Excelent work!

Thank you for getting right into the subject and not having a 30 second extravagant title intro followed by 2 minutes of the history of how you came to discuss the subject.

It's unfortunate the manufacturer (Creality in my case) doesn't provide a reliable tension spec for something like this. It always frustrated me that the tension recommended was something like "tight but not too tight"... How am I supposed to measure that LOL

Excellent findings! 👏 25Nm is lower than what I was expecting, but I also noticed there's no benefit of over-tightening them. Great idea with the Luggage scale!

Thank you you are one of the few that is looking at this logically instead of saying I have xx years doing this so I'm right. One thing I can add is what works for me may not for you. Except in this instance. Great Job.

My life as a mechanic has taught me that timing belts usually should be tight enough to not have slack, but not so tight that they ring when plucked. On my Ender, I just propped my thumb against the tentioner to hold it taut and clamped the bolts down.

I completely agree , I wondered about that some months ago and did some research that suggested between 6 to 8 pounds of tension on 6mm gt2 belts to avoid premature wear of the nema 17 stepper motor and your findings confirm this .

The object here is accurate movement of the stage. The way to measure that is with a dial indicator on the stage itself. Prevent the motor pully from rotating. Push the stage away from you and take a reading. Pull the stage to you and take a reading. A force of approximately 1 kg or 2 lb is adequate. Starting with a loose belt, as you tighten, the total indicated reading will approach zero. There is no reason to tighten beyond that point. As you have found, more tension results in more friction. As stepper motors are open loop devices, increased friction is to be avoided.

Thanks for validating basically everything I've told everyone on reddit since I started printing.

Only getting into 3d printing recently but having watched a many videos not one touched on this except one that used a tuning app. Very well documented presented. Thanks mate!

This all makes so much sense and is more generally specific than other methods. I've never liked belts being too tight

That's a lot of practical engineering ! Thankyou for sharing your experiments and enthusiasm

Thank you for confirming what I already thought. Not to sloppy that there is play without movement but just tight enough without putting real force on them. That worked for me for as long as I have this Ender 3V2. Excellent investigation and thanks again for sharing this info.

This is awesome, thanks I have a Creality printer and suffered from lack of information. My prints currently are very good but I may have over tightened the belts, now I can verify. Thanks again.

Very nice! I used your page and my phone to tune my belt. I then used that belt to calibrate a printable "meter" that I can use across all 4 of my printers.

Spot on! Finally an objective measurable reproducible repeatable method for tensioning belts! Excellent!!!

This is great info! Would have never thought that the friction would be transferred into the stepper, though it totally makes sense. Also the increase in noise from tighter belts actaully has a lot to do with the vibrations resonating through the aluminum extrusion. Stepper dampeners, corner braces from the bottom to the gantry arms, and high density foam under the whole printer with the feet removed almost completely kills frame resonance even with tighter belts. I'll definitely be checking my tensioners still!

Great video, and great information. I'm on my second printer, started with the prusa MK3S at work and now I have an Ender 3 that I'm modding up. When I put stuff together I really just go by feel. But you've definitely got me thinking about actually measuring the tension and setting it to spec. It would definitely seem to be a great way to remove one unknown while I'm still trying to diagnose and learn how to get good prints out of this.

I came across this at the exact right time, since I'm in the process of one-year maintenance on my printer and belt tension was a bit of a concern. It is no longer a concern.

Great video as usual!

Nice vid good sir I have just recently gotten the ender 2 pro my first printer with belt tensioners up till this point I have just made sure they are snug and not a lot of slack but it is very easy I believe to over tighten with the tensioners hence my research on this subject . So far best vid I have seen so far ty

I just got tensioners for my printer, this video was very helpful. Thanks!

I needed this video! Thanks

So info. Much learning. This is now my second favorite LiT video!

When I was working in design engineering and using timing belts, the belt manufacturer specified the force at exactly center of the belt. I never saw any issue after the initial break-in. Life expectancy was specified and was a normal part of maintenance. I eventually went to Kerk Motion Products with their zero backlash multi thread lead screw. I then left steppers behind and went to servo systems. Inertia had to be calculated more precisely for smooth acceleration and deceleration. At that time, I also used Ametek linear rail bearings that had an encoder / reader built into it. That gave us the precision with repeatability required.

