Duet 3 Mini 5+ VIN under-voltage issues in SBC mode

dc42

All, please try the new firmware at https://www.dropbox.com/sh/0mwfe22wepk5efh/AAB6Lp0rwBPjVK2R4SxbKhLya?dl=0. M115 reports it as 3.2beta2+2.

This version avoids using the ADC DMA sequencer completely, so I am hopeful that it will solve the problem. I was worried that the additional ADC interrupt load would be excessive, but with careful code optimisation I have got it down to 0.5% of total CPU time.

Serbitar

@dc42 Looks like this version (and probably the other recent ones) breaks duetcontrolserver.

With 3.3b2 the logs are nice and quiet, with 3.3b2+2 the duetcontrolserver log is a fast rolling spam like this:

Apr 13 20:58:38 ender3 DuetControlServer[1585]: [error] Failed to merge JSON due to internal error: {"key":"","flags":"d99fn","result":{"boards":[{"mcuTemp":{},"vIn":{"current":0.4}}],"fans":[],"heat":{"heaters":[]},"inputs":[{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"},{"feedRate":50.0,"lineNumber":0,"state":"idle"}],"job":{"build":null,"duration":null,"filePosition":0,"layer":null,"layerTime":null,"pauseDuration":null,"timesLeft":{"filament":null,"file":null,"slicer":null},"warmUpDuration":null},"move":{"axes":[{"machinePosition":0,"userPosition":0},{"machinePosition":0,"userPosition":0},{"machinePosition":0,"userPosition":0}],"currentMove":{"acceleration":0,"deceleration":0,"laserPwm":null,"requestedSpeed":0,"topSpeed":0},"extruders":[],"virtualEPos":0},"sensors":{"analog":[],"endstops":[null,null,null],"filamentMonitors":[],"gpIn":[],"probes":[]},"seqs":{"boards":0,"directories":0,"fans":0,"global":0,"heat":0,"inputs":77,"job":1,"move":2,"network":2,"reply":0,"scanner":1,"sensors":0,"spindles":0,"state":0,"tools":0,"volChanges":[0,0],"volumes":0},"spindles":[{"current":0,"state":"unconfigured"},{"current":0,"state":"unconfigured"}],"state":{"atxPower":null,"currentTool":-1,"gpOut":[],"laserPwm":null,"msUpTime":101,"nextTool":-1,"previousTool":-1,"status":"off","time":"2021-04-13T20:58:38","upTime":83},"tools":[]},"next":0}
Apr 13 20:58:38 ender3 DuetControlServer[1585]:    System.InvalidOperationException: The requested operation requires an element of type 'Number', but the target element has type 'Array'.
Apr 13 20:58:38 ender3 DuetControlServer[1585]:    at System.Text.Json.JsonDocument.TryGetValue(Int32 index, Int32& value)
Apr 13 20:58:38 ender3 DuetControlServer[1585]:    at System.Text.Json.JsonElement.GetInt32()
Apr 13 20:58:38 ender3 DuetControlServer[1585]:    at DuetControlServer.Model.Updater.Run() in /home/christian/Duet3D/DuetSoftwareFramework/src/DuetControlServer/Model/Updater.cs:line 157

root@ender3 ~ # dpkg -l duetcontrolserver
Desired=Unknown/Install/Remove/Purge/Hold
| Status=Not/Inst/Conf-files/Unpacked/halF-conf/Half-inst/trig-aWait/Trig-pend
|/ Err?=(none)/Reinst-required (Status,Err: uppercase=bad)
||/ Name              Version      Architecture Description
+++-=================-============-============-============================================
ii  duetcontrolserver 3.3-b2       armhf        Control server application for Duet 3 series

I was hoping to do another batch of small sequential prints with the new firmware

Serbitar

FIRMWARE_NAME: RepRapFirmware for Duet 3 Mini 5+ FIRMWARE_VERSION: 3.3beta2+1 ELECTRONICS: Duet 3 Mini5plus WiFi FIRMWARE_DATE: 2021-04-09 14:40:21

This is the first version that causes the DCS log spam

Serbitar

Actually this is a bit weird. I just went through the remaining versions in sequence and 3.3b2+2 isn't spamming the logs.

dc42

@serbitar, I've updated the binary at https://www.dropbox.com/sh/0mwfe22wepk5efh/AAB6Lp0rwBPjVK2R4SxbKhLya?dl=0 to avoid DCS reporting those errors. The build time reported by M115 is now 21:29:10. The previous version was 18:55.

Serbitar

@dc42 said in Duet 3 Mini 5+ VIN under-voltage issues in SBC mode:

With the 18:55 version I had a print paused over night and the printer didn't have any VSSA or VIN troubles.
Print has now been resumed and will upgrade again later today after it finishes.

dc42

Thanks. I am 99% certain that my rewrite of the ADC driver has fixed this issue, so I will mark it as solved. Feel free to revert that if the problem recurs.

For others who have experienced this issue, thanks for your patience. The fix is in the Duet3Firmware_Mini5plus.uf2 file at https://www.dropbox.com/sh/0mwfe22wepk5efh/AAB6Lp0rwBPjVK2R4SxbKhLya?dl=0. The 3.3beta3 release will include this fix.

Serbitar

@dc42 Yep, looks good. No issues in the last 24h.
Many thanks for sorting this out!

Fletcher

@dc42 Running a print successfully as we speak, after 24hr with no issues. Thank you for your work sorting this out!

droftarts

Many thanks to everyone who helped resolve this issue, for your patience, persistence and providing vital testing and feedback. I hope you didn’t waste too much filament!

Ian