3D Printer Belt Tension: Its ruining your calibration! Ender 3 V2 Print Farm Accuracy Issues

[@MakerMeraki]

Any time you adjust belt tension, you are going to change dimensional accuracy. And since their wear parts, over time the belts are going to stretch. Checking for deviations should be part of your monthly maintenance cycle.

[@qumefox]

I just stumbled on this video and was going to say this. The specific belt tension doesn't matter providing it's not too loose which leads to inaccuracy, or too tight, which leads to premature belt stretch and failure. The main thing is to calibrate the steps/mm for the tension you use, instead of of trying to figure out a way to reliably set an exact tension, which is doable but way more complicated than just checking the dimensional accuracy occasionally and fine tuning steps/mm.

[@JamieHarveyJr]

I’m thinking of designing a linear-rate spring preload tensioner. Unlock tensioner, let spring add preload, lock tensioner. Same tension every time.

[@Redo3D]

That would be an excellent add-on! Only thing left to do is find the right spring rate!

[@JamieHarveyJr]

@@Redo3D a cheap fish scale will get you in the ballpark quickly. Just pull on the belt until you are happy and read the scale :) Amazon: $5.99 Polychrome Gear Digital Fish Scale

[@Eden_M]

Thanks for the all the tips you share here man, I guess you understand the true nature/tendencies of these machines only when you run a printer farm, so your advice is really valuable, thanks again!

[@Redo3D]

Thanks for commenting and watching +

[@arminth]

It's aprinter, not a guitar! These belts don't need to be tensioned like a guitar string. Just enough to take any slack out of the entire belttravel. The problem of these new belt tensioners is, that due to the screw-meachnism which acts as a leaver, you don't have the right feeling how stromg you are actually really pulling on that belt! Never tension to elastic (or even permanent) deformation of the belt!

[@dodgyaz]

Well said.

[@MDigitalProductions]

Learned that the hard way 🙃

[@mmd7774]

I don't say this often because normally people say this to me but you're a really smart guy!

[@Redo3D]

I appreciate that!

[@VitorLima-yr9ti]

Use a guitar tuner app to measure the freaquence of the belt, a gt2 belt needs to be tensioned at the freaquence of 28531hz per mm, you need to divid the frequence by the lengh of the belt, like in ender 3 it gonna be 250mm, and the tune freaquence gonna be 114,1hz for the belt tension on the ender 3 for exemple, but you need to mesure it by yourself.

[@Redo3D]

I'll definitely try this out! Thank you.

[@bettis211]

This is better than measuring the torque on the tensioner because it actually measures the tension on the belt which is what you care about for the issue you found. I have been using the 28531 hz/mm calculation to tension all of my ender 3 belts for a while and it is working well.

[@anwarahmad4354]

I've been using guitar tuner app for my delta printer for years. It helps to get the correct tension on all 3 axis. Just make sure the length of the tuned belts are at the same length.

[@INZ4NEDISCIPLE]

@@bettis211 is there an existing tutorial on this? I've never heard of this method but I'm very intrigued

[@zg9xUmvV]

which app do you use?

[@foxprojects247]

I really haven't found too many videos that really delve into these topics, thanks!

[@eldricliew6223]

I'm honestly surprised 3d printers even work this well. When i was working on my cnc machine calculations, i noticed that just the belt reducers for the ballscrews can cause non trivial stiffness loss. Short/thick belts were required.

[@alarson1981]

Cogged belts do not need much tension to work correctly. All you need is enough tension to keep them from skipping teeth - much more than that and all you do is add friction to the system. Particularly with 3d printers where there isn't much force on any of the moving parts. I have personally pushed over a thousand pounds of force through a system using that type of gt timing belts that would be considered too lose by most. IMHO you got goofy step per mm measurements because your belts were too tight and causing the motors to miss steps.

[@85rengi]

That's what I thougth too. The movement should only be determinated by the circumference of the pulley. But I think he took out some backlash, because his case got bigger.