Wondered about under-tightening contributing to belt wear. Thanks to you, I now have a deterministic way to set the tension.

Just got a Sovol SV06 and had zero trust in my ability to judge how far to turn the belt tensioning knobs, so this was very useful. Thanks!

Thanks for this very informative video, now i understand why my steppers run HOT on my ender within a few minutes of printing. I am going to redo the belt tensions immediately and check the temperature again. I am confident that it will be lower

Great video. I have changed my extruder on the Ender 3 V2 to a larger (more heavy) one and started getting belt skips mostly on x-axis. I guess I will approach it like you do by measuring the tension with a small scale of some sort. Many thanks!

Great stuff. There is so much misinformation out there, even from printer manufacturers. I started down this rabbit hole myself a couple months ago and came to essentially the same conclusion you do. I don't have a tool to measure, but set belts to be just tight enough to "ring" at a low note.

Exactly right! I experienced this also on my direct drive MicroSwiss hot-end. I was getting some "skipping" (due to too low a temperature and too close (0.12mm) to the previous layer), so I turned the thumbscrew all the way in. Over a relatively short period of time, the bearing failed on my hot-end filament drive motor. Yea, I care. Good advice/video. Thx

Great video! Earned my subscription.

I suspect a lot of the print artifacts that people assume are due to belt stretch under dynamic conditions are actually due to inertia of the moving parts and flexibility in frame members or guide rails. There's also going to be a component from the motors themselves- the step and microstep positions of the rotor/shaft are defined by magnetic fields which are quite "springy".

Currently trying to help a friend calibrate/tighten their belts, this video defenitely helped but I have the feeling something like belt tension is still with a pint of experience I'm glad we personally have a 3D printer with auto belt tension

Nice ! Thanks !

This is completely valid, especially for Cartesian printers. Things do get more complicated with CoreXY, but the same principles apply. You want reliable and repeatable motion for your given system without introducing unnecessary friction (which leads to premature wear)

I have often wondered why people insist on tightening belts so far. Bearings don’t particularly care for perpendicular force. and quality belts are not very elastic. And I assume if you tighten them enough to play a guitar solo you now introduce the harmonics of the motor and bearings into the movement of the machine.

I tensioned my belts on the Prusa MK2s ones when I assembled it. I am a mechanic and engineer, so I tighted them pretty loose but enough. Now 5 years later the belts are still fine 👍 So I have not thought about belt tensioning :) The prints are also still very nice, so it is not to loose.

When I first got my ender 3 (first printer ever) all my circles were coming out as ovals... this really ruined all the screw holes and stuff. I found out later that it's cos my X belt was too loose. I have zero arm strength and that was as tight as I could get it apparently. So when the screwing system came in I immediately tried it out and I feel like I might be over tightening now, because of how bad it used to be for me. I'd suggest testing belt tension on cylinders rather than cubes, you can *really* see the difference there.

As a mechanical engineer for 20+ years designing precision XYZ tables for the semiconductor industry and new to 3D printing…. Thanks for the trigger warning.

Very interesting...thanks

I've been adjusting belts on all sorts of machining units since 1987 and you just get a feel for how tight they should be. Most people probably have them too tight.

Belts are made to a specific tooth spacing at a nominal tension. Overtightening a belt actually spreads the teeth out further than nominal. The simple consequence of this is that your accuracy is thrown off, Meaning you travel further than you expect to. The bigger consequence is that your backlash increases. Belt backlash is hard to see because it’s hidden by friction. You can’t feel slop unless your belts are really loose. But friction doesn’t hold the position the teeth do. And when your belts are overtensioned the teeth will no longer mesh as well with the pulleys because they becomes stretched out. So when you reverse a load direction and turn a few revolutions the teeth all slide over to the opposite side of the pulley grooves and that backlash shows up.

fascinating. my belt tensioning protocol is as follows: loosen up the tension point, then manually by hand run the axis back and forth while increasing pulling on the tensioner (to make it more taught), i basically tighten the system until i get smooth movement with as little wiggle between the pulley teeth and belt as possible. its also a good idea to find a way to pin the pulley in place, and then wiggle the system (bed or print head) back and forth to see how much play is in your belts/pulleys. if tensioned probably it should be minimal (talking like sub 1mm amounts) or none at all.

