Thanks for recording this. It helped me set up my machine. I have a fundamental difference in my machine setup. When the nozzle moves over my bottom camera, I have it go to the height I am going to place components at, ie the board height. it is much lower than yours. I have found this helps with the Z-axis not being 100% vertical.

Thanks for putting this together, it's really well done! I dropped all your gcode commands in my comment here for the benefit of future openPNP warriors. Might be worth adding em to the video description to make em easier to find. I realize these shouldn't be blindly copy pasted, but copying your values and then modifying makes getting the fine syntax details correct. M150 S{Index} P{True:30 R255 U255 B255}{False:0} M150 S{Index} P{True:50 R255 U255 B255}{False:0} M10{True:6 S255}{False:7} P{Index} M3426 A0 G4 C{Index} I1 ^.*V:(?-?\d+).*

Good work mate! Saved me some time. Thanks!

For my kit MB, the I2C VAC sensors are using A6.

FYI currently To change the current of steppers IJK, you've got to send ABC instead.

Hi, great video. I want to assembly one machine like this for me. I'm buying some stuffs on aliexpress, but I got a few issues. Is the solenoid valve to control the vacumm pump Normally open or closed? My machine will inicianlly assembly 0402 components. What Juki tip should I buy for 0402 components, 502 or 503? Thanks

Thank you

At 1:10 how did you get home fiducial, primary and secondary fiducial values?

Excellent to follow tutorial! Did you ever made a corrected version? I would love to see what "edit pipeline" was about. What controller are you using? If you use TMC drivers, are you using sensorless homing?

To change the I, J, K currents use A, B and C. Those are the mapped letters in the Opulo Configuration.h. No idea why they show up as I, J, K in the M906 command.

Did you cover the fisheye calibration anywhere? I didn't see it and once I got to the "Vision" step it's seems apparent that you've already performed the fisheye calibration based on how square everything is. Seems like that should be done before the pixels per mm is set.

Thank you very much for recording this. It is exactly the help I need to get unstuck. What do you use for the initial .xml files? Do you use the default OpenPnP files, or the ones Opulo provided for the Lumen? When you were doing the nozzle tip calibration looking at the pipeline starting at about 1:24:15, the vision pipeline window was not visible. I know that was a test to see who was the first one to watch the video all the way through, right? Now I need to follow along starting and stopping the video as I replicate what you did. This is excellent!

@quertymodo, you started out nicely, then unfortunately strayed from the Issues & Solutions path. Many things you do manually can be done automatically, and with proper guidance, many hard-learned sanity checks and most importantly in the right order when using Issues & Solutions. Please, everybody, do not delete your virtual axes, just learn how to use them properly and benefit. If there really are Lumen specific issues with the auto-generated actuators, please report to the discussion group, so this can be improved, though I have not seen anything in the video. Of course the G-code needs to be specific, but this can just equally be entered for the auto-generated actuators, any nicely guided by Issues & Solutions. By deleting actuators, you are causing much unnecessary work (and invite human error). Soft limits should be set by jogging the machine to the actual limits, as I&S suggests, so you would not have had the crash in backlash calibration later, and no restoring of (apparently faulty) soft-limits necessary later. Similar for Safe Z. There is also a misunderstanding in the axis units per millimeter discussion. OpenPnP does not set these on the controller, it only needs to know them. So you should enter the 39.88 you measured there, in addition to configuring the controller using M92. Please use Issues & Solution to calibrate the camera units per pixel etc., don't do this manually (in fact you do that next in the video, so that was completely unnecessary). Neither the virtual Z axis, nor the secondary fiducial calibration have anything to do with "auto focus cameras only", but with camera focal length i.e. objects appear larger when nearer in Z, and OpenPnP needs to know how much. So the secondary fiducial calibration needs to be done at a different Z. Do not skip the nozzle offsets steps. Your guide of matching up cameras and nozzles in cross-hairs is prone to camera tilt errors, certainly much less precise, and strangely more work than the automated I&S way. I doubt you can pick and place accurately based on that. Also do not calibrate the nozzle tips using an outer feature up the tip, it needs to be at the lower-most Z of the nozzle tip, ideally the bore. I&S will again guide you. And zoom in the camera (scroll wheel). Please, everybody, proceed as recommended by I&S and use the blue info buttons to jump to the Wiki for more information, about the needed calibration rig. The version of OpenPnP recorded here is already strongly outdated, the current test version covers many more setup issues, including actuators, camera device settings, nozzle tip calibration etc. all with Issues & Solutions guidance and Wiki links. Finally, the screen recorder did not record the pipeline editor in the nozzle tip calibration sequence (near end of video), which is actually good, because users should not need to edit the pipeline at all.