Yes, the slowdown is (probably) not caused by incorrect gcode, but rather some issues with segment interpolation within Marlin. One of the two issues (incorrect segment length calculation) has been corrected already within the marlin source. The other issue we found in the thread I linked on github. Feel free to comment in that thread about this issue so that we can keep all the information in one place.
1 Like
quick note though arc_support is enabled in my Marlin 
I will make future comments on the Github thread
1 Like
Just wanted to say Thank You for making this plugin. I've been chasing some Z-banding issues on my Hypercube for a couple of weeks now. I thought I had it completely solved until I tried to make a cylindrical lead screw coupler earlier this week.
Quick info: I'm printing with Overture PETG, using Cura 4.6.1, on my custom built hypercube running a 32-bit EZboard from TH3d, firmware based on Marlin 2.0. (235 nozzle, 65 bed)
My normal geometric prints had great walls, even ones with curves, but complete cylinders showed some really weird artifacts. I tried printing directly from the SD card, and the print had the exact same artifacts in the exact same locations as the one that came from Octoprint, so I knew it had to be a G-Code issue
Then I tried disabling "Compensate Wall Overlaps" in Cura, which helped a bit, but also introduced some other weird artifacts. Then I tried your plugin, and immediately could see the difference.
Here is a photo (I'm a new user, so I can only post one image)
Nearly perfect walls again, but there is a weird issue I noticed that I think is happening during layer changes.
It looks like the print head pauses for a bit (or just travels really slowly) at layer changes, which can create a blob. The first print with Arc Welder lead to some bad blobbing at the corners where layer changes were happening. I relocated the z-seam to better hide them, but the pauses still happen during the printing.
I was able to mostly hide the blobs created by the print head pausing by moving the z-seam around.
here is a video showing the pauses happening on the final printed cylinder.
1 Like
Well, mostly excellent!? Why don't you join the party over in this issue in the Arc Welder plugin repository. That thread was originally started by someone with Prusa firmware, but others with Marlin 2.0 started to chime in (I may yet separate the two). Some issues were found with the G2_G3.cpp file that have been fixed for some users. One of those two issues has already been committed in the official Marlin 2.0 repo. Check out the gist I posted about that file and feel free to chime in within the thread.
Also, for anyone else having issues with the Prusa firmware (no stuttering, but small curves come out flat), I'm working on getting a pull request approved now that will hopefully help.
2 Likes
Just spotted this in the Marlin GitHub...
Not helpful... 
I intervened.
4 Likes
Someone is always trying to raise the difficulty level it seems Thank you for your support quashing that issue.
I understand the purpose of that pull request, but in my opinion targeting old severely constrained boards with the default settings is not an appropriate path for growth.
It isn't, and neither is refusing heartbeat messages on serial because they would "spam humans connecting through minicom" but that's how things are done here... 
Great plugin and the first impression looks very good. Thanks for your work!
1 Like
I have been playing with this on my rather unique printer with S3D, Idea Maker & Prusa Slicer. Like many I did not have any issues before or after. What I did notice is that, not surprisingly, items with lots of arc were much smaller. The Makers Muse Tolerance test has a saving of 75% for instance.
It was interesting that on the Marlin board it says Arcs should work better in 2.0.6, although didnt mention specifics.
Lots of changes were introduced to the arc commands through some recent changes to the marlin source. It a majority of cases they will improve arc printing speed and precision, and will sometimes eliminate stuttering due to a few easy to miss bugs that were recently fixed. I'm hoping these improvements continue until G2/G3 commands are as reliable as G0/G1 commands!
1 Like
Any idea as to when the next version of Marlin will drop?
I thought it already had, honestly. I will do some research tomorrow.
According to their GitHub releases page 2.0.5.3 is the latest.
I’m looking forward to the new version because the arc stuff had issues for me that might be resolved with the new version.
You already opened a ticket at marlin's github concerning this?
It will not fix automagically...
No. But I’ve seen tickets that were resolved that should fix the issues that I was having. I honestly do t even remember what the issue was since it’s been a few months, but I know that there are fixes that I’m waiting on.
You can always just download the bugfix-2.0.x branch from github and use that to compile yourself.
https://github.com/MarlinFirmware/Marlin/archive/bugfix-2.0.x.zip
I will say that in my own testing, these crunching-of-the-segment-commands resulted in faster printing speeds on my rig (columns). Like in the case of the maker mask, it was zooming around there so fast I was actually worried at times.
Could that be because I have a high junction deviation already set, plus the curve feature in Marlin already switched on and fairly agresive default acceleration but not much higher max acceleration. i.e. my printing is already set as fast as it is going to go. Which is limited by it frame.
In a few months I should have a CR6 SE to test it on, possibly after tweaking the firmware.
I couldn't tell you why your speed is maxed out but I suppose I'm envious.