Advanced Search

tuxcnc wrote:

pb12 wrote:

Had to drop back to V2.7 to get a reasonable jitter number on my old Athlon X2 CPU. (20,000)
 
Very bad idea...
Wheezy with RTAI has low jitter on some motherboards, but today this is time waste only.
This is 32bit Linux, Python 2.7, no realtime ethernet, and old LinuxCNC not compatible with modern, even at configs level, so you get many useless experiences...

Yesterday I installed Debian13.5 with LinuxCNC 2.10 on my Lenovo M725s SFF Ryzen 3 8GB 128GB M.2 NVMe (not newest but still modern), jitter is 20,000 and all works well.
This model has LPT connector at motherboard, but you can buy cheap second hand PCIex to LPT card.
In my opinion, LPT is obsolete, but if exist, you can serve some auxilary signals by this.

There are cheap 7i92 clones on Aliexpress.
I have model similar to original MESA, and chinese version with I/O drivers, both works well.

In my opinion, the future is ethercat and modbus. 
This is more cost and more dificult, but works like in dreams...

 

Yes. I have come too this conclusion too. Linuxcnc has passed the old hardware plus parallel port plus BOB stage, it seems. My old Athlon X2 might have to be reporposed as file server. That will free up an Intel i5 machine.
Although I didn't realise Linuxcnc 2.7 was 32bit. The process did give me understanding that the config was different enough so as not to be useful for the future..
So now booting iso from USB stick and testing with SIMs.

Phil

 

Hello everyone,I wanted to give a small update on the current state of Lathe EasyStep.A lot has changed internally since my last post. While the panel may look similar at first glance, a large part of the project has been refactored and modularized. The original large Python files have been split into many smaller modules (G-code generation, preview, contour handling, persistence, tool logic, UI, etc.), making the codebase much easier to maintain and extend.At this point I consider the project to be at Version 0.7.0.Development is still ongoing, but over the last weeks the focus has been less on adding new features and more on improving the internal architecture, modularizing the code and refining the overall workflow. The goal was to build a solid foundation that will make future development easier and improve long-term maintainability.General workflowThe user interface is intentionally based on the workflow of conversational turning systems. Anyone who has worked with ShopTurn or similar conversational programming systems should feel familiar with the basic workflow.The Program tab is where all global settings are defined, for example:

• stock dimensions
• safety distances
• retract positions
• tool change and safety settings
• other program-wide parameters

After that, machining operations are created step by step using the corresponding tabs.Reusable machining stepsOne feature that was important to me is the ability to reuse operations.Every machining operation is stored as an individual Step file containing all parameters of that operation.For example, if I create a facing operation for 30 mm stock, I can later reuse exactly the same operation in another program without entering all parameters again.Over time this creates a small library of frequently used machining operations."Save Changes"A new Save Changes button has also been added.Instead of recreating programs from scratch, existing projects can now be edited and updated.Depending on what already exists, the application automatically updates

• the individual Step files,
• the complete program,
• and any previously generated G-code.

The graphical preview is rebuilt automatically as well.PreviewThe preview system has been improved considerably.Besides the existing side view, there is now an optional cross-section view. The section position can simply be moved with the mouse inside the side view, making it easy to inspect the generated geometry at any Z position.ContoursContours are created by entering coordinates, with the option to use either absolute or incremental values.Each contour receives its own name and can later be reused in any number of roughing operations.Inside the Roughing tab, the desired contour is selected, a tool is assigned and the remaining machining parameters are configured.Available operationsThe project currently supports, among others:

• Facing (roughing, finishing or both)
• Roughing using user-defined contours
• Thread cutting
• Grooving / Parting
• Drilling with different drilling strategies
• Keyway shaping (currently experimental)

