Advanced Search

RotarySMP wrote:

"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. 
 

Looking at it in this way, the Regler and Bremse being two distinct entities, what you're suggesting may be the case.

However, I do not think I have a signal for the brake itself. The only output signals from the controller to the drives I have are — Reglerfreigabe X-Y-Z Achse, Reglerfreigabe Hauptantrieb, Anwahl Werkzeugspanner, Kühlmittel, ...many empty ones..., CNC-Fehlermeldung, ... many empty ones..., CNC-Betriebsbereit. So that's essentially 5 signals with 29 unoccupied ones.

Do you think there may be a way to interface the brake separately? Right now, it seems that is handled internally by the drive, considering that it releases essentially immediately along with the drive enable — so much so that I had thought they were both one singular operation. An internal servo parameter? But I didn't find any brake related parameters in the doc. I do see some BR1 and BR2 signals in the Vorschubantrieb page (attached). I will investigate.

I did not see any files from you. But at this point, I am relatively certain that something is wrong with the 7i85 I am using. I had planned to just get the basic config running and then do the rest manually, but since I cannot get to a point where all cards are shown in HAL Show, I am a bit at a dead end.

I am seriously considering getting a 7i97T, as I have spent a week on this now and don’t really see where to go from here.

One way to maybe get this running would be to connect a 7i74 to the P1 connector of the 7i92, but there is no firmware available for that combination (7i92T + 7i74 on P1 + 7i85 on P2), and I am not nearly knowledgeable enough to create it myself.

"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 

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.