QTPYVCP+VISMACH+LINUXCNC ROLLED INTO ONE?

I found my self in a situation where i could end up getting sucked into the lathe department at work and had an idea to build some kind of test simulation for testing macro programs and testing of programs in general. So i thought it would be neat if i could create some kind of virtual machine for this. i have never tried anything like this in the past so who knows how far i can push the limits with what is able to be done with the tools we have here to work with? Being a mill guy for the most part and looking at lathes with a main and sub spindle with XYZ with a 4th linear axis on the sub spindle for part transfers and phase snyc on both spindles with some of these lathes having a second turret on top of this kinds makes you take a step or two back to get your ducks in a row when your in a situation where there is no one to ask questions because the last guy that knew these machines ends up retiring. So that's basically how this project came to light

what i have managed to do so far is get all 3 of these programs to play nice together open up and render my model and create a 3D orbital view with a zoom feature with a custom viewer so far. i almost gave up on this project quite a few times but got things working this past weekend. i am not just getting into the custom UI with designer which i am just now starting to understand a little bit.
added a screen shot of the model to give you and idea of what is being worked on this is by no means final or even close to being done. i had no idea what i was getting myself into when i started this and thought it would not be to much trouble until i started and that went out the window pretty quickly once i dove into it  

 

i really don't have any questions at the moment, from what i have seen with the Vismach models it usually gets tethered to axis and has it's own space to run alongside Linuxcnc what i am trying to do is roll everything into one screen if that makes any sense? i have not seen anything like it so basically all been uncharted territory for me. i would say it would end up being something like what you might find in the machine builder of fusion 360 if it would actually come together 100% when i get done.

There is a qt version of vismach in 2.9+. In theory you should be able to add it into any qt5 based screen.
I don't know much about qtpyvcp, maybe asked their devs if it's possible, or maybe they can port a version specifically for their screen.

linuxcnc.org/docs/stable/html/gui/qtvcp-vismach.html

this has already been implemented in qtpyvcp and does not require vismach. the model can be setup to be shown and run in main vtk ui backplotter natively and respond to gcode and subroutines etc. and is quite easy to setup.

 

some examples below of things already done. to setup once you have your stl/step file components, to do this follow the below steps:

step 1: add new folder to your machine config directory, name is "machine"

step 2: add your stl or step files to the "machine" folder

 

step 3: create a yaml file, can use the below example as a template, to set the base component and orient and configure each component relation in space, and assign an axis for motion.

 

step 4: add a setting in the ini file for [VTK] to activate simulation by loading the new yaml file on startup.

 


Have fun!

Chris

 

i have everything opening fine i just don't have the hal complexity for everything i want to have so it just a matter of building that out and using the qtpyvcp ui to wire it all together in the digital sense. this is basically a project that started out as me thinking it would happen quickly and then turned into a major project for me. now i got so much time in it i just have to keep going with it to see what happens i think i got past the really hard part so now i can start adding stuff to a bare bones setup test and if nothing crashed add some more stuff. i keep making progress with it it's just slow moving because i never did anything like this before. on the up side of the set backs i have not spent any money on real stuff lately so that ended up being a plus for me in that sense

Lcvette wrote:

this has already been implemented in qtpyvcp and does not require vismach. the model can be setup to be shown and run in main vtk ui backplotter natively and respond to gcode and subroutines etc. and is quite easy to setup.

 

 

if i would have know this last week when i was about ready to smash the pc because it fought me so much i would have been all over this in a heart beat i probably should have posted something before i started this to see what other have done  

to build what you see above once you have the machine files would take under an hour to be up an running natively. my suggestion would be to give it a try. it can be configured pretty flexibly for any functionality you might want to add. and instead of using hal you could simply trigger things like toolchange or part passing etc directly from the gcode with some simple commands. configuring vismach and hal for something like this while powerful, seems cumbersome and unnecessary in my opinion (that said, i'm not the hugest fan of working in hal, so im very biased on that front..lol). I would be trying to make this as gcode dependent as possible if your intentions are to use as a simulator for proving out programs.

if you are trying to make a configuration tester that would run off core signals and hal, that would be a different story. even so this could definitely be configured for that.

alternatively, for machines such as your talking about, if this is meant to be program proving sim and if you are programming with any decent commercial cam system they typically have excellent simulation built in with many more sanity checks that what you would likely have patience to build yourself in a stand alone sim. not trying to discourage you, its fun to have projects, just wanted to make sure you were aware of other existing options.

i have options for sim at work with different cam packages but one of the things i noticed right out of the gate once i started looking at the lathe controllers is lack of memory on the controllers since these lathes are maybe 15 years old with the exception one he new one that was just bought. the problem with cam if your trying to do something with live tooling without the Y axis the program size just gets way to big pretty fast. so macro programs seem like a good solution to keep the program sizes in check. Gibs cam might have something for macro's but i don't use it very much. i'm mostly interested in something to test lean programs with and maybe add some quick code for small run jobs for protype work we seem to get a good bit of that with the production work.

ok, s im not certain what your implying regarding memory? are you wanting to retrofit the machines?