Thank you for the quality analysis and observations. It seems I may have the belts set too tightly on my Ender 5+. As a side note - My life partner and I got a LOL from your pause after you said "to a very specific...... tension" around 2:13 in the video. It reminded both of us of the dramatic pause in Frank-N-Furters monologue regarding "relieve my tensions" - dialogue around "Dynamic Tension" song. :)

probably 80% of people invested in this topic said wow out loud at 5:40. thanks for confirming my suspicions about this whole topic

Actual engineering. Excellent. Read the data sheet for the parts and use them within spec. Test and verify good results. Perfect.

Thanks for the video. I just engage the stepper motor and tighten from trying to move the bed and hotend. Once there is no slop from slack I go a touch more. The results look near identical to what you show as manufacturer recommend. I’ve always tried to get all stepper movement as easy as possible to avoid binding but with absolutely no slop. Btw, just put a multi meter in line with stepper power to measure the current through the meter. Set the meter to capture min max. Bit more accurate measuring what’s causing the heat. Heat can be effected too easily by outside influences.

Quality contents! Sub for the record .. in my first experience as a noob i printed (vyper) with the bed packaging band for a week without a problem

now we just need a printer that can both tension and level itself

Steppers are just motors for what ever project you want to put them in. I think the application a stepper is used in can add additional needs over what stepper specs consider. Just like race cars people push 3D printers as far/fast as they can. Car manufactuers don't recommend any of what the racers do to their vehicles/engines but they get higher performance and strike some middle ground on parts longevity. Here is my process on my IDEX bed slinger: I print at the highest speed and accelerations I would use and look at ringing, adjusting the belts a little at a time until I see no improvement in X/X/Y. Then I run the resonance tuning in Klipper. Finally I run the other tuning steps recommended by klipper. It can be a bit tedious but you won't need to do it often. I think your idea of going by stepper specs as a starting point is sound advice. I like your idea of using a luggage scale and will adopt it into my process for initial belt tension, would save me a bit of tuning time with a more accurate and repeatable starting point. Good video !

Your absolutely right. There is no need to go crazy with belt tension! For example, on my ender 3 pro, I had replaced the bed stepper motor mount with a 3d printed petg one a while back, and then I put the printer in an enclosure. After a while, that petg motor mount began to deform and the belt has loosed up over time. It's definitely not floppy loose, but not in tension like it was either. It's made NO difference to print quality. Eventually I'll get around to putting the factory metal stepper mount back in place, but that requires me to remove the printer from the enclosure so I haven't done it yet.

I don’t know why I never considered looking up the manufactures recommendation’s for these things. I recently replaced my idler pulleys with the turn knob style tensioners, I realized that I had no idea how tight to make the belts. Thank you for doing the research and sharing your findings.

Thankyou, the practical visual helps alot! On the Ender 3v2 (unless you have upgraded to a fully metal tensioner) too tight is easily possible... My belts were only slightly tighter than your ideal ones and on the y-axis it somehow pulled the tensioner knob through the bracket/sheath, shattering it and sending shrapnel onto the belt, sending the printer into a loud clanging panic.

Good, sensible advice, especially when one considers that most machines don't use any name-brand belts with any kind of spec for optimal tension. Motors frequently don't have specs for radial load but you can safely use the spec from a similar size motor that does have such a spec. One additional consideration when looking at machines built using 20 mm t-slot or long runs of even larger t-slot is that the belt tension can cause the frame member to flex. In my large sand table built using 45 mm square t-slot, ordinary belt tension on 6 mm belts caused the 2m long frame members to bow. Additional consideration for corexy: the minimal pulley configuration requires 5 pulleys in each belt loop. The common practice of adding additional pulleys to allow positioning the motors in a more convenient location, greatly increases the load on the motors. Don't believe it? try sliding the X axis along the Y rails manually. Now try sliding the extruder carriage along the X axis. The difference in force required is due to the additional pulleys that the belt moves on when moving in X. Now imagine what adding more pulleys is going to do. Finally, there are a couple myths about corexy belts. 1) they have to be exactly the same length. 2) they have to be at exactly the same tension. Both are complete nonsense. The goal of tensioning the belts is to end up with the X and Y axes square. What good is a printer that has perfectly matched belt tensions if the X and Y axes aren't square? The tensions are secondary to that goal, and belt length has nothing to do with tension. When you install the first belt and apply some tension, it will pull the X and Y axes out of square. When you install and tension the second belt, you do so until the axes are square again. As you tension the second belt, the first belt is going to get tighter, so when you tension the first belt, you want to leave it a little looser than you want it to end up.