[@dirediredude]

Agreed here

[@SianaGearz]

@@85rengi the belts are made undersized and stretch to length under correct tension. It's easy to notice, when the belt is engaged but not fully tensioned, it's a particular length right, but as you tension it, the tensioner moves by about 1-2mm or so, maybe more depending on length. But the number of teeth isn't increasing, so there's about 1% of difference in real steps/mm. This isn't a missed steps issue. Because any offset that endures throughout the print that you can achieve will be multiple of full step, so either 0.2 or 0.16mm depending on pulley type. "Lost" microsteps are a misnomer, even when microsteps are not expressed, they still advance the counter that rotates the magnetic field and it will just snap to the correct position at the next quarter step. The result are not dimensional issues but surface quality issues, salmon skin.

[@TimothyAllenCampbellGoTimothy]

@@SianaGearz as my model is printing, about every 2 hours it moves over 1.6 mm or about .065 in. Then continues printing. I'm not sure if it's a loose belt, missed steps or missing tooth. Just got the printer as a rebuilt. They must not have tested it. I tried tightening the belt we'll see what happens.

[@hilaryturner6607]

Your on the right track. Your belts don't have to be really tight. They just need to be tight enough to prevent movement side to side or vibrations. As far as the belts are concerned they are gear toothed belts. Any gear is measured from center of the crest of the gear so to measure multiple gear tooths it should be from center to center of the crest of the number of gear teeth you choose. The big thing is that you get both the X and Y axis's as close to the same measurement as you can get it. For example if one or the other belt is greater than the other your movements are going to be rectangular not square as it should be. Just remember that if your model is large the accumulated error will be made by the number of layers your printing. Then there is the round which will be made oval the same way as a square is made rectangular by mismatched tension on the belts. Also you can mark the smooth side of the belt by the number of teeth you choose to gauge your measurement because it is a belt and will stretch the same regardless of which side you measure just measure from the same points on all belts. Once you have done one to your satisfaction repeat with all your machens. If you just had one write your numbers down so that over time you can gauge if the belt has stretched to the point it should be changed.

[@abyteuser6297]

This channel deserves way more views. Your calibration tips are super useful. Much appreciated, thank you!

[@sumitkhachane1956]

There's no need to tighten the belts so much. It ruins the bearings in the stepper motors due to the cantilever tension. The tension should be just enough that the belt doesn't skip.

[@jasonwoody8041]

I believe Gates high temp belts have wire in them like a car tire does. That should help belt stretch a lot as well. My belts are not near that tight.

[@zg9xUmvV]

do you have a part number? or amazon link? thank you.

[@rexxx927]

100%

[@beauslim]

I don't think you'll get all your machines to be the same even with a lot of work. A belt should be just tight enough that it never slips. Definitely check belts for wear and keep them snug, but they stretch over their life. Too tight and they will stretch prematurely or even bend a stepper motor's shaft. Print a calibration item once a week and tune each machine's settings to correct for deviations.

[@paulwills7902]

Another helpful video thanks. My ender v2 is being returned tomorrow as I haven’t achieved a single print with it yet. It really has been the most difficult and frustrating printer I have ever owned and currently have 5 others that just work. I have noticed the bed is not flat but can’t afford to waste anymore time on the machine. Incidentally Creality support has been non existent such a shame because I so wanted to like this ender 3.

[@Redo3D]

Sorry you're having so much trouble Paul. Sometimes I wonder if the ender 3 was the right printer for the farm. But I tell you what it has taught me a lot about 3d printing!

[@zk_6312]

Quality seems a lost art these days in favor low quality products for more money. Products from China seem especially bad in my opinion though they are improving. I have an Ender 3 Pro and it took me over a year to get quality prints out of it. Part of that was my learning curve I am sure, but some of it was the printer itself. Putting dual lead screws helped considerably. My point is that unless you are willing to pay several thousand dollars for a professional 3d printer, expecting a cheap 3d printer from China to work out of the box in a consistent manner might be unrealistic. They can and do sometimes take a lot of work before they work properly.

[@hathzorz]

One thing about the Ender 3 v2 style belt tensioners -- if yours are like the ones that came with mine, their construction is garbage. If you ever take one off and look at its mechanism you'll see what I mean. The tightening knob pulls against a *press fitted* bolt into the second plastic piece. If you over tighten the belts you literally pull out the bolt's head from its press fitted hole and then the tensioner is irreparable. I suggest replacing the belt tensioners with better ones that have better mechanisms.

