Anyone using CNC Controller software by Microsystems

[mcmmach]

Is any one out there using software called CNC Controller made by Microsystems of Buckhannon ?
Dean

[crackstar]

I use it! I'm learning it anyway... I have run other peoples machines before using programs like Scanvec tool path, and Gerber artpath (nasty). I picked up a PracticleCNC table and packaged with it was WinCNC and BobCAD. I've been having a difficult time making the switch from propritary programs to straight code.

chuck

[mcmmach]

Hi Chuck,

I just wanted to let everyone running this software to be aware of a bug that can cause problems.

If you have your windows power saver set to blank your monitor and the machine is running, when you touch the keyboard or mouse to wake up the monitor it will cause the axis to stall.

You must go into "start", "settings", "control panel", "power management", make sure power scheme is "always on", and turn off hard disks and monitor should be set to "never". If you want to turn your monitor off just use the power switch on the monitor.

This bug has been driving me crazy for the last few months because it was so random. My mouse sets on the machine and the vibration would sometimes cause the mouse to move and wake up the monitor and if the X, Y or Z axis were moving they would stall and lose position and you know what happens after that (ie, scrap parts and chopped up spoilboards).

Other then that problem I really like the software, especially the constant countouring, it never slows down in curves, only when making a 90^ turn.

If you have any programming questions I would be happy to try to help you.

Best of luck, Dean

[rherman]

Is there a CV, or constant velocity feature in WinCnc? Perhaps in the WinCnc.ini settings for vel and accel. My table will only make a real 15ipm when I tell it to cut at 90ipm on my 3d part work. On long straight cuts or long arcs it does close to what I tell it up to 120ipm. I believe the small stepover and small segments are creating the delay in the accel/decels. Thanks.

Rob

[mcmmach]

Originally Posted by rherman Is there a CV, or constant velocity feature in WinCnc? Perhaps in the WinCnc.ini settings for vel and accel. My table will only make a real 15ipm when I tell it to cut at 90ipm on my 3d part work. On long straight cuts or long arcs it does close to what I tell it up to 120ipm. I believe the small stepover and small segments are creating the delay in the accel/decels. Thanks. Rob

Hi Rob, wincnc has the best CV, look ahead built in and its automatic.
I have one job that that my old controller would stop at every segment change 4,515 times over the run and took 120 minutes to run. When I switched to wincnc it cut the time to 88 min.
The problem is probably in the acceration factor on short moves.
On my machine which has 7-10 amp steppers the acc speed is 800 inches per minute per second. It will accelerate to 150 ipm within .2 inch.
Have you setup your machine to the highest, safest acc speed ?
Dean

[rherman]

There are many settings in the wincnc.ini file. I have played with the following:

At the top of the file, the accel=50 I changed to 100

For each axis, I changed the r,f,h values to 400 except the z-axis. These are 45 for the z. I tried to go higher but the steppers whine at around 500. Besides in a rapid the table is still only making 150ipm max.

I believe I have smaller steppers. They are PacSci ones. I bought the table June 04.

I switched each axis's a=1000 to a=1200.

No luck.

Rob

[mcmmach]

Hi Rob,
I think the problem is the accel=100 setting. No matter what you set each axis accel value to (1000 or 1200) it can only run at 100 maximum because of this setting.

Is this a factory built machine or home built ?
I would be carefull about making changes to accel and max speeds without testing first.

The proper way to test a axis is as follows;
1. Set the accel to a fairly low value (100-200 should be okay)
2. Put a dial indicator on the axis to be tested and zero out ( make sure to move the axis toward the dial indicator before zeroing to remove the backlash).
3. Make a rapid move away from the dial indicator long enough that the axis can reach full speed (20-30 inches usually works)
4. Rapid back to the dial indicator and make sure it zero's
5. Repeat @20 times
6. Increase the "r" rapid value (I usually increase by 100) and repeat the steps above until the axis starts to lose steps (ie will not return to zero or you see or hear it stall)

It helps to write a short program with short pauses to read the dial indicator to do the moves and be ready to hit the estop if you see the axis stall.

Once you reach the point the axis starts to lose steps take that value and multiply by 0.7 and set the axis for that.

After you do all the axis for rapids go back and do the same thing for the acceleration.

On wincnc if you have the 7200 interface card set Maxstepv=50000 and Accel=50000 so they don't limit each axis's settings when testing.

Also if you have a microstepping driver card the steps should be set as low as possible and still have at least .001 resolution, the reason is 1/2 or 1/4 stepping has tons more torque than 1/16 or higher.

I know it's a lot of work to do this but you don't want to raise speeds or acceleration or change a motor or driver without testing first.

Let me know how it works out.
Dean

[rherman]

No, my a value under the axisvelspec? is 1100. There is an accel variable at the very beginning of the file. It is for the increments that can be used to step up the acceleration. I have made the r,f,h values for my x and y axis to 485. I left the z at 250, since it is a ballscrew with direct drive. No, this is a vanilla PCNC table bought in 6/04. I'm thinking the steppers are undersized. I'll have to look. The other variable is that I am cutting 3d surfaces in a raster toolpath. Meaning, I chose to run across the x-axis the router only, since it has less mass than moving the gantry and router in the y-axis. So in effect I am simutaneously cutting x and z until like a typewriter it stepsover in y at the end of each run. My stepover is .015" and each x move is about .020" to .040". I am going to experiment more today, but I am effectively achieving 17-22 ipm linearly. I am not calculating the z moves 0"-0.125". Thanks.

Rob

[mcmmach]

Hi Rob,
Are you doing wood engravings also ?
I do them on my machine and because of the short moves it runs slow as a snail.
Most of the one's I do are 9-10" wide and 7 1/4" tall with .015 stepover on X & Y axis and take about 1 1/2-2 hours to complete.
Check out www.mcmmachine.com for some pictures of some I have done.
Dean

[rherman]

Yes, I do. 9-10" wide by 7-1/4" tall with a .015 stepover would take me about 2-4 hours. I wrote my own software to turn the photo into a 3d relief. My depth of cut is anywhere from .095" to .150". I notice the deeper I go, and the more detail in the photo, the longer it is to cut. I just put up a site for this: www.thewoodenimage.com Let me know what you think. I've also cut lithopanes, but I am steering towards wood only. Thanks.

Rob

[rherman]

Success! I removed the idler shaft and spring on the z-axis of my PCNC and it alleviated some of the motor's load WITHOUT sacrificing accuracy. On the y-axis, the belt would sag quite a bit without the idler shaft and bearing; however, on the x-axis it overstrained the motor. I gained approx. 5-10ipm depending on what I'm cutting.

[mcmmach]

Hey Rob,
Thats a great looking wood carving, I double checked my ipm based on the amount of time to cut and I get @ 30 ipm on the average.
Dean