Advanced Search

Hi
I have seen recent topic of @tommylight about filter for pts files. This help me to start writing the postprocessor for apt files.

I can manage simple words and convert to gcode.
Next to steps :
1- process circular interpolations
2- make it working in LinuxCNC as program filter, actually only works stand alone.

I am sure that’s python code is not as beautiful as it should be but that’s working.

Hmm, this might work
parport.<p>.pin-<n>-out-invert 

 
I decided to build a parallel port LinuxCNC config cos I'd never done one before. All went very easily after firing up stepconf which got motors moving straight out of the box.

But now I got problems with the PWM to control the spindle because the volts were running the wrong way eg 24000 rpm = 0 volts and 0 rpm = 10 volts. I thought it would be easy to invert the signal but to  no avail.

I tried to adjust with a sum2 component so -24000 + Commanded RPM = - desired RPM. This turned out to be a dry gully too.
I figured if I had to  convert it to an absolute number I could do that but halshow promptly showed me that -24000 + 12000 = 0
I got rid of the negative and used a gain of -1 and I still got the same wrong answer...

So what is the trick to invert the PWM value?

Most people on this forum in your position have ditched CommandCNC and moved to QTplasmac with a Mesa THCAD to measure torch voltage and let Linuxcnc manage the THC function internally.

Indeed, tracked this to Python deprecating 'locale.format()' and replacing it with 'locale.format_string()':

github.com/LinuxCNC/linuxcnc/pull/4250/changes

Note that this issue is also present in some other places (eg the offsetpage)

Since this affects python code you could patch the files locally.
I think you would find them in:
'/usr/lib/python3/dist-packages/gladevcp/'

You will need sudo rights to change those files.

Time to dust off this thread with a small update.

I finally had some free time, so I decided to make some progress on the electrical cabinet. We created a dedicated section for the two braking resistors (2 × 1.5 kW) and the transformer for the axis servo motors. We also installed a steel mounting panel where all of the electrical components will be neatly arranged and secured.










And here's the donor machine that I'll be salvaging some of the smaller parts from. ?




This is probably a good time to mention that if anyone is looking for Fanuc servo motors or drives, feel free to get in touch. ?

 

DMoneyAllstar wrote:

Running CommandCNC on Linux.

 

I believe you might want to contact CandCNC tech support.

DMoneyAllstar wrote:

Was setting up a run, zero'd my XYZ, loaded my file, and pressed Run. The torch made the first pierce but sat there without any XY motion. Air flow shut off a few seconds later. Recreated the .tap file a couple times with same result when running. When I rebooted and restarted CommandCNC, I got an error message for LinuxCNC. No matter what I do, I can't get out of this loop.

LinuxCNC does not use .tap files, so what gives?

Running CommandCNC on Linux. I've had this system since 2017, but the mini PC is 3-4yr newer. Everything worked fine 3wk ago. Machine is in an air-conditioned shop.

Was setting up a run, zero'd my XYZ, loaded my file, and pressed Run. The torch made the first pierce but sat there without any XY motion. Air flow shut off a few seconds later. Recreated the .tap file a couple times with same result when running. When I rebooted and restarted CommandCNC, I got an error message for LinuxCNC. No matter what I do, I can't get out of this loop.

I saved the error message ZIP file. What do I do with it?

I pasted the error message in AI and it told me there's a idle-wait setting that I need to delete. Does this sound normal?

Can I boot with my factory files on the original USB and get around it?

Thanks,
Dan


PS: here is what AI is telling me to do based on the error message:

Your log output reveals a specific configuration issue right at the start of your boot line.
text
[ 0.000000] Malformed early option 'idle'
Use code with caution.

The Cause
Your Linux kernel parameters are fighting with each other. Look closely at your GRUB command line options from the log:
... isolcpus=1 idle=mwait intel_idle.max_cstate=1 ...
The argument idle=mwait is invalid (malformed) for this kernel architecture. Furthermore, explicitly forcing idle=mwait overrides and conflicts with intel_idle.max_cstate=1, which is the crucial setting LinuxCNC needs to stop the Intel CPU from falling into deep power-saving sleep states (which ruin real-time latency).

How to Fix It
You need to edit your GRUB boot configuration to remove the bad option and fix the kernel parameters.
Step 1: Temporarily edit GRUB to test the fix
Reboot your PC.
As it turns on, hold or tap the Shift key (or Esc) to bring up the GRUB boot menu.
Use the arrow keys to highlight your normal PREEMPT-RT kernel option and press e to edit it.
Locate the line starting with linux /boot/vmlinuz...
Move your cursor and completely delete idle=mwait from that line.
Press Ctrl+X or F10 to boot.

The 7I84U needs its own power supply.

You must supply VFIELDA and VFIELDB (Usually the same)
These set the output voltage and the input thresholds.
Normally VIN comes from VFIELDB (W1 LEFT=default)

 VFIELDA and VFIELDB are supplied via the ORANGE connector (TB1)
on the 7I84U.

There is a small 5V supply needed for the 7I84Us isolated serial interface
This comes from the RJ-45 connector, and in turn from either the 7I74 or
7I97TA.

