Strange depth of cut behavior with V2.01

Discussion/questions about software used with your CNC Shark and programming issues

Moderators: al wolford, sbk, Bob, Kayvon

Post Reply
KevinO
Posts: 67
Joined: Mon Feb 27, 2012 7:25 pm
Location: Long Island, New York

Strange depth of cut behavior with V2.01

Post by KevinO »

I'll try and explain what's going on here and maybe someone can tell me why it's happening. (running V2.01 Release 001 and firmware 2083)

I'm cutting a profile in plywood, using an 0.062 inch end mill. I have the tool depth of cut set at 0.03 inch. I'm cutting a piece of 0.120 inch plywood so the cutting depths (z axis) I expect to see are -0.030, -0.060, -0.090 and -0.120. What I actually see is -0.025, -0.055, -0.085 and -0.115 as the part is being cut. I can see the Gcode as it is flying by on the screen and it has the correct values for the z-position.
At the end of the job, the cutter returns to +0.805 inch, even though my safe height is set at +0.800 inch.
It's almost as though the program is compensating for a virtual zero correction, even though I did not run virtual zero before cutting the material.
I've thought about backlash being the culprit, but that doesn't seem to make sense.
It's not a major problem since I can compensate by defining the material thickness in VCarve as 0.125 instead of 0.120 inch, but there's an annoyance factor here so I want to get to the bottom of it.
Any one seen this problem or should I try NWA support.

Thanks.
Kevin

jeb2cav
Site Admin
Posts: 1524
Joined: Thu Sep 30, 2010 7:04 pm
Location: Kentucky
Contact:

Re: Strange depth of cut behavior with V2.01

Post by jeb2cav »

Hi Kevin,

I'm not confident I understand.

So, you set z0 using a touchplate (or not?), job starts at 0,0,0 (? - did the tool move to z0 = 0.000 at the very start?) - and while it is running, you see in the tool coordinate window -0.025, but in the tap file it is -0.030.

When the job finishes, the tool coordinate window says 0.805.

No virtual setup run for this job. Have you used a virtual setup prior to this?

If my playback is correct - then I agree, it is as if it is compensating for a virtual correction that you did not choose to use.

Joseph Poirier
Posts: 107
Joined: Fri Nov 08, 2013 10:03 am
Location: Toledo, Ohio
Contact:

Re: Strange depth of cut behavior with V2.01

Post by Joseph Poirier »

1. Virtual has to be on and checked for virtual setup to run.
2. The virtual file is not the loaded file unless virtual is on and checked.
3. Return position... max of height plus safe height. However, it should match the start position Z height. or the safe height plus the top of material from the last cutting position, whichever position height is higher.

KevinO
Posts: 67
Joined: Mon Feb 27, 2012 7:25 pm
Location: Long Island, New York

Re: Strange depth of cut behavior with V2.01

Post by KevinO »

Jeb2Cav,
You have described the situation I'm having exactly. Yes, I did set z0 using the touchplate, and the tool moves to z0=0.000 at the very start. I verified that no virtual file from a previous job was installed and virtual is turned off, so theoretically virtual file should not be relevant.
It does it consistently on several files so I don't think it's one quirky file. I never saw this problem in V1.5 by the way.
It's strange that the same 0.005 depth undercut suddenly gets added to my safe height of 0.800 at the end of the job.
I take it you have not seen this on any job you have run?

I guess the problem can be summed up with this question: If the tap file is calling for a depth of say 0.120, what would cause the Shark to cut at a depth of 0.115 as shown in the coordinate window?

Thanks.
Kevin

Joseph Poirier
Posts: 107
Joined: Fri Nov 08, 2013 10:03 am
Location: Toledo, Ohio
Contact:

Re: Strange depth of cut behavior with V2.01

Post by Joseph Poirier »

send the .tap file into support@nextwaveautomation.com and send copies of vtc.nwa, vtemp.tap and wtemp.tap and CPPref.ini files if they exist in [My Documents\Control Panel 2.0\Startup] folder along with a screen shot of your [Help About] window.

KevinO
Posts: 67
Joined: Mon Feb 27, 2012 7:25 pm
Location: Long Island, New York

Re: Strange depth of cut behavior with V2.01

Post by KevinO »

OK Joseph, they're on their way to you as requested.

Regards,
Kevin

Danwood
Posts: 44
Joined: Sun Nov 17, 2013 9:46 am

Re: Strange depth of cut behavior with V2.01

Post by Danwood »

One of the prices you sometimes pay for a lighter less expensive machine is accuracy.
I run a Legacy Artesian 24 X 72 inch machine with a 5 HP industrial power head.
I notice some of what you are experiencing. The difference is my error is 0.0002 to 0.0003.

What you see is caused primarily by two things.
First is the machines capability to repeat the programmed path.
Smaller lighter weight machines often cannot stop or change directions in time.
The other factor is the tolerance allowed by the controlling software.

CNC shark uses proprietary software from an unknown Source. This keeps the cost down but sacrifices accuracy and capability.
My machine is run by Mach3 software which is the most popular software in the industry.

I was at a local Rockler store the other day and saw part representing Washington State cut out of 1 inch thick material on A CNC Shark.
It was done in six or eight passes. One thing I noticed was no two passes were the same. Some passes were offset by as much as 0.020 in places. The repeat ability was not consistent.

I suggest you contact The Shark manufacturer and ask them what tolerances you can expect from your machine.

Good luck
Danwood

KevinO
Posts: 67
Joined: Mon Feb 27, 2012 7:25 pm
Location: Long Island, New York

Re: Strange depth of cut behavior with V2.01

Post by KevinO »

Danwood,

Thanks for your input. However, the problem I'm describing is a new one (for me) that I believe is related to Version 2 of the NWA software. I did not see this issue for the two years I was running Version 1.5
Also, my problem is not related to accuracy. Accuracy and repeatability are affected by freeplay, servomotor accuracy, repeatability and hysteresis, linearity and backlash of the leadscrew, machine bed flatness, etc.

My issue is that the GCode commanded depth and the position feedback in the servomotor (as shown in the coordinate box) do not agree and are off by a fixed amount of 0.005 inch. Those two values should agree exactly, even though the other factors I mentioned will cause physical inaccuracy between the commanded depth and the actual physical depth of the tool.

By the way, the source of the software is not "unknown". It is a proprietary program produced and sold by Next Wave Automation, who also manufacture the Shark CNC.

Regards,
Kevin

P.S. Do you own a Shark?

Post Reply