Advanced Search

So I have a Precix machine I picked up a while ago. This thread has been a HUGE help in finding some information that is kept secret. If you own a Precix machine, you may know what I am getting at.

Anyways, I wanted to join in on getting more information out there for it, mainly focusing on the encoder side and how I am managing it using some DC drivers. These drivers are a bit older, but using the Gecko G320 DC servo drive units along with the C46 Differential Single-Ended Converter. Taking the encoder that you have established for the wires coming from the machine, converting the singles to a DC signal for the Gecko Drivers to be able to understand. Again, this is a bit older stuff, but for anyone retrofitting the Precix machine, this will be a key factor, I promise you!

I did go a different route using a C11G breakout board instead of the Mesa Card using the LinuxCNC, but that's because I already had the board for another project and was just using what I had on hand. Check out the photo I included to help anyone else out there.

I will be doing a full wiring diagram on the machine once I sort out some of the other items suchs as spindle wiring, limit switches and such. I'd like to be able to use a RapidChangeATC (rapidchangeatc.com/) for this machine, but not sure the collet and nut would match up correctly. 

randypetersen wrote:

As always thanks for everyone's input!

I updated my wiring diagram per PCW's recommendation for the drive interface. I also took a guess on the encoder interface with the details I got from Magtronic. Hopefully someone can tell me if it make sense or not.

I took off the CAP of the encoder and would have to start to take it apart further to find details PCW mentioned.

I am thinking with the manual I found and the drawing Magtronic sent I can be fairly confident in the wiring, what do you think?

Anyone want to see more pictures (why I don't know) can see them here:
photos.app.goo.gl/wBgjA9fuxgXsEZX96

don't rely solely on the machine setup files, also look over the docs config section that is more kept updated and relevant, the machine setup files are a very very basic setup and may actually be out of date, i need to look at that when i have a moment to sit down at dev next.

use the docs those have been recently updated and contain the more current necessary items you may be missing.

probe basic itself has requirements beyond axis so again only rely on axis as a starting point, simply changing display is not going to work without the remaining items being present.

the post gui hal file CANNOT be removed the base items in the supplied postgui are requirements not optional if you do not use the stock items in the probe_basic_postgui.hal file, it will 100% have issues.

If you are using qtplasmac, I would upgrade to Linuxcnc 2.9.10 using the ISO. Seeing Python 2.7 errors raises alarm bells for me
Also, you may find with qtplasmac you will get better results with master branch so it would be good to learn how to build LinuxCNC from source. Its pretty simple these days. QTplasmac is still being pretty actively developed.

Please  try using a HOME_OFFSET of 0.1 inch. You want the axis to be off the home switch after homing. Not sure whether this should be positive or negative. If it fails, change the sign and try again...

2.9.10 is the current release so that is all good. Iwould expect you are therefore running Debian 13 if you used the LinuxCNC ISO

Well you got further than me years ago, you were able to boot the Baytrail PC (but mine was a J1850).
Isolating the sibling core is recommended. If you are using a NIC with an RT app that uses networking, its recommended to pin the NIC IRQ to the same core as the RT thread. Latency may be worse  but the network latency becomes the bottleneck. I could not see what you were pinning.

For sale new atc changers Deta International.
Location Poland
Model R3DE4  Cat 40
Price 3500E
www.deta.com.tw/For-VMC/30-disc-typeFeatures of Tool Magazine

1. High-Precision, Low-Noise Drive System
   Features a dual-roller Geneva mechanism combined with a worm gear reducer for precise indexing and positioning. Ensures stable operation, reduces motor load, and achieves high energy efficiency.
2. Highly Durable Tool Tilting Mechanism
   Employs a slim pneumatic cylinder on the rail to withstand side forces during tool tilting and provide a cushioning function.
   Equipped with a locking pin mechanism to prevent the tool pot from shifting in the event of air source interruption.
3. High-Strength, Lightweight Structure
   Key components are made of high-strength aluminum alloy, optimizing both rigidity and light weight.
4. Modular Tool Arm Design
   Modular design separates the main body and the gripper for quick and simple specification changes.
   Utilizes German spring steel wire with galvanization for superior anti-rust performance.
   Integrated drainage function prevents coolant accumulation, ensuring reliable operation
5. Components for conversion to smaller tool holders—such as BT30 or SK30—can be purchased from the manufacturer.
6. 
   

NWE wrote:

@tuxcnc, (...) I don't understand why the message is useless.

.

tuxcnc post=349633 userid=3011
but it doesn't tell you where that empty string came from.

@tuxcnc, I agree python 2.7 is ancient. I don't understand why the message is useless.

It says file ./plasmac/plasmac_gcode.py line 39
Popen() apparently returned an empty string to int() which needs some character. 

My small contribution compared to all the help I’ve received.

does the log show anything?

Yes, but since the hang doesn't happen at exactly the same point each time it's a bit murky.

After the first few tries the startup attempts have been with the postgui.hal commented out of INI completely.  I tried starting with every non-mandatory feature/setting commented out - no dice.

And the Axis good start-up is the same config.  Just switch [DISPLAY]DISPLAY = probe_basic to [DISPLAY]DISPLAY = axis.  No hal-related issues for the same exact LCNC connections.

As for the other probe basic INI stuff, I copy-pasted from the machine_setup_files INI.  You're welcome to have a look at my INI to see if I fat-fingered something.

if the sims don't hang, it is almost 100% in your configuration somewhere. may be missing an ini setting somewhere.. i would recommend going through your ini line my line and checking against the docs for configuration to see if something may be missing. then i would go over your post gui hal file and make sure nothing is wonky there, then check your main hal file and make sure everything is correctly spelled, i have had a typo plague me before and it took me weeks before i finally found the issue.

having a working axis config and a working probe basic config with all the extras is pretty different.  probe basics functionality is vastly larger and requires more things be in place at startup.  using axis as a base test is a good start but it is not a definitive proof the issue lies elsewhere than the config.

the sims use pretty much a complete running simulation of everything needed including atc and probing, so if the sims work consistently you know the software and your install are good.  that is proven so the changes between the axis config and the probe basic config is where to focus your attention.

does the log show anything?

You might plot the switch  hal signals with halscope when homing to get a better
idea of whats going on.

(it might be contact bounce or EMI)

 

Worse than what you describe, apparently I've gotten a pre-release version? The grub.config line says its booting vmlinuz6-12.95+deb13-rt-amd64
the words debian 13 show on my screen after boot for awhile then go away, I got the iso file by searching the linuxcnc site because the link to download wouldn't work for me .
on the other hand, linuxcnc says its axis.ngc - Axes 2.9.10 on LinuxCNC-HAL-SIM-AXIS
and it will simulate some parts, I'm still checking. It seems to be working very good so far, internet works, playing mp3's work, terminal works, selecting applications seems to work, usb drives seem to work.
probably too early to submit bug report here but thanks for the reply.