ฝัง
- เผยแพร่เมื่อ 24 พ.ค. 2024
- I've been testing a selection of modifications that you can make with relative ease to your Prusa i3 mk3 3d printer in order to reduce the appearance of the effect of the issue known as 'Issue 602' or inconsistent extrusion. 10 modifications were tested presenting a number of modifications that can be made, some that could be made, and some that probably shouldn't be made.
Affiliate Purchase Link:
Buy Prusa Mk3s+: shop.prusa3d.com/en/51-origin...
You can find the alternative idler design i created on thingiverse: (thingiverse upload currently broken will upload as soon as available)
For access to all the additional materials available for this video, become a Patron at: / vector3d
You can find the 3d printable parts for the Prusa mk3 here: www.prusa3d.com/prusa-i3-prin...
Gear: www.vector3d.co.uk/my-gear
Blog: www.vector3d.co.uk/blog
Website: vector3d.co.uk/ - วิทยาศาสตร์และเทคโนโลยี
Thank you for putting in so much time on the upgrades and documenting them. I'm in no rush to put any of them into place now, which cheers me up, as I hate taking the printers to bits! Great Video and good to see someone from the UK making such good content.
Thanks for the testing and conclusions and big thanks to everybody in the community chasing this issue!
Thanks for all this work!, great informative video.
Holy crap. A professional effort and an all information video on TH-cam.. kudos and cheers and respect.
Thx for all your hard/thorough work on this, and your other videos. Keep up the great work...
I am amazed by the effort you invested into this! Fantastic video.
I realize it is kinda randomly asking but does anyone know a good website to watch new movies online ?
@Joel Brandon Try flixzone. You can find it on google =)
@Travis Rodrigo Yea, I have been using Flixzone for months myself :)
@Travis Rodrigo Thank you, signed up and it seems like a nice service :) I appreciate it!
@Joel Brandon you are welcome xD
Very well done. What I would like to see in the future is what the mods were supposed to resolve and then again did they.
HOW have I not seen this. This is a Warkocki-class video. Great info and thank you.
Someone buy this guy a beer 🍺. You tested your theories from previous videos and gave completely honest results. Very informative. Thanks man!
great video, very informative, thanks for your research!
Thanks so much for all the hard work. The detailing you have gone through to show at each stage and their results will help many people all over, and it won't be just for those with a MK3. Your awesomely meticulous testing of the various upgrades will no doubt be used as reference. Now take a rest and do something less straining on the brain.
In regards to the 603 issue itself, Prusa himself has stated the problem is with the Bondtech's drive gears (th-cam.com/video/CLCpNIx3O0M/w-d-xo.html), and I think that's how it will be swept under the rug.
really good video on testing upgrades
Thank you!
Dear C.R.T.! Wonderfully educational video, I have been battling this issue for about 6 months as well, but my research has led me down a completely different line: If I print the cube at a 45 degree angle to the the bed, issue 602 is very highly mitigated for me. This seems to indicate that this is not the extruder's fault, but the fact that there is a repeatability issue in the movement of the axes.
Try printing your single wall cube at a 45 degree angle, so that both X and Y are in motion at all times, you will be amazed by the difference!
beherith66 what exactly do you mean by print the cube at a 45 degree angle? I’m still not sure what you mean by that? Thx also if you have any more updates on this issue that would be great cheers
Thank you - very intersting investiagtions 👍👍👍
Love the overview, very thorough! I'm a bit confused about the X ends and the Z top and bottom - those are someone's modifications to the Prusa, or you just reprinted the stock parts? Can you link the models you tested if not the stock parts? If they are the stock parts, what explains the improvement?
Nice, thanks for sharing😀
Impressive work.
So Informative !
Awesome work
Hi C.R.T, good job on the video. I woluld like to know where can I find the updated X Ends to replace as mention in Test 10? & I got my printer right before Prusa release the R3 extruders,
if not mistaken mine is R2 & did Prusa release any updated for the X Ends since launch of Mk3?
