How to Read and Analyze Input Shaper Graphs Generated by Klipper

Рет қаралды 29,539

Күн бұрын

This video is supposed to be a brief overview on what the input shaper graphs can tell us. It is the first video of a series of videos which will give a more in depth view on the kinematic system and input shaping in Klipper.
As always: feel free to jump to those parts, that help you with your problems:
00:00 Intro
00:57 What we can learn from the graphs
03:17 About resonances in a printer (slightly theoretical)
10:17 What does the graph mean
16:32 How perfect graphs should look like
19:30 How different issues can be identified inside the graph
24:46 Looking at some example graphs
30:07 About smoothing and maximum accelerations
31:07 Announcement and giveaway
Announcement: We started our shop for 3d printed ABS parts for RatRig printers, where we will regularly will drop batches made in our print farm. Visit the shop at cirr-it-printing.com
To take part in the giveaway, a full EVA set in ABS with heat inserts preinstalled, visit our unofficial RatRig community discord server:
/ discord
(edit: the giveaway will be up within the next few hours)

Пікірлер: 47

@MihaiDesigns Жыл бұрын

I love these in-depth explanations. Thank you!

@michaellindborg1510 Жыл бұрын

Very, very helpful! These graphs have been a mystery, now I understand how to read them. Great work!

@eMadman Жыл бұрын

This is absolutely fantastic. It helped me identify some some issues I didn't realize I had with the machine.

@firelion98 Жыл бұрын

I was looking around into what this graph means and thank you this is exactly what I wanted so much information!

@MikePilk 6 ай бұрын

I just want to say thank you for this video! I helped me figure out that my machine was not sitting perfectly flat on the surface I moved it to after building! My graphs on x were 100% indicative of having a wobbly table or the machine moving around. After repositioning the machine the graphs are pretty dang clean now! THANK YOU!

@monkeywrench1951 Жыл бұрын

This is amazing, thank you very much for this intro. It’d be nice to have a video with how the input shaping evolved and where it came from.

@nixxin_prime Жыл бұрын

This is an excellent video, instant subscribe for me! Thank you for explaining all this stuff.

@jakefromstatefarm8545 6 ай бұрын

This is such a great video and explanation. Thank you so much!

@shawnr8164 Жыл бұрын

Thank you, excellent content and explanation. You're very appreciated.

@jtrmal 2 жыл бұрын

Pretty cool explanations. Thanks!

@HardCoil Жыл бұрын

Thank you. That was a really good explanation

@BrianVoelker Жыл бұрын

Thank you so much for this! This really helped!

@HerrickAaron Жыл бұрын

Well done, and more like this! I am especially interested in why diagonal axis might have different amplitudes (~25 minute mark)

@petermeier3153 Жыл бұрын

All of your videos are great. Please make more :)

@petermeier3153 Жыл бұрын

Thanks for the nice explanation.

@frankdearr2772 14 күн бұрын

great topic, thanks

@fish3dp372 Жыл бұрын

Great video, subbed

@olenapazukha1015 Жыл бұрын

Thanks

@drumsmichael Жыл бұрын

Great video! Hope the series continues soon 😊 I have issues with my printer and I can narrow it down to a loose bolt and a wobbly table. The latter I’m aware of as I placed the printer on such a tabel To work on. However I’m not sure where to look for a loose bolt or even what is defined as loose? Like what torque is needed to tighten the gantty parts, mgn rails etc… Do you have any pointers, please?

@errdaytaylor5634 5 ай бұрын

this video is educating af! thank you very much. Small question to the diagonal movement for Corexy. Is it better to run 0° infill instead of 45° to run both motors and belts?

@philip_fletcher 7 ай бұрын

A very informative video - thank you. One quick question: would there be any benefit to having an accelerometer on both the bed *and* the print head? The print extrusion is always one relative to the other.

@ruizhang9469 2 ай бұрын

thanks，this is really helpful. and for one motor one belt moving，is it good of 2 peaks in gragh？

@michaeltaylor9505 8 ай бұрын

loved your vid plus schooling us, but I have one Question in the Power spectrum density if your number increase (peak) from say 3.2 to a whopping 8 what does that really mean?

@Vez3D Жыл бұрын

thanks man ! I learned a bit more with your video :)

@svenbaum6227 Жыл бұрын

HI. If you have loose screws/bolts can you say roughly whether it is the frame, XY gantry or print head? According to your video, I have a small peak in Y before normal and in X a very small peak after normal. When I compare it to your video, it's "loose bolts" every time.

@laukan 5 ай бұрын

Very informative video! Aren't the arrows for -1,1 and 1,1 wrong way? -1,1 is moving the top right motor.

@dap3d63 4 ай бұрын

oye que buena idea, pero y el linear advance? tiene que ir acompañado del input shaping.

@izartech Жыл бұрын

Im not sure, but I think you didnt talk about the probe points. Where should we have them and how many.

