Advanced Search

"Reglerfreigabe" is not brake interface, it is drive enable. That is the signal to enable the muscle power of your servo drives. Could it be that you are releasing the brakes before powering up the drives output stages? That would explain there being no holding torque. Linxucnc sees this Ferror but can't do anything about it as it is commanding a turned off power stage. Seems lucky there is some internal emergency mode in the driver which is catching it after only 10mm. Maybe it does that by ignoring the inactive drive enable, and powering up the output stage to arrest motion?

On my set up, the screw up I made was to enable the drives output stage with the "machine on" signal, but to release the brake when I release my finger from the machine on button, so I had drive fighting the brake for the duration that the button was pressed. 

Haha, well, that didn't work. I setp the signal to be 1 at all times. I think I will leave it like that, as I definitely observe the other machine with the original controller NOT dropping it at any moment, even on an emergency stop press. But either way, it didn't help.

I even tried connecting the wire to the 24V power supply directly, so that the signal will always be high no matter what. No result.

Lately what I've been noticing is that the machine is usually capable of starting up the first time, but the axis falls through on trying again after an emergency stop. Is this a clue of some sort?

Very interesting.
I don't see any errors in the configuration.
Can you please post the following outputs?

ethercat master
ethercat slaves
sudo dmesg

Our set up with 5 Mesa cards is a pretty unusual one, so it might be beyond the scope of PNCConf to configure it. I only used PNCconf to get the basic frame work set up, and then did all the rest by manual editing the HAL, or in my case a bunch of hal files, as I found it easier to break the HAL into single topic files, and work through them one at a time.

Try the HAL's I sent you? Since we have the same cards, LinuxCNC may start with them (although it is more likely it faults out due to cards connected in a different order than mine. For some reason my second 7i84 mounts as 7i84.0.3, rather than 0.1 for example.
Cheers,
Mark

What I found in a Heidenhain manual is that MDI is there meant as small programs which you can run. And for those you need the start button:

 
[content.heidenhain.de/doku/tnc_guide/pdf...01/en/533_188-20.pdf]

Maybe there are other options which I don't know. I never used a Haidenhain control.

radom wrote:

stenly wrote:

Oh, also, my braking/release signal is a single one for all 3 XYZ drives. Is it possible to try and separate them electrically? Most likely. I'm not sure that's a good idea, though.

Now I'm wondering if you are looking at the correct thing
is there anything in the electrical diagram as "Z achse bremse"?

Yes, I probably worded myself confusingly. The *signal* for the brake is a single one for all 3 axes, but the brakes themselves are separate.I am attaching pictures of the electrical diagram that mention it. One is Motorbremze Z-Achse and at the other instance is Reglerfreigabe Z-Achse. Assuming the Regler (regulator) is supposed to interface the brake?

Regarding your other post, thanks for pointing out the relevant part of the docs of the controller. LinuxCNC triggers a following error if my FERROR value is set accordingly, yes. But there seems to be no relation between the following error and how much the axis falls — less than 10mm or so, which the encoder tracks correctly. The drive itself catches it and locks it. From what I see with a multimeter on the Mesa, LinuxCNC tries to compensate for this by sending a crisp 10V to the analog out of the axis to try to get it back into position, but it obviously can't.

If your machine is similar, I'd be curious how you have it set up to power it on in such a way. AC power on does not latch if 6K2.2 CNC-Betriebsbereit (CNC Ready for Operation) is not on. I have said CNC-Betriebsbereit signal linked to halui.machine.is-on so that it triggers when I press F2 power on. After which I can press AC power on and it latches, after which the brake is released.

Is this wrong? Thinking about it, maybe it'd be a good idea not to peg that signal to F2? I will try having the CNC-Betriebsbereit signal linked to something else on LinuxCNC startup, so that F2 can be nothing more but an analog out enable to the drives.

To Norbert:

I agree with you. It MUST be configurable.

@ZZ912
Implementing the Hal signal for the execute button ist possible like you described. But if you want to forbit the Keyboard Enter Key to submit the MDI command, you need to modify also the MDI History widget to ignore the enter signal. If you do that, I am sure many users will ask you to implement the old behavior again. As it concerns to me, I use the Enter key to submit the MDI commands.