Great video !
ok first off: amazing work. secondly i think cube 7 is important: look at how consistent it is horizontally. the layers may look different from each other, but each layer is the same entirely throughout.
When you performed the idler change did you also (re)calibrate the steps/mm for the extruder?
I heard recently that issue 602 is down to the stepper drivers prusa are using on their boards. The Prusa Mini for example, does not have the 602 issue, and is using other stepper drivers.
Thanks for you hard work, that was a very interesting video. How much consistency do you see for a given solution? I'm wondering if you are really looking at print-to-print variations instead of
the consequences of some of the changes.
Thanks. I was only commenting from a general scientific experimental design angle. I'm sure you know what you are talking about :)
You may try to put rubber washers (standoff or the like) on the filament cooler fan. I have seen that fan vibration (specifically from a cheaper blower) to affect print quality by adding noise to a layer. This can be a bit chaotic as it does not always manifest itself the same way. But when removed (or replaced with an expensive blower) the issue is removed. The effect is close enough to warrant a try as it may be an element of the problem and not the entire issue.
do you have links to where you purchased parts such as the belts and bearings?
WOW...lists of effort...nice work...I thought you would have removed the PLA bearings before doing the R3 upgrade as they had a negative effect and may have impacted any improvements of the R3. Most likely wouldn’t have though
I think you deserve more attention in TH-cam
You had also recommended using very expensive Misumi bearings. Will you give those a test as well?
First Craftbot models also used direct drive extruder like the Mk3 and also displayed 602 issue specifically the woven basket look of the walls . Those were using Tmc 2208 drivers and proprietary firmware which leads me to believe drivers and firmware have nothing to do with the problem but that the culprit is the filament driving mechanism . Indeed in Bowden setups or geared extruders like the Titan or Bmg the 602 issue is almost non existant or at least greatly diminished . I suspect that a very short and constrained filament path from gear to nozzle and the fact that direct drive means no backlash to absorb some of the drive shocks means that all vibration results in extrusion pressure altering which leads to the inconsistent layer problem .
Hi CRT, are you using different belts than the prusa MK3 uses? If so, which belt are you using?
Awesome peace of science
Nice video! Do you ever consider updating to the Moons’ stepper motor for the extruder?
Here's a link about the Moons' stepper: prusacommunity.com/to-the-moons-and-back/
Great video.! BTW, not sure why anyone would swap out the recirculating ball bearings with pla bushings - that is just crazy talk.
nice vid.
what was your belts tension values?
I think main problem is low resolution + 256 microsteps. One rotation of the bondtech gears have only 200 full steps. Microsteps do not have much torque. Direct-Extruder do not have the "spring" effect bowden extruders have. So "randomly" only fullsteps tranport the filament, but sometimes also parts of the steps do so - it's not consistant.
By higher resolution of full steps you reduce this problem a lot. I tried thing:3165551 and an original Bondtech BMG (you may also use a Dual Drive BMG from china, as you can change the gears to the original Bondtech ones you already have). Due to the 3:1 ratio extrusion inconsisty seem to be gone. You also can change to a pencake stepper (to reduce moving mass) or a 0.9° stepper (for doubling resolution again) or both (if you can get a 0.9 pencake).
Also the prusas biggest problem is that the vertical linear rods are attached to the frame with those printed plastic brackets. I Think removing those and replace them with linear rails that are attached to the vertical frame would improve the printer a lot.
602 to me looks like a calibration problem with linear advance which makes sense since linear advance is trying to cause the extruder to extrude t the same rate during accelerations. If its messed up, the extrusion will be messed up. Have you just done a linear advance calibration per the book?
you are right, r3 upgrade is for better bridging
When you were making your comparisons, did you modify your printer back to its original state before testing the next mod? That will have major effects on how each mod affects your prints. It looks like you just went from mod to mod in comparisons which acres your results.
