Getting back to LinuxCNC plasma table

[acourtjester]

Well I was able to do a work around for my problem and get the parts cut today. Had to click on the ignore Arc OK and no THC.
Next step when I get the time is a complete install following the above link with the newer version of QTPlasmaC.

[plasmanewbie]

Gotta do what you gotta do to get things done. Hopefully you can get everything sorted out, pretty impressive how deep your into this Tom

[acourtjester]

Well me and LinuxCNC are doing battle again, trying to finish a new software change. Working with the procedure from the post above, having a problem with opening up QTPlasmaC the other GUIs seem to work. Working with Phill from down under, he is a great help we just have not put a finger on why its not working. One thing I did find with with this Dell I am using the disc must be set to Legacy in the BIOS. And the above procedure links to a newer version of Mint and that did not work, has to be 19.4 version from another download site.

[plasmanewbie]

Good job sticking it out Tom, I would have given up long ago, that stuff is beyond me!

[acourtjester]

Thanks I need to keep going the plasma section of my table is down until this gets worked out.

[acourtjester]

Well after going over some of the commands in the setup again and the network info it is now working. I will be moving the controller back to the shop for operational testing and work.

[kn612]

If I had known MX Linux was going to give you so much trouble I wouldn't have recommended it. It's been rock solid on the hardware I have. It's odd you are experiencing bugs with an identical setup

Looking forward to seeing if your arc ok issue is resolved. Are you using qtplasmac 2.10 now?

[acourtjester]

No sir not your fault, Problems I had were my miss-steps and that the site had updated version of Mint. Your recommendations put me on the path to get the newer version of QTPLasmaC, which is what I wanted. My posting is just to show the experiences and things can be worked out by sticking with the process. Phill (linuxcnc blog) from down under gave me the site for the older version which is this one, for those who want to setup their controller with it. And thanks for your last post about the arc voltage THC page.
https://sourceforge.net/projects/mx-lin ... o/download

[acourtjester]

Update I was having a problem with the pause command (M01) SheetCam would not complete a G-code when I tried to use it. This is part of the Center punching routine I do, so I can change the amp setting on the plasma unit after doing the dimpling in the G-code. Les sent me an edited Post Processor which corrected the problem. This is for the LinuxCNC PlasmaC post processor.

[weldguy]

Thats great. Wish customer service like that existed everywhere!

[acourtjester]

I am having a very odd problem, I am running QTPlasma and using the built in THC. I can cut parts and sometimes the torch takes a dive at the very end of cutting. The attached images show what I get, if it were not for the break away torch mount I may have broke my torch.
While cutting the THC works as it should and stays at a good constant height until the end. I have attached a PDF with the G-code for the 2 cuts that failed. It was the first and 7th parts. the G-code runs fine with out the THC on. I have both lead-in and lead-outs enabled.

short cuts.pdf

(36.52 KiB) Not downloaded yet

[kn612]

Do you have velocity anti dive enabled on the main screen?

[tcaudle]

Ahh...the dreaded End of Cut dive.

VAD does not help much with end of cut dives. Depending on your motor tuning (acceleration) settings it doe snot slow below the THC cut off speed until its too late.

There are a few solutions:

1. Take your acceleration settings down for the XY motors . it will start to decelerate sooner and maybe hit the VAD point sooner.
2. I don't know how they do it in the Linuxcnc version but there should be a setting as to when VAD cuts in. If its in % raise that number
3. If you use SheetCAm then use the Cut Rules and the EOC (End of cut) to disable the THC at a set distance before the end.

Since SheetCAM will let you insert a rule (M code) in a cut segment its a lot easier with a cut rule than by doing it by hand.

[weldguy]

Awesome suggestions tcaudle, curious to here if any of your suggestions work for acourtjester

[acourtjester]

Ok I did some work and first off I added a cut rule in SheetCam for .250" stop THC before end. Well that did not go over well as LinuxCNC gave me the middle finger see attached screen shot.

G-code error.png (30.14 KiB) Viewed 111 times

I did turn on the Anti-dive button but still have the problem, it may cut one of two parts good then crap out. I did also lengthen the lead out but not any real help.