[@t3chninja_official]

Good catch! I mean I tighten my belts and then run a calibration cube to make adjustments. Things wear out over time, belts slightly stretch. Shoot at work this last week a new Tech teacher wanted to know why I was printing a calibration cube on her new printer she borrowed. I said to calibrate and she was like these printers advertise they are super very accurate. I did the math and proved her wrong, made adjustments and then it was 20 x 20 x 20. No words....

[@ExplodingWaffle101]

very interested in the results of your solution. belts are one of the weakest links in 3d printers- it's why cnc machines use screws instead.

[@jakeengland1430]

@Lassi Kinnunen 81 agreed and chief problem among them is they are slow as hell. my cnc mill uses acme threaded lead screws and with a small nema 23 i can run them at 2000mm/min which sounds fast but 3d printers run in mm/s which is aprox 33mm/s which in 3d printer terms is damn sloooowwww

[@moo00se]

The best prints use belts and linear rails.

[@DavidtheSwarfer]

I use a fishing spring balance to tension all my ender 5 belts to the same value, 8kg iirc. 12kg on my cnc router. I was going to suggest the guitar tuner as I don’t think a torque wrench will be consistent due to friction, but I see someone already mentioned it.

[@Redo3D]

Torque wrench definitely won't work. Thanks for commenting +

[@wewyllenium]

This issue drove me nuts. Same thing I did. Measured the belts as per spec and when tensioned. The thing is that tension keeps prints clean but excessive tension offs the dimensional accuracy of prints. So I ended up buying belts that give me correct dims whether they are tensioned or not.

[@EGGSHL]

What belts did you use?

[@wewyllenium]

@@EGGSHL I got measured belts from creality.

[@bluemonxterDIY]

That's an eye opener. This goes to show that you absolutely need to calibrate the x & y steps/mm. Most guides show e steps/mm but skip over the importance of the x and y. And we probably need to re-calibrate the x and y over time as the belt stretches permanently.

[@TechKnightOne]

Hopping if you come up with a solution you make a video. I always thought it was weird the we just tightened the belt we felt comfortable with how tight they are. For instance the band say at has a gauge the tells you when it's to loose, right, and then to tight. And if it not to the middle ot the gauge which is where it should the cuts are off and the longevity of the blad differs. Thanks for the info.

[@madr8b]

I would think that any thing under tension would be prone to stretching. Also I would think that it would also depend on how old the belt is and how long it has been under tension. The question is how much stretch and how much tension is acceptable.

[@davidturner5418]

Pardon my ignorance, but would auto belt tension arm work. A toothed belt does not need to be that tight it just need to mesh correctly at the idles end and drive end.

[@glowpon3]

Another reason I like my leadscrew/ballscrew axis mod instead of the belts I started with.

[@Pfuschers]

I've wondered about the dimensional difference between the X and Y axes on my Ender3 Pro. This seems to be actually the belt. Thanks for the tip!

[@Lifesbeengood]

Thank you for this video. Been having problems with accuracy and didn’t realize the belts were causing it. Logically this make the most sense. Does anybody know what the proper tension would be?

[@josecabrera5632]

What is the accepted error tolerance? The melted plastic is also non uniform.

[@dodgyaz]

There will always be stretch in belts. Car fan belts are the same. Swap the thumbscrew with a Nylock lock nut. You will have to use a socket to adjust them, but it won't move. The belts will stretch over time as well. Just add belt checks to your weekly maintenance regime.

[@TheNoiseOfGold]

I WONDER IF CHAINS WOULD WORK?, OR STEEL BELTS?

[@lifehackertips]

This is actually shocking and I don’t understand why no one talks about it. A 0.8mm difference in a 117mm span is huge. Most of us are printing at 0.2mm layers but our prints could be almost 1mm off? They have all these auto levelers, why isn’t there more being done to make sure belts are accurately tightened? Feel like there should be endstop switches underneath the bed exactly 200mm apart, and the printer should move the bed to make sure it’s hitting both. If it doesn’t, it should have an error message so you can check the belt.

[@stephandevriesere3667]

You have problems because you are overstretching your belt. It does not need to be that tight, just enough tension to keep it running smooth is enough. You do not have to stretch you belt, it will only ware down way faster. And because you over tension your belt it will stretch to a point it will stay that way. See it as a elastic band, if you keep stretching it and keep it under tension it will stay stretched and to loose after some time.

