Upgrading to Linear Rails a Waste of Money? - An Evidence Based Answer

[@SpencersDesk]

Hey! Thank you for taking the time to watch this video. I just wanted to quickly point out a mistake I made in the video! When I show the grease I use, I present a grease that I claim does not have PTFE in it. Funny enough, as I find the link for it to put in the description, I discover that the Syncolon it has in it, is indeed PTFE. So, you may not want to use this specific grease... Sorry!

[@VikingRul3s]

What a quality test, thank you for this! Suggestions: Check the motors temperature, the added weight of the rails should, in terms of newtons law, require more energi (W) to run as fast as V-slot (edit: on many boards, even stock, you can up the driver voltage). Mine are all custom, so i can't give you any pointers other than; warmer then body temp = watch it, to warm and a little fan is needed. For Y with with glass bed and 2 linear, my stock Ender 3 motor needed cooling to handle properly

[@digitaltoaster]

I specifically buy that specific brand of super lube because it has PTFE in it!

[@chrisgill5692]

@@SpencersDesk the balls inside the carriages of miniature guide ways MGW/MGN are designed to circulate around little tracks. To allow this circulation, they need to be oiled, rather than greased. Definitely a good idea flushing and cleaning them, then applying the correct grade oil upon receipt. They require occasional re-oiling - there are tiny holes in the plastic end caps/wiper seal plates to allow this.

[@nicoleibundgut534]

@@chrisgill5692 Yeah my friend told me the same. You could use grease in the Z axis but in the y and x axis because of the speed an somewhat thicker oil would be good.

[@xXKisskerXx]

@3:23 its not normally 'grease' applied at factory, but a light machine oil - and yes, it is to prevent corrosion because your rails might sit in a damp, hot warehouse for months or years before you get them. It 100% should be removed and actual grease be applied. infill is overrated. more walls will get you a stronger part, almost every single time. This has been tested, pretty well by now. a good 7 or 9 wall thick part, while bits of it will be solid - the main 'gap' area, will be infill only, which doesn't need to be high %. (3d Honecomb and Gyroid being some of the best, strongest) So I would be curious if amount of walls, change your frequencies or not. I've personally come to the conclusion that our aluminum extrusions should be filled with like.. sand and capped off. This provides weight, stiffness and some dampening, in the frame itself. and I think your mostly right though - the rails vs wheels is.. kinda dumb for performance, it is more related to keeping the printer running smooth. Adding or replacing grease every few months, is so much less effort than cleaning plastic bits outta the rails, and preventing it from going onto prints while they print - not to mention eventual replacement of wheels, even PC wheels, as all will wear down as they ride along that 'sharp' aluminum angle they ride on. They are designed to be the expendable part - because when plastic and metal rub together, the plastic wears, not the metal. ball bearings on rails - work differently, which is why they don't wear out as fast and are used in most industrial machines.

[@MarkusOsmers]

secondthat transport oil is not ment for operation on linear rails. Clean and grease properly.

[@Shinobubu]

I heard this oil is Cosmoline. It's just there for rust protection.

[@DuyLeNguyen]

@@Shinobubu They're most certainly not cosmoline. Cosmoline is a type of petroleum wax, which is usually quite thick when fresh and then solidify over time into a hard wax. The stuff that typically coats metal parts to prevent corrosion during storage is just simple light machine oil or mineral oil.

[@SpencersDesk]

Hey, thanks so much for this comment. I’d read it multiple times and found lots of valuable info from it. I definitely agree with you on the convenience of the rails after the initial setup. I’m getting back into 3D printing after being on and off for a while, so I’ve forgotten about the walls vs infill stuff. Thanks for the reminder! It will definitely come in handy in the future.

[@Tezza120]

I think there's more friction with linear guides as the balls are preloaded against the races - why it's so rigid. This friction might be what limits speed in the steppers as they have to overcome more resistance. As for the lubrication - that is a whole degree. Just look at a SKF bearing selection process. I think for 3Dprinting you'd want to use grease as it's simpler but what I would do is clean the rails down first getting rid of the old grease. Then pump grease into the guide block until you see only clean grease come out - wipe away excess. Now this gets rid of the dirty grease but probably overfills the rolling elements but as the loads are small I don't think you'd overheat them. Any excess will just work it's way out - hopefully. Thankyou for spending time doing these experiments.

[@SpencersDesk]

That's a great point about the preload causing more friction. I guess there is a bit of a tradeoff between the rigidity and the friction. I think in my case, it's an imperfect setup, coming up a bit short on the cleaning, and using the wrong lubricant. I definitely plan on making a video about proper setup/maintenance of the rails. Right now I'm leaning towards giving the rails/carriages/bearing a soak in IPA for a little bit then properly lubing them as you described! Thank you so much for your support.

[@Tezza120]

@@SpencersDesk If you do, make sure they are submersed. The steel rails will flash rust when exposed to air if there's no lubricant to protect them. What I do to clean bearings is flush with contact cleaner and moved them to get grit out. Then dry with compressed air and then lubricate imediately - might work with linear guide block too but not sure the plastic seals will like the contact cleaner. Maybe flush with mineral oil instead?

[@J.H.hhhhhhhhh]

I don't think it has that much to do with preloading. I think the maximum speed the rails move is highly dependent on the grease you are using. My guess is, the balls start sliding at a certain speed and stop rolling. It would be interesting to test different greases and dry behavior. And better still, use a closed loop system or measure the amps at the motor.

[@VikingRul3s]

I'm sry but i'm going to have to specify this a bit. Rails are, exactly like bearings, sold with a certain tolerance (in case negative = pre-load). Unless it specifically said pre-loaded or a ISO standart letter+number was at the end of their name, they aren't pre-loaded or even outside ISO "stock" spec. They do however very unlikely have all the balls with that stock spec (uneven grind, slight oval and egg forms, easy to see under microscope). In my experience, you can get unlucky and need to buy a bag of balls fairly fast (low material quality), tho this means you can buy good ones and kind of upgrade the setup that way SKFs guide section are for industrial use and the quality of their goods can't be compared to cheap small rails. Among other things their seals are significantly better, allowing lower viscosity, which suits their superior tolerances. Super Lube is an absolutely outstanding choice for these cheap rails. Grease (like EP2) can be used, if you have nothing els, go for it! But it certainly isn't the proper choice, as it's designed for low speed, with a high friction and certain weather resistance, neither of which are pros in the case of this application

[@JunoStella-cj8kg]

love the vid, nice to see people take on these 3d printing questions with data

[@SpencersDesk]

Thank you! Many more ideas in the pipeline (:

[@TheOfficialOriginalChad]

Another great video! I agree with your sentiment on linear rails for quality and speed, with a huge BUT: As you pointed out, V slot wheels can and do cause problems when theyre not adjusted perfectly; this inevitably leads to a problem that you have to diagnose. I think it's extremely important to reduce the number of things that can go wrong, especially if you have a cheap printer.

[@SpencersDesk]

That is such a good point. I think the biggest upgrades in 3D printing are ones that make your head hurt less. I wish I would’ve talked about it in this video! Thanks for the support!!

[@croustibat682]

True, but there is a limit to what people should do with a cheap printer if they are budget conscious. I didnt much and bought a lot of cheap parts to "fix" and improve my ender 3, and while I learned a lot, i ended up spending bambulab money. And the printer does not have bambulab performance nor precision. Much better than stock ender 3, and the journey to get there was interesting, but still ... I could have just bought a better printer and call it a day. anyway, jump in the rabbit hole with that in mind :)

[@--3D]

I've been waiting for a video like this for ages, thanks for your time and dedication 🥳🙌🏼

[@SpencersDesk]

Glad to have delivered =)

[@iseolake]

Nicely done! I studied mechanical engineering and really appreciate they work going into creating a methodical approach and multiple test runs. Regarding the linear rails, I'm glad you are going to look into it further, but it is really valuable to find out that a particular upgrade does not necessarily deliver better results.

[@OrionAerospace]

Awesome video again! Loving the in depth content- as someone who spends a lot of time upgrading printers it’s useful to have it in well made videos! Keep it going

[@SpencersDesk]

Sick channel man, just clicked on it. I haven’t watched a video yet and I’m already jealous of your rockets. May be binging your videos tonight!

