Advanced Search

你需要去HCFA官网去下载对应你型号的伺服驱动器说明书 根据说明书改伺服驱动器的控制模式

Hi guys!!! Thamk you Tommylight, Zmrdko and PCW

I managed to solve the latency error by isolating one cpu of the processor. Now it doesn´t trigger the message, but there´s a new visitor....

Now the annoying "Error finishing read" appears too often, sometimes even while running the small "Linuxcnc" test program G-code.
The computer is a basic Lenovo tiny M600 with an Intel Celeron (2 cores). The ping values doesn´t seem to me to show crazy high msec values.

One more time, I´m completely lost in the journey!

Hi, the Mesa 7c80 board is a host for the Rpi 4B.  Looks like: desertshedstudio.com/wp-content/uploads/2023/01/7C80_RPI4_1.jpg  So supplying DC to the 7c80 (orange terminals) powers up the Rpi.  Which means that as soon as power is applied, it starts booting.

Then when Debian is shut down, the 7c80 has no way to unpower the Rpi, so it stays running.

We can control machine "on" power easily from a seal-in relay.  But has anyone devised a way to shut down the 7c80 automatically when Debian terminates?

你需要安装cia402 并把CIA402中的HAL 文件和XML 文件拷贝到你的机床文件 然后根据伺服XML的描述来更改CIA402中的XML文件 然后更改HAL文件

ok Benb, you are not from germany

We distinguish between emergency off and emergency stop.

Yes, you need a physical switch, but it can be handled by software. It even has to be in many cases, as a safe shutdown/braking of the machine must be ensured.

de.wikipedia.org/wiki/Notausschalter

in the english version are many more types:
en.wikipedia.org/wiki/Kill_switch

You could launch LinuxCNC with a script like so


and connect the pin gmoccapy.h-button.button-9, which exits gmoccapy, to an external button.

How do you have your Z axis configured on your machine? Specifically where do you have the machine Z zero position set up to be? If the command G53 Z0 moves your machine down to the table surface, that would be your problem.

The normal convention for a milling machine setup is to have the machine Z zero to be at the top of the spindle travel (or bottom for a moving table like the knee of a Bridgeport.) In other words the tool as far away from the table as possible.

Not that you can't configure your machine to have Z zero at the table surface. You can, but you will have to be careful of default assumptions often used in most g-code. You will have to watch out for commands such as G53Z0 often used at the beginning or end of files and replace them with something more appropriate for your machine. And for tool changes you will need to set up and configure a safe tool change position for G30 and add the line "TOOL_CHANGE_AT_G30=1" to the [EMCIO] section of your ini file.  (And even then Linuxcnc might try to move to G53Z0 before going to the G30 position, so I'm not 100% certain this will even work as I've never tried to do it.)

But it is easier and safer just to configure the machine to the usual convention of machine Z zero at the top of Z travel. Then use work coordinate offsets and or tool offsets to move the Z zero position to wherever you need it to be for your milling file.

(Also, yes the "TOOL_CHANGE_QUILL_UP = 1" means that the machine is going to move to G53Z0 at a tool change command.)

As luck would have it now we get to running this on a real machine (and a sim) it fails.
It seems the G91 is being inserted too early. It should be inserted after the G53 so an error is not generated
I've attached the filter we worked out earlier and a new gcode file.
 

Any help to fix this would be appreciated.
 

RLA wrote:

Hi...new to linux here ...bought 7i96s and have powered but when trying to open Mesa Configuration tool...which I think I need as everything looks so foreign coming from mach 3...but when I open in debian I get error " Dependency is not satisfiable: Python 3 packaging.??

Thanks for any help,

Rick
 

I removed the dependency on python3-packaging and still don't know why it's not in the Debian repo when using the LinxuCNC live CD.

JT

I don't know how to do this

It changes when I press the Set key, but I cannot connect it to a variable in the gui.

I think it is not possible to change this without going into the hal file. I tried every way I could think of but it didn't work.

