Okay....soooo

[gregrnel]

now that I'm back up and running and back at square one, I've been just trying to get a perfect circle on my machine. Background.....I'm going from Rhino 3.0 to Sheetcam 4.0 into Mach 3 machineware with a DynaCNC 4 x 4 table. I think I've tried every export option (dxf) in Rhino and still get a circle that is perfectly round. There is always a notch or flat spot somewhere on the circle.....I'm , any ideas from any of you pros?

[PowerCNC4u]

There can only be one of a couple of things that will make it so you can not get round circles.
1) Backlash can be set to high or too low. Mach3 has backlash settings you can adjust easily. Try adjustments with a pen instead of wasting money on material.
2) You may have CV (ON) at the same time as the backlash enabled. Use one or the other but not both at the same time.
3) Something is lose. A lose key on a pinion gear, a lose pinion gear to the rack gear, a lose belt, anything lose can do exactly what you are explaining.

I had the same problem on my DynaCNC table and a couple of adjustments and tightening up of a couple of items and I get perfect circles every time. Check all of your alignments, leveling, and gears. If everything is tight and true, then try adjusting the backlash. Any mechanical machine has thing come lose and needs adjusting once in a while.

[gregrnel]

Well, tried everything suggested, all the belts and gears seem fine, tried cv off with various levels of backlash but still have a noticable imperfection on the circle on the same spot everytime. Anything else this could be? Thanks!

[gregrnel]

Here's an image, the same defect happens in every circle in the same spot, regardless of size.

[ImanCarrot]

Just a thought:

You have definately programmed a true circle haven't you? I know it sounds daft to ask, but if you'd programmed two semi- circles with different tool compensation (right and left) it might do this effect.

What happens when you try, say, 3 circles, 1 above and two below... do you get the same effect? if so that would rule out backlash and motor problems on the X axis (assuming your X axis is left to right on the picture).

Hope this helps... just throwing a few ideas at the problem to see what sticks.

[InspirationTool]

I'd take the programming out of the equation...

Get out a dial indicator, connect the base to the table and the tip to the headstock/router (or vise versa) and jog the machine in .001" increments..... If it stops moving for a number of jogs when you reverse directions, you have backlash. Count the number of jogs it takes, and you have how MUCH backlash.

-Jeff

[ger21]

Post your g-code. But, it looks like a lot of backlash in one axis. Backlash Comp can't do miracles. If it is a lot of backlash, you need to fix it. The circle does appear round in Mach3, correct?

[PowerCNC4u]

I own a DynaCNC table and as stated before.

Hands down you have a lose X axis....period!
Something is lose on the "X" axis.
Either a pinion gear, rack gear has moved, the key is missing or lose, the timing gear is lose, the belt is lose, the key in the timing gear is lose. Something is lose....period! end of story.

Tighten down your machine and it will run round circles. I told you that before in my post and you keep stating that you have checked everything. I am telling you.... something is lose on the X axis. The "Y" axis looks fine.

Case in point: I HAVE SEEN THIS BEFORE ON MY TABLE AND FIXED IT IN 5 MINUTES! It has nothing to do with G-Code or Electronics. Machines work their way lose from vibration, belts get lose from wear.

That is your solution! There are no others. Everything else you try to correct from this point on to fix that problem will be a waste of everyones time. I know from experience!

and one more point: I have nothing but GREAT things to say about my table and my experience with DynaCNC. I have read your posts hacking them and I can plainly see what your problem is; and it isn't DynaCNC.

Sid

[gregrnel]

Thanks for the advice, I'll go through X again, however, I am not the lone ranger around here when it comes to hacking DynaCNC. This has been the least of my problems with their customer service and product. I'm far from being an expert, but when you pay thousands of dollars for a machine , it shouldn't be much to expect it to work properly. Hopefully once I get all the axes up again, Z won't inexplicably lock up like it has so many times in the past.

I own a DynaCNC table and as stated before.

Hands down you have a lose X axis....period!
Something is lose on the "X" axis.
Either a pinion gear, rack gear has moved, the key is missing or lose, the timing gear is lose, the belt is lose, the key in the timing gear is lose. Something is lose....period! end of story.

Tighten down your machine and it will run round circles. I told you that before in my post and you keep stating that you have checked everything. I am telling you.... something is lose on the X axis. The "Y" axis looks fine.

Case in point: I HAVE SEEN THIS BEFORE ON MY TABLE AND FIXED IT IN 5 MINUTES! It has nothing to do with G-Code or Electronics. Machines work their way lose from vibration, belts get lose from wear.

That is your solution! There are no others. Everything else you try to correct from this point on to fix that problem will be a waste of everyones time. I know from experience!

and one more point: I have nothing but GREAT things to say about my table and my experience with DynaCNC. I have read your posts hacking them and I can plainly see what your problem is; and it isn't DynaCNC.

Sid

[gregrnel]

I've disassembled X, everything is tight and the belt is good. There is nothing mechanically wrong with X. The machine has very low hours because it's been down for most of it's life. Here's the code I'm running for those of you who requested a look-see, as usual thanks to all of you for the advice even if I am the problem...

[ImanCarrot]

I imported the 8 inch diameter circle into my AutoCAD and noticed one thing... the X zero wasn't at X zero when I stuck a Grid on it, it was moved to the left by 0.2878 inches (the Y0 was at Y0).

Now I don't know if this would or should matter, but it's one thing to eliminate just in case: so I did a new 8" dia circle in AutoCAD and exported it as a dxf file with X0 and Y0 at the right place. Please find it attached below.

It just seems suspicious to me that the error on your circle looks about 0.3" sag over the arc... and the X0 is moved to the left by roughly the same amount... hmmm.

Unfortunately, I work in mm, not inches, but I don't think it should matter because the AutoCAD file is arbitrary units (I beleive?) and it's only your CAM software or the machine hardware that "knows" it's inches instead of mm.

Also, the small circle X0 was out too, but I didn't measure this one cos I'm machining like mad at the moment and dead busy

Be careful please and dry run with Z out of the way. If it tries to machine an 8mm dia circle instead of an 8" circle gimme a shout and I'll redo it at 203.2mm (8")


Please let me know the results cos this is getting intriguing!

Cheers!

[InspirationTool]

When you verified the X axis as being mechanically sound, did you use indicators? Did you do so at both ends to verify you didn't have any racking going on?

-Jeff