Scanning Z Probe
-
@Tortellini please post two or three images of the height maps you see after running G29, so we can see what sort of differences you are talking about.
Duet WiFi hardware designer and firmware engineer
Please do not ask me for Duet support via PM or email, use the forum
http://www.escher3d.com, https://miscsolutions.wordpress.com -
@dc42 Hey! Thank you for your response! It wont let me add three images to this response as they are too large, but I will add them to that original drive folder. I will also add their corresponding csv files. out of three tests, one after the other, the first two were quite close. However, the 3rd is quite a bit off from the previous two.
heightmap1.csv heightmap2.csv heightmap3.csv
Like mentioned before, please check the drive link for the corresponding images if needed.
Thanks again!
-
on second thought it is now letting me add them to the forum so here are the screenshots as well.
-
@Tortellini maybe add the csv files as well...
-
@oliof those are in my original response
-
Do you know, if the z-axis looses steps during probe moves?
The SCARA arm isn't the easiest to set up. At the far end, when the arm is extended, the load might be higher for the Z-motor. -
@o_lampe I don't think it's impossible, but I figure I would then see larger discrepencies at the far end of the scan. However, between scans, the difference between the points is more generalized across the entire bed. I'll be sure to bring that up with on of the mechanical engineers though
-
Anyone have any further ideas on this issue? I am still experiencing the same problem
-
@Tortellini I think you need to do some work to establish if this is a problem with the sensor or the motion platform. My own tests using an e3d toolchanger have shown the sensor to be very repeatable. Perhaps you need to mount a conventional dial gauge on the arm and create a map using that, and see if the results using that sensor are repeatable and how they compare to the map generated by the SZP.
It is hard to tell from the pictures you have posted, but it looks like you may have something like a bl-touch mounted on the toolhead if that is the case how do the maps produced from that compare? Also how fast are you performing the scan? Do you get different results if you slow that scan down?
Perhaps posting a video of the scan in progress would also help folks understand what may be going on.
Have you also checked what the sensor reading looks like in terms of noise. So for instance if you place the probe over various fixed locations you should see in DWC the current probe reading being displayed. How much does that reading change? I've found that using the shorter cable linking the coil to the probe board gives more stable results. You might at the same time want to consider moving the sensor board away from the stepper and fan motors. I'd also be tempted to try cleaning up the wiring a little that may not be helping things.
-
@gloomyandy Fair Enough. It may be a few days before I'll have access to the printer again, but I'll be sure to post a video when I do.
Also, we have been using the BLTouch to collect a "control" mesh, but since then the physical bed has been adjusted, so we'll have to do it again. Initially however, it seemed the SZP's results were pretty far off the BLTouch. I wonder if that is solely because of the PEI over the top of the steel sheet. There has been discussions that we may try to replace the printer bed to a more compatable material.
We'll also try the shorter cable as well as finding a better mount for the sensor board. cleaning up the wiring might prove difficult as it wasn't our team that did the wiring in the first place. so I'll leave it as a last resort.
Thank you for your help!