The only possibility I see, is to leave the default as it is and implement a setting parameter to change that behavior.
I am not sure, but if I remeber well, the Heidenhain control does allow the execution of entered commands with the Enter key.

Norbert

I will not be in the shop today, i will provide pictures of the wiering tomorrow.

Fyi to clear this up I was able to fix it. It likely wasn't a remora-based issue

please refer to.
forum.linuxcnc.org/49-basic-configuratio...without-error#343443

That said I'm still on a 2024 commit with my current setup, as it seemed more resilient in regards to SPI Comms on F103, but i'm pretty sure upstream would work too.

Hey guys

I figured it out, as far as I can tell. I didnt put Remora in the title because the states/signals that trigger iocontrol from remora side were, as far as I was able to tell working correctly.
That said all it took was to make it explicit that my linuxcnc instance that it should run with iocontrol v2, which I did by adding

to the EMCIO section. From reading the docs I had assumed that in newer versions of LinuxCNC that it would be the default anyway, hence i never bothered. Also I never had errors.

Now to close this, whats the default today for a standard version of iocontrol in a deb-install of linux-cnc?

Thanks,
Alessandro 

qtDragon: Automatically Return to Manual Mode After an MDI CommandThe ProblemIf you use a pendant with qtDragon, you've probably run into this annoyance: after executing an MDI command, LinuxCNC stays in MDI mode. Before your pendant will respond to jog inputs, you have to manually click a movement button (or otherwise trigger a mode switch) to get back into manual mode. On a busy machine this gets old fast.  I use a VistaCNC P4-S.The FixThe solution is a small addition to your. We connect to LinuxCNC'sSTATUS signal, which fires whenever the interpreter finishes executing. When that signal fires and we're still in MDI mode, we automatically switch back to manual.Step 1 — Get a local copy of the handlerIf you don't already have ain your config directory, copy the system default there. LinuxCNC will automatically prefer the local copy over the system one.Step 2 — Edit the handlerOpen your localand make two small changes:In the method, add the signal connection:Add the callback method somewhere in the class:That's it.How It Works

• interp-idle

  fires whenever the interpreter becomes idle — MDI command finished, program finished, etc.
• The

  STATUS.is_mdi_mode()

  guard ensures we only switch to manual when we're actually in MDI mode, so it won't interfere with the end of a normal G-code program run.

• ACTION.SET_MANUAL_MODE()

  does exactly what it says — same as clicking the Manual button in the UI.

Notes

• Tested on LinuxCNC 2.9.6 with qtDragon.
• This should work on any 2.9.x build without modification.
• If you're on 2.8, the same approach should work but

  STATUS

  and

  ACTION

  import paths may differ slightly depending on your build.
• This does not affect program (auto) mode — when a G-code program finishes it will stay in auto mode as expected, since

  is_mdi_mode()

  will return False.

CreditFigured this out with a little help. Posting here so the next person doesn't spend 3 hours on it.Hope this helps someone!

It's supposed to be a 7i85 not a 7i85s. I'll have to check it in the morning.
I'll also have to check the piece of patch cable I used. It's possible the cable could be suspect.

And so IMHO it does not really make sense to have a hardware button to send MDI commands which you need to type with the keyboard.

It makes sense. But the reason is not the one described by BertaA.

My dream is to build a CNC machine, controlled by LinuxCNC, that will meet all the requirements (standards and laws) for a CNC machine that the European Union requires. I think that it is possible to compete with brands like Heidenhain, Siemens, Fanuc ... in a certain category of machines.

When I studied the safety standards for CNC machines, I found a requirement there:
MDI commands or sequence commands must be executable only with a physical button. I think that the standard even described what the button should look like. I don't remember exactly because it's been a long time since I read it.

Therefore, it would not be bad to make a configuration that would ensure the execution of MDI commands only with a HAL signal.

Yes, it may be a cable issue.

Just verified that:

7i76eu_7i76x1_7i85x1d.bin from the distribution zip file.

md5 checksum:

485114d3d22390a7e5b44d77588981c5  7i76eu_7i76x1_7i85x1d.bin

Works correctly with a 7I85 on P1 (all 5 sserial ports on the 7I85 work correctly)

Is it possible you have a 7I85S rather than 7I85?