But you weren't doing just one test so your argument is fairly invalid. You're making combinations with upgrades which create new outcomes by themselves instead of testing the individual upgrades.
I must agree is this, the effect of multiple mods, makes the results unpredictable. Fx the any good from the R3 extruder might have been negated by the pla bearings. Or just improved the same thing as another mod did that was already on.
I really appreciate the work, but the methodology is a bit flawed it seems
The issue is still there. One mod could hide or make other mods worse. The test is valid in the sense that if you make the exact combination of mods as you, but not as useful for the result of each mod by it self.
If you test from base mk3 each time, you might end up with same results, and you might not. But the methodology is flawed if you want the result of each mods impact by itself.
All mod's must be seen as an example of things you might try to improve print quality. There is no guarantee that your situation is particularly similar to the situation Sanlander tried it on, and neither would it if he had made the test more reproducable for his particular machine. That being said, there is one test that I think is invalid as per your argument. That is the belt replacement. A good test would be to tension the existing belt very carefully, and position the belt very carefully and have "Settings" report a tension that is very near the recommended tension. After replacement the exact same situation should be reproduced, including in particular the tension. Then it makes sense to compare the prints.
I suspect that the outcome would be that it makes very little difference. A logical recommandation would be that if you use your printer enough to discover wear on the belts, or if you need frequent readjustment of the tension, that would be a sign to replace the belt with a high quality one.
One of the things I suspect is the most relevant to print quality is the force with which bearings are screwed. After I did that with much feel, my vertical surfaces are an absolutely flawless regular pattern of 0.2 mm layers, way better than what I see in this video on e.g. the cube. I can't imagine I could or would want to improve on it.
This would suggest an experiment with a controller momentum screwdriver set to different tightness.
Hi, what replacement belts did you use?
I Think the idler mod pretty much solved your 602 problem. Your print looked way cleaner however since you have other problems they appeared more prominent. You should try it again and play a bit more with the tension maybe. Oh and there is one more mod that can help: reverse bowden tube.
Thank you for all your hard work. Could please explain or point to a reference on how to do the recommend fixes. Example realign z nuts. Once again thank you very much I enjoy your channel.
Looking for an answer to my question concerning temperature setup for ABS.
When I start my setup for ABS preheat is set to the filament of choice which in this case is ABS. The temperature ramps up and settles in to match the target temp. When I select the item to print the printer changes the target temp for both the nozzle and bed to that of the a PLA setting. Ideas why or ifyou have seen this happen?
Thank you for the definitive answer. You nailed it. Again thank you.
You said adding the soft/felt feet made the unit louder, did you mean quieter? (can't think how it would make it louder)
Original feet are soft rubber
How much percentage in field and which material used...?
Try the Bondtech BMG extruder for prusa mk3, they say that it eliminated the 602
could you please print the same test objects with your Ender 3 so we can compare the Mk3 to the best value printer...
Can you print and review the Skelestruder design that Jltx has made? Understand if you didn't want to go there, but it would be good to see if gear reduction on the extruder makes any real difference to the 602 issue.
Try to change the entire extruder to E3d Titan. For some reason i think the Bondtech firm grip on the filament is not helping to have a fine consistent extrusion. I did that on my prusa and i had a significant improvement on my prints.
I had unconsistent extrution on my mk3s tried many things.new titaium heatbreak . Bondtech extruder , nozzle ,checking belts nothing help ..but when i changed out the original Main board whit a duet2 wifi board it went away
Can you post your alternative idler design?
why are you calling changing the tightness of fasters upgrades? shouldn't that just be maintenance and always to manual spec/snug?
the 602 you have there doesn't look like an extruder problem, the light and camera work you did was really good, I was able to notice some things: 1st, the extruder steps are not smooth, as noted by the vertical lines you can see on cube 13, 2nd the frame seems to wobble a bit, maybe by .05mm or so, causing those bad layer alignments, 3rd there's an artifact on layer alignment about every 5mm, than can be a mechanical problem, also the rods or bearings could not be true, some h7 tolerance rods and -.03 tolerance bronze bushings or encased igus bushings could be the solution, or maybe it's some printed part that just wobbles, and the wobble will be axially in-line with the moving part's direction, the rod holders for the bed look flimsy for such a high caliber machine
checking everything's tolerance is a good idea, also the extruder itself could have a bad design, or maybe the settings you're using aren't perfectly tuned to that machine, because I've seen some amazing mk3's out there, whose prints don't even look printed, they look like renders!