[@Damjanhd]

what if you use spring to constantly tight your belt the same force.

[@Redo3D]

This is what I'm looking into. A linear rate extension spring. I'm working on a solution now with this setup.

[@dananorth895]

Springs will also vary with temp/barometric pressure.

[@twinturbostang]

Great info! Had no idea the belts stretched that much just due to tension.

[@Viceray4760]

Also I noticed on my printer, the chrome piece of the tensioner, made the belt move to the right side whenever the y-axis would go back and fourth . It occurred to me that inside the tensioner it was wobbling and turning so I put painters tape on the inside- side walls to make it more grounded and ever since then I’ve been having perfect prints . Has anyone had this problem?

[@SP4CEBAR]

I don't think it works like that: the diameter of the drive wheel doesn't change, so when the stepper motor rotates, the belt should move a fixed amount independent of the belt tension (as long as the belt isn't really loose)

[@Redo3D]

I suspect any backlash might be causing the difference.

[@TroyMackay]

Not a fan of those free spinning belt tension knobs. Give me a locking screw any day. How much would just be movement from vibration over time?

[@Redo3D]

Very true. I was thinking of installing low friction teflon washers, but will for sure have to have some lock nuts or else it might back out on its own.

[@johnboy1045]

Is the belt round the wrong way as I thought the coged part of the belt went face down not up

[@wk5199]

@ 2:07 he said he flipped it for the experiment

[@TroyMackay]

Did you just demonstrate a calibration method using calipers? Mark a 20 teeth reference and adjust tension to a predetermined length. I would imagine those belts would be manufactured to a specific spec for pitch and tension.

[@Redo3D]

Interesting idea, That did cross the mind, but I believe the gear also has some play in it. Would be a worth while experiment so you don't have to adjust e-steps but rather set proper tension!

[@HuckThis1971]

You have to take temperature in account as well. Things expand and contract! Unless it's carbon fiber or super invar.

[@cours458]

no wayyyyyyy, so when I re-tension my belt I need to redo the estep configuration of the axis that uses that belt :o, thanks man.

[@Rebar77_real]

Lots of guitar tuner suggestions but how about a drum skin tuner style of jig? They work by pressing on a point in the drum skin and measure its push back resistance. You could maybe measure between a spanned length with a jig. Just a thought. The inch/lb tuning should help a lot too though. It would be interesting to know how loose a belt could get and keep a decent accuracy, eh? For longevity to keep them as not-overstretched as possible, maybe. gg on the brain food!

[@UnderDuress]

Got a follow up video for this?

[@SamFerlauto]

Use a belt tension tester. Either an app on your phone, or a real device that measures belt tension like a Texa TTC or a very expensive optical type like an Optibelt TT meter. I general set mine up to 100hz and calibrate the X and Y. Then periodically say every 3 months test and reset to the 100hz frequency. If you do this then you know all your machines are equally setup and the step calibration on all your machines should be very close if not identical of each other

[@Redo3D]

Definitely going to have to put this into the maintenance schedule. Thanks for the suggestion!

[@BeFree0_0]

I just came across a issue I think might be because of this same thing. I printed a hot end cooler for my ender 3 v2 and noticed my y axis is 0.5mm shorter than my x axis

[@BeFree0_0]

I just watched another video on belt tension and it said 25 newton's is recommend, did you give the nut and Teflon washer a try

[@lifehackertips]

I’m actually here because x axis belt broke on my ender 3 v2. It’s been printing fine and then it just broke. Maybe I had it too tight? Maybe I should loosen my belts a little.

[@infinitewars6373]

torqueing sounds good but i think a lot of other factors will still play its own game in here....let us know what you get to...

[@karlosp75]

So if you stretch a rubber belt, it gets longer... weird...lol

[@frankcastle2814]

Did you ever use springs or just those silicone thick washers? I am considering trying them out. My bed seems to drift quickly after some prints (although I wonder if it isn't my x gantry sometimes. I added a second z axis, but that doesn't seem to fix issues 100%. I did just order two of those digital probe gauges and plan to use them to really get the gantry spot on level for the future. I had been using a square and some meter feeler gauges to eyeball the level. I also like the idea of using a gauge to tram my build platform instead of the CR prob (which is slow as F).

