Advanced Search

Hallo,
die gute Nachricht ist, es fängt an zu funktionieren.
Über die vermutlichen Ursachen werde ich demnächst noch etwas schreiben, ich habe momentan einen anderen PC verwendet, ich denke das das nicht das Problem ist. Die für das Projekt vorgesehene Hardware (MSI Windbox) hat scheinbar ein Bios Problem, Linux macht da Bootprobleme. Das ist aber eine andere Baustelle.
Der Reihe nach, geflasht wird der NVEM über einen STlink Adapter, ich verwendete hier eine Software unter MXlinux zum Flashen, ich grabe das jetzt mal nicht weiter aus. Interessant war das die Software das Flashen dieser Firmware hier zeitweilig verweigerte:
github.com/scottalford75/Remora-NVEM/tree/main/Firmware
Es gibt aber ein weiteres Verzeichnis hier:
github.com/scottalford75/Remora-NVEM/tre...-NVEM-STM32/Firmware
Diese ließ sich flashen, und die hatte ich bis gestern verwendet. Linuxcnc konnte mit dieser Firmware zumindest den NVEM finden, das Einschalten der "Maschine" fuktionierte, jeder Bewegungsversuch endete in obengenannten Fehlermeldungen.
Gestern hatte ich nun noch einmal den NVEM zerlegt, um den STlink anzuschließen und noch einen Versuch mit der erstgenannten Firmware zu machen. Das Ergebnis war dasselbe Flashen verweigert mit einer nichtssagenden Fehlermeldung. Noch ein Versuch, auf einmal Flashen erfolgreich. Keine Ahnung was da schiefgelaufen ist.
Heute den Testpc mit Linuxcnc und der Anpassung nach der Youtube Anleitung vom Scott aufgesetzt, Controller dran und siehe da: Alles gut.
Die Ausgänge liefern die Signale, keine Fehlermeldungen etc..
Es ist also so das der NVEM beim Flashen bereits eine Grundkonfiguration bekommt, die dann auch ansprechbar ist.
Ein Hinweis, bitte, das ist keine Kritik, vielmehr ein Hinweis auch die Akzeptanz dieser Software beträchtlich zu steigern, vielleicht kann Scott mal das Installationsvideo hier etwas anpassen:

Hier verwendet er alte Verzeichnisstrukturen und es fehlt ein wesentlicher Schritt, der halcompile Befehl beinhaltet hier nur das Compilieren der eth Komponente, es fehlt das Compilieren der Remora-nv Komponente, ohne diese läuft Linuxcnc ebenfalls nicht mit dem NVEM.
Ich bin also der neuen Steuerung und vor allem dem Handrad (GRBL hat das nicht) ein Stück näher. Ich denke das war noch nicht der letzte Post, danke auch an Luca bis hierhin ,
Gruß
Mario

 

Do you think the MPG XHC WHB04B 4 model will work well?

Wireless, wired, or does it not matter?

depending on the type f mpg it is possible. some usb mpg versions require some conflict resolution in the postgui hal but its not too bad.

Current mesaflash version is 3.5.16
 

use bookworm per the docs

I did that as well (think it ended with version .12 2-3 days ago) not sure. But not on both machine and at that point not systematic as there was no difference. But now I will do that again with the current setup... and epp...
I'll update... have a nice weekend...8)

NWE wrote:

 
I know the motor's phase inductance or voltage rating can make a huge difference in top speed attainable. These motors do not have either of those specs listed, making it harder to compare specs, but when I get a chance, I'll try to measure phase inductance.

Higher inductance motors need higher voltage to produce the same speed and torque. This is the first thing that comes to mind, seeing what these motors can do.
sed with steppers, the ones I previously used would never have got this speed on 24V, hardly even at 48V.


 

During COVID, I worked with Marshall, a US engineer as we were going to design a plasma table together. He (mostly) and I (the tester) developed an insanely comprehensive spreadsheet model that benchmarks a stepper drive for current consumption, voltage requirement ambiant heat generation. We wanted to utilize stepper motors insane low down torque for acceleration which we thought trunped a geared down servo (which is a lot easier to design for).
If you can find these specs, mainly inductance, I'll run it over your drive.
 
I ran countless NEMA 23,24 and 34 motors through this model.
I then validated his model by replacing motors on my plasma table with those that passed the spreadsheet test. 

