Fixed! Thank you. Now I have a beautiful cylinder.

@bernardomattiucci The time integral of acceleration gives velocity and the integral of velocity gives you distance, so the calculation is mathematically possible. I think that such a calculation is part of inertial guidance for rockets. However, any errors accumulate over time and so the position calculated that way becomes more and more inaccurate unless there is a way to repeatedly anchor it to position determined more directly.

I have a new application for tapering extrusions that might make this feature request more popular. Moves that taper z height with fixed extrusion width will facilitate the production of invisible seams.

I have proposed a new feature for PrusaSlicer whereby it would make seams that I call scarf seams where the seam is spread out over complementary tapers. The feature request is here.
The scarf seam could be made entirely b y the slicer, but it is my opinion that the firmware should be responsible for the tapers. That would allow me (and others, presumably) to easily use tapered extrusions for other purposes.

My proposal is for an extra parameter in moves to allow the extrusion to start at one value and end at another.

@phaedrux Thanks Phaedrux. The software versions are not entirely current (3.2.2 for the Duet, I think). Now that I've re-started the PanelDue and also stopped PrusaSlicer from overwriting my maximum rates it is all back under my control!

(I'm not keen to update the firmwares right now as I'm printing parts for a new printer. Updates once I am done!)

Your point about defining the centre relative to the ending points is probably exactly why I am getting the fluted cylinders! Thank you. I'll give it a burl with fixed up points.

(I am indeed using CCW coordinates for the G2 move, but that's just because I was unable to upload all of the examples I tried. They include CCW coordinates for G3 moves.)

You might be interested in a method that I developed to make seams just about invisible. I have documented it in a feature request for PrusaSlicer here: scarf seam

The method requires the outside perimeter to be printed first. The layer starts at zero layer height (i.e. the nozzle is at the same z-height as it was for the previous layer), then tapers upwards to the full layer height over some arbitrary distance, and then at the end of the layer there is a complementary taper over the first few points before the nozzle moves away to complete the layer in the normal manner. The result is analogous to a woodworker's scarf joint and so I'm calling it a scarf seam.

The results are quite impressive, as you can see at the linked page, and with my hand-written gcode it is easy to achieve. The whole process can be done by a slicer but, in my opinion, the tapers should be processed by the firmware. I asked about adding moves that taper in Z for ReRap firmware some time ago, but I'm not sure that I had this particular use-case worked out at the time. https://forum.duet3d.com/topic/32474/tapering-extrusions I guess it might be worth re-visiting that request.

I am playing around with tapered extrusions, both where the layer height varies (think of non-planar printing) and where the width of the extrusion varies with constant layer height. Both of those things are fairly easily accomplished with segmentation of the moves into multiple G1 or G2/G3 lines. That leads to my hand-made G-code being extremely long and harder to parse than I would like.

Is it possible to tell RRF to make a tapering extrusion? Would that be a sensible extension to ask for?

@Phaedrux Thank you. Yes, that seems like it will be the problem. I wanted the extruder to have zero idle, not the others... I'll change it and my silence will indicate that the problem is fixed! Thank you.

Worked for me on two printers with Duet 2 Wifi. One was update from 3.1.1 and the other from 3.2. Updated and restarted and both printers appeared on the net.

@MJLew OK, I found the hidden interface element (see the rant below) that Phaedrux mentioned, but even though it promises when the command key is down to open the link "in a new tab behind the current one" is does no such thing. It just scrolls the current tab to put the selected command at the top.

When I hold the control key down I can get the popup that Infiniteloop showed (and it works), that is not how I have to use proper links in the browser.

• <rant> Hidden interface elements seem to be all the rage (even with Apple software) but they are terrible things for core elements of human computer interface. People should not have to hunt around for random hotspots that do surprising things, and people should not have to RTFM to work a web-page!

It sounds like it might not be possible to use the flow information to modify the print speed soon enough to work for you, but maybe you could use the flow rate information to modify the extrusion amount if you have a servo motor for the extruder instead of a stepper.

The inevitable delay between extrusion and the measurement of flow will be suboptimal for the control of the servo, but I dare say there are ways to deal with it.

@vbtalent2 It looks like either the nozzle is too far from the bed or the extruder is not extruding enough filament. The symptoms of both are pretty much the same. Have you confirmed the nozzle height and the e-steps?

@Macgyver This might be an old bug reappearing.

See here: https://forum.duet3d.com/topic/1816/config-file-vanishing

And here: https://forum.duet3d.com/topic/11266/config-file-disappeared?_=1722326054025

And another old one from me that I can't easily find.

Is Cura using absolute extrusion values? If so then the retract code will unload the filament by unwinding the extruder all the way back to -2. My guess is that putting in an M83 (extuder to relative) immediately after the G91 and G90 lines will fix your problem.

@jens55 Looks to me like there is a directional problem with the part cooling. If so then you should be able to rotate the part in the slicer and have the blobbiness move to a different place on the print but stay where it is relative to the printer.

If it's just generally inadequate part cooling then slowing the print down will probably help as it gives more cooling time relative to extrusion.

@DonStauffer The LiveCode development environment in which I mostly work has automatic variable typing, which is VERY convenient, but you still need to test for arrays using something akin to "if the keys() of [variable] is not empty then". That works without a hitch, but I avoid it by being diligent in appending "A" to the name of any array variable.

@adhanabal How?

@penguinya Yes, it's operating on Prusa MK4 and XL, but I had to turn it off on my machine because it was falsely triggering on the first layers when they had low layer height. I suspect that there might need to be some exception incorporated for those circumstances.

They have described it in these release notes.

@cgrosbeck What @oliof says about belts might well be true, but my next printer is likely to be a dual-gantry like the DuellingZero.

My plan is to stretch it along the y-axis to give a printing area of about 150 by 300mm. The short x-axis will allow it to be light and to use some nice THK rails that I acquired second hand a couple of years ago. I find my Duet-powered Voron 0 to be a very nice printer for my purposes, so a larger bed is not important to me.

@bernardomattiucci The time integral of acceleration gives velocity and the integral of velocity gives you distance, so the calculation is mathematically possible. I think that such a calculation is part of inertial guidance for rockets. However, any errors accumulate over time and so the position calculated that way becomes more and more inaccurate unless there is a way to repeatedly anchor it to position determined more directly.