Advanced Search

stkvit wrote:

I converted the bcl-amp output to a differential signal (max485), at a long length it works much more stably and reliably. You should do the same (I understand that your frequency is linear and depends only on the height). A couple more questions, your sensor have temperature stabilization?

I was thinking about differential signals and if there is such a need I will introduce them, we will see after the tests. There was only one obstacle - I wanted to leave the size of the device exactly like most BCL type sensors so that it would fit perfectly in the place designated for it. I did not find a suitable connector convenient to use to accommodate so many cables. I thought about RJ45 but it gave rise to other problems. 

Fortunately, such a converter is not a problem and it can be added even externally along with galvanic separation, which is missing here. 

Of course the temperature stabilization is needed - now introduced very basic and weak, it requires refinement. This still requires a lot of time. I have to collect a lot of measurement data and refine the algorithm but I am prepared for it. The current implementation should be completely sufficient for tests. There are other sensors apart from temperature that are more difficult to control and that spoil the measurement during cutting (dielectric parameters of the gas surrounding the nozzle).

We have built a CNC machine using open-loop stepper motors. The motor drivers are connected to a computer via a breakout board, which is linked using a DB25 cable. Most of the wiring is done using jumper wires, and we have a common ground for all electronic components. The Y-axis is driven by two stepper motors, one for each arm.To improve signal stability, we have added pull-down resistors to each Step- pin of the motor drivers. However, we have observed that the potential difference between the 5V pin and ground pin of the breakout board (powered by an 18V power supply) is only 1.6V. Additionally, when the Y-axis motor stops working, the potential difference between the 5V pin and the Y-axis direction pin drops to 0.06V.Thus, we suspect that the breakout board is damaged. The 5V pin on the breakout board should provide a stable 5V, but since it only outputs 1.6V, the board’s voltage regulator or power circuit may be faulty. The fact that the Y-axis direction pin drops to 0.06V when the motor stops further suggests a power issue or weak signal.Additionally, we suspect that electrical noise is affecting system stability, leading to random motor movements. Interestingly, noise issues only appear when the PC is connected to the breakout board. What could be the reason for the issue? Should we also connect breakout board with a usb cable to the PC for extra power. YOU can view our breakout board specifications in the file attached below. 

Hi Royka in the past you give me a link to a kernel.. the link no more work.. is there a new version of a kernel for the latest linuxcnc? i am on mesa.. rtpreempt 6.1.0-31.


the kernel you link me was :

royka post=299706 userid=31446
Which kernel do you run at Debian 10? "uname -a"
What is the result of "cat /proc/cmdline" on Debian 10 and 12?
What is the result of "cat /proc/interrupts" ?

This kernel I've compiled last week by the way (6.9.0-rc5):
drive.google.com/drive/folders/1L28lG0zD...2XQJx?usp=drive_link




can you please help me.. i make fresh install on the same machine and done everything but the kernel is needed to get it work perfect.... i see that the kernel which came with the latest linuxcnc iso doesnt work like my machine before...


thanks

I've never used rtai, but run latency-history and htop, sort by priority (F6?) and you'll see. I think you could use the same script.
Bad RAM memory could be a cause, usually using 2 ram modules (of the same) is better so it can use DDR. But if one of the two is bad removing one or the other might give a better result.
HDD can also be a cause, if you still use that buy a cheap SSD.
Disconnect any CD/DVD drives, SD-reader units.

1. Your ini is missing the spindle section while your hal references 'SPINDLE_9' ini entries. Looking at your hal file makes me think that you are not actually using a spindle setup. If that is the case then you might want to completely remove this section from your hal:

Yeh, that's the one! I made some enhancements for it and Les added it to Sheetcam. There are some settings at the top of the post to check and turn off. I made it so Sheetcam was the master seeing it was for his system

@ihavenofish
Wow that's nice. Having 3 ton machine moving like that also with the current jerkness is too scary definitely not good for my heart.
@LCR
I have similar issue with non-ethercat drivers. you can check it out this thread. It may help
forum.linuxcnc.org/38-general-linuxcnc-q...s-creeping-up#309973

I converted the bcl-amp output to a differential signal (max485), at a long length it works much more stably and reliably. You should do the same (I understand that your frequency is linear and depends only on the height). A couple more questions, your sensor have temperature stabilization?

