Duet 3 Expansion 3HC sensorless homing via CAN
-
@dc42 Can you tell us if there is at least a rough prediction of when the update we need will be released at least in beta? We would love to take on testing it and actively share the results.
-
@dc42 Good day!
I noticed the release of beta3 in the project changelog.
Could you please let me know if the sensorless homing issue for the Duet3 Extension 3HC board is planned to be fixed in beta3?
Thank you!
-
@dc42 Good afternoon!
We join the question from Cognirit about issue solution in beta3. We also are very much looking forward to a soon release in which this problem will be fixed. Thanks! -
I didn't know this was a limitation of the expansion boards. I ignorantly went out and got the 3HC, converted my RailCore to AWD and sensing works. My 3HC is CAN ID 1 and I see messages in the console that tell me driver 1.0 or 1.1 stalled.
The issue I do face, please don't discuss here, I'll start a post and not steal this thread, is that the steppers become increasingly unsynced after homing each axis.
-
-
@dc42 Great to hear that! Thank you!
-
@Cognirit this has been implemented in 3.6 beta3:
https://forum.duet3d.com/topic/37289/software-version-3-6-0-beta-3-now-available
-
@T3P3Tony Thank you for writing! @dc42
I can't properly convey how glad I am that I came across your developments and Duet2 about 7 years ago - and I still use them primarily in projects and research.
In my opinion it is currently the most flexible and user-friendly system for enthusiasts who are trying to create something new in the world of CNC machines.
Thank you and your team for all the work and effort you've put into this ecosystem and I hope it thrives for a long time to come!
-
@Cognirit thanks for the kind words!
-
-
@dc42 @T3P3Tony Hello! We installed firmware 3.6 beta3 and tested sensorless homing on the Duet3 3HC board. Individually all Z-axes perform homing without problems. However, when we tried to run homing with independent stopping (M574 Z2 S4) the system worked the same way as with S3 mode: when the first axis reached the end, all axes stopped. We tried running homing with the same settings, connecting the motors to the Duet3 6HC main board and it worked as it should: each motor stopped at the correct time. Is it possible that this is another unresolved problem with the Duet3HC Expansion board? I am attaching our configuration files. Thanks!
homez.g
config.g -
@tru_ann thanks for testing and reporting this, we will investigate further
-
@tru_ann I confirm this is a bug and I am looking into it.
-
@tru_ann this is now fixed in the source code. The fix will be in a new 3.6 beta, RC or internal build soon.
-
-
@tru_ann there are now unofficial 3.6.0-beta.3+4 builds containing the fix available at https://www.dropbox.com/scl/fo/wyqfqmzj0v7otx6wzqfdl/AJXLaai4nWowqfLtA4gePUk?rlkey=jpyla60xf2tqql0h4nx4sn9yv&dl=0
-
@dc42 Thank you very much! We successfully implemented unofficial 3.6.0-beta.3+4 build.The functionality we are interested in, including independent homing in S4 mode when connected to Duet3HC, works fine with it. We have not found for any other bugs, because we didn't do a full test. If we find any, we will write to you.
We are very satisfied with the use of sensorless homing for the Z axes, as it allows us to completely get rid of the end stops. We have been using this functionality for a long time in our previous 3D printers and have now implemented it in the new one. We would like to share our experience with those who may also be interested in this option, so that it is easier for them to configure and try it out. We attach a video of how the system works as well as our config files.
Triple Z axis sensorless homing.mp4
config.g
homex.g, homey.g, homez.g -
@tru_ann I'm glad it's working for you. We've released 3.6.0-beta.4 now.
-
@tru_ann Are your Z axes belt or leadscrews? As noted here https://docs.duet3d.com/en/User_manual/Connecting_hardware/Sensors_stall_detection#limitations-of-stall-detection, stall homing is not as precise or repeatable as endstops or probes:
The stepper drivers only update the stall detection state every 1 or 4 full steps, depending on configuration. So the actual position of the stall is uncertain to either +/- half a full step or +/- 2 full steps. This means that when using stall detection to replace endstop switches, the position defined by the stall is much less accurate than with typical endstop switches.
Ian
-
@droftarts Leadscrews. If you're interested, we've attached a video in the post above.