[@Redo3D]

I use silicone spacers and have not had bed issues since!

[@stephan.scharf]

I don't agree. Movement distances are fixed by teeth of motor pulley. Any stretching of your belts should not influence anything in your printng precision. This is the theory.

[@Redo3D]

I would agree as well, but the pulley and belts are not perfectly match, there can be some backlash as someone suggested. This backlash can be the reason for the variation in print size. Ive been testing this morning and noticed after a certain amount of tension, the size does not change. What I want to get to the bottom of, is how much tension will remove backlash

[@stephan.scharf]

@@Redo3D theory and practice will never match, haha Backlash phenomenon could be the reason for your printing tolerances.

[@surronzak8154]

So does it skip tooth ?

[@woodwaker1]

A very interesting video. I have 5 printers that I use to produce trinkets to give away. I'm so glad I don't do this for a living.

[@Redo3D]

It's definitely a lot of work. Especially if you need accurate prints and quality from a cheap printer.

[@scottydice1426]

Why would my prints have smaller/tighter dimensions when printed? it seems like everything I print lately is smaller by 2-3mm? I've done several calibrations but did not fix. Could it be something in my slicer? Or because I changed from a Bowden tube to Direct Drive? Any thoughts would be greatly appreciated. Thanks

[@Redo3D]

I would for sure check your steps per mm calibration. If that doesn’t work try a new slicer. 2-3mm is a lot so something is definitely off.

[@scottydice1426]

@@Redo3D I've calibrated steps a while ago and increased my steps from 93 to 96 and a flow rate of 80 to get the correct wall dimensions. But I thought maybe if I redo the firmware would fix it to. I did that this past weekend. I will try that again and reply back. Just doing tests. Thank you!

[@TimothyAllenCampbellGoTimothy]

3D printers only have an accuracy of about .040 per inch. When I make my models I print them at 103% to get a more accurate size. But then the length is 3% long.

[@scottydice1426]

@@TimothyAllenCampbellGoTimothy Thanks Tim! I'm actually getting about 1.5mm in my prints being to small. and sum parts I'm printing really is a big difference. Yeah I can sand but its every wall and hole. It's weird. I'm currently re-installing cura. I thought about trying ideamaker but maybe eventually. would my extruder steps change alot switching from a bowden tube to a direct? I did a calibration and went from 93 (default for Ender3) to 95.6 steps? with a 80 flow rate. But don't think I should have to tweak it that much? I'm still wrench'n on it off and on. I'll reply when I get close. Thanks!

[@CaIasanctius]

have you compared print results before and after tightening? has consistency improved? did you try different belts to see if consistency of the dimension of the prints changed at all? just curious if doing these changes has improved consistency, or theyre all within certain margin of error inherent on these types of machines.

[@HayabusaRydr]

I'm glad you made this video. I never knew this was a thing. I tested this on my ender with a dial indicator and was amazed the difference in measurement. Most people in general won't care about accuracy but i need all my parts to be. Thanks again

[@bitosdelaplaya]

A really really simple and good video. But im afraid that even you adjust at a defined torque the belt will continue to lengthen. But if you use same belt on all printer you can suppose all the belt lengthen in the same way and so you can establish a sheet with torque/steps. Another way, little youtuber (dont be obfuscated), talk about real problems. Another problem (i have to verify) is that you have to redifined your steps each time your change extruder temp. it seems to me that if you make extruder calibration at 200°c and after you change for 210°c extruded lenght is not the same. So i think that temp tower that all youtuber utilize to determine the good temp are bullshit.

[@zk_6312]

Tension on your belts is important and can cause holes to be more oblong if too loose, but I don't think really tight is a good thing either. Belts stretch that is just their nature and I am not surprised your printers were no longer calibrated properly once you tightened your belts. Probably you should have only tried it on one to see what the effect was rather than adjusting the belts on all the printers at the same time.

[@thelightspeed3d712]

Use a guitar tuner… but the measurement in X and Y don’t matter. As long as your motor isn’t losing steps everything is relative. The belt stretch only matters if it happens during the print

[@plc-cnc3230]

There are special torque knobs for that purpose. The only question how much torque needed.