[@OrionAerospace]

@@SpencersDesk Ayy appreciate it! Been loving your content! We have some new content coming soon so stay tuned! Lmk if you ever want to do a printing collab too :D

[@MichaelODonnell]

Thank you for all the effort you put in to this! I think I posted another long comment on one of your other videos a few months back as well. Was planning on making a video last year with some before and after results but got busy and never did anything with the clips. I did some ghosting / ringing torture test prints with accelerations ranging from 5000 to 15000 mm/s2 or something, on a highly modified Ender 3 V2. Similar to you, studied input shaper plots and calibrated everything before each test. The overall finding was that after a certain point (maybe it was around 6-7000 mm/s2?) the linear rails were doing a lot better. You could visually see less ringing in the test print at those levels, especially the Y axis. My theory is that changing direction simply took less energy. An interesting observation is that input shaper actually recommended LOWER maximum accelerations with the rails compared to V-slot wheels, which left me scratching my head. System stiffness should be higher, no? Used machined metal mounts. So I did another test after adding diagonal Z braces and relocating my filament off to the side, instead of mounting it at the top of the printer. Input shaper graphs looked even better and I got rid of a secondary peak, which was obviously the decoupled mass of the spool. The whole study was plagued with non-conclusive results but my results with the torture tests ultimately confirmed more consistent performance across a broader spectrum of accelerations. But one could argue that this is copium. Another interesting case was my giant 400 x 400mm converted Ender 3. Hilariously large bed, really limited in accelerations due to this, even with a 48mm NEMA stepper motor with higher VRef. The bed is still mounted to the extrusion on the center of the Y-axis, which isn't ideal at all for a large bed like this. In this case, I got way more consistent prints simply due to the higher stiffness. Adjusting the wheels is really annoying with a bed like that, and they were wearing out left and right. As dimensions increase, the higher precision and stiffness starts to make more and more sense. Used ASA printed mounts for everything. Last little tidbit... for axes where the rails would "over define" or double-constrain the system (for example, parallel rails in Y and Z), I added small rubber O-rings on the screws, between the mounts and the rail carriage to introduce some very slight angular tolerance. If they aren't perfectly aligned they will bind, and most rails aren't even perfectly straight when manufactured. So I would have 50% of the axis super smooth, and then there would be resistance toward the ends. These small O-rings completely solved that without giving up too much stiffness.

[@SpencersDesk]

I’m not sure how I’ve missed this comment for so long but wow this is awesome! I really appreciate the discourse. First, I’d like to point out the comment you made about the accelerations going down while the physical results were better. I work in a field where our background is physics (having equations for everything and expecting exact answers and consistent results) but we work in medicine (our models are often incorrect). My point being, there comes a point where relying on the numbers is a poor choice. While the resonance tests are great, what matters are the physical results. We should, at the end of the day, default to the physical outcomes rather than our arbitrary numbers. While I love the resonance graphs, I’m not sure we can rely on their recommendations. Just use them as diagnostic tools! I love the comment you make about leaving a tolerance. People are so bent out of shape about perfectly setting them up when the rails themselves are imperfect. I’m planning on making a proper linear rail setup/care video where I investigate 2 main things. 1, does the type of grease really matter, as there are a lot of interesting opinions out there… 2, does the setup matter that much? I.e. is investing in a dial indicator going to give you that much of a performance boost? I’m not sure, we will see. Thank you again for your awesome comments and continued support! Would love to see you in the channels discord!

[@jackblack7132]

Having experience in metalworking, I can say that rail guides should work in pair. One rail will always give friction. Even with preloaded cariage.

[@VikingRul3s]

Was thinking the same thing. Proper setup would include a high precision "arm measuring" tool to align them (no idea what that tool is called in english), but hen again, a 3D printer isn't a CNC mill designed to handle 100s of kg

[@digitaltoaster]

yeah, usually installed in pairs, and the grey bearded (decades of experience) toolmakers would spend hours adjusting them to be parallel to within a few thousandths of an inch along their entire runs. (I worked programming premium quality custom industrial machines that made luxury goods and medical products, we used very expensive misumi linear rails.) oh.. i remember how they measured the parallelism, you mount a dial indicator on one of the rails, and move it back and forth, adjusting the other rail so that you get the minimum runout along the length of the rail!

[@_Xantras_]

Huh, so every X beam on coreXY machines should have 2 rails ? Is triple Z still ok ?

[@jackblack7132]

@@_Xantras_ If you need reliability and long lasting with rail guides, they better to work in pair. It’s reduces radial friction of every single carriage. But for fdm it would be a very excessive precision and there are much lightweight and more reliable solutions. Btw a rail carriage with preload, (which used in A1 for example), although it reduces trembling and friction, also leads to more frequent replacement of both the rail and the carriage, cause preload = excessive wear. It’s a frequent problem with cheap chinese laser machines.

[@jackblack7132]

@@digitaltoaster I'm not gray bearded toolmaker yet, but had an experience with adjusting rail guides through 3 meters along. I agree with you that it's a real pain in... you know))

[@Scrogan]

What about linear guide rods? Some bedslingers come stock with them on some axes.

[@SpencersDesk]

I think they’re better than wheels, but they’re not quite as good as linear rails. They’re also not as easy to add to a printer that wasn’t designed for them whereas the linear rails are somewhat straight forward.

[@chrisgill5692]

Great job..... quality of linear rails measured by their cost (tolerances, grading, carriage pre-loading etc) and they also need hours of set-up using a dial test indicator. Not worth an upgrade IMHO. Tweaking the V-slot wheels can get tiresome, but being polymer they self-align and dampen to some extent - well adjusted they perform perfectly well for FDM. Your analysis of the mounts was super interesting!!! Really appreciate the time and effort you spent on this - subscribed.

[@SpencersDesk]

I’m definitely planning on doing an experiment to see whether the setup needs to be that precise or not! I thought it would be cool to use a dial indicator and compare it to a janky eye balling it. Thank you for your comment!

[@FrancisSt-Laurent]

Hey in general Linear rail (if you use a dial gage to set a low enough runout) will allow your machine to be rigid enough for higher acceleration. At low speed and acceleration you won't see significant quality upgrade. Also when you have in your graph a second spike it is usually indicative of a loose part. I would verify the alignment of your rails first. (X axis tend to have weird second high frequency spike due to the fans and sometime due to heavy cables)

[@ifell3]

I've just printed a new fan duct, stripped the printer then realised I need this and that

[@thomaswiley666]

Some notes, if I may: 1) Your premise is restricted to bedslingers. But I believe your question about liner rails would apply to CoreXY machines, too. As to the former, there are so many variables associated with harmonics on a bedslinger. One question I have, does it scale with the size of the bed? 2) Motherboard/hardware: I don't know which Anycubic printer you have (it looks heavily modified at the print head). However, I think Anycubic has only two motherboards, so the stepper drivers are either TMC2209s or worse,A4988s. If your board has the latter, then stepper rails are not worth it. The '88s are barely able to push basic printers and when encountering splines or parabolic curves they are going to barf, linear rails or not. The TMC2209s incorporate better kinematics with onboard implementation of StealthChop (making the driver-motor coupling more quiet on straightaways and SpreadCycle for more dynamic curves. Linear rail question should only apply to these printers with at least TMC2209s or equivalents. 3) Software: This assumes from your coding that you are using Marlin to translate GCODE to stepper driver signal. Yes, Marlin has resonance frequency adaption available in the code but it is not as efficient as Klipper in terms of firmware abstraction. Additionally the '88s are too basic for Klipper so any advantages of Klipper are automatically dismissed. In all, depending on your model of printer, linear rails may just end up being a cosmetic addition instead of an improvement. EDIT: Is this printer you modified in a previous video? If so, then my previous comments are applicable to anyone with a standard, unmodified printer. For yours, I guess my first question is how much of this is a 3d printer and how much is this a machine to tinker with? I ask because the time and effort, along with product costs, could easily put you into the realm of a simple CoreXY machine, a more stable printing platform.

[@SpencersDesk]

Looks like you answered most of your questions haha. Resonance is inversely proptional to mass. So, larger bed = heavier bed -> lower resonance frequency. Also, TMC2209’s don’t switch between spread cycle and stealth chop by default. I think they can switch during prints, but it is generally not recommended. As for the printer, lot’s of people say this. I enjoy the tinkering and modifying a lot. Printing functional parts to be used for that stuff. Not really printing random stuff. Also, I didn’t buy all this stuff at once, so it’s not like I could have bought a printer instead (ignore that saving money exists). Hope that answers, thank you for your comprehensive comment!

[@testboga5991]

Any linear bearing approach is good enough if it's having good tolerances. The main advantage of linear rails is that it's easier to get good tolerances with them.

[@croustibat682]

Also a major upgrade to solve printer stability at higher speed and acceleration is just ... firmly anchoring it into a slab of concrete. Your best 10$ spent (after a cht nozzle XD) check CNC kitchen vid for that. It is a bit old but still up to date. that was a very interesting video, please keep it up !

