Imperfect circles

Plas4u
Vectric Apprentice
Posts: 55
Joined: Fri Dec 05, 2014 10:38 pm
Model of CNC Machine: Multicam Classic 1000

Re: Imperfect circles

Post by Plas4u »

Excellent idea Glenn

I will try that in the morning. Thank you.

Plas4u
Vectric Apprentice
Posts: 55
Joined: Fri Dec 05, 2014 10:38 pm
Model of CNC Machine: Multicam Classic 1000

Re: Imperfect circles

Post by Plas4u »

Thought I would post an interim update for anyone interested. I tried Glenn's suggestion of using the circles in the cut utilities but they were also out of shape. Multicam loaded some software updates to see if that would make a difference but strangely it made things worse.

As of today they will be sending out a complete new Y axis to have installed ( sending from Texas to Australia). Not sure how they came to this conclusion but I am just going with the flow.

Once that is done I will post the outcome to hopefully provide closure to this thread.

CarveOne
Vectric Wizard
Posts: 1479
Joined: Sun Nov 28, 2010 5:10 pm
Model of CNC Machine: Home built designs - 3'x4' and 5'x12'
Location: Eastern NC, USA

Re: Imperfect circles

Post by CarveOne »

I don't know which machine control software the Multicam machines use, but I see no mention of X and Y calibration mentioned. I use Mach3, which allows tweaking the two axes to give perfect circles. My smaller homebuilt machine uses precision 5 start ACME lead screws for X and Y axes and the calibration tool correctly produces equal numbers and it makes perfect circles. If the circles are perfect, but the diameter is off a little, both numbers can be edited to correct the diameter error.

My big machine uses rack and pinion for the long axis (X) and the precision 5 start ACME lead screws for the shorter axis (Y). The Mach3 calibration tool produces different numbers in this case. If imperfect circles are produced, one or both numbers can be edited to produce perfect circles of the correct diameter by trial and measurement using a dial caliper.

There is also a way to calculate the numbers needed for my drive mechanisms and enter it into the Mach3 calibration tool menu. This assumes that the drive mechanisms move exactly the amount that are claimed for them. That isn't always the case for low cost mechanisms. But it can be corrected for in Mach3 software as long as it is a consistent linear error over full travel.

Mach3 also has a software compensation scheme for the cases where even the nonlinear errors can be alleviated for the most part.

If the new Y axis doesn't correct your problem it may prove that there is a small discrepancy in the Multicam machine control software settings.

CarveOne
CarveOne
http://www.carveonecncwoodcraft.com

glenninvb
Vectric Wizard
Posts: 1129
Joined: Thu Jul 18, 2013 4:58 pm
Model of CNC Machine: homebuild / VCP 8.0

Re: Imperfect circles

Post by glenninvb »

I glanced at the owners manual and it appears the 1000 is a basic stepper motor, rack and pinion drive setup. (except Z-lead screw) I'm guessing somewhere in the MC controller software a steps/per setting is adjustable. If the tech didn't find any backlash, not sure what a new axis is going to fix? Did you by any chance measure a cut square for perfect dimensions?

Anyway, glad they are standing behind the product and hopefully you'll be going soon.
Please update

armbrusterco
Vectric Craftsman
Posts: 128
Joined: Tue Feb 14, 2012 5:26 am
Model of CNC Machine: Roland MDX-40A
Location: Lake Barrington, IL
Contact:

Re: Imperfect circles

Post by armbrusterco »

Michael,
Can you ask Multicam to email you the file they cut into the acrylic sample with the map of Texas and shapes? Then you could run the same file with another piece of acrylic and compare how your machine cuts now.
Bob
Bob Armbruster
www.ArmbrusterCo.com

Plas4u
Vectric Apprentice
Posts: 55
Joined: Fri Dec 05, 2014 10:38 pm
Model of CNC Machine: Multicam Classic 1000

Re: Imperfect circles

Post by Plas4u »

Your comment about the map of Texas etc Bob caused me to go and have another closer look at their sample. It is very hard to measure but I am now sure that their circles are out of whack as well. Looks like they should be about 38mm ( or the imp equivalent) but are coming in at between 38 and 37.26 depending on where you measure and how much pressure you put on the vernier. There is also a large dot over the 'i' in multicam which I hadn't scrutinised before and it is definitely a bit squashed but that could be the font.