This is a topic that needed to be addressed and you came to the rescue. Great job for sure. Or instead of having everyone go out and buy luggage scales maybe figure out a common item that most people own around the house that they can tie onto the belt tightener to give give the same results?

Here's a test for you to try. Use your luggage scale to pull the bed at various belt tensions. Many people worry about print quality based on a slightly loose belt. However the force required to move the bed goes up substantially if you over tighten the belt. This can cause a different sort of delay. And the concern can be amplified at higher acceleration and jerk speeds. It can also vary by axis, so the orientation of the part on the bed will dictate where and how the printing artifacts occur.

Thank you SO much! Yes, I obsess over this, for all the reasons you state. Logically, the tighter the better... until it's too tight and you're screwing up your bearings or overheating your steppers or burning out the drivers or God knows what else.

Great video! I was following a blog post about belt tension, the one mentioning Mersenne's equation 22 to calculate the required belt frequency when strung (3kg of tension target). For the Ender 3 those values were calculated to be 86 Hz on the X, and 112 Hz for the Y (both strung when homed). When measuring the frequency from the video when you strung them at 25N, I can tell you it was much much lower than what was suggested.

I deal with belts on manufacturing equipment every day. Each application has its own tension spec some loose some tight. I think you nailed it on the head. Over tension can cause pre mature ware on machine parts.

You can use Klipper firmware and a ADXL sensor and measure the ringing with it. To stiff isn’t good to lose either but Klipper can compensate the ringing in both instances.

Nice video. I have found the best way for me to tension belts is to put the palm of my hand as slight as I can on the carriage or bed and move it backwards and forwards. Feeling for vibrations through the carriage. (You can feel the teeth vibrating in the carriage if too tight) I tension or loosen the belts by getting to when it feels it has the least vibrations running through the carriage. I find this gives me the best results and when plucked it does sound like a loose note, possibly on par with the note you got when tensioned in spec. I also make sure my motors are aligned properly and also adjust the toothed cog on the motor until the belts stays central as I can get it when the carriage is moved from side to side. Hope this helps people if they have no way of actually measuring belt tension.

On both bed slinger and coreXY printers, I have the best results by pulling the steppers or pulleys with one finger. As long as the belt isn't floppy, looser seems to be better. Putting additional strain on the steppers by over tightening makes it more likely to miss steps

Being that my CR-6 SE and CR-6 Max have the knob tighteners on the X and Y, and I could find nothing on how to properly tension them when I first get them, I spent a lot of time worried about this particular issue. In the end what I did was to tighten the belts slowly with the steppers engaged and the carriage halfway down the rail. I would tighten slightly and then attempt to wiggle the carriage back and forth by hand. If there was any movement in the belt or carriage, I would repeat until there was no movement. This provided no slack in the system with minimal force on the belts. The only true challenge was the CR-6 Max which has a through-shaft stepper with a gear on each side and two belts. I would loosen one belt, adjust the second for no slack, mark it and loosen that belt and repeat with the first, then tighten them both to where I marked them. They ended up with pretty much even tension between them.

tight as hell!

I will say you are very correct. I am new to the hobby and started tightening the belts until the tension adjusting knobs starts to get tight. Problems started after that. Just take out the slack with a little bit of a pluck tone just like you demonstrated. That's how my printer came but out of ignorance I proceeded to tighten the belts without having any issues in the first place.

I used to fix industrial scanners (banks, etcs). They actually had a third wheel that was hooked to a spring so they would always have to right amount of tension (usually 500g).

Helped someone recently in discord who over tightened the y axis belt to the extreme then started a 2 day print. About 50% in the motor shaft sheared in half which left the y axis gear spinning in place. Cleanup wasn’t bad as he caught it fairly quick after failure but couldn’t figure out why his y axis wasn’t moving at all anymore. After helping them figure out the issue also figured out their “brand new” CR6 printer was a refurbished unit with replacement motors and belts. the mainboard was overheating due to a dead psu fan too. With the right circumstances apparently snapping a motor shaft is possible.

I tension to a few mm deflection under light pressure, probably similar to your 25N example. That being said, the additional noise from your stepper will either be from lack of damping (lower tension higher hysteresis) or because the load is starting to cause bearing issues and core misalignment in the stepper, resulting in weirdness.

