[rnm85]

Hi All,

I've started to do some test cuts on my new machine. First just some simple straight cuts on mdf that seemed to go pretty well.

Today I tried some more complex shapes cut out of plywood that didn't go so well.

I was cutting 18mm pine plywood using a 1/4" 2 flute downcut carbide endmill. Cutting speed was 3600mm/min (140 inch/min) at 18000rpm cutting 4mm passes. It started out cutting well (photo attached) nice and clean and each pass was in the exact same line as the previous one. After about 10 minutes of cutting I noticed each pass started to step over slightly on the last. This got worse and worse (photo attached) until I decided to stop the g code so I didn't break the cutter.

The culprit is the Y axis that isn't returning to the right spot each time. The Y axis is running CRP R&P, Gecko Nema 23 motors and Gecko G540 Drive.

Any help would be very appreciated!

Cheers,

Rory

[CarveOne]

Your Y axis is losing steps.

Does this axis have one motor, or two motors slaved in Mach3 (or Linux CNC)?

CarveOne

[louieatienza]

Pulleys slipping? Spring tension too tight/too loose? Accel too high? Feedrate looks about right, though it takes a bit more force to push a downcut spiral versus an upcut spiral.

[rnm85]

Thanks for the reply!

Carve One - My Y axis has one motor, machine is controlled by mach3

louieatienza - I'll check the tension on the spring and the pulley belt. I don't really like the downcut spiral but its all I have for now. It really clogs up the slot after cutting and with no dust extraction and doing multiple passes it must put a lot more load on the cutter! I have been following behind it and trying to clear the slot as much as possible before the next pass.

Thanks for the help,

Rory

[ger21]

Check if the G540 is getting hot.
The downcut bit is fine, but I'd lower the rpm to 13,000, and maybe lower the depth to 3mm. the bit will last a lot longer.

[CarveOne]

Other things to check are:

1. Motor shaft coupler slipping
2. Crosstalk in the wiring from the router's AC cable.
3. Step signal not meeting less than 0.5v (logic 0) and greater than 4.5v (logic 1) at the parallel port input connector of the G540. You would need to use a male to female adapter that allows access to the wiring if your parallel cable has molded connectors.

CarveOne

[rnm85]

Thanks for the info, I'll have a look at all these points tomorrow and hopefully get it sorted!

Thanks again for the help.

Rory

[adprinter]

Hi All,

I've started to do some test cuts on my new machine. First just some simple straight cuts on mdf that seemed to go pretty well.

Today I tried some more complex shapes cut out of plywood that didn't go so well.

I was cutting 18mm pine plywood using a 1/4" 2 flute downcut carbide endmill. Cutting speed was 3600mm/min (140 inch/min) at 18000rpm cutting 4mm passes. It started out cutting well (photo attached) nice and clean and each pass was in the exact same line as the previous one. After about 10 minutes of cutting I noticed each pass started to step over slightly on the last. This got worse and worse (photo attached) until I decided to stop the g code so I didn't break the cutter.

The culprit is the Y axis that isn't returning to the right spot each time. The Y axis is running CRP R&P, Gecko Nema 23 motors and Gecko G540 Drive.

Any help would be very appreciated!

Cheers,

Rory

Have you calibrated all of your axis, and tested for back lash? (Using a dial indicator to determine actual movements to compare to the commanded movements?). I do not have any experience with a R&P system.

However, when compared to a screw-driven axis system, I would expect the amount of back lash on a R&P system would probably be significant in comparison. Check the repeatability of commanded movements against a dial indicator. (By commanding the axis to move the same distance, multiple times). Then in the opposite direction. Is there a difference? -BACK LASH!

Could also be the holding torque of your motors are insufficient. When the router plunges into the work piece, this exerts a lot of pressure on ALL axis points. Think about the Jerk you feel on the handles of a router, if running one by hand, and free-hand carving on something.

This "Jerk" is felt by every axis on your machine. And if the motors are not up to the task, (of staying put, in the position they are commanded to be at), then all manner of "drifting" events can occur. If any sort of mechanical binding occurs, at a moment a commanded movement is being executed (I.E. machine fails to actually move to where it was commanded to move to), then the machine has now "Lost it's mind".

Because the tool is no longer physically located, where the control software "Thinks" it is! In the event of lost steps, you might want to lower your feed rate. When first starting out, you will learn that it is largely a trial and error process. Until you find the "Sweet Spot" that your machine can handle with no lost steps!

Tip: Once you find the speed thats works well in Pine, or MDF, you may have to revisit this process again, the first time you try carving Oak, or Maple! Write down the feed information which works for each, somewhere for future reference. And refer to it, when creating your tool paths, or setting up your machine.

