HowTo useADXL345 on FLSUN V400
ฝัง
- เผยแพร่เมื่อ 22 พ.ค. 2024
- This video shows how to use the Fysetc ADXL345 with USB-C Connector on the FLSUN V400 running original Github Klipper. What is it about, how do you install, what is needed, issues you may have... all in this video.
Again it was much more work than expected... the next video will for sure have less size. Any comment is appreciated.
Relevant links from the video:
BCZAMD ADXL345 Amazon link... sorry this one is already sold out.
This is similar on Aliexpress
de.aliexpress.com/item/100500...
3D Print files for the ADXL345 Mount (FLSUN V400)
www.printables.com/de/model/2...
Github Manual for Klipper install, also including manual refered to in my video on how to setup the Fysetc ADXL345
github.com/Guilouz/Klipper-Fl...
Klipper Official Manual - Resonance Compensation (Manual Measurement)
www.klipper3d.org/Resonance_C...
Chapters
0:00 Intro
1:55 The Device - Fysetc ADXL345
3:44 Installing Software
5:44 Firmware for ADXL345
7:36 Issue 1 - Mount and Copy Firmware
9:52 Issue 2 - Get the USB Serial
12:16 Issue 3 - When size does matter
13:49 Measurements
15:44 Results
17:57 Graphs
22:37 Up against ringing and ghosting
24:30 Final Thoughts
27:16 Remarks
Music - credits to:
Song: Alex GL - Dreaming Ahead
Music provided by Tunetank.
Free Download: tunetank.com/track/4285-dream...
Song: Alexey Anisimov - Turn Up The Volume
Music provided by Tunetank.
Free Download: tunetank.com/track/5831-turn-...
Song: AHOAMI - Evening Chill
Music provided by Tunetank.
Free Download: tunetank.com/track/3154-eveni...
Song: AliveSound - Through The Sky
Music provided by Tunetank.
Free Download: tunetank.com/track/3878-throu...
Song: RomaYoung - Funky Hip-Hop
Music provided by Tunetank.
Free Download: tunetank.com/track/5585-funky... - วิทยาศาสตร์และเทคโนโลยี
thx for your guides! now "everything" from the wiki is running and calibrated.. it was a wild ride that wouldn't be easy without your vids!.. 😎 big big thx!
Thabks for making such simple instructions, its exactly what i was looking for 🎉
You saved me from insanity. I could not figure out why it couldn't mount the ADXL. After you showed me to just hook it up to my Windows laptop...WOW! that was it! It popped right up and the file was in the out folder. Fortunately, I didn't have the same issue you did where it wouldn't output the string. All I did was shut everything down, plug in the USB cable w/adxl attached, started it back up, and it spit it out. The next thing I have to figure out is why the camera with crowsnest only gets 1 or 2 fps and you have to restart it again if you reboot the system. By the way...what's the automatic input shaper button for? I
Hello Matt.
Perfect this video. I only had issues when executing the actual input shaping on the Y-axe. The processor overloaded and the procedure ended in an error.
In the Gilouz instruction on Github I read the solution:
Change the microsteps value in the printer.cfg file from 64 to 16
The first time I did that, I only changed the first occurence of this value. Not smart since this value is set for each axe seperately.
:(. So the error occurred again. The second time I changed all of the microstep setting is the printer.cfg and the procedure ended flawless :).
When changing the setting back to 64 , I also accidentily changed the microstep setting for the extruder to 64.
I realised in time that I my edit command chaning all microsterp settings from 16 to 64 could cause some hypercorrection and after checking the printer.cfg backupfiles I concluded that the extruder microstep setting has to be 16 and not 64. So I changed it back to 16 before I started printing !
Don't know what the effect would have been, but I am not going to test the effect.
Thanks for the video. Been a great help.
brdi
My hands Clapper, clapper clapper for you... tanks "gran finale"...
Wieder super Video von dir !!!
Danke für Deinen positiven Kommentar - freut mich, wenn der Inhalt ankommt! Ich hab noch einiges vor... mein Problem ist die Zeit, davon habe ich leider zu wenig.
Thanx for this video. I've used the same input shaper and official Klipper on my V400, but the ghosting and results got worse, so I had to reverse back the X and Y values to the default Flsun V400. Btw, Flsun has an input shaper (also for their version of Klipper), which can be purchased by asking them directly
Thanks for commenting!
