Different layers not lining up when cut

[acourtjester]

Some thing that will help you is with the sharpie slow the feed rate down as you only want to see how it marks with no changes to the speed for metal thickness or cut size. Looking at the circular cut shapes you may be off with the steps per calculations. If you had slop in an axis you would see a notch at the 180 degree areas, as the axis transitions it direction. From your image I see problems in different areas which could also mean crud on the rail or other glitches in the axis moving at random positions. Try creating a simpler series of circular cuts as a text and watch the sharpie marking to see where things happen. Use another sharpie to mark the axis rail where these things happen to go back with power off hand move the axis to see if you locate why things happen there. Troubleshooting requires using different methods to find the problems, simply slow moving be hand to feel things could show you what is the cause. One thing to add is I am not a fan of bolted together tables, with fast transitions in directions and weight of components and a water table, things get loose and wobble. IMHO

circular marks.JPG

[adbuch]

circular marks.JPG

[adbuch]

One thing to add is I am not a fan of bolted together tables, with fast transitions in directions and weight of components and a water table, things get loose and wobble. IMHO

Tom,

All bolt-together tables are not created equal. My Plasmacam DHC2 is a "bolt-together" table with water table, servo motors with encoders, etc. and nothing has come loose in the 7 years I have been using it. It all depends on the particular design, weight of individual components, fasteners, etc.

David

333.jpg

[adbuch]

If you had slop in an axis you would see a notch at the 180 degree areas, as the axis transitions it direction.

Tom - I will have to respectfully disagree. What you say would would most likely only be true if the pierce lead-in location was at the 180 degree position. I suspect that his lead-in locations are at the locations you have marked with your arrows.

David

[acourtjester]

Not a problem but what I am talking about with the 180 marks caused by loose drive components would appear where the axis changes directions.
The OP has not reported what his problem is so I am just putting out things for him to check.
My comment in this post is what I am talking about.
viewtopic.php?p=149591&hilit=notch+in+cuts#p149591
The bolted table objection is mine (IMHO stated), Plasmacam DHC2 is a better quality table then Premier plasma tables.

[adbuch]

Tom - I assumed that the OP was talking about the areas on the circles where you drew your arrows. When he returns to the discussion perhaps he will let us know for sure. I did take a look at the post you cited. Yes - I agree that the Plasmacam DHC2 is better quality than the Premier, as it should be at over twice the price.
Thanks,
David

[@metalman23]

I took a look at the Premier Plasma site and specifically at the 4x4 specs. It says 4 stepper motors, and for that price I doubt it would have encoders. Perhaps show us some photos of the motor/gear rack interface with the covers off so we can see exactly what you are dealing with.

David

They're steppers, no encoders

[@metalman23]

circular marks.JPG

Yes, lead in, and lead out,

Whats happening is the circles are oblong in the X axis, it's literally not coming back to the same spot it started, it doesn't seem like a lead in or out problem,

The circles are also out of square to each other, and it is exactly the same amount off. Everytime I cut out that part

[@metalman23]

hole pattern off 1.jpg

Yes, that was the direction of the X axis,

[acourtjester]

I am going to make some statements (these are my theories). If there is slipping in the drive (loose setscrew, gear slipping, loose belt) you will get a out of round circle or notches 180 degrees apart. This does not seem to be the case as you only have one notch at the start/stop position. What it looks like to me is the torch does not finish in the same position it started. This may be due to loosing steps in the motor as the G-code should show the same location for the start and finish, with no lead in/out involved. Have you tried to do a single step through a single circle G-code??
Here is a job file for a single cut circle at 1.5" diameter with no offset and no lead in/outs. Do the post processing on you system (SheetCam) and do a single step operation to see what it shows. You may want to change speeds to see it that effects to outcome. Just table moving no plasma.

single circle.job

[@metalman23]

