Hello Guest it is December 04, 2021, 03:52:02 PM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - eforer

Pages: 1
Hi All,

I'm working on a registration mark sensing macro that will allow for print cut registration (setting the origin), coordinate system rotation to account for imprecise placement of the graphic and stretch adjustment (scaling).  I have already discussed this topic a bit in some posts in this thread http://www.machsupport.com/forum/index.php/topic,26268.msg188768.html#new.  

I have already completed an arduino based mark sensor setup that will output +5V upon detecting the edge of a mark.  This can be easily wired to an input for mach3.

I'm not a VB guy, but I do have some programming and scripting experience in C++, Java and Actionscript.  I'm looking for some help.  I'm going to outline what the program needs to do below and perhaps you gurus of Mach3 can help!

Graphics will be printed with registration marks (L shaped blocks) of a known size printed at the corners outside the bounding box.
A cut file will be created in advance which will be offset from the origin according to the size of the registration marks.
The operator will place the graphic on the machine and position the photocell in the interior angle of the registration mark.
The operator will start the macro.
The macro will move the machine X+ slowly until the edge is sensed.
Once the edge is sensed, the machine backs up on the axis 50% the length of the (known) reg mark size and advances more slowly, recording the X position when the photocell senses the edge again.  The slow down is to ensure precise detection.
The machine again backs up 50% of the reg mark size and moves in the Y+ direction.
When the sensor detects the edge, once again the machine backs up and moves more slowly in Y+ recording the Y value of the registration mark.
Now the X and Y coordinates of our origin corner are recorded.
The machine will center itself between the x and y leg of the aforementioned mark and proceed to travel rapidly in X+.
Upon detecting the edge the machine will stop, and backup (a greater distance due to speed, this will have to be experimented with) and advance forward slowly.
Upon detecting the mark while moving slowly, it will record the X coordinate.
The machine will move forward slowly past the registration mark edge by 50% the registration mark size on the X axis and then proceed to move Y+
Upon detecting the edge, the machine will back up and then proceed slowly on Y+ again until the edge is detected.  Upon detection the Y coordinate is recorded.
Now with 2 points recorded, the rotation can be calculated with an arc tangent (used later)
The machine will return to the first corner defined (the corner nearest the origin) and run a similar routine to detect a 3rd mark at either remaining corner.

These 3 points should be sufficient to calculate rotation and linear stretch on both axis.  Can someone help me get started with the VB?  I'll be sharing the simple arduino reg mark detector design once it is mounted to the machine and working reliably.  If we all work a little on this I think we can quickly add some incredibly valuable functionality to Mach3 that many of us could really benefit from!


General Mach Discussion / axis acceleration
« on: August 22, 2012, 01:31:23 PM »
We retrofitted our Multicam cnc router to mach 3 some time ago after finally losing it with multicam support for the electronics, pendant etc.  One thing we have struggled with the whole time with this machine has been issues with chatter and cut precision when changing direction.  I discovered that while the motors weren't stalling, the accelerations have been too high creating shake in the spindle during abrupt starts and stops.  Is there a formula to figure out safe minimum acceleration speeds for a given feed rate.  We never cut much above 150 ipm.


General Mach Discussion / Missed steps??
« on: August 22, 2012, 01:28:26 PM »

We have been running a Multicam 5'x10' table retrofitted to mach 3 for quite sometime.  We are having an inconsistent issue where on multiple depths/pass jobs we sometimes have passes that don't line up by a few hundreths of an inch.  I'm having a hard time telling if it is because the material is shifting or if the machine is losing steps.  We are using an ethernet smooth stepper board with cnc4 pc C32 boards.  The machine has pacific scientific stepper 6410 stepper cards and some pretty beefy motors.   

General Mach Discussion / Unusual problem
« on: May 22, 2012, 10:13:48 AM »
I'm having a strange issue, for some reason I can no longer jog from anything other than my shuttle pro based mpg.  We have changed no settings anywhere and when I click the "jog" mode select on the on screen mpg, none of the options (step, cont. etc) illuminate.  I can't use the virtual buttons or my keyboard arrow keys anymore.  Any thoughts?  This machine is used in a production environment and gets a full shift of use daily and no one ever changes any settings so this is a little odd.

General Mach Discussion / trouble cutting circles
« on: February 29, 2012, 03:50:05 PM »
Hi Guys,

I recently converted a Multicam I purchased to mach 3 using an ethernet smoothstepper, and a c 32 breakout board.  I purchased the unit refurbished from multicam, they failed to properly test it and all the internal logic boards were toast so a mach 3 conversion was my best bet (they refunded some of my money).  The conversion has gone smoothly and we have cut a lot of parts.  We do a lot of sign work etc. so we have been getting usable output for big stuff in spite of the problem I'm writing about. 

The issue I am having is with small, detailed cuts.  Initially, they were out of round erratically, but we found a ton of slop in the Y axis and that cured it from being inconsistent.  Now it is consistently out of round.  It sort of looks like it could be caused by backlash, but I wanted you experts to look at it so I can get some other opinions.

The machine is rack and pinion and feels pretty tight, I have not gotten around to measuring the backlash yet.


General Mach Discussion / Flatbed prints, G68 and Copy Cat
« on: January 10, 2012, 10:05:25 PM »
Hi All,

I'm new to Mach 3 (retrofitting my ginourmous multicam).  It is part of our large format shop and we cut out all sorts of big prints made with our Gandi Jeti flatbed.  The flatbed is very precise, more so than the router.  So we don't need to compensate for stretch.  Rotation is another matter.  The pin registration is quite good, but on a 10 long board things can go a little out of whack.  The print will be square to itself but it might not be perfectly square to the board.  Now that we have the added flexibility of mach 3, especially the coordinate system rotation, this could work out really well for us.  I'd like to permanently affix a camera to the side of the spindle and use it like a visual probe and then use those coordinates to make the calculations to rotate the coordinate system automatically and set the origin.

Has anyone implemented anything like this?


SmoothStepper USB / More dual axis homing issues
« on: January 09, 2012, 10:38:15 PM »
Hi Everyone,

I have just finished retrofitting my multicam router (big sucker 70"x140") rack and pinion.

Obviously at that size it's dual drive. It has prox switches on both drives to square the gantry when homing. I have everything hooked up via my C32 and my ethernet smooth stepper.

I have my A Axis slaved to X and I have the the direction pin active hi on the A axis and low on everything else. I have the homes set as negative and all the axis normal (ie not reversed). All motors are wired the same. Everything jogs great and the Z and Y axis home just fine BUT.... The XA / XB is a mess.

Here's what happens. Whatever axis hits first does its little bounce off the home switch just fine, the other axis (usually the A) might jump around a little, might not do anything, but it will not move off the axis and the home sensor will remain active on that axis.

If I look at the prox switch locations they are inconsistent, the second switch which hits will always be more "over" the dog where as the first axis will move off the dog so that they are not inline.

What's going on??? I need to get my machine back up and running so I'm desperate.


Pages: 1