[@SpencersDesk]

I’ve seen that vid! I definitely intend on doing it one day. I’m planning a video about eliminating resonances around the printer and that would likely be part of it. What’s the cht nozzle? Thanks for your comments!

[@SpencersDesk]

I think they know about that (it's in the CNC kitchen video) but that's a great reminder! I actually don't know what kind of foam to go with when I eventually get around to doing this.

[@croustibat682]

@@SpencersDesk Nice ! I wanted to do it on mine, but the enclosure i built is from a couple of ikea tables called lack... and they are basically cardboard, so I am not sure it will take it. I already had to reinforce it toprevent twisting. Cht nozzles are the ones that split filament inside, usually in 3 parts. It increases flow like crazy without the need to heat more, use specific hotends, or a stronger motor for the extruder. You could make a video to compare cheap ones and the original for example :)

[@DavidMulligan]

Alex Kenis has a pair of videos about how to clean and improve low quality linear rails. Regarding speed and steps, your rails were probably binding a bit from not being quite aligned.

[@Queracus]

The results depend on the quality of the carriage on the rail itself. Diferences between manufacturers are huge. Pretentioned carriages have a positive impact as well, reducing any backlash/movement on rails. The amount of friction that motors have to overcome with rails is neglegable, if rails are larallel to eachother and dont snag. If you go for accuracy and repetability, rails definitly beat any wheels. Even if its on the cost of 5% speed drop. You can always up the motors :) Allways best to use grease made for thes use, even tho its a tad more expencive :)

[@TheElectronicDilettante]

Informative video. Very well done. My only input is the lubricant used. With all the particulate matter frothed up by the movement of print bed and extruder, that sticky grease is going to catch a lot of it a end up causing an increase in friction. Clean all that goo off, go to auto parts store and get dry silicone spray lubricant. Spray it evenly on the rails with the bits in motion. It’ll be slicker than snot. Don’t sweat the haters and those over critical of your content. They probably couldn’t perform for their daily session of self abuse and have to lash out. Thanks for the video! Oh, my vote for video topic would be for linear motor upgrade to your printer. That’s next level. Cheers

[@SpencersDesk]

Hey, thanks for your kind words! I’m currently working on a video all about the proper setup/maintenance of rails! For now I’m sticking with grease per a HIWIN recommendation, but if the plastic junk gets to be to annoying, I will have this recommendation in mind!! Maybe if the channel sees a little more success in the future I’ll spring for some linear motors😅

[@rokoksekerat88]

Mayb tune every part like guitar body. They tap the wood and measure its sound. N choose which wood to use. So every part will resonance at the same frequency. Every note they will resonance at a harmonic. So they are severel frequency for a same note will reacts one another. But we only able to choose one note. What note are better for the resonance in 3d printing that you introduce d to me just now. Thanks. Sorry if my idea bads ;p

[@MontegaB]

Medical physics? That's a pretty awesome field! I worked for the University of Florida for a time supporting their proton therapy facility. It was one of the first in the county and is an amazing machine.

[@SpencersDesk]

Hey! I agree, it’s a pretty cool one. That’s so cool! I heard that proton therapy requires lots of around the clock engineering and physics support. Such amazing technology though

[@Sadlife_t8]

neat video from a fellow v roller enjoyer, will you try this but on x/z axis? and maybe z1 vs z2 preloading if you have silly amounts of time

[@SpencersDesk]

Haha, I do not have silly amounts of time. I wish I did. I would get real silly. I will be moving to linear rails on my y and z axis, hopefully soonish. I don’t believe I will be investigating different preloads. I think that our printers would be fine with Z0 or Z1 on any axis. I don’t think we require anywhere near the rigidity that Z2 would provide. I’ve found through producing these videos that people have some seriously strong opinions on these things without much to back it up. I think we often forget that we are 3D printing, not operating industrial machines, which are 2 very different applications requiring vastly different considerations… Anyway, off my soapbox, haha. Thanks for your comment!

[@magnumoppai159]

the smaller the machine, the less linear rails will help. I would not bother with linear rail on most designs smaller than 250mm. on larger machines, belt stretch starts to matter and vslot bearings will start failing to hold things in place. resonances will be worse in the 400mm/s regime

[@TS_Mind_Swept]

9:16 Voice typing often confuses those two as well (and sometimes I just put it on purpose because funi) SuperVinlin I feel like the biggest argument for linear rails over V rollers is the reduced maintenance, and not having to replace the wheels every 5 minutes; esp if you just happened to tighten them a little too much and they decided to explode faster than usual.. not to mention having to adjust the tightness on them every time you change them, which may take a bit longer initially with linear rails, but it's not something you ever have to repeat

[@spyrit35]

The nerd in me liked this video, good job. I'm a big advocate for linea rails... the truth is that adjusting wheels is always a guess, its an excercize in experience that Im not interested to learn and they'll always need readjusting. I just dont want to include my carriage systemm in my troubleshooting process... the Y axis is particularly annoying for wheels... I would literally bybass a slightly superior printer to get one with extrusions that allow me to use rails. Lightning speed isnt interesting to me because I dont print PLA anyway and ASA and PETG just wont print at high speed... rails, set it and forget it for the most part (apart from occasional lubrication).

[@CarloVaccariPlus]

I just upgraded my Ender 3 to rails. No regrets, simply because I don't have to worry about flatspotting anymore. I only paid $60 for both x and y axis upgrades, though.

[@talbech]

Excellent content. Thanks for sharing.

[@SpencersDesk]

thanks so much!

[@RaoulEvilD]

Very good exploration video. I applaud the intent and the effort. 👍👍👍 One thing though, I find it very confusing to compare the resonance graph shape using different vertical scales. The final recap using the max power density thanksfully normalizes your findings, however there we have pretty much lost the info regarding the shape of each graphs. Overwhole, thanks very much for sharing your process and publishing of your results, good job there 👍👍👍.

[@SpencersDesk]

Thanks for the feedback! I’ll look into generating my own graphs in the future!

[@croustibat682]

steppers and drivers may be heating too much during the speed test. You may want to up the stepper voltage reference too once cooling is sorted if it still cant go faster. Also, losen the belts, more often than not they are way too tight and put too much strain or steppers

[@SpencersDesk]

That was another idea in my head, although I attribute most of it to my lack of experience. Belt tensions definitely weren’t the same for both tests which is something I need to figure out how to control better in the future!

[@croustibat682]

@@SpencersDesk i saw a video recently made by "lost in tech", about belt tension,while trying to find why my steppers were so hot i could not even touch them. He explains how they need to be barely tightened to avoid going over stepper specs, and how tightening them more only puts more strain on steppers. Dont worry about lacking experience, experimenting is how we learn :)

[@digitaltoaster]

this problem has been solved pretty well in the automotive field in a few ways, when we discuss damping resonance from a 3d printer motion, why don't we consider internal combustion engine mounts? my friend's car was incredibly uncomfortable to ride in because he'd replaced the soft mounts with hockey pucks for a harder rubber, this however gave it much tighter cornering. adjusting the density of the mount material is definitely a route to dampening. I'd expect for printer mounts we'd want a very hard rubber with just a little flex to it, preferably a material with decent insulating properties to prevent heat creep out of the bed too.

[@nathanblanchard8897]

We kind of do? But I don’t know why they aren’t used very often. There are stepper motor dampers out there, but I think there just isn’t as much resonance from the motor itself and more of the resonance visible in prints will be from the motion system (gantry, bed, etc). I used to build the MatterHackers Pulse (it’s okay that you haven’t heard of it, there’s a reason I USED to. It’s a Prusa clone with arguably better hardware imo), and every motor has a damper on them. It definitely made the printer quieter with the non-silent driver board, but it actually started to cause some problems with the Y axis on occasion. In combination with the PETG blend printed mount, the pulley would get pulled out of alignment with tight belt tension. It wasn’t a big deal, since that meant the belt tension was too high anyway, but removing the damper DID help reduce that twist. Similarly to squishy motor mounts, squishy bed mounts or gantry mounts could help reduce big spikes, but I’d imagine it would introduce more issues than it would solve in dampening resonance. Tackling resonance at the actual source with input shaping just seems like a more effective intervention. It’s like tuning and balancing an engine, rather than covering up the problem with more vibration isolation, I guess. I don’t know enough about resonance to actually speak on these things, but these are my observations. What I do know is resonance is very complicated, and that input shaping is magic.