[rnm85]

Hi everyone,

I ran through the points you have given in this post before doing another test on my machine. Everything seemed fine before cutting. This time I tested on a piece of pine board using a 3mm 2 flute upcut bit. I was cutting to a depth of 6mm in four 1.5mm passes at about 2500mm/min (100 inch/min). I was cutting letters out of various sizes. It cut 5 separate letters out really well and then all of a sudden from finishing one letter and moving to the next it was way out with the y axis again which would have been about 10 mins in to it. I stopped the code went over and checked the G540 and touched the back of it which was extremely hot. I couldnt hold my hand on it for more than a second or two.

What are the issues that would cause the driver to get so hot and how can I stop it? I'm not sure if this is the problem as the X and Z axis seem to work fine?

I haven't checked the step signal as CarveOne suggested as I don't know how to do this.

Anymore help would be great!

Rory

[CarveOne]

Hi everyone,

I ran through the points you have given in this post before doing another test on my machine. Everything seemed fine before cutting. This time I tested on a piece of pine board using a 3mm 2 flute upcut bit. I was cutting to a depth of 6mm in four 1.5mm passes at about 2500mm/min (100 inch/min). I was cutting letters out of various sizes. It cut 5 separate letters out really well and then all of a sudden from finishing one letter and moving to the next it was way out with the y axis again which would have been about 10 mins in to it. I stopped the code went over and checked the G540 and touched the back of it which was extremely hot. I couldnt hold my hand on it for more than a second or two.

What are the issues that would cause the driver to get so hot and how can I stop it? I'm not sure if this is the problem as the X and Z axis seem to work fine?

I haven't checked the step signal as CarveOne suggested as I don't know how to do this.

Anymore help would be great!

Rory

The G540 needs some heatsinking and maybe a fan to help cool it if the motor currents go above a certain level. The GeckoDrive website has an article about heatsinking the G540 here.

If the cut does not drift sideways until some distance into running the gcode like you say, then the step or direction signal voltages may not be in error. It sounds like the G540 is getting too hot. Try putting a household oscillating fan blowing at the G540 to cool it and see if it completes the cut job.

CarveOne

[rnm85]

Thanks CarveOne I'm off to the electrical store. I have the driver mounted in a thin piece of plywood in a wall in a 200sqm warehouse with no other heat producing things around so hopefully a small fan will do the trick to get a bit of air flowing around. I'll soon find out!

[CarveOne]

Thanks CarveOne I'm off to the electrical store. I have the driver mounted in a thin piece of plywood in a wall in a 200sqm warehouse with no other heat producing things around so hopefully a small fan will do the trick to get a bit of air flowing around. I'll soon find out!

Just replacing the plywood with a sheet of aluminum that is 2" or so larger than the G540 case is better than nothing. Mount the sheet to the back side of the G540 case using screws and standoffs. Use white thermal grease from Radio Shack between the sheet and the back of the G540 case. Adding a PC case fan blowing on the sheet is even better. The cooler the G540 runs while cutting projects the better.

edit: Oops, I guess you don't have Radio Shack stores in Australia.

CarveOne

[rnm85]

Hi everyone,

The issue was the G540 over heating. I went out and bought a small desk fan for $9 and mounted it to the wall blowing directly on to the back of the driver. I also got 2 second hand cpu heat sinks that I was going to mount to the back but I ran out of time.

I did a test with just the fan and it 100% fixed the problem. I am still going to mount the heat sinks as well and mount the driver in aluminium as CarveOne suggested.

You can see the machine cutting here if you would like. In total this sign took about 40 minutes to cut and afterward the G540 was warm but not hot with the fan on instead of 10 minutes of cutting and being almost too hot to touch as was happening before.

Thanks everyone for their help this last little bit of the build has been slightly frustrating but its worth it now everything seems to be working!

Thanks again,

Rory

[CarveOne]

I figured that ger21 was correct about the G540 overheating. Putting a temporary fan on it just verified it. Do put a heat sink on it though. Even mounting it onto to the center of a 12" square sheet of 1/16" aluminum will help a lot.

I have a CPU heat sink mounted to a piece of 1/4" aluminum flat bar, that is mounted to the G540. There is white thermal grease on both sides of the flat bar where the heat sink and the G540 contact it. I also have a 4" square 120vac fan pulling air across it. The 48vdc PSU and G540 is mounted inside of an old 900 VA UPS case. There is a rectangular opening in the side of the case for access to the connector side of the G540. There is a rectangular opening on the floor of the case behind the front panel for an air intake. There are 1/5" thick rubber stick-on feet to let air come in under the case. It has been working pretty well so far.

There are lots of other ideas for mounting the G540 shown on in the build logs on this site.

CarveOne

[CarveOne]

Keep in mind that you aren't having fun if it isn't frustrating enough to fix the @#^%*& problem.

CarveOne