After testing countless motors, For all that is said about low inductance, We busted that myth. Some of the better performing motors had quite high inductance. Rotor inertia seemed to be more imortant.
Some failed because they needed too much voltage (beyond what you could buy a power supply for,
Others failed because they drew to much current (above the motors rated amps),
And some failed because the estimated ambiant temperature eise would have melted the wiring (which is  usually rated at  about 130 deg C)
This particular Sanyo Denki has insanely high rotor inertia and weighs accordingly! I speculated that the reason this was limited to 2 amps was because there simply wasn't enough room left in the drive for heavier wire! On my plasma table X axis, I achieved 60000 mm/min rapids and 8000 mm/s/s acceleration with a 30mm dia pinion and 3:1 gearing. I paid about $20 AUD per drive on eBay and another $10 for the wiring harness. After this testing, I saw the guy had a few left and bought the lot. the RRP of the drives is still around AUD $250. They run silently as smooth as silk.

So I had a few left and thought I'd better use them. I have some DM542a stepper drives here I could have used but my pc doesn't have a parallel port so by the time I bought a card and breakouts or even a Chinese 7i92, I might as well buy a multiaxis Ethercat drive.

Anyway, that's why I am here. My next step is to check I have phase polarity correct. I have a feeling I might have changed something early in the piece trying to get things to work because I know these drives deliver.

No fault but seems like something very basic is missing/wrong
maybe in the hal file?
 

Hi guys, I know this thread is old, but I bought an equal body tool setter and... of course I could not managed it to work... so I opened it.
Just in case someone buy it and want to know how it should work.

Attached some photos (and so the correct, I think, wiring):

This wiring has two contacts switches only:
black and red (long springy contact)
yellow and green (short fix contact)

Under the metal button there are 2 plastic parts, the first being the top of the body of the setter with the hole to insert the setter pole and the spring between them, the second is inside and (should be) attached to the metal button pole. Inside there are other two poles as guides to the sliding plastic (and to impede rotations...).

The first plastic (the top) has a pcb with two exposed parts = contacts, the second plastic has the two metal contacts (one springy).

SO, it should work like this:

IDLE: when not pressed, the two black plastic parts are one near the other, so both the contacts are closed.

TOOL MEASUREMENT: when the tool press a little, the central pole (that should be attached to the bottom black plastic part..) goes down and moves the bottom black plastic part away from the upper black plastic part, so the fix contact (yellow wire to green wire) opens and the tool is set. At this stage the springy contact remains closed.

OVERPRESSED: if for some reason the green/yellow fix contact does not open, than the black plastics parts keep on moving away from each other and also the springy contact (black/red) will open.

Where are the problems?
1) The lower plastic that should be well attached to the button pole... it's not well attached (can slide!). Mine had a little spring-fixing washer but it was so poorly fixed that it came off... If it can slide, contacts distance changes.
Being not fixed, the total height of the probe can changed from measure to measure... and being different from the height programmed in the software... the tool results not correct...

2) The spring does not always push the button back to the idle position (where both contacts are closed), so the yellow/green contact remains open like when the tool is already pushing the button.

I will try to fix both problem but for sure this is an unreliable probe... My idea is to put a thread M3 on the bottom of the button pole
and put a nut over and under the plastic that should not slide on it, so to block it on the pole. To do this I will need to drill (mill?) a little inside that plastic because the pole is too short and there is no space for the nuts. I don't have an M3 female threading now... but.. in the future.. let's see.

I opened it because when I received mine, the yellow/green contacts were always opened (as in problem n 2).

Photos attached if interested.

Have a nice day, Alberto.







 

Konstantin wrote:

Hello,

I have made the decision to try LinuxCNC for my plasma machine that has been running with Mach3 so far.

Thebplan is to use EtherCAT servos for the motion of the machine. I installed LinuxCNC from the official site Downloads section, the latest stable version - 2.9.8. Then followed MrRodW's video about installing the EtherCAT master and CiA402. This is the - Video.

The computer I try to run LinuxCNC EtherCAT is a mini conputer a few years old with an HDD of 1TB.

The problem is that when I reched the step that requires using the halcompile command, it executes without any error but there is no cia402 folder in /linuxcnc/configs. Without having the folder, I cannot continue the procedure for CiA40w installation.

What could be the reason that prevents creation of the cia402 folder?

Best Regards
Konstantin Kolev
 

I just quickly checked the video you mentioned and I think you know have a working Ethercat/LinuxCNC. The next step will be to create your machine configuration: you'll need at least a "chooseAName.ini" file, a "chooseAName.hal" file and a "ethercat-conf.xml" file. Put them in a folder that you create in your /home/linuxcnc/configs/ folder. As 0x2102 said, there is an example for the .hal and .xml files on dbraun1981's github. For the ini file, if it is your first time with LinuxCNC, just start one of the simulation that has similar properties as your machine (3 axis? Gantry?) and have a look in the files that are generated automatically in the "linuxcnc/configs" folder. 