 

That's a very detailed answer, thank you! I'll read through that a few times in detail, much appreciated!

I flash the 7i97t with the 7i97t_7i74d2.bin.
now i have a bit of a trouble cause i looking to connect the 7i84u, but the trouble is i can´t understand correctly if i need to supply 5v to the 7i84u or is it supply by the 7i74 or its switched by VIN (and its need to be 8 to 32V) right now im working with all the parts in the desktop and dont have a 24V line here. i need to supply the voltage to VIN or im passing somenthing by.
Thanks

Good afternoon eveyone!
First the video of what i have been working on and dealing with, smashing my head against the wall, too much coffee and a lot of thinking but the results are spectacular!

youtube.com/shorts/jgv-_DtneDw?si=rTl5A_pSrwuul9d2


Time for the photos:
 
 
 

Now the description:
We were presented with a blown $iemen$ controller and our customer wanted drivers + cnc fully replaced. Machine is a huge vertical lathe, 1200mm chuck with automatic chuck changer, real C axis (with cartesian mapping and cylindrical mapping functiones needed), X and Z axis with tool changing system for lathe or mill tools. 16 tool magazine and tool probe system. To all off this add low pressure and high pressure coolant, hydraulics, tool clamp/release system, gearbox for main spindle and tool spindle, c axis brake, gearsbox oil chillers, lube system for guides and a bunch of other systems. Also dont forget the alarms and conditions for all this systems!!!
So, in summary, he complexity was not on cnc side as most of the functions we have already solved BUT the plc logic was a no go with the actual classicladder we have been using, we needed a more robust and scalable solution. What you can see on the video is what we got!

Now to the internal workings!
First needed to solve the communication with the cnc<->plc, i wanted something similar to $inumerik where a DB exist for the NC, then DBs for each channel, each axis, each BAG (mode group) and so on...Thats exactly what was done.
After that a way to send this 1500 bytes and receive 1000 bytes from the CNC, the way i found is using Codesys NVL (network variable) system. A udp based way of sending/receiving data amongs codesys plcs, nothing that a little reverse engineer cannot solve and make it communicate with PC<->PLC. 
Next problem: linux network stack, to slow for handling upd message burst so solution was given by Xenomai and RTNet! Not much info about it on the net but after some nights, coffee and wall-head-hitting, eveything went smooth.

So now i got this arquitecture:
 


Best regards,
Franco

G'day,

Since I updated from 2.7 to 2.9 on the lathe, with the latest GMOCCAPY, I can't enter lathe tool angles anymore. 


Any suggestions appreciated.
Cheers,
Mark

240VAC line rectified and filtered via DC capacitors to 340VDC to power a ~200VAC motor via inverter should be fine as long as the parameter(s) for setting motor rated voltage and v/f ramp (if applicable) are set accordingly. You just can't run such a motor across the line with a switch or contactor.

The few times I have seen this done using VFDs or servo drives, it was working well. I am not familiar enough with stmbl to tell you whether they have the same capability.

1. First confirm your 220VAC power. In my area our 220VAC usually measures around 240~244VAC, just make sure you're not using something like 380VAC.

2. The basic idea for your DC bus is AC through appropriately sized fuses or circuit breaker into the AC side of the bridge rectifier, the DC terminals of the rectifier feed the DC capacitor bank.

3. In practice you will also need a precharge circuit and a discharge resistor.

4. For the bridge rectifier I might use a GBPC3510 or comparable. It has to be bolted to an aluminum or copper heat sink, the required size of heat sink will depend on the ambient temperature and quantity of airflow. It is possible to calculate the heat sink requirement but there are so many variables, I simply use what I think and monitor the temperature during the first runs. If it gets hot the heat sink is too small. An IR temperature gun works great for this. For an application this size I might simply attach the bridge rectifier to the back or sub-panel of the steel electrical box housing the project.

5. I tend to size my capacitor banks by looking at comparably sized (Kw) VFDs or servo inverters and using about the same total uF. It can also be calculated using formulas. I keep calling it a capacitor bank. You might just need one single capacitor for this size project.

6. For the precharge circuits there are three popular ways. If your 220VAC power source has a high enough resistance you might do without the precharge circuit but I would not try that.

6.1. The cheap simple way is to insert an NTC thermistor in series with the AC input or DC output of the bridge rectifier feeding the capacitor bank. I have not attempted doing it this way but I see it a lot on mass produced inverter motor drives. The catch is appropriately sizing the NTC. I would tend to buy a couple different sizes so I can try each, measuring the start-up surge and running temperature. These WILL run HOT by design, probably over 100C.

6.2. I always install a single pole contactor or relay on one AC power lead for the bridge rectifier. An on-delay timer powers this contactor coil. A power resistor across the contactor points trickle charges the capacitor. After a few seconds the timer activates the contactor, allowing full power flow.

6.3. A third precharge method worth mentioning which I consider out of scope for a project of this size is the active rectifier.

7. A relatively high ohm resistor across the + and - of the capacitor bank will suffice for the discharge resistor, to ensure the capacitor is discharged in a few minutes after power is removed.