Several common thread presets are already included, but all parameters can be modified as required.For parting operations it is also possible to define the X position where feed and spindle speed should be reduced, helping to make the final separation more controlled.Keyway shaping is currently still experimental. The preview already works, but I have not yet been able to validate it on a real machine.What's next?My next focus will most likely be the G-code generation itself. I am sure there is still room for improvements and additional machining strategies.Up to now, development has naturally been influenced by my own way of programming parts. This is exactly where feedback from other users would be very valuable.I'm especially interested in how other machinists or LinuxCNC users would approach certain operations. There is often more than one valid way to generate the same toolpath, and in some cases there may even be a better solution than the one I currently use.Constructive criticism, ideas and suggestions are always welcome.My own lathe is unfortunately still not fully assembled, so I have not yet been able to validate every feature on a real machine. That makes feedback from other LinuxCNC users even more valuable, especially from people who are willing to try the panel and share their experiences.I also hope to create a short video or a small series of screenshots soon, as I think the workflow is much easier to understand visually than from static images alone.

Overkill...
For purposes like this, GRBL is more than enought...

7i97t_7i74d.bin and 7i97t_7i74d2.bin differ
in that 7i97t_7i74d.bin has just one Sserial port
with channel 0 on the 7I97T and channels 1..7 on the  7I74

7i97t_7i74d2.bin has 2 Sserial ports with port0, channel 0
on the 7I97T and port1 channels 0..7 on the 7I74

The reason for the nearly identical firmware images is that 7i97t_7i74d
matches the older 7I97 firmware, and 7i97t_7i74d2 ( with 2 Sserial ports and
all  7I74 channels available ) wont fit in  a 7I97 with its smaller FPGA.
 

My latest brain fart

If you use a recent pncconf it will use the INM pins rather than the GPIO pins
The INM pins have the advantage that they are debounced. Standard 7I96S
firmware has INM on all isolated I/O pins (the GPIO pins work but the built-in
debounce is often an advantage)

You should still be able to install it without internet or broken installer but apt will point to the CD only. I think it will get confused in some locales because the installer defaults the mirror and in some countries, this fails.

There is an issue with Debian Live which we use for the installer. Debian moved to using debian deb822 format for sources. This places the sources in /etc/apt/sources.list.d/debian.sources instead of /etc/apt/sources.list

Debian live has not caught up with this change.
so assuming you do get a bootable system, what does
cat /etc/apt/sources.list display? is it just the CD or DVD?

You could create Debian.sources in the folder above that contains
Then delete /etc/apt/sources.list
and then type sudo apt update
 

I think we are working at seperate parts of the code base. I suspect the info you want is probably thrown away after its buffered but I've never looked. You could look at adding your own NML as described here
linuxcnc.org/docs/stable/html/code/code-notes.html
The diagrams are also pertinent

What is visible in real time (motion) is limited to what's in state_tag.h I linked to and its populated in interp_convert.cc.

There is a high level API based on my ideas using gRPC but we've not tested anything. Its limited to what's available in the Python API
github.com/joco-nz/linuxcnc-Headless-UI
This is designed to address the authentication and security constructs the modern connected world demands.

There is also an effort to use a standards based methodology used by Haas and others but I have not been able to discuss it with the author at this stage. But we will schedule that in the short term.

For me it really simple. I rely on my machines for my livelihood. Once I had it working satisfactorily, there was no need for any updates. These controls function exactly the same as the original controls, an update isn't going to change that.

So all is working now....thanks again for your help...I am not sure what INM pins are...would I need another board?

Thanks again!,

Rick

what are the differences between the 7i97t_7i74d.bin and the 7i97t_7i74d2.bin
What should i use?
flashing direcly from mesact is a good idea?

I managed to get the machine to power up and boot into LinuxCNC, and I’ve put together a basic configuration just to test it on the bench.

What would you do next? I was planning to load the driver for the 7i74 onto my 7i97T so I could configure communications with the other boards.

I have downloaded and flashed Debian Trixy Linuxcnc from the download page twice now. And every time I try to install, it gets stuck at configuring the sources. I do not think it is a network problem because when I run the live instance, I can browse the internet. Please can someone offer a solution, or tell me where I can download the Bookworm version, which I know works.
Thanks
Aaron

if I go to V- instead of power supply ground IN2 lights...is that ok? or does it indicate a different problem?

I forgot about that.....no