see on my test part, when I cut down the Z-Axis at a fast rate, you can see the lines come close together then spread appart every .01 in. On the curves I turned the feed down for a final cut and as you can see thoses waves are getting to me. I was looking into getting a DRO with feedback for MACH 3 to solve this problem. I am not sure what will happen if and when errors will come up on that end and if it will correct it or halt the program.

I did notice an option called screw mapping? I tried to click on anything in that screen but did nothing, I guess I got to read up on that as well.

[skillakali]

Ok I think I found out the problem. After a few hours of tweaking I decided to mount the boards on something else so I can have all the electronics inside my computer and use the 12v pwr supply inside to power fans for cooling the drivers (I never got to checking if they make higher voltage fans). Also it's all in one box for space.

After that my motors were performing accurately on the gauge but it still was cutting wavy. So after being satisfied about the motors being on point, I thought to do some jog cuts and adjust the feed and depth. The waves were still there no matter what I did. So I thought how about cutting in the other direction and to my surprise it came out perfect. So I also have to get a ballscrew set up or add ball bearings to the stock screw because I think that my stock acme setup is binding in one direction at a certain point thus missing steps. This is only on the Z-Axis, my X-Axis is working perfect now.

If I run into anymore problems I will post 'em. Thanks now back to g-codin, backwards lol.

[eSilviu]

[QUOTE=skillakali;695048]I think that my stock acme setup is binding in one direction at a certain point thus missing steps. This is only on the Z-Axis, my X-Axis is working perfect now.
QUOTE]

You use acme threads? How do you support-it on fixed end?
On my first CNC iteration, I've used 2 nuts and 2 bearings to connect the acme. This proved to be a mistake, since nuts were not parallels with acme axe. So, because my whole construction wasn't rigid enough, I was getting waves on the weaker axis.

[skillakali]

I was going to use bearings on the block where the motor is bolted to but decided to go to ball screws since I found some good priced screws and nuts.

It will take about a month before I get it finished because I need to take care of other things first. I will post up some pics after its done.