Advanced Search

Bumping an old thread… Can these g37 & g37.x macros from Pathpilot be used in LinuxCNC and is the code available? I’m also looking to use a tool setter to automate tool length measurements after M6.

Linux noob. Downloaded latest Live image. Burned on DVD. Put in old Compaq tower running Vista. Booted from DVD, everything worked. Installed to hard drive. Worked. Poked around in a few directories. Set time zone. Went to bed. Now it boots, gets to login screen, enter id and password, logs in and loses display. No clue where/how to troubleshoot. Any help greatfully accepted! Thanks,
Darren.

Hi Rod, Thanks for the feedback. Unfortunately it does not have THCAD5, and this is what I am using. So calculator does not allow me to double check. It does have THCAD10 and THCAD300. So I have set to 10, and the Voltage scale is 1/2. So hopefully this makes sense. The setting I have are now: THCAD2: Low (i.e. Down) 5V, Frequency divider: 1/32. 0V 95.4, 5V 896.0. Gives me the following as set by pnconf: Voltage Scale 0.005996, Voltage Offset 2981.25. Plasma machine: Unimig Razorcut 45 CNC port is set to 30:1 ratio. I set the Plasma to 30:1 ration as a precaution. Do people think this is necessary, or would 20:1 be fine? It is a 45A plasma cutter.

Hello everyone, I am working on an upgrade for my old lathe. In short, I want to replace the old main motor with its multiple belt stages as well as the change gear system for the lead screw with two servos (more detail below). I am wondering whether or not LinuxCNC is the right way to go about it. I am a mechanical engineer by trade, but have little to no experience with control systems or LinuxCNC, so I do apologize in advance for any potential stupid questions coming up. 

The long version:
I want to upgrade my lathes drive system. It's an old machine, build somewhere around 1945, and it uses an induction motor followed by two v-belt, one flat belt and an optional gear stage to set the spindle speed, giving me a grand total of 12 possible spindle speeds to select from. Changing speeds therefore also isn't the quickest process. Even worse is the feed gearing driving the lead screw; It's set up as a change gear type system, so to set the feed rate I have to disassemble and then reassemble the drive train, which makes for painfully slow changes in feed rate. While the machine is certainly mechanically beautiful, my shop is not a museum and my time is quite precious to me, so I want to change that. The plan is to replace the main spindle motor and drive train with a big servo, and the feed gearing with another, smaller, servo. That would allow for easy changes of my feeds and speeds, as I just need to set whatever value I want on the controller. The feed rate per revolution would then no longer be controlled mechanically through the feed gearing, but electrically, by the smaller servo acting as a virtual reduction gearing if you will.

I already sourced the servos, a 1kW for the spindle and a 400W for the lead screw. They use EtherCAT for communication. For the controller I was thinking about either a Raspberry Pi 4 or a Kunbus RevPi that I have lying around. To run the machine I want to use LinuxCNC and an open source EtherCAT master, as done by Marco Reps for example.

I am however unsure on how to go about it. I don't yet require the full CNC potential of LinuxCNC; That might come at a later point, but for now, all I need are the following two features:

• Set the spindle's rpm
• Set the feed rate

Both of those options I would like to have on an easy GUI, ideally with as little unused "clutter" as possible. Depending on how hard that is to pull off, I would not mind learning how to program/code/design that myself, it might however quite likely be far above my current skill level. I am a mechanical engineer after all, not a software wizard. 

For the actual servo control I would like the option of having soft start and stop options by controlling the acceleration and jerk, so that the servo doesn't just violently try to reach the set spindle rpm.
Ideally the feed servo would be set up as a slave, following the main spindles rpm precisely, i.e. perfectly mimicking a mechanical gear train. So if for example the main spindle bogs down during thread cutting, the feed servo should ideally not just run it's course and lose synchronization. That should certainly be possible via EtherCAT, I just don't know how hard that would be to set up with LinuxCNC. It would however also be acceptable to solve that via a maximum allowable tracking error for the servos, if full on synchronization doesn't work.

I have a lot of ideas and things I would at some point like to add/have, from motor load indicators, a DRO via linear encoders, small easily programmable cycles all the way up to full on CNCing the lathe. But for now its baby steps, i.e. getting the lathe functional with its new servo drive train.

Is what im having in mind possible with LinuxCNC? Does anyone have pointers to a similar project? Or any advice or resources on how to set this up in LinuxCNC? How do I go about setting up the big servo as a spindle motor? Can I actually use the smaller servo as a slave synchronized to the main spindle motor? Is it possible to create a simple custom GUI showing only spindle speed and feed rate? Or would you guys recommend an entirely different route to set this up?

Any advice is greatly appreciated. I apologize if the post became a bit "rambly" in places and for any stupid/basic questions contained in it.

Cheers,
Nik

The 7i92TF is the best fit in my situation.