[@SpencersDesk]

I would somewhat disagree with one point you made. I think tackling the source of the problem is indeed removing resonance through the build of our machines. Rubber mounts, more rigid connections, etc. First we eliminate the resonance the best we can. Then we use input shaper to mitigate it. IS can only compensate so much and has inverse effects. I’m planning a more in depth video on all this soon!

[@SpencersDesk]

I wouldn’t say solved. Resonance is a huge issue throughout the world, big industry. However, there is a lot we could learn from automotive to apply in printing! ( maybe that’s what you meant)

[@nathanblanchard8897]

@@SpencersDesk great point! I’d love to see more content on input shaping, linear advance, etc, specifically Marlin’s implementation since I still haven’t switched yet. It’s much less… easy? than Klipper’s sensor based approach and I still feel like I haven’t gotten a tune that works well on my machine. Do you think looking into the mechanisms behind the different code implementations might offer some insight? I feel like you’d be crazy (and actually capable) enough to do a deeper dive than I’ve seen in other videos on that aspect.

[@josefjelinek]

One thing to check that you are not introducing some accumulating errors into the consecutive tests using the same setup is: run the first test (petg?) as your last test as well. If nothing significant changed, the results would align. If they do not, something unexpected happened. Maybe linear rails got looser, some geometry moved, screws got loose, etc

[@SpencersDesk]

Very good tip, I’ll make sure to implement that in my upcoming videos!

[@NickBR57]

Given the cost of these rails and adding the mount parts costs, you may get better value by just replacing the printer with a BL A1... Nice though, despite some negative comments

[@SpencersDesk]

You’re not wrong at all. Bambú is about to have a sale where the A1 is $300 (less than I paid for my AnyCubic Vyper 3ish years ago) I hope to get an X1 in the future for when I want to print instead of experiment!

[@Karaon]

Just like I said in previous videos. This channel blows my mind if we speak of 3D printing brainiac

[@brunoyudi9555]

i feel like there is a bit of misticism surrounding how much better linear rails are lmao. if you got a well maintained printer you probably won't see much difference, but keeping it well shaped demands more work over time, meanwhile the linear rails gives you a more constant experience on the long term, the most maintenance it requires is grease and re-tightening the screws, which the machine vibrations makes it go loose over the months. despite not seeing much difference in print quality i feel like it was a worth upgrade to me, the upgrade that i feel like gave actual better print quality was an oldham coupler.

[@Amybnuy]

It depends on the use case and printer. Saying “linear rails aren’t worth it” is like saying “olive oil sucks compared to avocado oil” because it just didn’t work well in one dish

[@stefanguiton]

Excellent video!

[@spyrit35]

on a side note... the initial machine oil removal should look more like a half an hour soaking than wiping with a paper towel, especially for the carriage sled... I slide the sled while doing it too, any my logic says that my grease should actually aid in dampening because of it's higher viscosity, I use the one in the grey tube and I always slide it back and forth for a few minutes before full assembly... I also dont keep tools in my printer drawer, I just figure that the jingle and add resonances. (also diengage the motors if youre doing sliding of the sled to distribute the grease well to prevent electrical feedback to the motherboard)

[@SpencersDesk]

I definitely did this wrong… I plan on doing a more thorough video on it down the line. Thanks for your wisdom!

[@boxbawks]

Possibly try mounting the rails to be horizontal instead of vertical would make the speed better? Thats about the only other thing I could think to try out, but that might require printing out some L brackets to fit the sides for mounting. that would also allow for you to have more clearance from the switch and belt tensioner as well.

[@SpencersDesk]

I think my current plan is to get some aluminum extrusions to replace the y single y axis one, and the single one on the bottom of the printer connecting the left and right side together. I’ll then, like you said, mount them horizontally instead of vertically. I think this is a much better way for the rails to be mounted, but maybe it won’t make much of a difference? I think the best thing I can do is ensure they’re parallel, and grease and clean them well (which I didn’t in this video)

[@C-M-E]

Saying this based on principle and purpose of linear rails without watching all of your presentation just yet (but I will!), linear rail(s) implementation is almost entirely misunderstood for what they Should be used for: precision and repeatability. Naturally they can't do that alone, which also comes down to the precision of what is being driven on them, subsystems (steppers/motors/controllers, drivers, processors, etc). Getting rid of belts and degrading surfaces through wear/friction are another side benefit, and you Might be able to increase speed to a degree, but that's not what they're for on design principle. Providing a smooth and rigid support structure for repeatedly hitting a precise point of space, that's the bread and butter of linear rails, along with a well-surfaced ball screw and all the rest. To that end, side effects of motion systems that degrade with wear (rollers and belts for instance) induce certain defects just through operating principle even in good condition. A belt is essentially a long spring under tension, and when you repeatedly use said spring, it's dimensions and capabilities degrade over time. Add into the spring effect with general momentum and direction of travel changes, those effects have more to do with printing defects than frequency.

[@SpencersDesk]

I think I misunderstood that going into this project as well. This is a wonderful comment! I think rails are great for eliminating some amount of uncertainty in your extrusion. I think 3D printing is mostly held back (accuracy wise) by the actual extrusion process, not motion.

[@C-M-E]

@@SpencersDesk Oh, most certainly. It comes into play with the ol' economy of scale and production quality for x price bracket. A $300 printer will kick out usable models all day, but it costs money to hone in on micron precision. Electronics, software and stepper motors on most commercial models will only hit essentially the 8 ring of a target if you'll pardon the phrase, but getting into the 10 ring every time requires operator knowledge and better hardware. Naturally when you step into engineering-grade machines, tolerances and materials increase exponentially, as does the price.

[@davidjiannotti1537]

It depends. Fixed some issues i was having with me s1 plus. Bolts pressing into frame was causing my bed to tilt up as wheels ran past. It also got rid of side to side wobble. Completly eliminated first layer issues i was having. My other machines are happy with rollers.

[@yellowhouse88]

Rods with LMxUU bearings are the optimal option IMO, Prusa picked them for a reason

[@oleurgast730]

Definitly, but always high quality ones. Prusa had some not so nice ones in the beginning of the mk3. Personally I have good results using Misumi bearings. There are alternatives like brass bushings. However, they are much more precise and have lower tollerances. You need some way to place the rods more exactly in parallel- Trying them on my mk3s+, it moved smoothly in the middle of the x-axis, but bloccked at the end of it.

[@croustibat682]

For the cost of a high quality one, you can buy a whole printer from creality, so... yes it is better. It better be. Meanwhile my ender3 v2 prints petg at 150mm/s (everything at that speed, infill, outer walls and so on). without rails.

[@SpencersDesk]

One thing I see as a problem with them is the difficulty in setting them up. The linear rails were fairly straight forward, but I’d have to design parts to get rods. I definitely agree with Prusa design. I wonder what made them choose the rods?

[@thomaswiley666]

@@SpencersDeskStability across the bed. Rollers have so much variability across the bed (wear spots, tension, etc). Rods are… rods, as long as they are as straight as possible.

[@benhoffmann6870]

Great video - great analysis! I have machine tool design background. Big difference between linear rail trucks that have a light preload (much better damping, much more friction) and linear rail trucks with normal fits (slide freely, less friction, much less damping). Just something to keep in mind. Side note, your grease might be a bit non-newtonian, high friction at high speed. Expected operating temp of grease might be much higher (??) than temp seen at rails in printer ?

[@SpencersDesk]

Great input! I didn't know about the differing amounts of preload before this video, just that there was preload. Do you have a recommendation on which preloads are better for which axes? I imagine you'd want your Z's to have a preload and the X and Y to have less preload? I think lubrication is a whole other world so hopefully what I've got is good enough for now haha

[@Juiceboxmakes]

They are so handy for builds. Voron v0 the form factor of the mgn 7 rails allow for a tight build that makes for an easy enclosure. Rails are used in some of the harshest and cleanest environments. They are made to known dimensions. They just make sense. Brake clean on the rails, i soak the carriage in alcohol and air dry. Mobil ep2 and then amsoil mp on the rail and good to go. Higher quality v wheels climb in cost as well. Id give the overall edge to rails.

[@SpencersDesk]

I think you’re exactly right. If I could eliminate v-slot wheels from future builds, I would. However, maybe not from printers that are currently built (assuming the wheels are working fine)

[@polycrystallinecandy]

2:15 "I'll be experiencing linear rails" That's what my gf tells her friends when she plans to stay at my place

[@egemens]

Linear rails are made for linear motion of machine parts. Vslot is more like diy ish solution which really outperforms the expectations. It performs yes but if you are serious about what you do, you will recognize that it makes sense to use an industrially accepted and tested solution. That doesnt mean there isnt room for innovation. Carbon tubes like of bambu or hydrostatic bearings might be interesting to look at.

