Will we get real input shaping (with accelerometer)?  

RE: Will we get real input shaping (with accelerometer)?

What's about an accelerometer on x- and y axis? This will make the setup easier and also a drift during hundreds of hours working time. You can calibrate the input shaper today. Simply follow the guidelines explained in this forum. I have done this - and it works. An accelerometer might also improve the quality on very large prints. But then you will use larger nozzles (0.6mm) and higher layer thickness. Therefore the ghosting artifacts are minor anyway.
Bottom line: Input shaping on the MK4 works and is not a "bluff package"!

RE: Will we get real input shaping (with accelerometer)?

Thanks for your reply!

Placing an accelerometer and calibrate the input shaper myself is what I do with my other printers but they are not advertised as "with input shaping". So, even if it's possible, the marketing of the MK4 and the behavior of Prusa is absolutely not ok. That's just my opinion.

Could you please share a link to the thread where the procedure is described? I already searched but only found some very long discussions about the MK4 having an accelerometer or not. No description of the procedure so far. 
I will go on with my search but if you have a link, that would be very helpful.

RE: Will we get real input shaping (with accelerometer)?

I don't mean to be rude, but do you technically have any idea how input shaping works?

I would advise reading the posts here in the forum without prejudice. Maybe even the clipper documentary.

It's not that a slight shift in frequency will make it not work.

In my post I shared info and Gcode for the test print. Just try it in practice.

You have nothing to lose but a few grams of filament.....but gain your own insights 👍 

RE: Will we get real input shaping (with accelerometer)?

I've already done it on two other printers with Klipper (which doesn't mean that I really know how it exactly works in detail).
Yes, I saw and read the thread. Didn't read it at first because of the "Klipper" in the title.

Even if input shaping still works with a slight shift, that doesn't change my mind about the marketing behavior. I have the impression a lot of people are whitewashing everything what Prusa does and did.

Don't get me wrong, the MK4 is a very good printer. I had to decide between the MK4 and the X1C. My decision fell on the MK4 because Prusa is a long existing company known for good and long support. Still I don't like everything they do just because they are Prusa.

RE: Will we get real input shaping (with accelerometer)?

As @Rainer states, the algorithm is not that sensitive and manual calibration (via the Klipper test object) is a good alternative to an accelerometer. Different suspension does result in slightly different numbers (I put my Mk4 on a concrete pad with isolator feet), but the results of fine-tuning those numbers is barely perceptible in prints.

I agree it would be nice to have an accelerometer and an "automatic recalibration" option in firmware, as it would make recalibration a little easier/faster than the Klipper test print method.  But after some experimentation, I can say recalibration is something I would rarely do -- perhaps only if I made a major modification to the printer or its suspension.

RE: Will we get real input shaping (with accelerometer)?

Imagine you purchase a new expensive monitor. Naturally, the software package includes a standard profile for Windows/Mac.  Anyway, you want accurate color reproduction, because it’s a known fact that all the screens, even of the same model, have different color rendition, Therefore, you decide to perform a custom calibration with a good  colorimeter and then comes the surprise. This monitor can’t be calibrated because, for some strange reason, the colors are locked by the manufacturer and your colorimeter won’t work. The only way is to adjust the RGB values using the graphic card software to eyeball the colors by trial and error.

Would you purchase such monitor ?

The only difference with this example and the MK4 IS “one size fits them all” profile is that good colorimeters are expensive and accelerometers cost 5 buck.   

Time will tell when/if Prusa realizes that this absurd MK4 IS implementation is gonna backfire on them sooner or later,

RE: Will we get real input shaping (with accelerometer)?

That's the point. The example is just wrong.

An accellrometer doesn't make anything better, absolutely nothing.

You can set everything the same as in klipper. It's klipper input shaper. An accellrometer does nothing but measure the resonances you see on the test print. Is that so hard to understand.

Do you think Prusa is stupid. Just because others are making such a hype about it with sensors, lidar, camera AI, etc.

What worked when the printers started printing on their own? The super camera should have noticed that there was still something on the print bed.

And that's exactly how it is with the accellrometer. Measure before each print? Why?

Klipper advises against it because it puts too much stress on the hardware. Everything just for show.

Do the test print and see what kind of area the input shaper works. You can experiment with it because you can change everything. Try other input shapers and how it plays out in REALITY.

Nero3D once explained it clearly in a video. I'll see if I can still find it and post it here.

RE: Will we get real input shaping (with accelerometer)?

They advertised it as "with input shaping". And making a test print to measure the resonances myself is completely different from automated measuring with accelerometers. For the first one I have to use filament and do measurements and calculations the second one can run completely automated. 

And I would accept it the way it is if they advertised it correctly. Why exactly do you ignore this point? 

For me this is nothing else but whitewashing. And in my opinion this discussion leads to nothing. You will keep your opinion and I will keep mine.

RE: Will we get real input shaping (with accelerometer)?

Nero3d input shaper

RE: Will we get real input shaping (with accelerometer)?

Don't get me wrong, the MK4 is a very good printer. I had to decide between the MK4 and the X1C. My decision fell on the MK4 because Prusa is a long existing company known for good and long support. Still I don't like everything they do just because they are Prusa.

You're right. Just because Prusa does it doesn't make it right. And it's not just black and white.

But I find that there is a lot of nonsense being said on social media. Not only in relation to 3d printers. I grew up without anything like that.

If someone here gives the technical reasons why something works in one way or another, I'm the last one who can't be convinced.