The manufacturer claims the pinout of the parallel connection to the drivers/switch interface is parallel port compatible but does not state signal levels for each pin. Multiple manufactures do supply devices for it. The final test of cause is in trying it with a 7i92T module.

Thanks for the advise.

Kind regards.

Did you add it to the LinuxCNC repo or do you plan so?
Would be nice to have it in the sources if not already there.

If you want a parallel port equivalent interface, a 7I92TF or  7I92TM might
make more sense, as these both have 2x parallel port equivalent interfaces

If you wanted to use a 7I96 or 7I96S, you could use its P1 parallel expansion
to connect to the parallel interfaced device.

The real question however is what is the pinout of the parallel connection
to the drives/switch interfaces. If it's truly parallel port compatible, a 7I92T(F,M) or
7I96(S) will work. If it's some proprietary interface from the USB card to a
custom driver card, they will not.
 

Hi everybody,

I'm looking for an advise to by an Mese 7i96S card or not.

This is why the question came up. I'm leaving Window (for good) but I have a desktop CNC machine that is depending on it, sort of...
I've bought new hardware and it is running Debian Trixie (with Xfce) as the primary OS, dual boot for the moment with windows 11.

On my old hardware I have the following line of control to my CNC machine.
Windows -> (usb) --> proprietary controller --> (D25 parallel bus connector) --> power stage --> CNC steppers and switches.
There is some proprietary controller communicating with windows driving the CNC machine through a D25 parallel port compatible connection. A power stage electronics board is driving all steppers on X/Y/Z/A axis (and providing signalling in return.

I want to replace the proprietary controller by an Mesa 7i96S card. But also want to verify that it is a valid option to go for.
The new configuration will be:
Debian trixie (RT kernel) + LinuxCNC -> (ethernet) --> 7i96S controler -> D25 parallel bus connector --> power stage --> CNC steppers and switches.
I have found out that there is a one-to-one match between signals on the 7i96S TB1, TP2 and TB3 connectors and the D25 parallel bus connectors but can not find much information on the actual D25 parallel bus signal levels at the desktop CNC side. I expect then to be in 'normal' ranges for a parallel bus connection.

My questions are:
- Is this a 'standard' way of applying a 7i96S controller card?
- Are the Mesa 7i96S TB1, TB2 and TB3 connections capable of driving a D25 parallel bus?
- Are there any pitfalls that I should be aware of when I follow this path?

Kind regards.

 

no braking features.
noise, I'm not sure how to really tell if there was. the cables are all cat6.

Hello linuxcnc members,
I would like to use an AX5201 servo driver with LINUXCNC. The AX5201 is setup with TC3 and works fine within the TC3 enviroment. 
The firmware of AX5201 is 1.06 (Build0030) from 11.5.2015 and it seems it can not be updated to version 2.x.
In TC3, the driver goes only on stable OP mode, when the cycle time is 2ms. So the parameters in the ini file were adapted to 2ms cycle time.
The ethercat-conf.xml was created with the great conversion tool from Endian and is loaded and seems to have the correct DC, but the driver stucks in PREOP mode.
The startup-XML was exported from TC3 startup commands. 

How can I find out which parameter causes the mode change problem. 

Thanks for your help in advance!

 
Here the logs:

Trying to setup Probe Basic Lathe for use with an EMCO Compact 5 CNC lathe with my own electronics.
After the splash screen there is about a 30 second delay before the main GUI appears.
All looks okay, the tabs work but the E-Stop won't click...
Note: In AXIS the E-Stop is correctly asserted for the 7i96s, and the power button works to enable a 7i96s output.
I think it is a python thing, but don't know where to look.

Which version of Remora did you flash? Luckily, it worked with the latest beta. I've attached some files that work for me. Would you also like the pinout for the DM556Y drivers for the EC500?

I get it to work finally. Idk what im doing wrong. After flashing newest beta its working

Another thing that comes to mind, do your servos have braking feature? The clunking and the resonance could be related to that if you do have brakes, possibly they are not being properly managed? I use braking on my Z servo drive, and if I lock it when idle I have resonance as well, that's the only thing that comes to mind, also it does make a noise when it engages more of a click than a clunk, just my brain went there...

Never had any of these issues, could be something with your configuration, ini Hal XML or the servo parametrization.
What my patch definitely improved is phasing, what could happen before the patch, was wrong phasing where latency could happen to sit on the threshold of when the servos were expecting the packets, and they could arrive at different cycles, making the motor grind, but all your listed issues probably have a different root cause. If you shared your configuration I could see if anything stands out to me. Other than settings it kind of sounds like an electromagnetic noise issue, are you using shielded cables for the ethercat network connections? Are all components properly grounded? Mostly the servo motors, and if you have a spindle with a VFD those need to be properly shielded filtered and grounded.