[@Redo3D]

I have seen these. They look like T handle nut drivers. I have to experiment with how much torque. The Teflon washers will be to remove any friction from the nut and plastic bracket.

[@Badjujubee]

The only other ding is that anytime you make changes to the setup (change tension on belts, unbolt and re-torque anything in the extrusion path etc) you realllllly need to print a validation object to confirm your dimensional tolerances. It stinks, but the flip side of having such cheap and prevalent home manufacturing/cnc setups is now we get to do all the stuff I would have done as a line mechanic/setup person, but on 20 something highly irritable and slightly sketch melty robots, and on my own dime XD

[@Redo3D]

Definitely requires a lot of attention but little by little improvements can be made for longer lasting operation without maintenance.

[@Badjujubee]

@@Redo3D for real, it's the one downside to the extrusion based linear motion systems we use on most of these machines. It's massively overkill, but a 12mm 2gt gates belt might be less prone to creep over time, but you are not squeezing that in the foot print of a 20xx based rail easily XD

[@Badjujubee]

Heck, now I'm tempted to do a project to see if I can measure ring harmonic changes over time. Like bake it into the firmware to detect over shoots with something like Klipper and input shaper, and then have the printer let you tune tension just like an auto tuner on a guitar

[@dananorth895]

Great to hear these topics addressed. I'd like to point out the in tuning an instrument say guitar/god forbid 12 string.Variations in temp change pitch but most particular variations in barometric pressure. You can tune up and it will stay that way untill weather changes then its time to tune up sometimes significantly. Sometimes hour or mths. As long as pressure returns to initial value your in tune again. WHO TOUCHED MY GUITAR!!! LOL.

[@dananorth895]

PS...seems simplest solution might be min. pressure just enough to prevent slip.

[@girenloland]

i don't get it, if you tension the belt, the printer will always go to the end stop switch before every print. If not the printer does not know where is at. The stepper motor and bed will always move correctly. No matter how loose or tight the belt is. As long as you don't adjust while it's printing.

[@85rengi]

I could not get it, how a stretched tooth pitch should influence the travel distance. The pitch of the pulley won't increase and that's what defines your movement, so i've tested it by myself on my E3Pro and made a short video: "kzbin.info/www/bejne/pqW8hmRjfb2qaKM With higher tension I got a even shorter movement due to increased friction (needle moves slowly afterward), but all in all it is negligble .

[@TroyMackay]

Yeah, but the belt will be forced to conform to the pulley where it wraps around (by local stretching, not skipping teeth). Motor turns 20 teeth, so the belt moves 20 teeth but now down the length of the belt that equates to a longer distance travelled. Maybe your belt is a little less stretchy to measure it for such a small movement.

[@natebean151]

Good name brand belt helps like gates. That where known for timing belts on cars

[@Redo3D]

Half my machines have gates belts. I guess some shipped with them and some don't. Majority of my machines were purchased from amazon.

[@Mooblr]

Are the teeth on the belt not upside down..? Edit: nvm didn’t realize you flipped it purposefully.

[@Redo3D]

I explained in the video I flipped the belt for demonstration purposes.

[@Mooblr]

@@Redo3D ah my bad, skimmed video.

[@jodymorrison9495]

Ok so I am just going to say it your belt is on backwards causing slippage??

[@Redo3D]

I stated in the video I flipped the belt on purpose for demonstration

[@1607rosie]

Man you stretch the belt it going to change things. Never thought about it at all need a screw drive. There's got to be away to convert belts to screw drive. Thanks a lot for pointing it out.

[@viru52000]

Screw drive is VERY SLOW, that's why belts are used.

[@1607rosie]

@viru52000 I do agree I think then maybe run along a gear rack . Or have a linear encoder to show position on the gantry. If you want to do repeatable semi precision printing like he is you need todo quality control.

[@zenjaya101]

Btw in their manual, i re read it and says, should set belt the tenion between 0 to 3 mm; the protruding screw that protrude from the tension knob. Im not sure if that matters in the calibration(maybe even between printers?)

[@Redo3D]

Interesting. I have some as tight as I can go and it varies how much the screw is protruding.