[@sofascialistadankulamegado1781]

The debate about vslot vs linear rails seems almost absurd. Vslot looks like it has a lot more issues with momentum during acceleration due to the roller resistance and weight of the carrier, but then the linear rail carriers have lots of preload compared, the grease viscosity resistance and then the momentum delta due to carrier weight. I feel as if linear rods are the best overall just like Bambu Labs use on their corexy printers.

[@SpencersDesk]

I think it's obvious that linear rails are better than V Slot rollers, no questions about it. The question I tried to answer in this video is, are linear rails going to bring any substantial change for the effort/price? I think the answer is no. If I were desiging a printer, I'd go for linear rails every time

[@mastah39]

I'm in the middle of this problem myself, my printer is an old Tevo Tarantula, upgraded lots of stuff in the past 7 years, but never the kinetics for the x and z axis. I have a linear rail on the y axis because a friend of mine had a 300 mm Misumi one with absurd specs that he got from a machine that they were scrapping. The point is not on what is the absolute best system. Rails will win every time. The point is if the money you spend in rails is well spent. If it results in a improvement in print quality that's worth the money and effort. I don't print miniatures. Nor I'm trying to beat the Benchy speedrun record, I print mainly functional parts and the few aesthetic parts I print are usually sanded, primed and painted over. Is it worth to spend new printer amounts of money on rails to fix things lile ghosting that have no effect on the properties of the printed part? Are cheap Amazon no brand rails, which still are crazy expensive, any better than a good wheel design? Those questions and their answer are much more subtle than just picking the mechanically better solution on paper.

[@jerrysmith6798]

I mostly use linear rails due to ease of use when building printers as opposed to v wheels

[@SpencersDesk]

100% I would do the same. Linear rails are better than all the other motion solutions in my opinion (except for cost). But this video is solely about whether upgrading is worth it, not if rails are better or not (which they clearly are). Thanks for your comment!

[@daliasprints9798]

Yep. Linear rails are a staple of good design if you're making your own machine from scratch or making a major alteration that already requires a new linear motion system, but silly on their own as an "upgrade".

[@GremlineQPl]

If I understand correctly, you have anycubic vyper, I have a few questions about it: 1) Did you do a mod to make the stepper motor drivers work in uart mode, and is it on a stock motherboard, or did you replace it? 2) Did it bring any benefits if so? I'm asking because no matter what I do, when I exceed the acceleration of 209mm/s² on the y-axis, this axis starts to make terrifying sounds as if something was about to break, also if its important I use marnin for now (to much work to go from working marnin to klipper mayby some day)

[@SpencersDesk]

I sure do have a Vyper. I actually have a video on the change in electronics which will answer your 1st question! I will say, I enjoy the new motherboard a lot more for various reasons. I can't point to what your issue may be. I would check that your belts aren't too tight, motors aren't burning hot to the touch during operation, etc. Hope that helps! Also, I definitely recommend Klipper. It takes a bit of a learning curve, but it's super cool.

[@GremlineQPl]

@SpencersDesk I thought that my problem was related to the fact that currently stepper motor drivers use standalone mode (if I'm not mistaken) and at higher speeds it is better to use in UART mode, or at least that's what I heard, but I noticed that the problem is with the belt, but I doubt that it is weak or too strong tension, because I managed to observe that at these high speeds the belt somehow starts to vibrate at the end of the rail, so probably something is crooked somewhere. Anyways thanks for your answer.

[@SpencersDesk]

@GremlineQPl I don’t believe there’s any difference in stepper performance between UART and standalone. Standalone means the TMC drivers are in SpreadCycle mode naturally which is better for higher speeds/accelerations. Having drivers that you can communicate to with UART doesn’t really change much IMO. Not sure what the problem is though. Good luck!!

[@GremlineQPl]

@SpencersDesk if you want to hear what it sounds like I once created a post on redit, I added a video there where you can hear this sound (normally this sound is louder than in this video) and you can imagine the horror when something started printing and suddenly you hear something like that. I deleted the beginning of the link because yt doesn't like it /r/AnycubicVyper/s/3LGTY505xF and finally, thank you for this conversation

[@GremlineQPl]

@@SpencersDesk ok thanks for this conversation, I would send a link to a post about it that I once created (if you wanted to hear the sound) but yt doesn't like it

[@YourBuddyDinec]

I know the LDO site recommends you soak the rails in high % alcohol for 10 minutes to clean them

[@SpencersDesk]

That is fantastic information! Thanks for the heads up. I’m planning on figuring out how to properly clean and set them up in the future!

[@TheEnigmaBlade]

@@SpencersDesk If you want to avoid soaking the rails in a large amount of IPA, you can shortcut the same process by moving the carriage centered over a hole in the rail and pouring the alcohol through the hole. Move the carriage back and forth and wipe off the rail. Repeat the process a few times until it seems clean-you should be able to feel the effect of cleaning in the carriage's smoothness of motion.

[@SpencersDesk]

Not a bad idea! I'm lucky to have loads of IPA from resin printing. Now if I could only find a container for the cleaning...

[@BeefIngot]

I think some now ship with proper lube from the factory rather than this storage gunk they put on. Im surprised at LDO doing that. Id expect better from them.

[@naasking]

On the material properties, even though PC and Nylon might have better resonance properties, they have a lot of creep. ASA and ABS have very low creep which is why Voron builds use those materials. A printer that forces you to re-tighten all of the screws every few days is not a fun time. Thanks for the video!

[@SpencersDesk]

Hey, great tip! Definitely helpful for me to keep in mind!

[@enmanuelhenriquez975]

Good video v slot need more maintenence that linear rail

[@paulminor4707]

the difference you felt in the rails is more likely than not a difference in the amount of the protective coating put on for shipping. you should have done the feel test again after cleaning and regreasing. the difference in feel will lower drastically.

[@SpencersDesk]

I definitely did not prepare these rails properly. Thorough video coming in the future lol

[@ErtsenPlayGames]

3:00 - you should clean them inside , cheap ones only have anti rust protection but are filled with dust etc from production line , clean them , take ball bearings out and clean too ,deburr the plastic ends (holes for balls sometimes have leftovers from mold) if possible to take them off, use ptfe oil (they will be louder but easier to move) It will make your prints way smoother , better quality etc. Next about infill... Meh with more, thick walls , less infill but only honeycomb , gyroid , cross3D (cross 3D is my favourite oposit gyroid , honeycomb is great in Z axis but in xy gyroid and cross 3D works better in my opinion with less %) Mostly to hold top layer / add stiffness for hollow thin parts Overall ....good linear rails and stiff frame are a way to go , but with cheap rails dont expect anything better , good linear rods / rly good vrails will work way better but need less accuracy to mount them but more work from time to time

[@SpencersDesk]

The cleaning is a great tip! I definitely intend to do this in the future. I forgot all about the infill rule despite watching videos on it lol. Thanks for your comment!

[@ErtsenPlayGames]

@@SpencersDesk PS. You can also check the ball sizes inside it they rly loose and try to find bit bigger size then for few bucks more you can get way better accuracy from them ....but with proper cleaning , more screews they should work fine

[@The-Anathema]

It would be fun to see how the aluminium bracket changes if skeletonized, what about something exotic like titanium? PCBway offers titanium last I checked (and again, skeletonized vs solid)

[@ifell3]

22:41, could this be down to your microprocessor's compute speed?

[@SpencersDesk]

I don’t believe so. They are very simple movements as far as I’m aware. The board + pi would be quite overkill for it!

[@sofascialistadankulamegado1781]

Grease in linear rails is ok, but don't load the carrier until it's full. I would leave the factory grease in. The friction from the grease alone limits the speed especially at the start of the print when the grease is cold. Use machine oil instead. One that has PTFE in it will be the best. The rail will be much noisier but however it will have very little resistance from the grease viscosity. The wear rate is slightly higher than a fully greased carrier, but for the sake of printing speed, it's a small compromise.

[@SpencersDesk]

I think all machine grease should be removed personally. Part of greases job is to provide friction. It’s like how oil in your engine becomes more viscous as the engine gets hotter! There is definitely a huge debate on PTFE vs no PTFE which I plan to investigate!

[@pavelkolar9543]

Main thing is to get a high quality rails. Like HIWIN with Z1 preload.

