Firmware 1.20 Release Candidate 1 available



  • David,

    I've just tried this release candidate with the two issues that I had been having.

    The first was that using M109 in a macro with only X and Y movement caused the Z axis to move to absolute 5mm. This is now fixed.

    The second was the behaviour of U and V on a CoreXY using the P3 hidden axes parameter in M584. This is different but unfortunately not fixed. What happens now is that if I use the P3 parameter, the U and V axes do not move at all.

    As a reminder, in config.g I have M584 X0:3 Y1:4 Z2 U10 V11 E5:6:7:8:9 P3. Then the first line of my homeall.g is M584 X0 U3 Y1 V4. When I run homeall.g, the U and V axes remain hidden which is new behaviour (they used to appear in DWC). Then when I run the next line which is G1 X-380 U-380 Y-380 V-380 F4800 S1 the U and V axes remain stationary. I get the same thing from the console too. i..e if I send M584 X0 U3 Y1 V4 then G1 X10, or G1 Y10 those axes move as expected but if I send G1 U10 or G1 V10, those axes remain stationary. If I remove the P3 parameter from M584 in config.g so that the U and V axes are visible at all times, then everything works as it should - i.e. all 4 axes move as commanded.

    To recap. With 1.20beta10, the issue was that the M350, M92, M203 etc parameters were not being applied to hidden axes even when the M584 command preceded those other commands. The work around was to use M584 without the P3 parameter at the start of config.g, then add a second M584 command with the P3 parameter near the end of config.g. This worked really well. If I put M584 X0 U3 Y1 V4 at the start of my home all file then M584 X0:3 Y1:4 Z2 U10 V11 E5:6:7:8:9 P3 at the end, then when I ran homeall.g the U and V axes would appear at the start of the homing sequence, then disappear at the end. I really liked that behaviour.

    With 1.20beta11 you made a change which should have negated the need to use two M584 commands - one with and one without the P3 parameter. Unfortunately, whatever change you made had the effect of causing the U and V axes to lose their CoreXY kinematics. At lease that was the observed behaviour as they seemed to be just using a single motor and moving at 45 degrees. The work around was to keep the axes visible at all times. i.e. not use a P3 parameter anywhere.

    Whatever change you have made with this RC1 candidate has had the effect that the U and V axes do not move at all when commanded to and when the P3 parameter is used in M584. The work around is again, not to use the P3 parameter.

    I really liked the behaviour of beta10 - it's just that I had to use two M584 commands, the first without the P3 parameter and the second with the P3 parameter.

    Cheers

    Ian


  • administrators

    Ian, thanks for the feedback. What happens if you use M584 P5 in your homing file(s) before you need to move U and V, and M584 P3 when you have finished doing that?

    The behaviour is intended to be:

    • If you use M584 to create new axes (that haven't been mentioned before in a M584 command), by default they are visible, unless you have a P parameter on the same command
    • Axes have to be visible in order to move them
    • You can change the number of visible axes at any time using M584 with a P parameter (at the same time as changing the motor mapping, or by itself).


  • Hi David,

    That works. With those changes (P5 at the start and P3 at the end of the homing files), the axes are hidden unless they are being homed and all movement is as expected (as long as the axes are visible). So it's all good.

    I have M584 X0:3 Y1:4 Z2 U10 V11 E5:6:7:8:9 P3 near the start of config.g (before any other drive related commands and before the axes limits) but no other M584 commands in config.g. That proves that parameters are now being applied to the hidden U and V axes.

    At the start of the homing files (all of them apart from homez.g because no axis mapping is involved) I have put M584 X0 U3 Y1 V4 P5. Then near the end of the homing files I have M584 X0:3 Y1:4 Z2 U10 V11 E5:6:7:8:9 P3. I guess I could abbreviate that final M584 to be simply M584 X0:3 Y1:4 P3 but it does no harm as it is and I'm less likely to make typos with copy and paste :).

    Cheers

    Ian



  • Hi David,

    Here are some things I have noticed, and they could be features or bugs, you tell me 🙂

    • if you issue a M913 X10 Y10 to reduce XY currents to 10 percent, the new current setting sets in only after a move on one of the axis
    • if you speciffy in bed.g coordinates outside the table, it will execute ignoring the table dimensions set with M208

    Cheers,
    Sascha



  • David,
    I have just tried to upgrade to 1.20RC1 on Duet 0.85, from 1.19 Stable.
    It seems I have a difficulty to connect to my Duet over the Ethernet. I have my Duet set up to get the IP address over the DHCP. It seems it cannot do it anymore for some reason. I must admit, that I have a complicated network, but it was and is working until 1.20 just fine.
    I have used the DWC to upgrade the FW, and after the flashing, I could not connect to the board anymore.
    I was thinking that flashing over DWC went wrong for some reason, so I have flashed the board with Bossa.
    It seems that the Duet reacts just fine with RJ45 disconnected, I can use PanelDue to home, and to jog the axes.
    However, as soon as I plug in the Network cable, printer starts to act strange. The PanelDue commands are ignored, or executed with a great delay. It seems like the CPU is struggling to get the IP or something.
    I also notice the SmartEffector LEDs are blinking very shortly once in a while. Like after a reset during boot up.