I built my own printer with that problem in mind. All motors with a belt is supported with a bearing. X and Y. Also all motor shafts are round. I file out a flat on the shaft where the grub screw goes. There for no nasty flat. Which I think a lot of the issues stem from

Mine started skipping so I added one of those spring tighteners and all is well!

Tightening through sound is a good repeatable way but there needs to be two things said before. 1. You need the belt length you are recording to be the same every time you gather a reading. That means putting the bed at max or min travel or a specific mm in distance. 2. You don’t use notes. Instead you use hrz in frequency. To get more accurate. Timing belts are usually tensioned this way and there is even an official gates spectrometer to make sure the belts are in tune. That requires a microphone of some sort. Repeatability is the key here. I believe voron does a great explanation of this for their gantry’s. As far at the bed someone will need to calculate load at certain distance related to the beds distance. Then you are Will have exact load and repeatability. The thing that verifies this for me is looking at input shaper graphs at each hrz reading and how they change. Although your method does indeed work as it is using another measurement device that is repeatable. Great video but the other methods work when done correctly!

i think when youre looking at belt tension the question is how much do they stretch under load? fast accelerations will put a higher load on the belt for a short ammount of time, causing it to stretch, imparting movement innacuracy. if you put more tension on the belts you limit the stretch that occurs during acceleration giving you better accuracy. the 2 solutions i can think of is 1) print slower 2) use higher quality belts that stretch less the better the belt the lower tension you need. when you look at some of the newer coreXY machines you seey theyre using 9mm and sometimes 12mm width belts, i assume to make them stretch less during fast movement. ive put 10mm belts on my corexy machine (i know its a wierd size) and i run them with barely any tension at all. anyway im not an engineer, i dont know anything.

Thank you so much for sharing the knowledge. Was just adjusting the left side, eccentric nut for the Z-bar. Rubber wheel was very tight and couldn't spin because it was smashing against the left, vertical bar. Have you posted a video on adjusting both eccentric nuts on the Z-bar for the Ender 3-Pro? Cheers

I didn't want to ruin the printer... for a change. You're the best thing to come along since box wine.

Speaking from experience, overtightened belts can cause much more problems than what you mentioned. The increased heat is there for a reason. More tension, more torque is demanded to achieve the same acceleration. The motors have a max torque rating and when that is surpassed it can cause issues that can go from distorted dimensions to small imperfections due to micro skips or dramatic layer shifts on major skips at peak accelerations. The worse is that it's not usually evident what is causing them. And that applies to all motions systems

I love the disclaimer!! I'm going to tell you wrong stuff, I know that, just don't tell me I'm telling you wrong stuff.

Instead of using a luggage scale you could use a 2.5 kg weight, some string and a stack of books. Stack the books to the height of the belt tensioner. Attach one end of the string to the tensioner and the other to the weight and let the weight pull the tensioner instead of a spring scale. Now you have both hands free to tighten the tensioner. If you don't like the stack of books idea, use a spool holder or something else.

For me, I just tighten them enough to stay straight, mine aren’t guitar strings but they shouldn’t hang through. Sweetspot for me was a straight (to the naked eye) line from the anchor point to the pulley. As you said, too tight will increase noise

On my Sovol SV06+ I’ve been running them not lose but not tight, makes way less noise running fast

I know a person that added belt tensioner knobs to their Ender 3 Pro, and instantly snapped their belt. I had even warned them beforehand about how easy it was to over-tighten as the knobs offer no resistance, and they still snapped it. That's how easy it was.

I ran my belts at max spec for the longest (16lbs or 3.6Kg). I found this greatly reduced ringing on my printer and i was printing at 180mm/s and 1500k acceleration with no visable ringing before Klipper and an accelerometers were a thing 😊 Quickly found i needed input shaper much more at higher speeds and the belt being that tight was showing ringing far past the square corners ar higher speeds. I dont know if that was the added friction showing the stepper motor inperfections, but it went away completely when i lowered the tention down to 7lbs/3.2KG. I chose to leave it just over 6.6lbs/2.8Kg because i found they would feel loose after the chamber heated up. Dont know if that was because the rubber is softer or the metal has more give, but the belts definitely get looser in hotter environments. I spent a good amount of time squaring and tramming the bed. Got dementional accuracy down to >0.05. about as good as a unmaintained CNC mill 😅 granted specs for Gates GT2 belts are >0.3mm every 92 teeth on a 20 tooth pully from the motor. All that means is my X-Y are very in line with each other, and Z is dead on because of the lead screws (twin stppers in sync with a belt, that are in phase with each other). 150mm part, 100mm, and 50mm, all messured -0.01 to -0.03 randomly. I give that up to slicer and measuring inconsistencies. You'll find that when you start putting really thick layers and it's larger nozzles that the slicer has a lot to do with where the walls land. You can see it in an orca slicer when you preview the sliced model and run the slider up and down. Then change your compensation.