No idea what they are opting for a hardware fix. They did try a couple of software fixes as mentioned which made it worse. The local tech is out of his depth and I am not privy to his discussions with the US so I am in the dark a bit. They assure me it will be fixed and that is my expectation. We will see.

As for the questions about perfection in other shapes, as I have mentioned before, I am satisfied that the machine is very accurate in every aspect apart from circles.

CarveOne
Vectric Wizard
Posts: 1479
Joined: Sun Nov 28, 2010 5:10 pm
Model of CNC Machine: Home built designs - 3'x4' and 5'x12'
Location: Eastern NC, USA

Re: Imperfect circles

Post by CarveOne »

Draw a circle in Vcarve Pro, create a toolpath for it, and cut it to see what you get. Cut it at a low enough feed rate that there is no flexing of the cutter under load, and keep the cutter as short as is reasonable. (Stiff cutter and low loading) Make the circle just under 6" so you can use a dial or digital caliper to measure it. The larger the circle the better.

This way you know for sure that the circle is not a slight oval.

One other way I can suggest is to look at the gcode file that gives imperfect circles on your machine and find the largest X axis and the largest Y axis position numbers. They should be the same or within the allowed tolerances if it is to create a perfect or nearly perfect circle.

CarveOne
CarveOne
http://www.carveonecncwoodcraft.com

User avatar
TReischl
Vectric Wizard
Posts: 4596
Joined: Thu Jan 18, 2007 6:04 pm
Model of CNC Machine: 8020 48X36X7 RP 2022 UCCNC Screenset
Location: Leland NC

Re: Imperfect circles

Post by TReischl »

Been following this thread for a while now.

The G code for cutting a test circle is very straight forward:

G90 'Set the machine in absolute mode
G0X2.0Y0.0Z.1 'Rapid to .1 above material Z zero should have been set to surface XY zero should be center of material
G1Z-.1F30.0 'Plunge cut in Z to .1 depth
G3 X0Y0I-2.0J0.0 'Cut a circle 4 inches in diameter, center is located -2 from cutter location. Circle will be less than 4 inches.
G0Z.5 'Rapid up out of material
M0

Ok, now for the dirt. Contrary to beliefs held by many cnc users. . . .cnc machines do not and cannot cut a true circle. The software breaks down a circular command into a series of very small steps. I will not bore everyone with the gory details but suffice it to say the control is optimized to handle the circular command. Attempting the same thing by programming a series of tiny steps will generally result in the motion buffer being emptied and then the machine goes into the herky jerky fits.

All that said, switching the control software was more than likely an attempt to to alter the circle algorithm. Probably not a good move but the easiest one to try first. They may now be suspecting that the pitch of the lead screw (if it has one on that axis) or the gear/rack pitch was goofed up. So that would be the next thing to attempt fixing.

Most of us check for motion accuracy by cutting a circle or square then getting out the dial calipers for a look see. But that is really measuring the result of cutting, not true machine motion. It includes all the mechanical slop and tool flex. To measure actual motion it is best to use a dial indicator with 1 or 2 inches of travel to see how each axis actually performs.

Sorry to see you are having these issues, but from what I have seen the company stands behind its products.
"If you see a good fight, get in it." Dr. Vernon Johns

Plas4u
Vectric Apprentice
Posts: 55
Joined: Fri Dec 05, 2014 10:38 pm
Model of CNC Machine: Multicam Classic 1000

Re: Imperfect circles

Post by Plas4u »

Thanks for your informative response Treischl. I may pass it on to the tech as you seem to have more of a clue than him. At this stage I am just going to form the wagons into a circle and wait till the cavalry arrives. Given up trying to solve it myself. I will say though that since they have played with the software other weird stuff has been happening so hopefully everything will be sorted in one hit in the not too distant future.

Post Reply