[@SpencersDesk]

I am of the opinion that you’re in diminishing returns territory when it comes to HIWIN rails. I think well taken care of and properly setup mid-priced rails are good enough for our use cases. Also, with preload, I’ve seen a lot of people in the Voron community saying that it really doesn’t matter. I think I agree with them. Our frames don’t have the same tightness of tolerances as the rails, and our extrusion system are definitely much less accurate than the positional accuracy of rails. So, that’s an opinion I have I guess. No real facts to back it up

[@Mattvbro]

I have a Bamboo A1 and a Ender 3 S1 and i have to say the v wheels have less maintenance to do as compared to the rails on the A1 as i find the grease and even the oil is just a dust magnet and within days of cleaning the rails and putting new lube on them they get all grindy feeling from all the dust that builds up on them. I have had the S1 for over a year now and have probably cleaned the wheels about 3 times, never adjusted them but feel like the bed ones may have been a bit tight from the factory, the bed wheels are currently looking like little fur balls but they still roll smooth as butter so im not phased. why fix what's not broke yet right hahaha

[@SpencersDesk]

I agree with this sentiment. The grease definitely attracts debris. I also never really saw issues with my wheels. I have heard many stories of others having issues though

[@Recycled]

Linear rods are the best because they look cool. Shiny, shiny!

[@bobbldr]

Great video, minor suggestion: lower the guitar mix. It cuts through like you want in a song, but over your narration it gets to noisy and with a single earbud in i felt like i had to listen over another video. Thank you for providing the info on the test!

[@SpencersDesk]

I’ve hopefully addressed this issue in my latest and future videos! Audio is something I found difficult to get right. Thanks for the feedback!

[@Xizax41325]

I've slowly been upgrading my ender 3 v1 trying to get to a decent speed vs quality thing. I think I might swap to linear rails for the Y axis simply because my bed wobble is insane. That single width aluminum for the V slot on it is a menace. I just wonder is linear rails would be good for it.

[@SpencersDesk]

In this case, I’d definitely recommend the rails if you’ve got the money and time! If not, do your best adjusting the wheels. Otherwise, rails will give you a much more sturdy setup. I can’t remember if I mentioned this being a case where I’d upgrade in the discussion or not

[@croustibat682]

it is known the v1 has issues with the aluminium extrusion not being large enough. Creality doubled its width on the v2. The stock springs are also very soft, replacing them limits wobbling too. If you are going to do something about bed wobbling, maybe you could try to have an even wider gap than the v2 by using 2 extrusions. There is a bit of work on the bed carrier, but it should be even better.

[@rodrigofarruguia1708]

Did you wash the rails and carts, cheaper priced rails will have all sorts of metal shaving and other junk from the factory ? and replacing the bearings for better tolerances helps.

[@croustibat682]

I second that. cheap rails often have loose bearings. There is a lot of precision to gain by using the right bearings inside.

[@SpencersDesk]

Hey, I did not thoroughly wash them, just rubbed them down with IPA, so the insides of the carriages were not ideal. I’m planning on doing a more thorough video on properly preparing/mounting linear rails in the future though!

[@EpicHeroSandwich]

linear rails add weight, and the grease required attracts plastic particles like no other, that's why V slot poly wheels are really the answer, you can clear the tracks with compressed air and when properly set up, you really don't have to think about them, just like linear rails, unless you're running exceptionally large/heavy prints all the time. Just my 2c. Really nice writeup!

[@christopherd.winnan8701]

Have you also considered a cable robot set-up as an alternative to linear rails?

[@SpencersDesk]

I can’t say I’ve heard of it!

[@christopherd.winnan8701]

@@SpencersDesk - kzbin.info/www/bejne/Zn26qZKPbql-hsU

[@christopherd.winnan8701]

@@SpencersDesk Do a search for high speed cable robot and you will amazed at the speeds available.

[@pavelshevchenko575]

Maybe you should mount accelerometer in the center of the bed. Right now you measuring z movements of the corner of the thin plate, mounted in the center. During the real prints, your model probably will be in the center of the bed

[@SpencersDesk]

Hey, this is a great point which I intend to investigate. Ideally, our print head and bed are perfectly rigid to the point where the point of measurement doesn’t matter. However, I’m not sure to what extent that assumption is valid or to what degree mounting it in different places is an issue… Thanks for the comment!

[@edriyin]

Thanks for the video, but i think you are focusing too much total vibration amplitude and missing point. Nearly every case at Y resonance measurement, vibration from Z axis is too dominant. You should add braces or something to add rigidity to z axis and fixing printer to more rigid surface to increase damping. Than getting new measurements from axes may give more intuitive results about linear rails. I love your presentation of videos, keep up. 🙌

[@SpencersDesk]

I agree with this a lot! I’ve actually discovered it’s my print head. It bounces in the z-direction as it goes along. Planning to address this before I do anymore resonance on the x-axis! Stay tuned and thanks for the feedback/kind words!

[@deltacx1059]

Ok but how much did those PCBway parts cost? Literally nobody ever says anything about that bit.

[@SpencersDesk]

Fair point! I’ll see about putting that in my future videos. ABS~$9 for both Resin~$14 for both Nylon ~$19 for both PC~$20 for both Aluminum~$88 for both (keep in mind I had them painted which added a significant cost) You can also upload files to PCBWay and get a price estimate for plastic parts pretty quickly! Hope that helps!

[@deltacx1059]

@@SpencersDesk that is actually reasonably useful, i have looked at other services other channels have recommended and they were hideously expensive for even small parts.

[@gsestream]

riding an I-beam with a mono-rail train is useful. self-balanced. also in 3d-printing.

[@leloctai]

What's about steel rod and bearings? They're a lot cheaper than the square rails and doesn't wear like plastic wheel. They can sag over long distances but the typical desktop printer are too small to worry about that. Especially on Z I see no point in spending on square rails.

[@SpencersDesk]

I think that they're a valid option. The only issue is they may be hard to install! Linear rails are easily mounted to aluminum extrusions. Linear rods would require some additional parts to be designed and created. I personally think for the effort required, some cheap linear rails are the better option if you're upgrading. Thanks for watching and for your question!

[@TheNamelessOne12357]

You're lucky not to spot any major difference after installing noname rails. Those rails are produced on low quality worn out CNCs and may have really bad tolerances. They can arrive bent, jamming or having too high backlash. The cheapest acceptable rails are made by HIWIN. Not the best, but OK for 3D printer. Anything cheaper might be garbage. And even if they seem to be OK, they may wear out very quickly.

[@SpencersDesk]

I think most of the community uses rails cheaper than HIWIN. Honestly, I think there aren't going to be huge differences between the cheaper rails and the HIWIN rails relative to the cost increase. I think the cheap rails require a lot more attention at the beginning, making adjustments, making sure the bearing are all good, etc.

[@Amybnuy]

Edit: technically your title says “ upgrading” so I take it all back. But even then it only applies to this one printer. Original: Tbh… I feel horrible saying this considering the effort you put on this video but you seem to be misunderstanding not just linear rails but just the whole concept? Linear rails can be bolted on and are usually bolted on something much stronger, therefore are better for very long distances, they are also better for some heavier loads at times. The list is long, everything has pros and cons, but they’re not something you slap on and expect magical improvements with

[@SpencersDesk]

Haha this is the big point I’m getting at. Linear rails are definitely a superior motion system. But yea, you can’t slap them on and expect magical improvements. Working on future videos that go more in depth to them, etc. thanks for watching!