[kn612]

It's my understanding that qtplasmac handles all z moves, not anything in gcode. Like tcaudle mentioned, increase vad threshold % if you haven't already tried. Maybe 95%. It's odd you're having a torch diving problem that is consistently at the same distance from the end of the cut on on different parts. Especially on that handle looking part that has lots of small moves, I would think it would dive much sooner if your settings were off.

[acourtjester]

Thanks for the info on the VAD, on the location of the dive I wondered if QTPlasmaC had a look ahead function in the software. And reacted to the end of the cut. One of the test cuts I did was a series of 7" line that folded back on and forth for 4 lines, not a problem, but that did not have a lead in or out.

[tcaudle]

Linuxcnc parses the entire g-code as it loads which yields some parameters that get set as variables . Then as it runs it keeps real time variables on several parameters. I does have a look ahead in the trajectory planner (or so I understand) but VAD runs off a REAL TIME variable of XY true velocity (speed over the surface) . The dive at the end is the THC reacting to the increase in voltage from the change at the end. its especially pronounced on a piece where the ending segment meats the beginning (closed shape) and the piece tends to start pulling away BEFORE the final small amount . If you use lead outs its worse because the piece has already dropped out before it start to slow at the end of the lead out.

I assume you are not actually positioning the metal as per the SheetCAM material display (red area) where you are cutting to the very edge of the material. That wont work with a THC unless you do put in specific cut rules .

I am not real familiar with the QTPlamaC setup but typically the Z is "shared" between the THC logic and the G-Code so that positioning calls in G-code can be used to do Z moves prior to the torch firing and the THC taking over motion DURING a cut.
I see three are parameters for pierce and cut height where as CommandCNC uses values stored in the tool set from either the SheetCAM side or the internal tool table.

Since in CommandCNC the tool table parameters include Preset Volts, Pierce Height, Pierce Delay, Cut height,Feedrates and (for HYT-Connect Users) Cut AMPS and Air Pressure settings), those values are stored in the G-code file and automagically load so the operator has little to do as far as doing any settings changes at the console. The CAM designer sets the material type and selects a Tool that is for a special Type and thickness of material and the cut current you want . Then that Job and that G-code cuts the same way every time you use it. You can manually override any setting after the code loads but is usually not necessary.

SheetCAm lets you change tools in a job so you if your table has a Plate Marker you can call a Plate Marker tool for some of the "Cuts" then switch to the Plasma and its preset parameters from code to cut out the parts.

I theory if you had a multi tool head you could select several tools and operations . Since a toolcna have preprogrammed offsets those are automatically applied with the tool then reset when its taken back to primary tool (torch in this case)

[caretech]

Hey acourtjester,

I've been tinkering with EOC AntiDive settings in Sheetcam here as well. See if this works for your End Of Cut Anti-Dive rule:

1. Some distance before end of cut (0.5" as a starting point?) execute code M62 P2 (this shuts DHTC off for QtPlasmaC)
2. At start of cut, issue code M63 P2 to turn DHTC back on

Personally I find Sheetcam's code editing process to lack quite a bit. You can think you have a code edited then it defaults back to its weird M67 or whatever it comes out of the box with for DHTC on / off code snippets. After monkeying with mine for weeks and muttering under my breath sometimes, I think I have it figured out. Let me know if you need some screenshots.

[acourtjester]

Thanks for your inputs, as to tool info for cutting Linuxcnc has a conversion routine to take SheetCam's tool list and create a file to use in QTPlasmaC.
Which I have done and it also will look in the G-code for what SheetCam puts in it for tool info. From what I understand Tom's (DTHC)electronics uses a version of LinuxCNC and SheetCam has a Post Processor coded to work with that. The same is for QTPlasmaC (Linux CNC PlasmaC) even though there are similar options the G-code commands are different. This is reflected in the post above about the end of cut snippet. And this is why SheetCam has so many Post Processors. I have built 10 tables and use different electronics on some of them (2 used the older version of Tom's DTHC) with Mach, UCCNC, and now LinuxCNC (Mesa Electronics). Your suggestion are welcome and appreciated, due to variations in electronics and situations may not fit, but continue. The more eyes on a problem the quicker it is fixed, one reason PS is so valuable.