Advanced Search

saxa wrote:

Now recently I have visited some exhibition and I saw today most of industrial things are on EtherCAT.

Yes, it does seem so...

saxa wrote:

My plan is to make linuxcnc act as an master node, and send signals to the motion controller which will then through some adapters transform the signal to pulse/dir and send to the drives.

That would be redundant as LinuxCNC is the motion controller.
In short:
-buy a Mesa 7i96S and a THCAD
-a used Entrprise PC
-A network cable
-Wire, install LinuxCNC, start wiring drives and motors
-those drives will have to be set in position mode, hence 7i96S
-yeah, wiring, you will need very good quality shielded cables for everything as cables that long will collect a lot of interference, not very important for oxy cutting, but a mist for plasma cutting.
-....well, that is it.

Very strange, Do the GPIO inputs work?
 (GPIO0..GPIO10 match INM0..INM10 on a 7I96S)

Hello,

I have a 7I96S that has developed a specific fault: the 11 isolated field inputs (TB3-01…11, the InM module) stopped updating during operation. They read the correct state once when the hostmot2 driver loads, and then freeze — actuating a switch or the E-stop no longer changes the value in software. The card's own input status LEDs do still follow the switches correctly.

Everything else on the card works normally: Ethernet comms are solid (no packet errors, no watchdog trips), all 5 stepgens drive motors correctly, PWM works.

Setup
- 7I96S, Smart Serial Firmware Version 43, BoardName MESA7I96 per mesaflash --readhmid
- LinuxCNC 2.9.10 (uspace), hostmot2 0.15, hm2_eth 0.2
- Raspberry Pi CM5, Debian 12, PREEMPT_RT kernel; card on a dedicated NIC at 192.168.1.121
- Standard 24 V field sensors; input common (TB3-12) tied to field 0 V

What I have established by testing
1. It worked, then stopped. Early on, inputs updated live (two limit inputs flipped TRUE→FALSE on halmeter when actuated). Later in the same session they froze and have not recovered.
2. Not comms/packet loss. With a stepgen commanded to move, stepgen.NN.position-fb updates live in the same read cycle that the InM inputs stay frozen. A lost/stale read packet would freeze position feedback too — it doesn't.
3. Not debounce. raw-input-NN (undebounced) is frozen as well, so the scan itself is not advancing (scan_rate/fast_scans/slow_scans at defaults).
4. Survives full cold power cycles. The card powers down completely with the Pi; it was fully powered off several times after the freeze and stayed frozen.
5. Firmware config intact. mesaflash --readhmid reads a complete, correct HostMot2 cM module (base 8500).
6. Adding hm2_7i96s.0.read-request to the servo thread did not change the behaviour.

Possible cause on my end: a contactor in the machine failed (fried) shortly before the inputs froze. Its coil ran on the 24V field power. I suspect a transient may have coupled into the input section via a shared ground.

My question: given the inputs read correctly at load but never update live, while thunctional and the firmware reads back intact — does this look like a damaged inputsection that needs repair/replacement, or is there something else I can test first?

Thank you,
Andreas

I corrected and tested it. A few words of explanation:
The 7i92 inputs are controlled by a short to ground. If none of the axis selection inputs is shorted to ground, then none of the axis moves, if one or several, the selected axes move.
I added the line "setp mux4.0.in3 0"
If both coefficient selection signals are shorted to ground, the axis will move, but at zero speed, so the effect will be the same as if no axis was selected. (You can change it to "setp mux4.0.in0 0", then there will be no movement when no selection input is shorted to ground).
I did this to have a wider selection of supported switches. It is possible that someone will have, for example, four axes and a four-position switch, then they would not be able to select the "no axis" option, which is necessary.
It is also worth mentioning how to control two inputs with a single-section switch.
You need to use diodes (anode for the input of the mesa, cathode for the switch). Then, a short circuit to ground of one input does not affect the status of the other inputs. I won't describe it in detail, if someone needs it, they will find it.

I went for the skookum SMPS on my Deckel FP2NC. 240VAC input, 160VDC 5KW output to match the existing Siemens servos. I found this item as cheap surplus, and bought two for insurance.

For the spindle I used a no-name import 220V-to-380V inverting VFD. Name-brand examples from ATO below:

www.ato.com/220v-to-380v-vfd

"-not" works for inputs, but not outputs
(without some fairly micky-mouse logic)

Hello all, I am rebuilding an old gantry type machine which I plan to use for cutting mostly with oxyfuel and eventually for thinner plates with the plasma source.  The machine is 10m long on Y (longitudinal axis) and about 5m wide on X (transversal axis).
I would like to use linuxcnc and I have bought many years ago the servo motors and drives from DMM Tech, more precisely the DYN4 
series. I have 2 motors for the Y axis which I plan to use in parallel (1kW each motor) and a smaller motor to move the torches on the X
axis (a 750W one). All motors have a single turn absolute encoder.

Now recently I have visited some exhibition and I saw today most of industrial things are on EtherCAT. I saw that LinuxCNC supports it 
in the latest versions, but since I have no experience with EtherCAT I would like to ask if I have to pay some special attention to certain things or there are no troubles at all with the compatibility ?

My plan is to make linuxcnc act as an master node, and send signals to the motion controller which will then through some adapters transform the signal to pulse/dir and send to the drives.

This is the general idea on how this should be working. 

Can I hear some comments from you guys , anybody have a similar setup , should I go some other way around the thing ?
Thanks for the time.

Make sure you test your kinematics for A=0 , C=0 first so you can confirm that the kinematic model returns proper results. Also make sure that the coordinate values switch between the same values when you change between trivial and RTCP kinematics.

have a look at the millturn simulation configuration:
github.com/LinuxCNC/linuxcnc/tree/master...xis/vismach/millturn

The limits are switched between two sets configured in the ini file.
See 'mcodes' for the switching scripts and the files in 'remap_subs' that actually call them.

Hello, I've encountered a similar problem. Where can I find an example of a milling machine with switchable soft limits?

Hello everyone, I'm a DIY enthusiast. My English isn't great, but I'll try my best to explain clearly. I'm encountering a problem while using LinuxCNC as the operating system for my machine tool. When I replace the existing kinematics system with my own custom-built five-axis kinematics system, the machine tool coordinate values ​​change when I switch to RTCP mode at the current position, causing a system alarm because the machine tool coordinate system has exceeded its limits. My understanding is that enabling RTCP changes the workpiece coordinate system values, not the machine tool coordinate system values. I haven't been able to solve this problem yet.

Additionally, my machine tool structure is A (head), C (table), with the distance from the C rotation center to the machine tool zero point being (Xrc, Yrc, Zrc). The tool length is set to dt, and there is an offset between the two rotary axes with offset values ​​of (Xos, Yos, Zos).

Please share zipped configuration folder, or whatever you are willing to share to shed light on the issue.

I'm not sure on this but I thought there was a change to support 64 bit floats for instanceslike this that managed the wrap but I have no experience. There is a big body of work pending in master branch around 64 bit pins. Message Grandixximo on Discord.

Alright, the 4th axis joint error inverting is 32 bit count roll over.

I am getting confusing info in my searches, but there seems like there is a hall function to manage this? Anyone have some insight as how you implement it?

thanks!

ok, the joint error is something else. it happens at 912.45 ALWAYS. it must and over run of something. we have like 800 million counts per turn of the axis, so its inverting at 2 billion encoder counts.

Ill put this in my other thread as it i dont think has to do with the ethercat jitter problem.