[@dirediredude]

Yeah I noticed that as well. All belts stretch, that's the nature of their design. It's a wear item so Creality is basically saying that when you have to set your tension to the point where that bolt head is protruding more than 3mm it's probably time to replace your belt. There's not a good way to check deflection of the belt which is the typical beat to inspect a drive belt ( like in a car ) so I think this is their workaround.

[@CourtneyRoberts1982]

@@Redo3D that's waaay too tight. You just need them snug enough to not slip or create ringing.

[@grathado28]

thank you so much for sharing

[@mrphysics2625]

Belts should be as tight as a guitar string. No more, no less. some of those beds like number 4 that you pushed were waaaaaaay too easy to push, or atleast it looked like it lol. the easier it is the more wobbly itll be. The tighter it is the more harmonically sensitive it is [vibration frequency] there's a sweet spot in between. If you tighten it too much the belt will snap. That being said, belt tension would have almost nothing to do with your esteps thats entirely to do with the extruder. At most its just going to be vibration interference within the motor. So limiting that in the belts would help but what really effects that is the quality of the motor and the board. there is only so much you are going to get out of an ender 3.

[@randomname3894]

Since you’re stretching the belt with that screw.. it’s pretty obvious that it’s stretching the print also. Question is, if we assume the pitch of the belt should be right under perfect stretch.. should we then always calibrate belt tension to match print size.. vs the old way of calculating the motor steps?

[@Redo3D]

This is my thinking. Im going to experiment with calibrating based on the tension of the belt, but I need to find a quick way to measure this. What I may do is measure the 100mm movement of the X and Y based on a fixed object, and make adjustments to the belt and see if i can get it dialed in. I will reset the e-steps all back to factory defaults.

[@SianaGearz]

Problem is you want to measure a 150-200mm move to better than 0.1mm, better to 0.02mm. how do you best do that?

[@igorchak48]

Switch to 9mm belts, I had the same issue on all of my Ender 3 V2's.

[@Redo3D]

I will have to look into this for the Y. For the x axis it looks like the extrusion won't allow a 9mm. I think it would rub. Thanks for the suggestion!

[@paulradford4100]

I'm not sure I understand your point.. you adjust the tension screw .04mm and the marks get .04mm wider.. The stepper still drives a certain distance (properly calibrated) and although the belt has stretched by .04mm the tension screw makes the belt travel .04mm further.. .04mm (belt stretch) - .04mm (tension adjustment) = 0mm difference.. Now, if your tension adjustment is loose the belt can slip on the stepper and prints will be quite poor. Overnight causes friction reducing quality as the stepper has to work hard to move the carriage. Tighten taught, not tight is key

[@Redo3D]

The belt pitch changes. So there is a difference in distance traveled. As someone suggested it might have been backlash that was removed that cause the change in the calibrated distance. Still need to do more testing but it is definitely not staying the same no matter the adjustment.

[@noway8233]

For big objetc (100 mm or more), dont use the typical 20x20x20 cube, yuo can multiply yuor acuracy if yuo use an object similar in dimension of the object yuo gone print.Lets say yuo print a cube of 20 mm and huo mesure 20,10 ,the error is 0.1 but in an object of 20 mm long, lets say yuo print an object of 200 mm long , so yuo can expect an error of 0.1x10 = 1mm, so its gone mesure 201 mm and thats cuold be wrong for some pices.To solve this print an objec 100 mm long and calilbrate yuor step, if yuo have 100,1 in 200 mm yuo will have 200,2 mm so an error of 0,2 vs and error of 1mm, i hope this help yuo with yuor tolerance problem,bye

[@kazolar]

This is due to cheap belts used in Creality machines, gates, or gates clone belts which have a reinforced core don't have this issue, rubber non-reinforced belts stretch. Using cheap belts from Amazon, I had trouble getting consistent movements, switched to gates, and issues went away.

[@zg9xUmvV]

do you have a part number for these belts? amazon link? thank you.

[@Suburbanwoodsman37]

Do your belts have teeth on both sides?

[@Redo3D]

They don't, I just flipped it over for demonstration

[@bernardtarver]

You would think you would do a some lifecycle testing.