Your results got worse? How can a filter make that worse? Strange to hear. I would expect edges be less "edgy" , but Ghosting should be suppressed at any setting (maybe more or less, but it should always improve...?).
Are you running at "ludicrous speed"? As mentioned in the video... I had to crank it up like hell to see some ghosting at all...
@@craftingmat5425 yes, I was also a little bit surprised to see that the ringing and ghosting got worse after the input shaping. My values for X any y were around 45 to 50 with suggested 6300 accl. Speed and MZV. After lowering the values back to around 31 an 34 (the default Flsun values), it seemed better again. I don't print faster than 300 mm/s.
@@3D-Amir you should read klipper docs, it says that the best input shaper for a delta 3D printer is EI and not MZV… also that you should disable input while calibrating and recalibrate pressure advance after. I hope this advices helps you in some way.
@@franlinkin18 Thank you for your advice. I've read the docs about input shaping but maybe I missed something. Any I'll try that again at some point but currently I have good results and don't want to mess around too much with the printer ;)
I do not have this printer (yet) but as mechanical engineer I know some things about vibration response curves, so this advice is coming more from general knowledge than from hands-on experience, but here it is anyways: Please note that mass is a _very_ important factor in this analysis. You have to do the measurements with the configuration as close as possible to the configuration you are going to be using when printing (i.e. filament loaded, etc). As you can see, the MZV vibration reduction goes over 1.0 for lower frequencies. This actually means that the vibrations are not reduced but enlarged in this region. You may have set your frequency values too high for the filter, which will worsen the performance. A reason could be that your printing head was too light when doing the measurements. The result would be a too high advised frequency, which in turn will make matters worse. You can easily test this by setting the filter values to 15 Hz lower and see if the ghosting gets better. If that improves things I would even try 25Hz lower and recheck. For this filter, having the frequency set too low will not matter too much, while setting it to too high a value will absolutely ruin things...
Danke dir 🎉
Danke Dir auch für den Kommentar. Ein Like und ein Kommentar sind immer gut für den Kanal.
Clapper powaaaaa!!!!!
😛
Considering how delta's move in X and Y, I am wondering how the resonance frequencies change with the actual X, Y and Z position, as the position of the arms will change significantly as well a the direction of the reaction force of the head on the rest of the printer. For a normal core xy there will likely also be some effect of the X, Y and Z position on the resonance frequency, but my "feeling" (known to completely miss the ball if not challenged with actual tests) is that a delta's relatively long "waving" arms would inflate the effect. I feel that you getting a 4Hz difference between X and Y while probably measuring around the center of the bed is already an indication of this. A delta has triple symmetry, so the movement in Y is likely directly towards one of the rods and a movement in X is at an angle (or vice-versa) and that can already cause 4Hz difference, so what when the current position of the head is very close to one of the vertical pillars or just the opposite as far away from it as possible? Finally, Z may also play a bigger role on a delta as the place where the arms excerpt the reaction forces on the linear rails moves higher and higher with increasing Z and that will likely cause the entire printer to shake more. It's unclear to me whether that would also affect the resonant frequency, but I would assume it impacts the magnitude ... Again, similar questions may hold for a corexy, but it's typically not that high and the forces are always delivered in the same directions, which makes me think the variation will be less than for a delta. But the head of a delta will often be lighter than the X gantry of a corexy, making the amplitude overall probably smaller ... Aaarghh so many variables ... Maybe CNC Kitchen should measure them all and let us know the result?
Thanks for commenting! I understand your request. Basically, measuring the resonances on other locations (not in Center) is possible. I agree with you that there ***should*** at least be a difference in the magnitude dependent on the position of the effector over the bed. Hm. I think it only needs a slight mod in the according macro to do that. Maybe really a good idea to check on that. Measuring 8 positions at the outside of the printer bed could already show if this turns out true.
I think about it once more. The next Video is already planned (belt tension) and I do not know, when this will be out, because I wait for some info from FLSUN for that. Hope I get this next week. So maybe this could be something to check next month. Had some ideas on Mainsail as well... hm... lets see.
I was a bit "afraid" that I will run out of ideas soon, but it seems there is a lot of stuff left I can do with my V400.