Probably just removing the spool off the frame would make a lot of those issues disappear.
In a video of a Prusa tour by Thomas Sanladerer, Joseph Prusa admitted that this issues may be due to the Bondtech gears. See minute 13:56 of th-cam.com/video/CLCpNIx3O0M/w-d-xo.html . If you can get your hands on another set of Bondtechj drive gears you can add that to your testing procedure.
I would have liked the video, but it’s currently at 602 likes...
nice one! for testing i have spend 4 kg of PETG)))
do you by any chance have a another youtube channel called Zerkovich?
Nope
one thing I've always thought, and many people say: their lm8uu mounting is the shittiest ever, NEVER just zip tie them, it's too sloppy, you need a proper block to hold them in place snugly, maybe this is the cause? also, the 8mm rods are just bad, they're too flexible, 10 or even better, 12mm rods give much better print quality
Hi Demb question. What is 602?
Thanks
Google: prusa github issue 602
This'll make you laugh. I spent hours in OpenSCAD trying to make single-thickness walled cube/cylinder/cone STL models that would slice with one path on the wall, without success for the cylinder and cone. Then yesterday it suddenly clicked - just use a trivial normal cube/cylinder/cone and set the slicer (Cura) settings to no infill, no top/bottom layers, wall thickness 1. *headdesk*
Try to remove spool from the top of frame to separated holder
Is issue 602 fixed on the prusa i3 mk3s?
i was also wondering this, i would imagine so as they built a new extruder for it
@@theaces3697 I found a solution! www.reddit.com/r/prusa3d/comments/c4wxmp/prusa_issue_602_actually_solved_mk3s_and_probably/
You should try increasing your extrusion temp about 20 deg. My mk3 prints look like that when it is set to usual PLA temps, the extruder seems like its around 20 deg less that what the screen says, I have seen others also with the same issue.
@@Vector3DP True, I had wondered if there were just some dodgy thermisters.
It would be much better to have the two images side by side to have a better visual for the audience.
Thanks, i'll keep that in mind for next time.
Just make a cylinder and put the bearings in, all knows that plastic bearings get deformed without a proper even mount. Correctly made they are way better, very easy to print as well.
Didn't Jo Prusa recently reveal the cause of that issue? It seems to be the Bondtech extruder gears. Cheers!
But did they actually fix it? Given the premium price of the printer that you're paying because it's supposed to be "tweak free" compared to the competition, a major issue like this that you have to fix yourself seems to defeat the point.
Great test procedures and very helpful video. Consider holding your hands more still - extremely distracting!
After struggling for months with the 602 and related print quality issues, after trying all reasonably cheap and effective mods I could do, I could not solve this... until I've built myself a Skelestruder (w\ Zaribo Pancake motor, the lightweight shaft, ..) by JLTX! Since the extruder rebuild/switch, It's been printing consistently easily and fast, the extruder (end even X axis) motor is completely cool at all times, I am speechless. And it all can cost just couple $. Here's the link: www.thingiverse.com/thing:2845416
Adam, if you've still haven't solved this issue yourself, please give at least the Skelestruder a try, but don't overlook the JLTX's Taurus axis upgrade "line".
As I’m a professional in IV&V, the test between 11 and 12 show that you did an accumulative testing ,
U said the PLA bearing (11) was not good and it seems u made test 12 with the „bad“ PLA bearing.
