C-Axis Calibration Steps per mm

[Christian Liebhart]

Hi,

I am currently calibrating a LumenPnP Machine using Marlin Firmware. 

I dont really know which controller is in use. 

Now the problem: 
It seems like my rotational axis needs a little calibration. It moves WAY more than the selected distance [mm/deg]. 

So I thought by adjusting the resolution [Milimeters] I could fix it. The problem is that by adjusting the resolution nothing seems to change. The C-Axis moves the exact same distance no matter what resolution is set. 

I have already looked through other posts but didnt quite find something helpful. 

Is this a driver-problem?  Or does anyone know what I am doing wrong? 

Thanks a lot for your help! 

Christian 

[Mark Patton]

Have you tried running the motor check instructions in the Lumen online manual?  It steps through checking out each axis using G code, which would eliminate OpenPnP being the issue.  If you were to type "G0 A90" into Marlin, the nozzle should rotate 90 degrees.  If it doesn't, then the issue is with Marlin or a wiring issue.  If it does, then make sure that the axis is defined as a rotational axis in Machine Setup -> Axis.  Opulo has preconfigured .XML files for the Lumen.  Did you start out using those, or did you set it up from scratch?

Motor Setup Instructions: https://docs.opulo.io/docs/testing/motor-check/

Using preconfigured .xml files for OpenPnP: https://docs.opulo.io/openpnp/calibration/import-config/

[Mark Patton]

It looks like the instructions for the pre-configured .xml files for OpenPnP is out of date and does not give you a link to the files.  They cam be found here: https://github.com/opulo-inc/lumenpnp/tree/main/openpnp

[Christian Liebhart]

Thanks for quick response! 
So I sent the suggested gCode command: G0 A90 and it moves not only more than 90 degrees but almost 720 or so. 

unfortunately I was not part of the machines building process so I dont really know what .xml file has been used.  (Setting up the Machine with OpenPnP is part of an university project in case you were wondering why I didnt set up the machine myself). 
Can I solve my problem by manipulating the .xml file? 
And if so, how do I find the file? 

I know this is might be a stupid question but a quick response from you guys could save me a lot of time researching :)

Thanks again, 

Christian 

[mark maker]

Hi Christian,

Be aware that just executing G0 A90 will not give you a reliable indication, because the rotation then depends on where the A axis was before that commend (unless you are in relative mode, which is usually not the case, and would be problematic in itself).

Instead execute:

G0 A0

to reset A to zero degrees.

Then execute

G0 A360

and look at some set screw or similar to see whether the axis does exactly one turn.

If not, you need to adjust the steps per mm (or per degree). But those are not in the machine.xml, this is a matter of the controller configuration i.e. in Marlin (I believe in a header file, to be compiled).

Note, you can enter the Steps per Unit in OpenPnP, but only as a means to let OpenPnP passively know the resolution of the axis, this will not configure Marlin for you 😁:

_Mark

--
You received this message because you are subscribed to the Google Groups "OpenPnP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/bf7989d8-ba25-4061-9d48-ca9f87e957a2n%40googlegroups.com.

[Clemens Koller]

Hi, Mark!

On 8/24/22 15:28, mark maker wrote:

> Note, you can enter the Steps per Unit in OpenPnP, but only as a means to

> let OpenPnP /passively /know the *resolution* of the axis, this will *not*
> configure Marlin for you 😁:

Where is the steps per unit used inside of OpenPnP?
To calc the Resolution, ok, but then, where is that used?

Clemens

On 8/24/22 15:28, mark maker wrote:
> Hi Christian,
>
> Be aware that just executing G0 A90 will not give you a reliable indication, because the rotation then depends on where the A axis was before that commend (unless you are in relative mode, which is usually not the case, and would be problematic in itself).
>
> Instead execute:
>
> G0 A0
>
> to reset A to zero degrees.
>
> Then execute
>
> G0 A360
>
> and look at some set screw or similar to see whether the axis does exactly one turn.
>

> If not, you need to adjust the steps per mm (or per degree). But those are *not* in the machine.xml, this is a matter of the*controller configuration* i.e. in Marlin (I believe in a header file, to be compiled).
>
> Note, you can enter the Steps per Unit in OpenPnP, but only as a means to let OpenPnP /passively /know the *resolution* of the axis, this will *not* configure Marlin for you 😁:

>> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/bf7989d8-ba25-4061-9d48-ca9f87e957a2n%40googlegroups.com <https://groups.google.com/d/msgid/openpnp/bf7989d8-ba25-4061-9d48-ca9f87e957a2n%40googlegroups.com?utm_medium=email&utm_source=footer>.

>
> --
> You received this message because you are subscribed to the Google Groups "OpenPnP" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to openpnp+u...@googlegroups.com <mailto:openpnp+u...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/openpnp/289156b5-2d8e-19bc-6608-81b09c307818%40makr.zone <https://groups.google.com/d/msgid/openpnp/289156b5-2d8e-19bc-6608-81b09c307818%40makr.zone?utm_medium=email&utm_source=footer>.

[Mark Patton]

Hi Christian,

If Mark's recommendations point to a problem with the embedded Marlin code in the controller,  you might want to back up a step and flash known good Marlin firmware.  The rotational axis should work properly with the version of Marlin that was programmed into the controller from the factory.  Was your Lumen a kit from Oplulo?  If so, then you should email support (https://opulo.io/pages/contact-support) and they can help you get the file and direct you how to download it.  They can also help you troubleshoot this issue.  If it was not purchased from Opulo, then there is a Discord server for Opulo Lumen related issues at https://discordapp.com/invite/TCwy6De   The people are very helpful in getting new Lumen machines up and running.  This issue needs to be fixed before you can hope to run OpenPnP. 

[Christian Liebhart]

Hey guys thanks a lot for your answers, I really appreciate it!

So problem needs to be fixed via adjusting Marlin. Thus I will seek help in the discord server linked above.

I asked the guy who assembled the machine I have to configure right now, and he told me not to simply flash a new Marlin Software because he had to change the Motherboards hardware a little due to the el. component shortage last year. Since he used a 16Mhz Quarz instead of 8Mhz he had to adjust the firmware and that might cause the problem with my c-axis. At least thats what he told me.

So thanks again for your help!

Christian

[mark maker]

> Where is the steps per unit used inside of OpenPnP?

It is used here:

1. When comparing axis coordinates, i.e. to decide whether they are the same. For example, if you have micro moves (typically from run-out compensation/vision iteration), this is used to decide whether to issue a move command or not (or include the axis in it, or not).
   
2. When doing backlash calibration, it is used for the step accuracy test (top row in the graphical diagnostics).
   https://github.com/openpnp/openpnp/wiki/Calibration-Solutions#backlash-calibration-on-the-axis--diagnostics
   
   

Possibly others, I would have to check.

_Mark