I am going to make some statements (these are my theories). If there is slipping in the drive (loose setscrew, gear slipping, loose belt) you will get a out of round circle or notches 180 degrees apart. This does not seem to be the case as you only have one notch at the start/stop position. What it looks like to me is the torch does not finish in the same position it started. This may be due to loosing steps in the motor as the G-code should show the same location for the start and finish, with no lead in/out involved. Have you tried to do a single step through a single circle G-code??
Here is a job file for a single cut circle at 1.5" diameter with no offset and no lead in/outs. Do the post processing on you system (SheetCam) and do a single step operation to see what it shows. You may want to change speeds to see it that effects to outcome. Just table moving no plasma.
single circle.job

What do you mean by single step operation? Is that just the one circle? I'll try it tomorrow if I get time

[adbuch]

I am going to make some statements (these are my theories). If there is slipping in the drive (loose setscrew, gear slipping, loose belt) you will get a out of round circle or notches 180 degrees apart. This does not seem to be the case as you only have one notch at the start/stop position. What it looks like to me is the torch does not finish in the same position it started. This may be due to loosing steps in the motor as the G-code should show the same location for the start and finish, with no lead in/out involved. Have you tried to do a single step through a single circle G-code??
Here is a job file for a single cut circle at 1.5" diameter with no offset and no lead in/outs. Do the post processing on you system (SheetCam) and do a single step operation to see what it shows. You may want to change speeds to see it that effects to outcome. Just table moving no plasma.
single circle.job

What do you mean by single step operation? Is that just the one circle? I'll try it tomorrow if I get time

I think Tom means to run the g-code program one line at a time (single step) if your system allow that mode of operation. I can do single step with my cnc machining centers, but not with my Plasmacam table. I could also probably do it with my cnc router tables that run Mach3 and Centroid CNC12.

David

[@metalman23]

I am going to make some statements (these are my theories). If there is slipping in the drive (loose setscrew, gear slipping, loose belt) you will get a out of round circle or notches 180 degrees apart. This does not seem to be the case as you only have one notch at the start/stop position. What it looks like to me is the torch does not finish in the same position it started. This may be due to loosing steps in the motor as the G-code should show the same location for the start and finish, with no lead in/out involved. Have you tried to do a single step through a single circle G-code??
Here is a job file for a single cut circle at 1.5" diameter with no offset and no lead in/outs. Do the post processing on you system (SheetCam) and do a single step operation to see what it shows. You may want to change speeds to see it that effects to outcome. Just table moving no plasma.
single circle.job

What do you mean by single step operation? Is that just the one circle? I'll try it tomorrow if I get time

I think Tom means to run the g-code program one line at a time (single step) if your system allow that mode of operation. I can do single step with my cnc machining centers, but not with my Plasmacam table. I could also probably do it with my cnc router tables that run Mach3 and Centroid CNC12.

David

Not sure how to do that., but I'd be willing to try

[adbuch]

According to the Premier Plasma website, it looks like the software they offer is SheetCam and Mach3. So I assume you are using Mach3 to control your table. In Mach3, select the Single Block mode.

mach3 single block.jpg

Toms circle g-code file.jpg

Then it will execute a single line of code and stop/pause until you hit the start button, then it will execute the next line of code and stop/pause - and so on. You would do this with the torch off and height control off. Set the feed rate to a very slow value so you can watch the motion of the torch as it moves around the cut path.

I'm really not sure how much good this will do you as it will be hard to tell how exactly where the torch is when it ends the circle compared to the exact location where it started. If the goal here is to find out if you have slop/backlash in your mechanical system, a simpler more straight forward method would be to simply command your machine to move 2.000 inches in the x direction, and then command it to move -2.000 inches in the x direction and see if it ends up in the same physical location where it started from. If you install a torch pointer into your torch cap and position it so that the point is about 0.060" above the material and make an accurate mark at the x location where it starts from, then you can see how closely it is positioned to your mark after it mover the 2" away from the mark and then the 2" back towards the mark. If you have no slop, then it should end up at exactly the same place it started from.