    I have downgraded to 1.19, and network and the Duet is working fine.
    But I get random Duet resets out of nowhere with 1.19… so I wish to get upgraded as soon as possible.


  • administrators

    @Teilchen:

    David,
    I have just tried to upgrade to 1.20RC1 on Duet 0.85, from 1.19 Stable.
    It seems I have a difficulty to connect to my Duet over the Ethernet. I have my Duet set up to get the IP address over the DHCP. It seems it cannot do it anymore for some reason. I must admit, that I have a complicated network, but it was and is working until 1.20 just fine.

    Thanks for your report. One of my printers uses a Duet 085 and gets its IP address via DHCP, so I know this works at least in some configurations.

    Please can you load RRF 1.20RC1 again, confirm, the problem is still there, then send M122 for USB and post the report here. Also post a M122 report from RRF 1.19.

    One possibility is that the configuration you are using is leaving your Duet short of memory, which would explain the problems you are having with both RRF versions. The M122 report may shed some light on that theory.


  • administrators

    I confirm that it is not possible to tune heaters in firmware 1,20RC1. This will be fixed in 1.20RC2. Meanwhile, use the same heater parameters that you were using with firmware 1.19 or a 1.20beta.



  • @dc42:

    Please can you load RRF 1.20RC1 again, confirm, the problem is still there, then send M122 for USB and post the report here. Also post a M122 report from RRF 1.19.

    One possibility is that the configuration you are using is leaving your Duet short of memory, which would explain the problems you are having with both RRF versions. The M122 report may shed some light on that theory.

    David, I have tried to flash it again using the DWC. The same behavior. It seems I also cannot sonnect over the USB, the Duet is recognized in Windows, but the drivers can not be installed for some reason, so I cannot connect via Comport. I am not sure why it is not working. The erased board is recognized just fine. I can flash it with Bossa.

    Any thoughts?
    Update:
    I have been playing around to gather more information, please let me know if I should open a separate topic.
    It seems that I cannot flash a newest FW over the DWC (1.19.3) anymore. The procedure works fine with FW_1.18.1, but with 1.19 and 1.20RC1 the printer doesn't seem to be able to reboot automatically, and gets stuck with the 'Update successfull! rebooting…' message on the PanelDue.
    With the 1.19 FW I can power-cycle the printer and it will operate 'fine'. But reboots once in a while, even with sitting still and doing nothing.
    With 1.20RC1 even if I power-cycle it, it doesn't seem to be able to find the network (see my original post).

    I will use 1.19 to get the M122 listing for you...

    Update
    I have the M122 listings, but I cannot find a way to attach it to this message from my mobile device...



  • @dc42:

    M307 now accepts an F parameter to allow the PWM frequency to be set. Caution: do not use excessively high PWM frequencies, especially with the bed heater, because you may overheat the mosfets.

    This feature I am especially excited to try out, I've had to use bang-bang for my heated bed because with PID a couple nearby lights in my house would visibly flicker.


  • administrators

    @MarMed:

    @dc42:

    M307 now accepts an F parameter to allow the PWM frequency to be set. Caution: do not use excessively high PWM frequencies, especially with the bed heater, because you may overheat the mosfets.

    This feature I am especially excited to try out, I've had to use bang-bang for my heated bed because with PID a couple nearby lights in my house would visibly flicker.

    I implemented it just for you, in response to your post in another thread.



  • I just installed RC1. I can confirm M307 Fx works fine at 1, 1.5 and 2Hz and the light flicker is a lot less annoying. Thank you!


  • administrators

    Only integer frequencies are supported, so 1.5Hz will get rounded, to 1Hz AFAIR.



  • I've just noticed something that may be a firmware issue but may be a DWC thing. I'm currently printing an object using all 5 extruders with a mixing ratio of 0.20:0.20:0.20:0.20:0.20 and notice that on the Print Status page of DWC, in the Machine Status panel top right, under Head Position I have Extruder Drives but it's only showing filament used for Drive 3 to 5 - Drives 1 and 2 are showing 0.0. (and yes, all 5 extruders are indeed moving and all at the same speed).

    I've no idea if this issue applies to other mixing ratios - it's not something I take much notice of. For the same reason, this issue could have been present for some time and not specifically related to 1.20RC1 release. I'm in the middle of a long print but if I get chance, I'll run some tests with other extruder drive combinations to see what happens.
    Copy and paste of software information is this:

    Firmware Name: RepRapFirmware for Duet Ethernet
    Firmware Electronics: Duet Ethernet 1.0 + DueX5
    Firmware Version: 1.20RC1 (2017-12-08)
    Web Interface Version: 1.19.3

    Oh, and following on from another thread, this is another instance of where Drives 0 to 4 are labelled as 1 to 5 (my pet hate rears it's ugly head again) 🙂

    Edit. Ignore that. I may have found the reason. I initially sliced this with a 3 extruder printer configuration and afterwards realising my mistake, I manually edited the gcode file. It's likely that there is something in the comments that Slic3r adds that DWC reads and which I haven't edited.



  • @dc42:

    Only integer frequencies are supported, so 1.5Hz will get rounded, to 1Hz AFAIR.

    OK that seems about right. Thanks again.


Locked
 

Looks like your connection to Duet3D was lost, please wait while we try to reconnect.