Software bundle 3.4.0 stable released!

dc42

@mike3d you can get Windows and Mac binaries of a fork of Prusa Slicer that supports QOI format from https://github.com/n8bot/PrusaSlicer/releases/tag/version_2.5.0-alpha0%2Bn8-release1. This version also has the option to retrieve the machine parameters from the Duet via http, if you have configured it for upload via http. Thanks to user @bot for implementing all of this.

chrishamm

@fractalengineer Have a look at the DSF upgrade notes for v3.3, it sounds like you missed them.

bberger

@dc42 said in Software bundle 3.4.0 stable released!:

@mike3d you can get Windows and Mac binaries of a fork of Prusa Slicer that supports QOI format from https://github.com/n8bot/PrusaSlicer/releases/tag/version_2.5.0-alpha0%2Bn8-release1. This version also has the option to retrieve the machine parameters from the Duet via http, if you have configured it for upload via http. Thanks to user @bot for implementing all of this.

This is big news. Especially on the Duet integration part.

zapta

Thanks @chrishamm, I was able to install the external input shaping plugin. Next will figure out how to use it.

BTW, if most of the external plugins are not DSF specific, maybe remove DSF from their path on github?

bot

@bberger I have submitted a Pull Request to PrusaSlicer to include those integrations: https://github.com/prusa3d/PrusaSlicer/pull/8087

n8bot opened this pull request in prusa3d/PrusaSlicer

open Fix RRF Time Estimates and Add "Get Machine Limits" button for Duet printer hosts #8087

ctilley79

Is there a roadmap for RRF 3.5 available? What will be its focus?

oliof

@ctilley79 Looking at the Github commit history, independent gcode streams seems to be the main feature being developed right now.

StevenL

Failed up to Upgrade from 3.2.2 to 3.4.0

Wifi card is stuck in Starting mode. no actual IP working.
Help!

Do I need to revert back to 3.2.2 or reset back to factory or what?
Please advise

Phaedrux

@stevenl said in Software bundle 3.4.0 stable released!:

Failed up to Upgrade from 3.2.2 to 3.4.0

Wifi card is stuck in Starting mode. no actual IP working.
Help!

Do I need to revert back to 3.2.2 or reset back to factory or what?
Please advise

You may need to reflash the wifi module. Try sending M997 S1 from the USB terminal.

ctilley79

@oliof I was hoping they would continue to work on input shaper implementation because I believe it still lags behind Klipper.

dc42

@ctilley79 said in Software bundle 3.4.0 stable released!:

@oliof I was hoping they would continue to work on input shaper implementation because I believe it still lags behind Klipper.

In what way? We are continuing work on the IS plugin to make it easier to use accelerometer data to tune IS.

Argo

@dc42

Differences I see compared to Klipper right now:

separate shaper frequency for each axis
automatic resonance testing (with a accelerometer): Klipper firmware generates vibrations from lower to higher frequencies and tests each shaper for it‘s effectiveness at each frequency. The results and recommendations are then displayed in a graph. It also recommends maximum acceleration values.
Example:

https://www.klipper3d.org/img/calibrate-y.png

dc42

@argo separate shaper per axis is not only unnecessary (because the supported shapers cover frequency ranges of up to 3:1), it is a bad idea because if you use different shape on different axes then the motion of the axes will no longer be synchronised during acceleration and deceleration. This will result in print artefacts, for example on corners that are not aligned with the axes.

I already said that we are continuing work on the input shaping plugin to make tuning IS easier.

zapta

@dc42 said in Software bundle 3.4.0 stable released!:

it is a bad idea because if you use different shape on different axes then the motion of the axes will no longer be synchronised

How about a single input shaping that is direction dependent? That is, some interpolation for angles between X only and Y only?

(I don't know how input shaping works, just thinking loud).

chrishamm

@zapta Applying two different input shapers on two axes has certain downsides so we decided it would be better to support only a single input shaper for two axes to damp both ringing frequencies.

We have a first beta of a new input shaping plugin ready but we need to discuss a few more details internally before we eventually release it. I expect we'll publish it sometime this week.

samlogan87

@chrishamm looking forward to this

zapta

@chrishamm said in Software bundle 3.4.0 stable released!:

Applying two different input shapers on two axes has certain downsides

I was thinking about a single input shaper but in direction of movement rather than direction of axises. Here is an example, let's say that you want to optimize a 45deg movement in a cartesian printer. You run the input shaping calibration process with moves in 45deg, analyze to determine frequencies and IS parameters, and then when you move in 45%, you use that input shaper. Now imagine that you test IS for direction in 45deg increment and when you move for example in direction of 50deg, you use some interpolation to determine the input shaping in that direction.

(I must say that I don't know how input shaping works but my understanding that it's some kind of modulation of the movement which can be applied for movements in any directions)

Does this make sense?

dc42

@zapta given that the input shapers supported by RRF have effective bandwidths of up to 3:1, it's not necessary to do anything as complex as that.

To get a 3:1 ratio of resonant frequencies on the X and Y axes, and assuming X and Y axes use belts of approximately equal width, length and tension, you would need a mass ratio of 9:1. That's higher than you are likely to get even in a large bed slinger.

zapta

@dc42, thanks for the explanation.

SputnikOC3d

@dc42 - this release seems to be breaking my Nozzle Heater on the 1LC.

my lonely forum post