A more accurate method would be to do this same test, but instead of the torch pointer use a dial indicator. The example shown below is for a cnc router, but the idea is the same for your cnc plasma cutter. In the example, the carriage with spindle is commanded to move 2.000" to the left and then 2.000" to the right. In your case - you have a carriage with your floating head/torch mount - but the same idea.

backlash test 1.1.jpg

Then use a similar setup to measure your y axis backlash/slop.

David

[adbuch]

The SheetCam Simulation mode has a single step (single block) mode much like Mach3, except it will only display on the screen whereas in Mach3 you can actually execute the code and run your table.

David

sheetcam simulation.jpg

[acourtjester]

Yes I an looking for a way to see if the table has a wobble and that is why it does not finish where is starts. Single stepping should be a slower action then running under normal cutting speeds, or running the G-code under a very slow speed. In my G-code if he changes line N0150 to have a F value to 5 or 10 speed that should be slow enough. The SheetCam example is only showing the path movements on the screen. David's suggestion of the dial indicator will show if the are steps missing in the motion of the axis. With the aid of TJ (from another post) running the actual G-code on his table with no problem will point directly to the table.

[TJS]

Yes Tom,
I am waiting for the file to be sent to me from Phil. He said he will send it as soon as he can. I think I will post it here if I can as well.
Hopefully more to come.
T.J.

[acourtjester]

Are you able to post some images of your table???

[@metalman23]

The problem is, I don't have the time required to sit down and do all the troubleshooting that I need to do, to get this thing to work, but now I'm stuck with it. I do regret buying this one. I should have saved up for a professional model

[@metalman23]

20240814_105257.jpg

20240814_105255.jpg

20240814_105249.jpg

[adbuch]

The problem is, I don't have the time required to sit down and do all the troubleshooting that I need to do, to get this thing to work, but now I'm stuck with it. I do regret buying this one. I should have saved up for a professional model

Since you are short on time, it would be great if you could find someone in your area with general knowledge of cnc plasma tables. They wouldn't even have to have the same brand, just a general knowledge of troubleshooting techniques to help you to pinpoint your exact issues. TJ has mentioned that you came to his shop to see his table run. How far away are you from TJ? Also - where exactly are you located?
Thanks,
David

[TJS]

I brought it into Inkscape. The big circle is off by a thou. The little circles are all the same size and are aligned. I am not used to bringing dxf files into inkscape. I get some UTF8 crap and it opened in MM. I then also noticed all the objects were not combined or grouped but that might not matter for DXF.
I saved it off as an .SVG cause that is what I am used to and CandCNC likes ,SVGs

Dave and others I am interested on what you guys come up with.

Instrument Cluster.dxf

[TJS]

Phil,
I blew up some of those pics. What comes to mind is the handheld torch and that Z car plate looks very long. This could flex when changing directions. Also what type of cable management do you have for the torch lead. I run mine thru my plastic chain, others do not as they have some sort of hanger set up for a strain relief.

[acourtjester]

I agree with TJ I have attached one image with things that could be changed to help. The torch mount is way to long and the lower part looks like a floating head assembly. If the blue section is repositioned down and the torch holder shortened it may help. The single linear rail that is used on the torch holder may not be steady enough. Her is another view from their YouTube post and the torch holder is positioned much shorter, it may be a newer design.

bad table.JPG

bad table 2.JPG

[adbuch]

I brought it into Inkscape. The big circle is off by a thou. The little circles are all the same size and are aligned. I am not used to bringing dxf files into inkscape. I get some UTF8 crap and it opened in MM. I then also noticed all the objects were not combined or grouped but that might not matter for DXF.
I saved it off as an .SVG cause that is what I am used to and CandCNC likes ,SVGs

Dave and others I am interested on what you guys come up with.

Instrument Cluster.dxf

Untitled.jpg