Accelerometer?
Hi, i am not up on the technicals on this, but i understand the XL has an accelerometer in the extruder(s) whereas the Mk4 which has the same extruder does not, i also understand there not expensive and the board in the Mk4 extruder has a port for said accelerometer, so whats the reason it doesnt have an accelerometer?
Many Thanks
Many Thanks
the XL has an accelerometer in the extruder(s) whereas the Mk4 which has the same extruder does not, i also understand there not expensive and the board in the Mk4 extruder has a port for said accelerometer, so whats the reason it doesnt have an accelerometer?
Apparently it doesn't need one. The M4's printhead only moves along one axis, the bed deals with the other. The XL moves in a full 2D plane and the accelerations must be complex. I guess the development Mk4s had them fitted and once the movement vectors had been properly mapped the accelerometers became redundant.
Cheerio,
RE: Accelerometer?
The mainboard has a socket for ADXL acelerometers (or any other type).
You have to test input shaping for X axys and Y axys. Usually you put an ADXL in the hotend and another in the bed.
I dont see any "accelerometer" in the mk4 version so, Im really wondering how they can claim it has input shaping.
For example: P1P from BambuLabs do a input shaping test BEFORE printing anything. I dont expect to be "hardcoded" from prusa without a real accelerometer in X and Y...... It does not make sense. (at least for me)
Hope some dev could clarify this.......
RE: Accelerometer?
The mainboard has a socket for ADXL acelerometers (or any other type).
You have to test input shaping for X axys and Y axys. Usually you put an ADXL in the hotend and another in the bed.I dont see any "accelerometer" in the mk4 version so, Im really wondering how they can claim it has input shaping.
For example: P1P from BambuLabs do a input shaping test BEFORE printing anything. I dont expect to be "hardcoded" from prusa without a real accelerometer in X and Y...... It does not make sense. (at least for me)
Hope some dev could clarify this.......
Input Shaper suppresses these effects by “learning” how the printer moves – this is done by placing an accelerometer in the extruder and interpreting the reported values. The MK4 doesn’t have a built-in accelerometer in the Nextruder, but there’s an accelerometer port on the xBuddy board. We use it to calibrate the Input Shaper and then you can use the pre-calibrated firmware without an accelerometer installed.
So basically it should work fine as stock, if you start modding it and change the mass of things then yeah, it’s not going to work without an accelerometer to re-calibrate.
RE: Accelerometer?
It might work like on Vorons with Klipper - as in, you attach the accelerometer to the printhead, do the measurements, save them and remove the accelerometer. Re-do the measurements if something mechanical changed, including the surface under the printer. Since this is a bedslinger, it will need separate measurements for X and Y axis. So - attach to Y, measure, save, attach to X, measure, save, remove and done. This would require a purchase of an accelerometer, since MK4 doesn't include any. Hardcoded values, especially on a kit assembled by the user, sound like nowhere near enough imo. Each printer is different, especially when it's meant to be assembled at home.
Afaik, Klipper recommends to not re-do the measurements too often for their input shaping implementation anyway due to resonances being harsh on the printer:
https://www.klipper3d.org/Measuring_Resonances.html#input-shaper-re-calibration
If Bambu is doing a full recalibration for some reason, then I guess it's overkill. I'm not sure what this input shaping test is doing.
RE: Accelerometer?
During Prusa Live yesterday; they mentioned that in their testing, that there was not a large variance in the accelerometer tuning between the farm MK4 they had been using to test. So the MK4 will get a firmware update with preloaded settings for input shaper data and run off of that.
Will then be interesting to see how you address that in PS. Do you need to check off that you have a certain FW to then unlock faster speeds?
RE: Accelerometer?
Wel;l this raises a question, as i think the Mk4 unit from factory has a different z frame than the Mk3s which can be upgraded, so was the test also done on an upgraded Mk3s (different frame) ergo, are the boards for the upgrade have a different part number in them so the firmware knows if its a NEW FACTORY MK4 or an upgraded Mk3s so it knows which pre-loaded settings to use?
Many Thanks
RE: Accelerometer?
That is for sure a very valid and interesting point
RE: Accelerometer?
They have to clarify this situation. The info given in the "live event" is not enough.
We need to know:
- Can we purchase an acelerometer?
- Can the Firmware Fine-tune the printer en X and Y with the acelerometer?
- Prusa Understands that the vibration in their labs is not the vibration in a lack table or in a granite+foam base?
RE:
There is something else. At Klipper it is said that input shaping always makes sense when you move the printer to a new location.
If the printers at Prusa are all on shelves or similar and these values are now imported to a printer that is on the floor or on a stone slab (or on a less stable shelf, or or...), they will yes clearly miss their target. ESPECIALLY with a kit, in my opinion, the sensor should be included(!)...
@Goyetus faster than me...
RE:
Thanks @L&OR ,you are right!
I asked prusa directly in twitter.
Please, FEEL free to give lot of LOVE and add more questions about this......
It´s awfull and disrespectful if they think the community is not worried about this.
RE:
Accelerometer is not needed for regular MK4 (in theory, due to all MK4s being the same) but what about MK3.5... or modified MK3.5 (if you have different hotend like revo, different weights etc).
