Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: billturnbull1976 on September 12, 2013, 09:14:41 AM
-
Hey all,
I reckon I'm missing something stupid, but we'll see...
I use a 3 axis router and do my design and toolpaths in Vcarve.
My problem is everything seems to be around 1% out. Say I cut a 50mm Square using an outer toolpath, the resulting cut is about 50.45mm.
So Im thinking, ok the step count is not quite right for each mm
But, If I cut a 50mm inner profile using the same tool etc, it comes out to 49.55mm, so its not the step count (sorry cant remember the proper term in Motor tuning section) Mach3 is at the workshop and Im at home now... :)
Now, I've examine the resulting G Code for both these cuts and it appears to be perfect... So Im guessing its something in Mach3...
I've had this problem for quite some time, initially I thought it was worn out tools as I was only cutting small pieces and my material was never the same thickness, so I was fudging it a bit...
But now I'm doing a lot more work that needs to be as accurate as possible, not to mention efficient in my workflow ie no fudging!
Any ideas?
Like I said, I'll bet its something silly I've missed or taken for granted...
Thanks,
Bill
-
I should point out also, that the problem is always the same, I mean, I can cut this 50mm square, then cut a whole bunch of other things and cut the square again and its exactly the same, so I dont believe its a backlash issue or a physical fault as such...
-
I've had this problem for quite some time, initially I thought it was worn out tools
Hi Bill, welcome !
At first glance, the cutter diameter being .010" undersize than what is described in your CAM would give the same effect that you are seeing.
Did you verify the actual tool diameter ?
Russ
-
A small amount of flex in your machine might do this as well.
Are lighter, slower cuts off the same amount as heavy, faster cuts ?
-
Yeah, that was my first suspicion also (Im a bit rough on bits ;) )
But yeah, I verified that the bit size is accurate.
Also, I have the same issue using different size bits.
Generally it hadnt been a problem as, say all the bits for a project were cut on the router so, the 'scale error' matched and parts generally fit. As my material thickness varied a little, I'd hit issues where something wouldnt fit into a slot, so I thought that was why...
As you can imagine the problem becomes more apparent the bigger the piece I'm cutting (You see, Jupiter for my sons school project just didnt measure up quite right! ;D )
I need to start being a bit more professional now, so I need to find the cause...
Are there any strange offsets that can be set in Mach3 maybe?
I didnt actually setup this config, but maybe it's time I get around to making my own...
yeah the problem is the same on high speed and slowspeed cuts etc
Having said that, the error margin may be enough that the problem appears the same.
Tomorrow morning I'll cut a nice light and slow cut and a deeper high speed cut and compare them just to be sure! :)
-
Double check the Backlash settings (if enabled).
If it is enabled and there is compensation values entered, the physical values may have changed over time.
Worth checking.
Russ
-
Fair enough, Will do!
and report back! :D
-
But yeah, I verified that the bit size is accurate.
How did you verify it? I prefer to cut a slot and measure the slot.
My guess is that this is a steps/unit issue, as the error grows proportionally to the distance moved. Changing the steps/unit will fix this.
Now, your smaller inside cut should not be scaling the opposite direction, but will probably be different from an outside cut by the same amount, regardless of the size.
The discrepancy from inside to outside is most likely mechanical. Either tool deflection, or deflection somewhere in the machine.
As a test, try the same cuts, but cut in the opposite direction. Most likely you'll see different sizes depending on direction. The reason is that a conventional cut will see the tool pulled into the workpiece, making it smaller. A Climb cut will push the tool away, making the workpiece larger.
-
What is your dial indicator telling you? The dial indicator with a gauge block will tell if the step motion is accurate. It wil also identify backlash, and deflection in the machine under load. Wonderful simple tool. For my machine I found a 9" x 9" x 9" granite angle plate with 4" thick legs. I put the dial indicator in the spindle and I can check all three axis in a matter of minutes. Works great and gives me confidence that the last crash didn't actually do any damage other than the broken tool, my wallet, and my pride.
-
Guys, thanks so much for your suggestions, I have a lot of testing to do! :D
I havent got back down to the workshop yet, but I check all these things and see how I go :)
I dont have a dial gauge myself, I'll harrass my old man and see if he has something suitable (it is his machine!)
Will let you all know my findings!
Thanks Again!
-
I havent forgotten... Just flat chat at the moment!
I've done some tests, but I'll finish first before I post em...
Feed rate definitely changes things, also climb vs conventional direction... I've not check whether thats tool deflection or flex in the machine yet (moneys on flex to be honest)
Also, gauge dialled 1mm and it comes to 0.99mm
Anyways, I'll check more soon...
Thanks, Bill
-
If feed rate changes things are you losing steps? Are your motors set to high. Im no expert by ant means. I do know with my CNC if I set my motors to high they are not accurate they loose steps I guess its called .
-
Velocity and acceleration are very important during set up. It is very easy to set either value to high and loose steps.
-
Hey all, Again, I havent forgotten, just school holidays, rush jobs etc...
Im pretty certain that Im not losing steps anywhere, but have noticed considerable flex in the body of the machine. I had originally written this off in my head as unrelated but since realised (after a dumb moment, one of those get caught up thinking the problem is elsewhere) it could be a serious contributor to the issue. Though I think I might be experiencing a couple of small things adding up to create a bigger overall problem.
So again, once things calm down, I should be able to get back into it and finish the diagnosis...
-
Set up the dial indicator an push on the machine frame. You may be surprised by what you find. By moving the indicator and measuring at different places you often will find that one thing is a big contributor to a position problem and that there are also small contributors. Fixing just the big one may be all you really need to do.