If Prusa brings an accellrometer upgrade I might even buy one. Just out of curiosity. I even have one for my Ender.

RE: Will we get real input shaping (with accelerometer)?

They advertise it “with input shaping” and it has just that.
IS is not IS due to accelerometers, IS is software
Nowhere does it state that it needs automatic measuring or something like that

And you do not have to do anything manually either
I’ve measured mine with the manual version, made adjustments and can’t even spot any difference in the prints and has since restored the values from Prusa

RE:

And I would accept it the way it is if they advertised it correctly. Why exactly do you ignore this point? 

For me this is nothing else but whitewashing. And in my opinion this discussion leads to nothing. You will keep your opinion and I will keep mine.

After all, we agree that we have different opinions. Which isn't too bad.

I don't think Prusa is cheating and Input shaper doesn't exist.

That's how I understand your opinion. Or at least an inferior input shaper.

In the end, however, this is not correct. The result will not differ. regardless of whether the resonance was determined with an accellrometer or with a test print.

I think we can agree on that.

I can understand that some might like it better with the sensor or that it gives them more security to do the right thing.

RE:

I agree. IS with calibration print is not inferior to IS with accelerometer. It is less comfortable to calibrate, yes but then this is not z-offset and will rarely need recalibration. 

So yes, accelerometer is superior as a whole package but not if you are looking at achievable results.

I am not sure why some believe that IS is only "real" IS if you calibrate it one way and not if you do it the other way. 

RE: Will we get real input shaping (with accelerometer)?

Not everyone is interested in how things work in the background.

That's why I opened my thread to explain it with the testprint.

There are many ways to achieve a goal. Prusa is more on the KISS side, others are trying to optimize things with technology.

Prusa tries to make it easy for the user. That is part of their philosophy. And the profiles in the prusa slicer mostly work pretty well. It will be the same with Input shaper.

What you can blame prusa for is the lack of information and time management. An article in the blog, a short video that briefly explains input shaper and why prusa chose this path.

In the future, this will no longer work in the aggravated market. Prusa makes it too easy for its opponents. The Chinese want to conquer markets and destroy opponents.

And let's be honest, most Youtube tests are bought in some form. But they have reach and influence people who don't know much. And it's a lot of money. 

Finally, I would like to go into a short video of the test of a Sovol SV07. It also has a default profile for input shapers. Can be recalibrated with an accellrometer. Which was also done in this test, unfortunately only in German. The result was that the standard profile was better. The differences weren't big but visible.

RE: Will we get real input shaping (with accelerometer)?

Posted by: @rainer-2

[...]

What you can blame prusa for is the lack of information and time management. [...]

Exactly this is my main concern. Why didn't they just state openly what they are doing? I still would have bought the MK4. 

And just to mention it: I don't really _need_ input shaping for faster prints. For me it's just a hobby and I don't earn any money with 3d printing. If a print takes 1 hour or 5 doesn't really matter.
But I really like to optimize things. Just because it can be done 😉 And because I'm interested in technology.
That's the only reason I flashed my other printers with Klipper firmware and did the calibration process. I did it in both ways, with test print and with accelerometers. Just to see how it works.

RE: Will we get real input shaping (with accelerometer)?

What many people don't consider is that input shaping is also a very nice feature for regular or only slightly faster print speeds. With challenging designs you have to slow down to less than a crawl to get rid of all the ringing and even then you might still see some. With IS however, the ringing is simply gone, also at regular speeds.

While ringing is not harming parts, many of us are in the hobby in order to print nice things. Things are nicer without ringing.

RE: Will we get real input shaping (with accelerometer)?

Sorry, should have mentioned that. 
Of course (especially my two older printers) didn't gain speed due to the input shaping. But they delivered prints with much better quality and without ringing.

Just had speed in mind because printers like the X1C use input shaping to deliver a much higher printing speed (while keeping a good print quality).

RE: Will we get real input shaping (with accelerometer)?

The Babulab's speed comes from the relatively low mass of the moving parts.  That's why the carbon rods, etc. Also, no moving bed. 

Input shaper does nothing but counteract the resonant frequency.  Similar to noise canceling in headphones.  This allows the speed and acceleration to be increased without the ringing that would otherwise be visible.  So the ringing is nothing more than visible resonances. 

The disadvantage is that with higher acceleration or higher damping through input shaping, the corners are rounded.  Therefore, klipper recommends checking the settings made with the accelerometer with a test print.  So if you want more speed with your old printers, you need to significantly increase speed and acceleration. 

The acceleration is also determined with the test print on the klipper documentation.  By the way, it's the same print as in my thread, only with different values.  But everything is described right there.

RE: Will we get real input shaping (with accelerometer)?

Low mass and input shaping work together on the Bambulabs. Why are you restricting the fast printing to the low mass? Because it makes the MK4 looking worse?

Input shaping counteracts the vibrations in general and not only the resonances.

And of course I followed the Klipper documentation. I'm not an idiot.

And of course you will answer again and "know it better" again.
And you only claim to have all the knowledge.

I really like to learn from other people and I'm absolutely open to admit mistakes and fix them with tips from others. 
The way this discussion goes really annoys me and it seems the best for me to leave. 

You of course are not the guy I get my information from. IMHO you are absolute _not_ trustworthy plus your know-it-all attitude paired with your slightly condescendingly way of speaking to other people.
Hey, there would be a way out for me: never contradict you. Always believe your statements. And this will clearly _never_ happen.

So, I will not only leave this thread but the whole forum. These kind of discussion really annoys me and that is not what I expect and like in a forum.

Over and out.