richards
08-04-2004, 12:33 AM
NOTE: This is a long, detailed post.
After using my new Alpha 120-60 for about 2 weeks, I've pulled out most of my sparse hair.
Don't get me wrong, the machine is wonderful - if I cut directly along the X or Y axis. If I cut curves, circles or diagonal lines that require movement in both axis, and if I set the move speed above about 1-inch per second, I get chatter, regardless of the VR settings (I must have tried dozens of different settings.)
All of the tests have been done in MDF at shallow depths (1/4-inch to not overload the router motor), with a variety of cutters ranging from 2-flute 1/4-inch spirals to 2-flute 3/8-inch straight and spiral cutters, and with both a PorterCable 3-1/4 hp router and a Bosch 2hp router at 16,000 to 21,000 rpm. Results were similar regardless of the cutter or the motor.
If I cut a 2-1/2 inch diameter circle at 0.5 ips, it is perfect, except for dwell marks at each quadrant. At 1.0 ips, it's still perfect. At 2.0 ips and above, it looks like I used a dull hacksaw to butcher the circle. Cuts along either the X or the Y axis at 0.5 ips up to 8.0 ips are identical except for increasing tooling marks as the speed increases; tooling marks that can be explained mathmatically due to the distance the cutter moves per revolution.
As a simple test, I ran cuts at 45-degrees, which should have moved both the X-axis and the Y-axis motors identically; therefore, the cuts should have been as smooth as cuts made along either the X-axis or the Y-axis. The 45-degree cuts were rough at speeds 2.0-ips and above, indicating that the motors were not ramping identically.
Out of curiousity, I made a spreadsheet that showed X/Y ratios at each degree from 0 to 45 needed to cut a circle. Unless my math is wrong, a circle is made up of eight 45-degree segments, with each segment having the same ratios, but with the direction of either the X-axis or the Y-axis reversed from the previous segment.
In my example, I made a circle that was 2,000 stepper motor steps in diameter; therefore, the first move would have been 1000 steps along the X-axis and 0 steps along the Y-axis. After reaching the diameter of the circle, the next step to move to the 1-degree position, required 1 X-axis step and 17 Y-axis steps (I'm writing this from memory, so the figures might not be accurate). To have a totally smooth move, the duration of the X-axis step should be 17 times as long as each Y-axis step. Anyway, moving 1 degree at a time, the X-axis steps increase in relation to the Y-axis steps, until at 45-degrees, each axis moves the same number of steps. What I'm getting at is that controlling both the the X-axis and the Y-axis properly requires a lot of computing power. Adding ramping to the mixture increases the complexity.
Although I'm a newbie to the CNC router world, I'm fairly experienced at designing, building and programming process control computers that incorporate stepper motors. I know how important correct ramp values are and how tedious it can be to formulate them. But, even after following a methodical pattern of testing different ramp values, I haven't improved the quality of the cut.
Bottom line: How smooth should circles and curves be in MDF at speeds greater than 2-ips?
After using my new Alpha 120-60 for about 2 weeks, I've pulled out most of my sparse hair.
Don't get me wrong, the machine is wonderful - if I cut directly along the X or Y axis. If I cut curves, circles or diagonal lines that require movement in both axis, and if I set the move speed above about 1-inch per second, I get chatter, regardless of the VR settings (I must have tried dozens of different settings.)
All of the tests have been done in MDF at shallow depths (1/4-inch to not overload the router motor), with a variety of cutters ranging from 2-flute 1/4-inch spirals to 2-flute 3/8-inch straight and spiral cutters, and with both a PorterCable 3-1/4 hp router and a Bosch 2hp router at 16,000 to 21,000 rpm. Results were similar regardless of the cutter or the motor.
If I cut a 2-1/2 inch diameter circle at 0.5 ips, it is perfect, except for dwell marks at each quadrant. At 1.0 ips, it's still perfect. At 2.0 ips and above, it looks like I used a dull hacksaw to butcher the circle. Cuts along either the X or the Y axis at 0.5 ips up to 8.0 ips are identical except for increasing tooling marks as the speed increases; tooling marks that can be explained mathmatically due to the distance the cutter moves per revolution.
As a simple test, I ran cuts at 45-degrees, which should have moved both the X-axis and the Y-axis motors identically; therefore, the cuts should have been as smooth as cuts made along either the X-axis or the Y-axis. The 45-degree cuts were rough at speeds 2.0-ips and above, indicating that the motors were not ramping identically.
Out of curiousity, I made a spreadsheet that showed X/Y ratios at each degree from 0 to 45 needed to cut a circle. Unless my math is wrong, a circle is made up of eight 45-degree segments, with each segment having the same ratios, but with the direction of either the X-axis or the Y-axis reversed from the previous segment.
In my example, I made a circle that was 2,000 stepper motor steps in diameter; therefore, the first move would have been 1000 steps along the X-axis and 0 steps along the Y-axis. After reaching the diameter of the circle, the next step to move to the 1-degree position, required 1 X-axis step and 17 Y-axis steps (I'm writing this from memory, so the figures might not be accurate). To have a totally smooth move, the duration of the X-axis step should be 17 times as long as each Y-axis step. Anyway, moving 1 degree at a time, the X-axis steps increase in relation to the Y-axis steps, until at 45-degrees, each axis moves the same number of steps. What I'm getting at is that controlling both the the X-axis and the Y-axis properly requires a lot of computing power. Adding ramping to the mixture increases the complexity.
Although I'm a newbie to the CNC router world, I'm fairly experienced at designing, building and programming process control computers that incorporate stepper motors. I know how important correct ramp values are and how tedious it can be to formulate them. But, even after following a methodical pattern of testing different ramp values, I haven't improved the quality of the cut.
Bottom line: How smooth should circles and curves be in MDF at speeds greater than 2-ips?