Advanced Search

Which driver should i go for?

The command "lspci -nnk | grep -iA3 net " 
tells me i have a "RTL8111/8168/8211/8411" , i read something about changing the driver to the 8168, but i'm unsure..

Other thing, if it is a bit thing to get it running, what do you think about these small computers for LinuxCNC?
They are not too expensive..
Lenovo tiny M92? Which CPU would you recommend at Least?
I learned the hard way to get a INtel-Networkchip, so i will definitly have an eye on this!

Hi Cmorley

I've done some more work on 3pcenter.
If you're interested, I have either a very minor fix for one function:

   def goCenter(self):
            c.mode(linuxcnc.MODE_MDI)
            c.wait_complete()
            c.mdi('g53 g0 z0')
            c.mdi('g54 g0 x'+self.dataX+'y'+self.dataY)
            c.wait_complete()
            self.w.pushButton_5.show()

or a fix for the entire file. It's more resilient to user errors.
X0Y0 can be executed at any time and from any position.
Test it if you like.

Hi!
I am trying to add the Mesa 7i97t card into my Anilam 1100 and I am so confused. I have (hopefully) wired the three servo drives and the rotary encoders correctly to the card, but I'm lost on what else needs to be connected. I'm specifically trying to figure out this group of 8 wires running from a PCB 801 board to a 90100289 REV A board. From my searching, I believe the 801 board needs to stay and the REV A board is the one I'm trying to take out, but I'm confused on where the wires go into the Mesa. I believe they go on TB5, but I'm lost on the order/what they do.
If anyone has advice or can point me toward a guide or something, it would be appreciated!!

I have learned so much about LinuxCNC, hal, and the linuxCNC community from this thread. Thank you.

In the end I was able to solve my problem with the following hal configuration...

# --- ESTOP configuration ---
loadrt and2 count=1
addf and2.0 servo-thread

# 1. Connect physical switch to the first slot of the AND gate
net estop-ext <= hm2_7i96s.0.inm.00.input-03
net estop-ext => and2.0.in0

# 2. Connect GUI toggle to the second slot of the AND gate
net gui-request <= iocontrol.0.user-enable-out
net gui-request => and2.0.in1

# 3. Feed the combined result back into LinuxCNC's safety gate
net estop-loopback <= and2.0.out
net estop-loopback => iocontrol.0.emc-enable-in

# 4. Use the loopback to fire your physical SSR
net estop-loopback => hm2_7i96s.0.ssr.00.out-00
setp hm2_7i96s.0.ssr.00.invert-00 true

Best,

Bill

The standard Realtek Ethernet driver has real time issues, this can be fixed by using the DKMS driver

No sure whet the definittive installation instructions are but doing a forum search for DKMS should help
 

Typical.
Your network card drops the connection.
Although it can be corrupted, more often it is caused by a bad kernel or a bad kernel module.
Try installing a different kernel version or dkms driver.

Hallo Forum!

Ich bin nun soweit die Maschine zu testen und die Achsen zu verfahren.

Mein Hauptproblem: Ich kann selbst im Dialog "Achsen tunen" die Kommunikation mit meiner Mesa-Karte nur für kurze Zeit aufrechterhalten. Die LAN-Schnittstelle blinkt dann nicht mehr und es geht nichts mehr.
Die 5V Versorgung habe ich schon gegen eine stärkere getauscht, hat aber leider nichts gebracht.

LinuxCNC habe ich noch nie gestartet ohne diese Fehlermeldung:
Operation Error: hm2/hm_27i92:error finishing read! iter=4730

Der Rechner ist ein alter AMD Athlon(tm) II X2 250 Processor. 4GB DDR3-Ram, alt, aber eigentlich ausreichend, hoffe ich zumindest.
Leider auch mit Realtekchip für Ethernet.

lspci -nnk | grep -iA3 net  gibt aus:
02:00.0 Ethernet controller [0200]: Realtek Semiconductor Co., Ltd. RTL8111/8168/8211/8411 PCI Express Gigabit Ethernet Controller [10ec:8168] (rev 03)
    Subsystem: Foxconn International, Inc. Device [105b:0e22]
    Kernel driver in use: r8169
    Kernel modules: r8169

Ich habe bereits einen Kern isoliert mit "isolcpus=0,",
meiner Meinung nach sind die Jitterwerte auch eigentlich ganz io..

Irgendwelche Ideen?

Vielen Dank schonmal, Gruß, Thomas

In my opinion, Ethercat is a great solution for small machines that can use integrated drives.
I have mills where the individual axes are connected to the rest of the machine via a one patch cord and one power cord, and things like homing limit switches  or other sensors are connected locally to additional drive inputs.
Such a machine is simple to wire and, in my opinion, more reliable.
However, if all the control is to be located in one cabinet and long and multi-core cables are to be run, Ethercat loses its advantages for now. It's just expensive and difficult to set up, and what's best about it becomes secondary...

First off, AI helped with website, marketing etc ... I dont think its capable of complex applied math thats required for a CAM and we did not use any AI for the same.

Second - GPL obligations attach to distributing a derivative work of GPL-licensed source. No GPL source is ever in our tree. The code was written against our data model, our kernels, our architecture — the post-processor library isn't a re-derivation of anyone's, it's config-driven dialects off a specified ISO base.

Every component we ship is enumerated and licence-checked by the build, and the build fails if one isn't.

I'd fall back on other more knowledgeable folks who use AI more extensively than we do if this does not answer your query.

So do you have any recommendations wich brand of ethercat drives to use for minimal setup and maximum stability? How big of an issue is hosst PC performance with ethercat ?
Is the importance of host PC performance comparable to a mesa card? Or more or less important? The thing is that for the first evaluation of Linux CNC we will use a mesa card, as it's the cheapest and quickest way to test it, but later we would like to use ethercat drives.

Sorry for the offtopic, but I have to ask, how can you be sure that AI won't use GPL-licensed code to write a commercial program? After all, AI does not write code from letters and numbers, it only modifies examples found on the Internet...

jjdege wrote:

Hi,
in the XML file you posted, the vendor ID (VID) is 0x00400000, while the EC1-57E manual says 0xAB5 or 00000AB5.
That could be the problem.
Hi

Use the values reported by ethercat slaves -v

Hi,
in the XML file you posted, the vendor ID (VID) is 0x00400000, while the EC1-57E manual says 0xAB5 or 00000AB5.
That could be the problem.
Hi

Thanks Peter for your help and quick responce.

I would strongly recommend against deionized (DI) water in any cooling system where you don't have the exact specifications of all materials in the loop (i.e. an off-the-shelf spindle cooling system).

DI water can be extremely corrosive to certain metals, plastics, and rubbers, especially at elevated temperatures.  I suspect most of the major components would be unaffected, but you can't verify what the MOC is for everything you can't see:

• o-rings & seals
• adhesives
• solder/braze
• hose & pump internal materials
• internal coatings in spindle

My recommendation is to use distilled water, not deionized.  Distilled will achieve the objective of eliminating tap water mineral deposits, but without the risk of some system component you can't see or control being corroded.

(distant) 2nd choice would be reverse osmosis (RO) water.  RO water is corrosive to certain metals, but should be ok (ish) with most polymers.

Don't take my word for it - a simple search for "DI water corrosion" should result in a trove of chemical compatibility charts.