The documentation explains every thing in details and I can only recommend you to read it, but long story short, the ini file will describe your machine, the hal file will load components, create links between them and attach them to threads, and the xml file will describe the element in your cia402/ethercat component.

Hope this helps.
Cheers, Luca

Hi everyone, i have an industrial cnc plasma table made by CR Electronic on wich the numerical contol has stopped working and i would like to try and retrofit it with linux cnc and QTPlasmaC interface..
it has 3 axis for the usual X,Y,Z movements and then an additional axis named "K" that is basically a slave to the X axis. it is a tray connected to the air extraction/filtration system that travels underneath the table to catch all the sparks and fumes made when cutting.
it has servomotors on all axis with analog/digital drives controlled with  +/-10V

i have a few question that i'd like to ask before getting started with the retrofit:

What board should buy?
since it has the analog drives i think the ideal board for this retrofit would be the the Mesa 7i97T  combined with a THCAD2 for the THC control.
the plasma currently has a proprietary board wich handled the thc control and the Z axis motion. also it has ohmic sensing and torch breakaway detection. the voltage signal comes from the plasma generator wich is a Hypertherm HD3070 and outputs direct plasma voltage to the thc control, so i'll have to put a high voltage resistor prior to the THCAD2 board.

Do i need 1 or 2 THCAD2 board?
i've read multiple topics about retrofits on cnc plasma tables on the forum but can't figure out if i need to get 1 or 2 boards to combine with the 7i97T Board. can the 7i97T take the input from the plasma voltage to control the THC? or do i need to use the THCAD2 board an then connect the frequency output to the 7i97T? also since it has ohmic sensing and torch breakaway detection wich are 2 wires that trigger two different things, when the torch breakaway circuit opens it senses that the torch disconnected from it's magnetic base and for the ohmic sensing it had another wire on wich when the torch touched the metal plate it closed a circuit and sent the signal to the old thc board that the metal plate was found and the zero was set. can the torch breakaway wire and the ohmic sensig wire be connected directly to the 7i97T board or would i need another THCAD2 Board to have these 2 functions?

Z axis motor:
the old cnc control THC board controlled the Z axis motor wich is powered by 24VDC and since i will be removing that board i won't have a driver for that motor to use, i wanted to replace it with a new servomotor and i've seen the "stepperonline" 100w 230Vac servomotor with his dedicated analog drive,  would that be compatible to be controlled by the Mesa 7i97T? i'll add a picture and the manual to that. 

How to wire the analog servo drives to the 7i97T?
i've read multiple times the manual for the 7i97T board but i'm struggling a little bit to wrap my head around how i should connect the motor drives to the I/O board. i have all the wiring diagrams of the machines wich are very detailed so i've been able to locate all the encoder wires plus the +/- 10v signal and 0V. i still have to understand how to connect the drives to the board correcty.. it seems to have a circuit on the X and Y drive wich is 24VDC generated by the drives for the drive enable signal. i'll add pictures at the end of this topic.

Are all of the current motor drivers on the machine compatible with the 7i97T?
the X and Y axis use the same model of motor driver wich is a TDE-Macno DMBL7 i'll add the manual for that at the end of the topic.
the K axis has a Axor industries mini magnum  MM230T 6/12 drive, i'll add pictures of that too.
this is the new Z axis motor and driver that i'm looking to get. Motor: A6M40-100L2A1-M17 Driver: A6M40-100L2B1-M17 i'll attach the manual and a few pictures of this too. 

Is a Raspberry PI5 suitable for the application?
i wanted to make sure before buying one that the Raspberry PI5 is suitable for my application or if it would be better to get a normal pc.

Thank you in advance for the help.



 

Once your EtherCAT master is working you will need:

1.) LinuxCNC EtherCAT:
github.com/linuxcnc-ethercat/linuxcnc-ethercat

2.) For CiA402 drives:
github.com/dbraun1981/hal-cia402

Some servo drives have native support in linuxcnc-ethercat.

Take a look at the linuxcnc-ethercat documentation - which also includes example ethercat-config.xml files.

I was suggesting wiring the encoder directly to the 7C81 pins
on a different 7C81 connector than the BOB.

(with series 100 Ohm resistors for protection from negative voltage spikes)

Thanks for your reply. But the cheap Chinese boards require 24v power supply to energize its inputs. So inputs are all not responding to 5v input of the encoder. Kindly help as how to wire the 5v inputs with 7c81.

How can i Check this fault status ?
this way ?