[@LuluTheCorgi]

My guess would be before watching the video is that it's good if you print Extremely fast and your axis are very long while being worthless for most i3/ender style printers Probably

[@suivzmoi]

first off thanks for comparing resonance for vslot vs linear rails (19:57). believe it or not, there is no other video on youtube for this. i have searched. there are tons of videos about linear rail upgrades but people don't bother with evidence testing, and if they do, they don't even show the results before the upgrade. my own finding though is that they do increase max accel just like you showed. brought my ender 3 from 3.9k->10k on x and 2.2k->3K on y (ender 3 y accel is limited but because of bed weight being too high and motor too small). if you get your rails from a reputable source, they will have a light machine oil on it. if you get them from china, you would be lucky if it was machine oil. sometimes, they don't even bother to remove the cnc lubricant. if you have washed chinese rails, you would find swarfs and chips in the wash. 100% guarantee. the best washing liquid would be brake cleaner, next best and nontoxic is 99% IPA. 70% is ok, but you need to force dry it with heat because you don't have much time to reapply lube before rust begins, which is instantly. my recommendation is to use a gallon ziploc bag so you can actually fully submerge and soak the rails with little waste and also minimized time exposed to the air. take them out and dry with a hairdryer and immediately lube to protect against the air. don't forget the ball carriages. these need to be taken apart and cleaned too. metal particles collect inside there and need to be washed out. the balls from one channel should not be mixed with the balls in the other channel. be careful about this as the ball sizes in these two channels could be 1-4 microns different, depending on the QC fitting process, which is done by manual labor. there are tons of rail and carriage cleaning tutorials on YT. anyone bothering to clean their rail and carriage from factory should heed this advice to spend an hour learning the cleaning process before deciding to upgrade. steel balls on rails should be lubed with oil, per Hiwin recommendation, but grease will not need to be re-applied as frequently because of much lower evaporation. if using grease, PTFE grease is not the correct grease to use but actually lithium based for metal on metal. for the home 3d printer application though, where lots of plastics are involved, PTFE grease is actually totally fine. rails and carriage will not slide under gravity when using grease instead of oil. this is also completely ok. what you want to watch out for is catches and binding, not carriages falling under their own weight like some of the linear rail cleaning tutorials brag in their before and after videos, because that is a sign of no preload and preload is what you want to minimize resonance at high accels. a non-falling carriage has a little more friction, but completely negligible against even the stepper motors we have on our printers. the balls only contact the two tiny side channels in the rails. this part you have to ensure is cleaned with a strong wiping action and lubricated. the rest of the rail only needs lube to protect against the air. dont forget the back of the rail, that will rust too. after lubing them, don't handle with bare hands because of acids in your sweat. use gloves. for a bed slinger, your final speed and accel will be severely limited by your Y axis. find the max y accel and use that for both axes (also applies to corexy). max speed is only important for travels. print speed you will be limited by extruder flow and part cooling well before 150mm/s on 0.4 nozzle, 0.2 layer ht. 200 maybe if you have aftermarket hotend. far lower if larger nozzle or layer height. anything beyond that is a racer build and will require parts costing about a whole new budget printer. zv vibr% comparison is an invalid comparison (20:24). the vibr% is the estimated residual vibration after application of the input shaper model and is dependent on the recommended accel for that model. you have several minutes of footage showing that the different mounting adapters have different recommended ZV accels with over 30% variation among the candidates. the comparison you were probably seeking is the vibr% with NO input shaping, which the graphing script does not give because it is meaningless. instead, plot the adapters against the accel value for the RECOMMENDED model, which would would all have vibr% close to zero but have the vastly different accel capabilities showing the true winner, even if they all use different IS models. still not worth using plastic over metal though. don't bother with plastic mounts for rail to bed attachment. it is pointless to go through the effort to buy and use metal rails and then have flexible plastic connecting it to the bed. but your results already proved this (19:57). even if you print the best plastic adapter, there is no way for you tell when they start to fail and it is simply not worth the effort to have to fix that later. the time and effort saved from not having to take the bed off and put it on again and possibly also re-align dual rails for the Y axis is worth far more than the few dollars you need to spend for the correct metal parts. yes you would probably have to realign the rails if you change adapters. because of mfg tolerances and the fact that the bed bracket itself may not be totally flat. for others reading this, buy the metal adapters. you can find them on Ali for your specific printer. (10:43) the Y extrusion rocking is beyond terrible. this is probably why you are seeing Z axis vibration in all your Y resonance graphs. like you correctly pointed out that this extrusion only sits at one spot on printer base, this is simply a design flaw. you may get very good results by adding 2020 or 4020 square brackets to more rigidly fix the Y extrusion on the printer base. i think this will help a ton.

[@SpencersDesk]

Hey, thanks so much for this comment! I wanted to wait until I had time to thoroughly read and respond. I highly agree with a lot of what you’ve said! I plan to make a video about the proper prep and lubrication of rails in the coming months. There’s so much argument over PTFE vs. Lithium which I find a bit silly, but hopefully I’ll provide something more concrete for people to argue over. I am curious, is mixing the bearings in a low end rail of any consequence? Maybe in a higher end ones, but I doubt it is so accurate for mine. I actually just got some ball bearings that had a higher tolerance to use in all the rails in the future, likely replacing the stock ones. I don’t think I’ll use vibration or acceleration or any of the numbers in the future. I’m likely to use actual printed results. While more expensive and time consuming, we are chasing something qualitative (how the ringing shows/how rounded the corners are) which the quantitative results, IMO, poorly represent. I definitely agree with the metal adapters. I wish I’d made this point. Why switch to linear rails for the convenience then add plastic adapters that might remove that convenience, haha. I would argue that they add some amount of tolerance to how misaligned your rails could be? But at what cost. (Edit) I forgot to mention, yes, I hate the construction of the printer. I plan to remove the y-axis extrusion and the middle brace at the bottom that it sits on. I want to add a brace at the front and back, then lay two extrusions for the y-axis to sit on. But, I don’t think it is the major source of my z issues. My print head actually appears to bounce when you watch it in slow motion/see ghosting on the top of flat prints (as opposed to the side which input shaper actually effects). I have some major mechanical issues which I hope to sort out over the next few months. Thanks so much for an awesome comment. I’d love to see you in the channel’s discord!

[@dylanlasky2389]

I would not us the nylon near anything warm. I believe it was CNC kitchen who used nylon for a voron build and all the parts creeped and the whole printer loosened.

[@SpencersDesk]

Really? I thought it was one of the more heat resistant filaments? I wonder if it just has a low crystallization temperature that has a very broad range? Thanks for the heads up!

[@livedeliciously]

@@SpencersDesk Apparently ePAHT-CF can be used as 3d printer parts with no concern of creep.

[@riba2233]

​@@SpencersDesk it is heat resistant but can creep a lot over time, unless you anneal it.

[@elliotschmidt1559]

​@@SpencersDeskI encountered something similar during a custom design I worked on. I don't remember the exact science behind it but nylon and certain other materials will "cold flow" and over time deform under pressure. Glass fiber or carbon filled nylons are less suseptible to it in my experience but it doesn't completely eliminate the problem.

[@nicoleibundgut534]

Carbon fiber would be interesting because its also very light.

[@SpencersDesk]

That’s a great point! The only concern is what plastic is carrying it. PLA, PETG, and Nylon (common CF filaments) would be good in the hot environment. Maybe there are others I’m unaware of?

[@nicoleibundgut534]

@@SpencersDesk i mean the classic carbon fiber. usually they are mixed with 2k Epoxy resin. Very light and very rigid. Much stronger then aluminium. But you need to find someone who made it you can't print it that way obviously lol. But for the ender there are Carbon fiber bed's.

[@SpencersDesk]

I guess what I’m getting at is the shape. Carbon fiber is great for tubes and sheets as they are long fibers combined. This makes it not so good for oddly shaped stuff, like these brackets. I do agree that carbon fiber is awesome though, I plan to use it in some upcoming DIY projects!

[@owned323]

The shapers that you are showing show a fundamentally flawed machine. There is a lot of noise in your graphs for both V Slots and Linear rails. It makes it very hard to compare anything when the machine has other mechanical errors. Linear rails are absolutely, 100%, without question a better solution to pom wheels. This is why all of the fastest printers like a monolith Voron or a VZ bot run them. Its also important to note preload used on the rails which i didnt hear you mentioned, but may have missed. For example a good Y rail would use Z0 preload, a good X rail would use Z2 and be wire EDM cut.

[@SpencersDesk]

I should have emphasized that, yes, linear rails are the better system in pretty much every way. But, not worthwhile if you have working wheels already (IMO) Hope that clears that up! Yeah, I know my resonances are flawed. This is actually something I deep diving into right now, so a video will be made on that!

[@MaciejWiercinski]

Great video. Your studio lighting could be better.

[@SpencersDesk]

Thanks! Current budget allows for some morning sun. Hopefully that may change overtime!

[@kristian95iversen]

Duuude i love your channel, but I have a core xy printer and it seems that your channel is very bedslinger oriented?

[@SpencersDesk]

Thanks so much! Yeah, for now I’ve just got the bed slinger. Hopefully I’ll be able to build a coreXY one day. Most things I cover should apply to all 3D printers though! PS, I love your profile picture

[@KenDilas01]

Lmao, love the honesty!!

[@ReneGaytan-i9y]

You just saved my credit card

[@SpencersDesk]

It’s only safe until that next printer shows up ;)

[@laszloszell8753]

Not waste of money, but in 2024 the resin printer to go. I know,for lot of people don't have good environments or enough place for resin printer, cleaner,wash,cure but I would never buy or build anymore an fdm printer . I still have a self made corexy, but I never use it honestly.