Im currently building a RC plane wing printer with a over long 380mm x 800mm bed, the twin belts will be long long, im yet to design the tensioners for the Y axis. Might be a knob screw threaded sliding clamp on the carriage rather than moving pulley mounts. Ive never had belt tension issues before, hopefully this long bed wont be a problem.

Thanks for the Vid... The inventor of the gt2 timing belt said 6lb (2.7kg) I believe... so pretty good with your motor spec... perhaps they designed them for the gt2 belts

miraculously, that plucking at 5:30 is exactly what i tightened my belts to without even knowing. It just felt right. I cringe at these printable belt tensioners.

NIIIIIIICCCCCCEEEEE!!!!!

A bit more tighter belts do result in a better print quality thought. At slower speeds it's not noticeable but as you print fast with klipper for example, over 3000 acceleration. even a little slack in the belts shows artifacts on the prints.

What you really need is the pull force meter on the actual belt tensioner. That way you see the readout while you are tensioning the belt. And if the belt stretches over time and the tension decreases you can see that. What you have shown is a force gage you have to pull with one hand, and probably have to hold the machine stable with another hand. So you need to have a second person to tighten the bolts while you struggle to apply a stable force! And even then how do you know the pulley doesn't move a little during the tightening process? For something that is as important as belt tension, it is all so frustratingly vague.

I find the V rollers tension is far more crucial in good quality prints. There's a lot of margin in the belt tension. If you have a loose wheel anywhere it will cause all kinds of banding and artefacts, and if one is over tight and has flat spots, it will do the same.

I've been wondering about this. Prusa has a printable tool to assist in belt tensioning. I wonder if it would be universal? I'm not sure how you'd go about using the scale to tension also, doesn't it require you to pull against the belt? Once you get it to 2kg, how do you then apply it to the belt with the scale attached?

Basically I only tell people to tighten belts slightly when they post image of layer-skips, since that is usually the belt or if they run it super fast

The way I do it is I loosen the belt so the bottom is clearly not under tension and then I just tighten it to the point where it doesn't sag anymore, then I turn it about a quarter to half a turn more, seems to be fine :)

I noticed on my old i3 clone (this style of printer out of acrylic plates or even wood) when you tension the belt of the gantry too much, you pinch the Z-axis rods together. gave me a few head aches until I figured this out 😅

I've watched this twice, and I now fully believe that the reason I've gone through 2 "Y" steppers is that the belt is too tight. Unfortunately, my 3d printer is horrible with adjustments, and I'm not sure how much tension I can let off. (It's a k8200 kit, so I may redesign something, if necessary)

My way is to print a small circle approx. 1cm dia, if the circle is oval then belt is loose. Tighten x-y axis belt accordingly. Print again and check if the circle is within +- 0.05mm tolerance x-y

Re. noise: a looser belt will damp noise from the non-continuous stepper motion, a tighter belt will transmit that torsional vibration into the frame of the printer.

Great walkthrough of the process 👍 Thanks for sharing your experience with all of us 👍😀

You could use Input shaper and measure how the resonances change with different tensions. See if tighter belts allow for more acceleration etc 🤷‍♂️

I just think it just needs to be firm to remove any possable play but not overly tight as that will only load up the bearings and increase belt wear, and if it's too lose you will see this in your print quality.

This makes me think that it would be interesting to add a load cell "weight" sensor to a printer's axis, and with something like HX711 you could measure the weight created by the belt during the actual ongoing printing on the printer itself. Much cheaper than an optical sensor like used in this video, and can be cheaply added for each of the axis on a budget 3D printer. Print a mount to hold the pulley on the load cell instead of having the pulley sit on an immutable position.