So test 12 had to fail!
IV&V = Installation, Validation & Verification
Yet those prints look worse than a properly tuned ender-3 :/
But this goes to show, frame rigidity matters most, and Prusa has been always crap on that. See how the feets change the results, if the frame was sturdy, it would have 0 impact.
PLA bearings... now I’ve heard everything
I believe PLA bearings were experimented with a fair bit early in the RepRap movement, but there's a reason that they aren't commonly used these days! 😹 (I mean, genuine Igus bushings are pretty affordable if you want to go down the polymer bearing route...)
602 on my MK3S+ Is really becoming a pain in the A@@:~!!
Hmmm. A lot of work but these better prints look worse than my Ender 3.
One note: every change and print is lowering the mass of your spool. I don't get why designs put the spool on the printer frame. You should try putting the spool beside the printer.
Why would anyone use PLA for bearings?
so... half of the "mods" are actually maintenance
Why in 4K? You don’t get any more details by upscaling. Instead you generate more traffic for nothing.
C.R.T , ok, I get it. 🤦♂️ You mean the actual photos you display in the video. 😜
blurry
no need for 4k of fullHD, just use 720p, for the unsharp picture that would be enough
@@Vector3DP I read somewhere you can adjust the lens of the raspberry pi camera manually. I did it with a small 6.5 mm plastic cap I had been laying around. Just do delicate 1° pushes, THEN the 1080p@10fps would be soo cool :-D
Hi, shooting in HD and thde uploading in 4 k does not increase the detail. How can it?. All that's happened is that some form of interpolation has taken place, but the apparent extra detail is all fake. Stick to HD, no one can resolve 4K TV with their eyesight unless just a few feet away from a 65 inch TV, and most people are watching on tablets or phone sized screens. It can be worth while doing it the other way round, shooting in 4K and down sampling to HD, but eve that is fraught with problems - extra artifacts can appear unless you use quality software to do it. BobUK.
The photos are much higher than 4k resolution.
I wish you mentioned the purpose of this testing earlier in the video. Kinda tough to watch
If it is not broken don't fix it.
Subscriber #5147
Looks like 480p to me
Uhm... wow?
Shot in 1080p, but uploaded in 4k? Am I tired or does that not even make sense? If that' was the case, why don't you just record in 320p to save space on your memory card, then upload in 4k? Wtf?
Upscaled the output from editing. The 1080p footage won't look any better, but the high res images will be closer to true resolution.
@@Vector3DP oooh, righto! Got ya!
Simply the Mk3 extruder design is shit. Mk2 was good and the only thing needed was cast parts vs printed for accuracy. Superglue rubber feet and you will see a big difference. Tomas just wont admit that his mk3 extruder design was a step back not forward. Especially the filament sensor and the dual knurled wheel feed. I think a felt seal to wipe the extrusion prior to entering the hotent is helpful as is partial bowden feed at least at the entry of the extruder. The rest is just assembly if your zip ties arent spot on and using a torque driver elsewhere for assembly gives uniformity. While prusa is a great printer. The price value outcome is not what it should be
Its not a prusa i3 mk3 its a prusa i3 mk3s
definitelya a nice video but your audio is a pain in ass and you talk much to fast for a non native to understand all of it. Baybe you are tinking on improving so non native speaker can enjoy your videos too.
Issue #602 solved! www.reddit.com/r/prusa3d/comments/c4wxmp/prusa_issue_602_actually_solved_mk3s_and_probably/
Please test!
PLA bushings work, but you have to take a few things into account. For a stock i3 i'd say only use them on the X and Y for noise reasons. Also those are the SHITTY pla bearings, id recomend you make your own, the ones on thingyverse are shit and badly designed, obviously not by an engineer
why does the mk3 have such bad extrusion!!! my anet a8 (although it has 2 years of my life of work into it) prints better than that! what???!
@Jim Bob mine does to lol, the mk3 isn't some magical printer, just very hyped