I think I've screwed up a Setting in my PlanetCNC software and I'm hoping that someone will recognize the symptoms and suggest a remedy.
I'm running a Mark 2/4 Controller and the old software so you may need to go way back in your memory bank.
My problem is that parts are coming out oversize. I've had problems with UNDER size parts before but that could always be traced back to slop in the mechanism or a loose set screw in a timing pulley.
For testing I'm machining a pocket (100 x 20 x 3mm deep) in plastic (Sintra). But, the finished pockets measure 0.8 -0.9mm TOO BIG in all 3 directions.
The machine seems to cut oversize ONLY when running a program (ie reading g-code).
Pockets cut using the "g-code Wizard" are also oversize.
If I move the cutting head using the Manual Data Input (MDI) panel of the PlanetCNC software, the head moves the EXACT distance specified. It DOES NOT overtravel.
Same with Calibration. The head moves EXACTLY the distance it's told to move.
A full description of my Test Procedure and the Technical Details of my Router are below.
Things that I've tried to solve this problem....
I've double checked my CAD model. It's accurate. Exactly 100mm x 20mm
I've double checked the DXF drawing dimensions as exported by my CAD software. Exactly 100mm x 20mm
I also imported the DXF drawing into CorelDraw and eDrawings and checked the Pocket dimensions with those 2 software packages. Again, the dxf is a perfect 100 x 20mm.
From the PlanetCNC software, I exported the g-code and manually verified that the g-code is correct.
I've double checked for loose pulley set screws (loose set screws generally result in undersized parts but thought I'd better check them)
There is no backlash compensation set in the Planet CNC settings because the PowerHouse GT belts and pulleys are virtually backlash free.
Retensioned all belts to be at 50 lb tension using a Gates Krikit.
Re-calibrated the X and Y axes in Settings/Axes at least 5 times. The machine is dead nuts accurate! If I tell it to move 1000mm, it moves exactly 1000mm. Not 1001mm.
The Calibration settings for Steps/Unit EXACTLY match the mathematical calculation for Steps/Unit (see Technical Details below)
Verified that Tool #1 in DeskProto, a 1/8" (3.175mm) end mill, matches Tool #1 in the PlanetCNC software.
In the PlanetCNC Settings, I've set ALL tool offsets and compensation to zero.
I've measured the cutter twice. I've changed the cutter twice. They're both exactly 1/8" diameter.
Have I missed anything?
I've attached my DXF file and the g-code for the 100mm x 20mm pocket. (in ToolCode.zip)
Also attached the Settings file and screen shots of each Settings page in case you don't want to load the Settings file onto your system. (in CNC Settings.zip)
Everything points to there being an incorrect parameter in the Planet CNC Settings, but I can't find it.
I've run out of ideas and things to check.
Can somebody PLEASE tell me what the heck is going on?
Thanks,
Neal Cooper
.................................................. ..................
Here is my Test Procedure...
Create a DXF drawing of a 100mm x 20mm pocket.
The DXF file is imported into DeskProto 7.0 Multi axis edition 64 bit, Build Date 2018-11-30 (rev 8267)
In DeskProto, tool #1 is a 1/8" (3.175mm) end mill.
The 3mm Pocket depth is defined in DeskProto.
g-code is generated by DeskProto.
My CNC controller is a Planet CNC Mark 2/4 board. SW: 10.1807.2601 HW: 2014.11.03
Tool #1 in the PlanetCNC software is also a 1/8" (3.175mm) end mill.
g-code is imported into PlanetCNC software and you know how the rest of it works.
.................................................. ...................
My Computer / CNC Technical Details...
Computer is a Dell Precision T3500 6-Core 3.20GHz W3670 Desktop machine running Windows 8.1., 24gb RAM, 900gb free space on HD
Communication with PlanetCNC Controller Board via dedicated USB 2.0 port.
USB cable is approx 16 feet long. (Shielded???)
3 axis CNC Router with the following travel...(X)1270mm x (Y)1270mm x (Z)155mm
Gantry style machine with X and Z axis driven by single stepper motor.
Y Axis driven by 2 stepper motors on either side of gantry.
X and Y axis are timing belt driven.
Z-Axis is triple start Acme lead screw with Derin anti-backlash nut.
All stepper motors are 200 steps/revolution.
There is a $65 Anaheim Automation MBC25081TB Stepper Driver on the X-Axis. There are two $6.50 eBay Chinese drivers on the Y-Axis. (Yep, it's a comparison test)
There is a 2:1 gear reduction on the X & Y axis (call it an engineering fudge factor or "just in case" engineering)(Made with PowerHouse GT belts & pulleys, so zero backlash)
Drive Pulleys are PowerHouse HTD aluminum pulleys, 20 tooth, 5mm pitch
Belts are 15mm wide PowerHouse GT belts. Neoprene and Fiberglass construction.
All belts are tensioned to 50 pounds tension using a Gates Kricket.
Currently the X Axis is micro-stepped at 1/8 step. Y Axis micro-stepped at 1/16 step. (and, NO, I don't remember why they're different)
Steps/Unit (X Axis) = Motor steps/rev / Pulley distance/rev * Gear Ratio * Micro steps/step
So, for X-axis...Steps/Unit = 200 / 100 * 2 * 8 = 32 steps/mm
And for Y-axis = Steps/Unit = 200 / 100 * 2 * 16 = 64 steps/mm
Are you sure that SPU values are correct for all axes?
Have you checked if scale factor is set?
Which drivers you have? Have you tried shorter delay values?
SPU values in Settings are correct. Everything on this machine is metric so the math is simple. Please see the last four lines of my original post. Also, if I I use the Calibrate function in the CNC USB software I get EXACTLY the same SPU values. Like I mentioned in my post, the machine is dead nuts accurate when making MANUAL moves. The problem only show up when machining an imported part. And, no, the imported models are NOT oversize. I have confirmed this by importing the g-code or DXF files into 3rd part software (DXF viewers, Alibre Professional and Corel Draw) and verified that the model dimensions are correct.
Not sure about the scale factor or delay values. I wish you had suggested that 2 months ago. I finally gave up messing with the MK2/4 and CNC USB software and upgraded to a MK3/4 and the TNG software. Just got the MK3/4 installed this week and got license for TNG last week. Still climbing up the learning curve for the TNG software, but I should have results in a week or 2. I'll let you know if things improve.
Meanwhile, I am hooking up the old MK2/4 to a single axis stepper stage. I'm going to connect the Step output from the Mk2/4 to my Labview Data Acquisition system and count pulses. I just want to verify that when the CAM software directs the Mk2/4 to machine a feature that requires 1000 steps, that hte Mk2/4 actually sends out 1000 steps, not 1050 or 1100 or some other random number of steps.
PROBLEM SOLVED!!!! It was a very amateur mistake. I had the wrong router bit installed. What I thought was a 1/8' end mill was really a 5/32', 2 flute, straight flute. Being a straight flute cutter, it visually looked much smaller from most angles. I should have listened to my son. That was one of the first things he suggested, but I KNEW I had the 1/8 bit installed. Learned an expensive lesson.. just like carpentry, measure twice, cut once.
But, on the bright side, I now have an extra 4 axis controller and licensed software. Looks like it's time to start designing a hot wire machine!
Reply with Quote
Parts machining OVERsize.