[@fanek706]

what springs are you using in your ender 3 v2

[@Redo3D]

I am using silicone bedmounts from Amazon. amzn.to/3z4zqM8

[@adamzapotok2681]

SUPER HELPFUL!

[@denvermatt]

very interesting thank you

[@fataxe1]

But some good gates brand belts with aramid fill. The Chinese belts could have steel belts in them or some other random construction method. The best way to control your QC is to buy repeatable, well known brands for motion parts. Gates makes the best belts in the business and they'll be consistent from belt to belt as far as tension and longevity is concerned.

[@Redo3D]

Thanks Anthony! Will check out gates for sure!

[@willl84]

My 3V2 came with Gates belts. Says it right on them

[@Redo3D]

@@willl84 you are correct. They are gates. I just checked every single one.

[@broderp]

Not seeing the issue. The stepper will move the belt tha same amount regardless of tension.

[@____5837]

The stepper will move the belt the same number of teeth per rotation (as the gear pulley has a set number of teeth), but as shown the actual distance between the teeth can vary slightly

[@2muchscott]

Solution would be to tighten taught but dont stretch the belt and then calibrate

[@michaelward9422]

Sounds like you need to change up to a lead screw some how to get that result. You could get it there abouts as you said but overtime your belt will slacken as it loses elasticity. Saying that, haven't tried it! Best of luck, look forward to your next vid, very considered content

[@patrickbodine1300]

All machines need servicing on a periodic basis. Failure to follow protocol will result in missed opportunities to excel in product quality.

[@TheLastChapter2023]

I've often wondered if you could drive the x and y axis with a lead screw + no belts.

[@alteredworks]

I've thought about this too. There's too many problems that need to be overcome and adds way too much cost. The motors will have to rotate at a significantly faster speed, the faster the speed the less torque they can provide and then you can start to loose steps. Lead screws also have to "float" and can't be installed rigidly or you'll get binding. Combine that with the backlash they have, it would introduce a lot of problems. I work with a lot of professional large format printers and plotter cutters, all are driven by steppers using timing belts on the tool heads. I'm sure it's for a good reason.

[@miguelcgamer]

It's not the nut that gets losse, it's the belt that gets stretched over tive. maybe investing in good quality bests will solve that

[@viru52000]

They're not likely to be using high quality belts (Gates), and you may be making them too tight

[@radicalreel]

You should only the esteps in the extruder... not for X, Y, Z.

[@Redo3D]

I would disagree here. Two printers with the same e steps on x and y might be far off enough from each other to make repeatable dimensions an issue. Calibration of the x and y is not for a single printers accuracy, but to match more than one printer to another in terms of dimensional output with the same file.

[@BerryTheBnnuy]

I'm not saying you're wrong, but you're using those calipers wrong in the video. To measure the edge of the teeth of the belts you were measuring you should have been using the other side so that you can get a reliable reading on the edge instead of eyeballing it like you were having to do.

[@moo00se]

I think your belts are backwards. The teeth should be on the inside.

[@Redo3D]

I stated in the video I flipped the belt on purpose for demonstration. Thanks for watching.

[@moo00se]

@@Redo3D fair enough. I didn’t have my headphone in the whole time. I kinda figured lol

[@nic0fpvnicolas606]

I think that you over tightened the belts to the point of permanent deformation. You should watch Eddie the engineer's videos about belt tensioning.

[@Redo3D]

I will check it out!

[@jimpifarre4601]

ordinary rubber belts stretch with weight on build plate, replace with steel wired belts..!

[@michaeld954]

Convert to lead screws

[@spoke4915]

Gates Belts. ... replace them all.. these belts are 2mm so all of the measurements should be a even numbers

[@PickleballThornton]

I don't know if anyone said this, but you belt is installed incorrectly the teeth should be facing down not up.

[@osiman84]

Kayış ters 😀😀😀

[@TheShorterboy]

this is the reason people use ball crews..

[@shaunmorrissey7313]

You need to ditch the crap standard belts for good quality items, best upgrade you can make to cheap Chinese printers

[@Redo3D]

Just checked my belts and they are all Gates. My understanding is thats a good quality brand.

[@MrJr298]

Your belt looks upside down.

[@Redo3D]

I flipped them on purpose