I bound all other settings to variables. However, since this value is brown, I think the status file cannot be intervened. Does this color mean that this pin cannot be intervened?

TheRoslyak wrote:

Strange. I have this kind of drive, and I didn't have any problems with it, although I didn't run it for long. Actually, when you buy it, it's initially sold as either closed-loop or open-loop, and the seller asked me about it when I purchased it. Well, I immediately chose the closed-loop one.
 

Yes I have its big brother ECT86 running perfectly. I am sure I had this running  in testing with no load but lost the config. I think my manual XML has been used by most people on these drives. Pretty sure its found its way into ScottLairds linuxcnc-ethercat driver too.

@besriworld, I was hoping you had nailed it, but not to be. Being a plasma guy, I forget about inertia! I  tried disconnecting the drive and see if it runs with no load  but its no better.
The fault is "Joint 1 amplifier fault"

@tommy, this ethercat drive looks after the encoder connection. I can see the DRO moving when I twist the shaft so its OK. It has automatic tuning by default and the motor and drive are a matched pair. It can be set manually, but I don't think I need to go there.

Thanks everyone for their ideas. I am still none the wiser. From memory, there is an internal error register I might publish a pin for and review.

Hi guys. Since i've worked about this topic, i wanted to say something. First of all, Benb you are right. I also think software e-stop should be implemented in software side. I've integrated emergency stop button like yours to my machine some time. It was actually a circuit which cuts the power of motors, so when you pressed it the motor movements were being stopped but with a problem. The problem is that user was being able to open e-stop via gui(change the pins iocontrol.emc-enable-in and iocontrol.user-enable-out) while the emergency button is still pressed. And of course user was being able to move motors in software, without actual physical movement. I wanted to stop this action. As result, i've tried to add some logic in hal layer. But it didn't work as i want. The thing i really wanted was updating iocontrol.emc-enable-in and iocontrol.user-enable-out pins with extra control, which is emergency button is pressed or not. I've decided to examine the iocontrol component's source code because of this is the component which does the software e-stop. I've changed the source of iocontrol component in the way i described before, and it worked. I've left it as it is. The thing i did is:
1. Intervention to EMC_AUX_ESTOP_ON_TYPE and EMC_AUX_ESTOP_OFF_TYPE with checking extra pin which shows the button status. If button is pressed, iocontrol.emc-enable-in and iocontrol.user-enable-out are not gonna changed. If button is not pressed, it is okay to open e-stop in software.
2. Check the pin which shows emergency stop button is pressed in iocontrol's while loop, if it is pressed, update iocontrol.emc-enable-in and iocontrol.user-enable-out pins to do e-stop.

This solved my problem. User is no more able to change iocontrol.emc-enable-in and iocontrol.user-enable-out pins with a pressed emergency button and when the button is pressed, machine actually stops the movement and does e-stop.

This is my solution. I've changed the iocontrol component and integrated it in the way i described. The problem is that the way i check emergency stop button is pressed or not can be late. Because i am actually measuring the voltage which is given to motor drives. When it is lesser from some value, i am updating my hal pin as button is pressed. So the software receives the button status little bit late. But it is ok, if the button is pressed and there is no power, motors are not gonna be moved. I think it is safe.


I hope i didn't get your problem wrong. This is what i understand from the topic. If you guys have any questions, i can answer it in this context.

Thanks.

Which of the 3 options? What didn't work?
The easiest route is to install the kernel from Google Drive by downloading both files and install with:To compile it yourself, the easiest way is to use the already rt patched source. Then you need to use the right config file, the one in the attachment will probably also work for yours. Assuming that you downloaded that config file in your Downloads directory: 

Ok….
Erst mal danke …
So wie ich jetzt festgestellt habe wenn ich den G-Code ohne
M190 in QtPlasmaC lade nimmt werden die Mat-Daten ( Mat-Nummer ) die eingestellt sind abgearbeitet.
Ich dachte im G-Code muss immer ein G 190 Pn stehen

Ok…. alles gut

Danke und
King regarde
Juergen