Back to topic: I can confirm, that the input Shaping is working really well - also if it is not directly centered print. As already said in my video, my printer is obviously already dialed in quite well. I am not hot for printing a 6 Minute Benchy, so with regular speeds I already have good results without input shaping. But using it on top seems to even further improve the results.
Another "issue" would be: If there is really a significant difference between the vibrations depending on the position of the effector over the printer bed, then this means that you need the next iteration of InputShaping - in an adaptive way, like "Adaptive Input Shaping"... I am a software guy, so maybe there is some possibility to pull this off, but I am doing this in my (rarely available) freetime... it would take me considerable time (like many months, not days or weeks).
Thanks a lot for your comment, it was really a good one - making me think again and reconsider the topic.
[EDIT]On the other hand the Effector and the arms together are like 400g of weight... which is really not a lot compared to the total weight... well... I think I will measure it. [/EDIT]
Not sure why but it does the x axis fine but has an issue on the Y axis at 80 hz and shuts down with an Internal error on command:'ADXL_AXE_Y' any idea what that could be ?
Sorry for my late response. I have not a lot of time left for this channel (close to zero).
Also I do not have a proper response for you, why this could happen.
Are you sure that you have mounted your ADXL tight enough?
Not sure if the ADXL macro detected some inapropiate value and did safety-stop the macro within itself - which could happen if the ADXL is not tightened to the Head properly.
Hello! I installed everything according to your video. The printer sees mcu PIS(rp2040), but when entering the ACCELEROMETER_QUERY command it responds with "Invalid adxl345 id (got ff vs e5)". I tried connecting with three USB wires...
Sorry I did not keep up with the comments. This is very old meanwhile - did you fix this issue? Strange error message (especially if it already worked for X).
I had some problems connecting the sensor to the PC and the speeder pad. :) Do not forget! Not every USB-C cable is a real USB-c cable :) it has to be a data cable and not just a cheap "power" cable for charging smartphones or for powering e.g. webcams
Isn't possible to use this module with the stock klipper?
If you can edit the printer.cfg file in stock mode, you can at least use the manual method described at the end of my video. TBH I did not spend too much time with Stock Clipper.
I read in another comment, that FLSUN has its own Input Shaper available - maybe this is an option for you then?
Good morning
Thank you for this tutorial, I have a problem I do the test on the X axis and it's ok, when I do the test on the Y axis I get an error message. You know why ?
The comment is old - sorry, I did not keep up with the comments on my videos for some time. What is the error message?
@@craftingmat5425 Hello, thank you
I found the solution, the ADXL was not fixed well enough on the printer
@@Duiling Thanks a lot for the feedback - this may also be a good hint for somebody else with similar issue. Thank You!
Bro could you try to make a video with a instalation of the chc pro just for fun bro???
Sorry for late response... I do not have a lot of time unfortunately. I barely have time for the things I planned - so unfortunately... no this will not happen in near future.
Either way - thanks a lot for your comment! Appreciated!
Hello, thank you for the tutorial. I am looking for a solution when I want to test the configuration of the y axis I receive this message each time at 82HZ.
I use Fysetc USB
07:46
Internal error on command:'ADXL_AXE_Y'
07:46
Internal error on command:'SHAPER_CALIBRATE'
07:46
Klipper state: Shutdown
So I have no problem with the X axis
This is really strange. I did not run the ADXL for a long time now. Is this error still persisting or did you resolve it?
I have the same issue
I had the same issue, it's caused by the immense amount of stepper instructions that are being sent across the usb bus overwhelming the cpu.
It's relatively easy to fix however. In printer.cfg there is a line for microsteps, by default it is set to 64, set it to 16 for doing the calibration. The original dev actually left a comment on that line to do so just for calibration, it's just not documented well. The other spot you might need to change is the hz rate per step in the adxl config. It's default 100, but changing it to 50 gives the printer enough room to breath to finish the test.
Remember to set the microsteps back to 64 after calibrating.
thank you so much@@alittlehammmer447
@@alittlehammmer447Thanks a lot for the input! Very appreciated! I suspected maybe the ADXL is not tightened down enough and producing values out of scale so the macro shuts down. But your explanation sounds much more legit! THX!
Russians say use ei filter 30hz
See video title Обзор FLSUN V400 - Где амазинг спеед