@ichbinrootunddarfdas Жыл бұрын

Richtig gut, danke! Ich denke, es würden sich viele über eine deutsche Version freuen... Einige haben es nicht so mit englisch ;)

@spartanfoxie 6 ай бұрын

Im having trouble getting these graphs off my k1, I dont know if mainsail or fluidd can access the file locations to download it and it doesn't have the script to turn the CSV file into this image either

@jorg1109 5 ай бұрын

Hi, first of all thank you so much for this nice video. It explains a lot..... But I have one question though. In minute 22:29 you explain that usually the X has an higher amplitude than the Y Axis. This is not clear to me since X has less mass. I'm using a Creality K1 Max and in all my diagrams actually the Y amplitude is higher than X. If you are really right than it would mean that my X-Axis (and probably the X-Axis from many other users) is overdamped. Could you confirm that you described it right?

@mikev4558 Жыл бұрын

love your video. one thing i don't really understand is the "over-dampening" isuue. What do you mean by this? can you give a few more examples. i though dampening was a good thing, as dampening absorbs resonance and converts the virbration energy into another form like heat. hence reducing resonance virbations.

@3dprintersawhiteboard401 Жыл бұрын

Think of the following example: We think of a linear oscillator mostly here. A very easy example is a spring with some weight attached. Now think about the spring being in a liquid, which will be dampening. We want to move the weight by moving the end of the spring. Now think of a second spring attached to the same weight, perpendicular, but with different length of the spring. If the liquid is very viscous, it will start delaying the movement of the weight. So if I pull on both springs at the same time, the weight will move in a bow instead of a line. Now, of course we can argue that this happens anyway with different lengths of the spring, but I think you can get the point. Over dampening has bad influence on the acceleration and deceleration phase and will create something that looks like overshooting in the corners. Worst case would even be, that it eats some tiny movements. It also stretches the belts and therefore reduced the lifetime of both belts and pulleys, and can also create belt artifacts when the teeth start to not perfectly fit the pulleys.

@stephaneclement5008 Жыл бұрын

@@3dprintersawhiteboard401 , thanks so much for your videos, they really help understand. Can’t wait to see the next ones! What could cause over dampening on x axis?

@quelixfenzer5108 11 ай бұрын

I just made a funny discovery on my Printers. they are both standing in a not very rigid ikea shelve and when looking at the heavier exis of my two completely different printers, the peaks were at almost identical positions, just heavier on the one with the heavier gantry. It seems like i just found the resonant frequencies of an Ikea shelve xD

@SCrappyX02 9 ай бұрын

what cause over damping?

@Stancemir 6 ай бұрын

Can i contact you for some specific questions?

@Fragmon1 Жыл бұрын

Hallo, bist du noch aktiv im 3D Druck? Wir (crydteam) würden uns gerne mit dir zum Thema InputShaper austauschen wollen.

@Dustinspeed1 Жыл бұрын

ARRRGHG I love the video but the one reason I was watching you said you were going to cover later lol.

@3dprintersawhiteboard401 Жыл бұрын

I have to find a way for people to watch till the end :P More seriously, what did I teaser but not explain? This video is a one take without script, so I might have lost that thought, sorry. I am working on improving my videos. The next ones will hopefully be better in that regard (as well as in audio and video).

@neeamradia3740 Жыл бұрын

@@3dprintersawhiteboard401 you said you would tell us in a later video why the two diagonal axes only need to have the same frequency and not the same amplitude.

@BestKosmakCZ Жыл бұрын

25:00 i dont really understand this. On the top the axis are almost perfectly overlapping in both frequency and amplitude, while on botom the frequency shifted up and only one belt is dominant. How is the bottom one better? I hope that your next video will help to explain this to uneducated people like me :[

@3dprintersawhiteboard401 Жыл бұрын

The difference in the height of the amplitudes might be a whole video by itself one day, but they are not a big issue. It even leads to some interesting things about clamping positions of the belts in relation to the teeth and such. With higher mass, the amplitudes will get closer together as the effects causing those get less important, so you mostly see that on x, not as often on y. What we only want to look at here is the frequency of where the peak occurs, and that's much closer on the bottom one. It might be a bit harder to see because of the sharpness and slight asymmetry of the x peaks, but on the bottom graph the peaks are withing a single Hz, while on the top one they are about 3.5Hz apart.

@BestKosmakCZ Жыл бұрын

Thanks for the good clarification! I was wondering. How is the belt frequency calculated? would it be possible to assume the belt as a string and use that to calculate the ideal frequency for belt tensioning?

@3dprintersawhiteboard401 Жыл бұрын

There is not really a good way to calculate the idea; resonance frequency of the belts. If you were to try calculating it ignoring a lot of stuff and trying to keep it simple you can do the following: - assume it to be a spring - you need the belt length (linear factor) - the hook constant (which is usually actually part of the data sheet of belts) - calculate an effective mass out of the combination of the mass moved and the strain Now you can try to compare the resulting load from the effective mass to some data sheets of the components involved and chose it to be below all data sheet mentioned limits and then calculate the resonance frequency from that. But it is actually pretty simple for corexy: just tension the belt until it stop flapping against the frame when it is resonating. For a v-minon for example I recommend a tension where you can push the two belt sides together with the print head being in the middle without much force. It is unfortunate, but we dont have a good metric for the absolute belt tension yet (that includes the tuning approach).

@liamventer Жыл бұрын

4 mins in....your directions are wrong. And I think the operation of input shaping is also incorrectly explained, (will explain when I can give it some more time) despite this the trouble shooting analses and advice is excellent.

@3dprintersawhiteboard401 Жыл бұрын

@@liamventer Well, along the axis is the important part Id say...