Had to share this... So happy lol.
Lower tower test print was pre upgrade, current print is after upgrading.
Night and Day!
@nuramori on klipper its a run once and remove from the system thing until you change something.
not sure if it will be used in the same way here or not
My understanding is also that it's a run once to characterise the resonance of the printer.
Input shaping is an open loop approach as far as I understand.
However, the accelerometer needs to be rigidly secured to the print head to accurately measure the printers characteristics so a dedicated mount IMO is a good idea.
That's not to say we won't find other uses for an accelerometer in the the future of course!!
This is my polar printer, it's down for a bed bearing upgrade at the moment, but she's definitely running and the duet's working a treat.
Working volume is 570mm diameter by 1150mm tall.
Here's some footage of her in action...
The YouTube playlist for the build so far is here:
My channel is www.YouTube.com/CNCModellerUK
I'm hoping to come to TCT so will drop by the stall. If you want some better pictures just let me know!
If I understand it correctly I believe the accelerometer is just a temporary fit of test instrumentation to gather the resonance data to set the relevant parameters. I don't think it's part of normal operation.
The multicopter uses the data to close a control loop which as I said I don't think is happening here.
OK so this is weird (and annoying I've lost hours to this fault), all I've done is checked the BLTouch and the signals from the duet with a servo tester and a servo respectively (as above), and now I've re-connected the BLtouch and everything works...
I guess I might have had a "sticky" BLTouch???!
Also, for checking probing I decided to use a sponge to help prevent crashes, works a treat would recommend!!
All the best,
@cncmodeller You can also try the iOS app.
The app does not allow any actions so should help with the view only aspect.
Thanks, I've downloaded the same app for droid, looks good
@phaedrux Thanks but I'm not sure yet. I'm planning to use 4 tools on my CNC machine. T0 - laser, T1 - Spindle Router (main tool), T2 - Pen Tool in Spindle, and T3 - Tangential Knife. I'm still getting my head around workflow, working out offsets and g-code generation for multiple tool jobs.
@EducatingSavvas If you're using fusion 360 for CAM you configure tool selection there and it can be different for each tool path in each setup. It's worth a look if you haven't already. I use it all the time.
@droftarts I'm using a Duet2Wifi and M150 results in a Command not supported error. This is with 3.2.2
Same here on Duet Ethernet... no baclight and M150 "Command Not Supported"
At least its working, I can now look at a custom menu set.
I built a little connector conversion board to address custom the loom requirements, see below.
Thanks for the help folks
@dc42 does RRF support the Maslow CNC router style kinematics? It's essentially what's been discussed here.
Been thinking about DIY building one of these for quite a while. Future larger workshop means it's a practical proposal.
This is the app, after an initial look I also upgraded to the paid version.
@cncmodeller IMO it's better to to a deliberate sharp movement, longer than the period of any resonance. See https://duet3d.dozuki.com/Wiki/Accelerometers#Section_Using_accelerometers but in your case I suggest G1 H2 Y moves covering about 90 degrees in place of the G1 moves in that example.
Looks like 18Hz it is then...! 60deg bed rotation move triggered from my 12864 display
Thanks for the insights @dc42
@cncmodeller IMO it's better to to a deliberate sharp movement, longer than the period of any resonance. See https://duet3d.dozuki.com/Wiki/Accelerometers#Section_Using_accelerometers but in your case I suggest a G1 H2 Yxxx move covering about 90 degrees.
Thanks @dc42 , will do when this print has finished.
I've found an droid app called Resonance, it records vibration data and then provides analysis.
I've put my phone on my bed and these are the results
I've plumped for 25Hz for my M593 F25, but I don't really know what I'm looking at.
Edit. It's a polar printer so X and y are components of the bed rotation.
Thanks @giostark, that's great. In fact it's just the sort of thing I wanted to know. So I take it from that the bed.g holds the mesh point locations.
I'm printing on a mirror, on a flat bed levelled at 3 points, I just need to allow for inaccuracies in the bed plane.
I'm sure there is a way to define a custom bed levelling grid, but I cant for the life of me figure out how. I just want to probe the circumference of my circular bed.
Any thoughts? I think I'm just missing something obvious... its late at night lol.