There was accelerometer question to prusa support:
https://www.reddit.com/r/prusa3d/comments/128mubs/if_i_modify_my_mk4_can_i_recalibrate_the_input/
Also accelerometer itself is not enough. There needs to be a way to teach firmware about new accelerometer data for input shaping.
RE: Accelerometer?
Accelerometer is not needed for regular MK4 (in theory, due to all MK4s being the same)
It's interesting that the sale of the acceleration sensor was mentioned. 😀
But your theory makes no sense! 😶
There are always small differences, not to mention the do-it-yourself kits,...
Are all the screws tightened exactly the same(?), are the 3030 profiles aligned exactly(?) < we have the MK3s (similar to my Voron) aligned for a while before it stood properly... 😫
And even if you put the same printer (extremely speaking) on Jell-O, or put the feet in concrete (like the mafia) Prusa's stats won't do you any good, because you'd have to buy a pre-built one and also use the exact same shelf as they + is it bolted to the floor or not?... 😒
Input shaper IS AND REMAINS Individual for each machine AND location! 😬 🤐
Greetings 🤗
RE: Accelerometer?
The printers are the same *enough* for input shaper to work correctly with the same calibration data for all MK4 - that's not my theory but that's what prusa people said multiple times in recent videos/articles.
That's why there is no accelerometer in each MK4.
RE: Accelerometer?
Again: It doesn't matter "how the same" the printers are...
Where a printer is located also has an influence on the vibrations that occur!
That's a fact, because you can see the differences in the printed object.
On a printer with Klipper firmware you can adjust the input shaping if you change its location.
And I'm not enough of a fanboy (honestly not at all!) to accept "what prusa says" as sacred...
It's just plain WRONG!... ENVIRONMENT affects vibrations!
Input shaping is nothing new, Prusa didn't invent it!
Here Prusa plays Apple and presents such a rough implementation as a notable achievement.
If you don't want to accept the facts, that's up to you.
However, anyone who deals with this will only recognize the existence of input shaping if the corresponding sensor is made available for individual calibration!
If you come around the corner again with an unfounded defense/apology for Prusa's ideas/actions, don't expect any further reaction from me, because I stick to facts... Sorry...
RE: Accelerometer?
I reiterate again. Prusa own research claims that one calibration data for all MK4 printers is doing it's job for their own input shaper implementation requirements.
It's not perfect I guess, it's not ideal but it is *good enough* for faster speeds and not loosing quality. What's hard to understand in that? (rhethorical question)
I'm sure they will sell accelerometer boards/kits for people who aim better results but currently it is what is is. I personally will wait and see results when everything can be tested outside internal prusa team.
"don't expect any further reaction from me" - good.
RE:
Good enough .. or bad enough ?
I concur with L&OR and other forum members. Although Prusa marketing may claim otherwise, they sure know that the whole purpose about input shaping is be able to customize it for every individual machine, once installed on a specific table/bench/enclosure/customized feet/dampening matt ... whatever.
If you consider that the MK4 will be sold also as a kit, shipping all the units with the same hard coded input shaping settings is pure surrealism. Depending on the individual setup of each costumer, sometimes you'll have worse results with that "built-in" input shaping than not using input shaping at all.
RE: Accelerometer?
I hope Input shaping is not enabled by default. Did not search into the firmware code yet in github.
Input shaping enabled in a wrong place, or with the printer kit, is nosense.
Anyways:
Remember Marlin 2.1 only supports Input Shaping measuring "by yourself". It dont allow accelerometers to get data.
So:
If Prusa Klipper is not finished yet, It´s loss resources getting the dev team to develop a "input shaping" for Marlin.
Klipper = input shaping + Acelerometer.
Marlin = Input shaping + print gcode to test, measure , and change it yourself. (no accelerometers allowed)
Maybe the main reason is this one.....
Good enough .. or bad enough ?
I concur with L&OR and other forum members. Although Prusa marketing may claim otherwise, they sure know that the whole purpose about input shaping is be able to customize it for every individual machine, once installed on a specific table/bench/enclosure/customized feet/dampening matt ... whatever.
If you consider that the MK4 will be sold also as a kit, shipping all the units with the same hard coded input shaping settings is pure surrealism. Depending on the individual setup of each costumer, sometimes you'll have worse results with that "built-in" input shaping than not using input shaping at all.
RE: Accelerometer?
One point made in one of the live interviews, is most of the people currently running input shaping are doing so on CoreXY printers, with long belts and more complex routing. The I3 style printers have much shorter belts, with simpler routing, and they made the point that this reduces the need for customizing the input shaping calibration to each individual printer. Time and experience will show whether or not the claims are true.
Where I think the one size fits all solution is definitely going to fail is in modified printers. Not having accelerometer mounts on the bed and print head from the factory are not a big deal. That problem can be solved with user mods easily enough, and it sounds like the mainboard has provision for connecting it. You only need to install the accelerometers and connect them temporarily while doing a calibration run so it doesn't need to be pretty.
The catch will be whether or not the factory firmware allows for a custom input shaping calibration. Otherwise, using custom firmware or needing to switch to something like Klipper in order to keep the input shaping calibrated when you have (for example) added a different fan shroud, or added a nozzle cam is going to discourage most users from making mods in the first place.
RE: Accelerometer?
Is input shaping even truly possible to do properly with the missing accelerometer(s)? It certainly can't possibly be active.