So, I dug out my 14-year-old 6040 CNC again. 
I already retrofitted that thing once, tossed almost all the original electronics and went with G540 and some old industrial PC. Still, I was strapped for money, and it was a hack. I only used it occasionally for a couple of years, then it just collected dust. 
A few months back, I found myself in a situation where I had a lot of time, and I'm no longer that strapped for cash either.
I got myself a new set of closed loop steppers and CL57T drivers from SteppersOnline, MESA 7i96S board, proper PSUs, a new mini PC and nice steel enclosure, and I'm trying to push hard to do it properly this time. I have a physical layout somewhat complete but still grinding the schematics.

Would somebody be so nice as to look it over?

I have the following questions:
1) Is it ok to use 5V to enable stepper drivers (OUT0)? Stepper drivers will be configured to use 5V, but I'm assuming that DIP switch is only for STEP/DIR pins.
2) Since I'm feeding 24V to IN COM, I have to use 24V to detect alarms from stepper drivers. Is it ok to mix 24V and 5V on the same driver for different functions (5V for enable, 24V for alarm)?
3) Can I detect driver alarms with just one input? Any practical benefit have them wired individually?
4) Right now, I have all limit switches wired independently, but I really need that? Current setup has them all in series on a single wire. Is there any practical benefit of having the wired separately?
5) Not sure about E-Stop. In the schematics, it kills the VFD power and sends a signal to the controller. But I'm not sure about that at all.
6) I have two input pins left, but I still need a probe at least. Am I missing anything else? 

Thanks a lot for any suggestion.

To the OP just for reference next time could you specify if you are using a RPi4 or RPi5 so if we need to look at the images it makes life easier. As both the 4 & 5 come in B variants.

Cheers
Rob

Your ini start up line has ag20 in it, change to g21.

Homing sequence is set in the ini, read the lcnc docs for how to configure homing settings correctly. All homing is stock linuxcnc motion. Go to home is for the go to home button. You can edit it to suit your needs, but has nothing to do with homing. All homing settings are in the ini under each of the axis sections.

I finally got good results of making a servo out of inexpensive bldc motor and drive using 6i25_7i76_7i74 and 1 micron resolution linear encoder. I consistently get repeatable positioning of the axis within 4 micro meters measured from an external device. Now I'm wondering if there's a way to offload the PID calculations from running on servo thread to the FPGA? I'm guessing this will open up options for adaptive control algorithms to run on the fpga down the line. How much processing can we push on the fpga and the daughterboards? Does the daughterboards also have some hardware for processing?

weiiird. is that mm for f error?
what axis?

Mine are omron, so no help there but when they stop, they are within 2 encoder counts (20 bit). I wonder if there is a feed forward thing where its just not bothering to catch up after stopping. I think i had that on my brother analogue servos when i chose some bad settings.


(also , side note, i need some help with ethercat setup, haha)

that's still very decent. my machine is "faster than most". yours is more in line with what most people will be doing with these servos i expect.

omron 200w servo. 1G, 3000rpm. (cant wait to get it to 5000rpm muhahhaha!


jmc motors in use. 0.3G, 12m/m (pulse limitation). The real issue with these is they are soft. like even though they track with feed forward, stiffness is so low that cutting forces push the axis off track by a thou or 2. Need to take a spring pass ore 2 for a precise cut.


My hope is that these lichuans will be a step well above the JMC but still "cheap and cheerful" for my production machines. Cause those omrons cost more than the machine frame
 

In your ini file check this line

Line 277 for joint 4
STEP_SCALE = 3.9307 # 1414.752 steps / 360° for 10-tool carousel

AI isn't worth the effort.