[@SpencersDesk]

It very much depends on what you like to print. If you're into making structural things, I would say resin is not the answer. Also, Resin definitely costs a lot more to operate over time and requires much more effort in my opinion

[@DJProPlusMax]

Yeah, I saw linear rails as a speed upgrade, clearly it is not - so wheels are going to stay on my printer.

[@SpencersDesk]

I think that they are still faster. They’re not intended for speed, but they are faster. I think my preparation of the rails in this video was subpar. Stay on the lookout for a future video on it!

[@tomtrevino8192]

I would say that the fundamental reason for linear rails is true movement and not having the inconsistency or "slop" of v-slots and wheels. All of your movement goes the direction you want it to. If liner rails were not a solid method, you wouldn't see so many of the higher end printers using them like Voron, Rat Rig, BLV, and other printers. Personally, linear rails have saved me time in maintenance and repairs as I am not chasing eccentric nuts or bad wheels. On my Ender 6, they are also allowing me to use almost all of my printable area making it almost a K1 Max.

[@SpencersDesk]

I have to agree with you 100%. Funny enough, after publishing this video, I found a wheel on my z-axis which freely spins, making 0 contact with the extrusion. Luckily I have those other rails haha

[@NWalker-zx6xu]

You should never do 100% infill but rather 50 outter wall count. its stronger.

[@SpencersDesk]

Why is that? Is the wall pattern better than the infill pattern? Is that true for all infill patterns?

[@whydoesyoutubehavehandle]

@@SpencersDesk I don’t really know the exact science word, but more walls = more strength because the majority of strength often comes from the outer edge A pipe and a cylinder of same dimensions will pretty have almost the same strength, despite the pipe is missing its center compared to the cylinder Also, I think that infill is more for supporting the top layer (and to help hold the walls from becoming wonky (I think, Im just guessing)) dont quote me Im not sure) and help with with print time. Because infill can be printed pretty fast, unlike walls I hope someone who have more knowledge can give us more details on this, but I hope I still helped you out :)) (Edited because typo)

[@SpencersDesk]

That's a good point about the different speeds that they're printed at. A good rule of thumb is slow prints = stronger parts. My guess is it's the orientation of the lines reative to each other. Having contiuous lines stacked is likely going to be better than a bunch of lines crisscrossed. Maybe certain infills like linear wouldn't have much of a difference with being all walls.

[@alfiegordon9013]

Would that be the same as 100% concentric infill?

[@whydoesyoutubehavehandle]

@@alfiegordon9013 I guess it would be pretty similar... But that will need to be tested for sure

[@BeefIngot]

They shouldn't be an upgrade so much as users should simply refuse to buy 3d printers with Vrollers. Linear rods and linear rails are both better solutions and penny pinching here will cost you a lot more in frustration later as vroller carriages loosen or the rollers themselves wear out or the dust that will accumulate from their use or the frustration of dealing with eccentric nuts. No matter which way you slice it, there isn't a budget for which this cost cut makes sense. Vrollers get a big 👎 from me.

[@SpencersDesk]

I totally agree with this sentiment. Past your first printer, I wouldn't recommend paying for rollers ever again. The first printer is a bit of hazing though ;)

[@BeefIngot]

@SpencersDesk I dont think a first printer should be a hazing at all. In the days when an a1 mini is 200 bucks regularly, there is absolutely no reason to experience these low quality, poor user experience printers. You get nothing out of it because they have problems you wont face on any mosern printer ever again. I think that people arent forced to learn useless maintainance and repair skills specific to poorly designed printers is great as it allows them to have fun learning what they want to learn instead of what they were forced to learn. To me its all about where time is best spent, especially something as valuable as your free time.

[@SpencersDesk]

@BeefIngot that’s very true. I guess I’m coming with the perspective of having a printer for tinkering, not for printing, which I know is an odd POV. You’re definitely right though. If you’re only interested in printing, you shouldn’t have to deal with all the silly issues

[@BeefIngot]

@SpencersDesk Id go a step further even. Im occasionally interested in printer tinkering, and one day Ill get back into it. but when I do, this round Ill have realized what actually gets me going. I dont want to tinker to fix problems that have already been fixed. I want to do something unique and new, like maybe implementing a good semi multi toolhead system, or a nozzle that isnt quite 5d, but does non planar 3d printing to higher degrees. New things that exercise the brain and that when Im done with I can say, hey, I really created this thing. Currently, I what I typically do with my printer is use it to help me prototype things. I like to think of the greater hobby as kinda hobbyists product design. Anyhow, thats the long and short way of saying that even for tinkering, I dont think one should start with anything except a good as close to hassle free experience as possible so they can focus on the aspect of fff printing they want to improve, whether thats speed, a new aspect etc. For that Bambulabs probably arent the printers for that considering their closed nature, and probably not Prusas either but maybe something like proforges newest machine, or the SV08 or similar printers which are decent enough out of the box but open to modificiation. A great example I think, is where Michael from Teaching tech is working on making a multitoolhead SV08.

[@JustinBuildsThings]

you should not have uploaded this video because it is essentialy pure nonsense with bad tests and a complete lack of understanding of whats actualy changing, especially since you didnt spec the rails. vwheels have the ability to adjust the tension and both rods/bushings and rails can come with different preloads, essentialy a modification to the ball size and how much its compressed on the rail there. this also has a pretty dramatic impact on print quality and resonance as the super smooth feeling rails might be very low (too low) preload and might give you issues like toolhead wobble or wider peaks on shaper. higher preload will still be smooth but wont necessarily gravity drop and this is completely fine and on certain axes substantially preferred like on the toolhead axis. there are also grades for deviation over length etc as well and some rails are stainless and some are not, different materials and so on. cheapest rails and mystery rails are also not necessarily straight laterally or vertically. good rails can be extremely expensive as you push further into accuracy from brands like hiwin, iko, thk, etc. there are decent cheap china rails but like, you should still be specifying the rail material, preload, etc. you also used the wrong kind of lube. dont use ptfe lube in motion stuff. its generally stupid to buy random rails without knowing the specs you are getting if the goal is performance testing on something like an accelerometer on a printer. when cleaning your rails, you also should have soaked the balls and carriages and flushed them in the ipa in a bowl and let them soak and pushed them through the raceways a few times. and then re lubed them, put them back through the raceways and relubed again, before putting them back on the rail. there is no shortage of guides about this stuff from other maker and printer communities. you also should always be measuring the axis of the toolhead from mounting the accelerometer to the nozzle- you dont x-ray your arm when you break your leg because they are both connected to your torso....one of the big problems with like for example voron kits, is that they ship with low preload rails (zf/0), so it is an extremely common mod to get better, straighter, higher preload x axis rails (z1.5 - z2) and virtually all of the diy printer parts shops sell these. v slots and t slots have different profile angles for roll-in nuts and you need to be careful when sourcing given this. there are also 2 major standards for t-slots, and this is problematic for example when trying to fit certain size parts like certain mgn9 rails. if the rails arent in a moving axis you probably want to bolt all of the bolts down since there wont be a weight penalty, as the rails can deflect and you wont have to worry about that weight/rigidity relationship. from a resonance testing perspective, it may not be apparant to you but belt tension will impact your charts too. rods can be more rigid than people realize and the rod bushings also have tolerance specs and preload. for example my stratasys printer uses rods and it is extremely rigid. my vorons and annex printers are on rails. the fact that you arent tuning the printers until you get clean charts is also problematic because its essentially not meaningful. dont even get me started on not making sure the rails were parallel. i run a decently large print farm (~30 machines) and i would never buy vwheels for them because they are inconsistent and a wear part with a fairly high wear rate and require frequent adjustment, among other reasons, but primarily just that alone. also, even cheap 100$ printers like kingroon kp3s's come with usable rails. there are much more durable variations of wheel guides and metal wheels and other stuff but thats a whole other rabbit hole to get into. this video is a giant waste of time, because your experiments are using objective tools in stupid nonobjective ways with bad experiment setup and you dont understand what you're doing or how or why. i dont want to discourage you from making videos, but like, do better. i want to see content like this but i want it to be with at least reasonably correct information. people will share and watch and reference this and now im gonna go have to help like a bunch of idiot noobs unlearn bad shit they saw on youtube in the discords where i help people. please dont make my life harder.

[@SpencersDesk]

Thanks for watching! Lots of great info in this comment. KZbin is available to you as well, so feel free to fill in whatever gap you feel is missing. But that last paragraph man. I hope life doesn't have you down...

[@The_original_Reth]

I must concur with some of the other comments. I think you present a very surface understanding of the motion system.

[@riba2233]

Haha bro about that last part, I totally get you

[@joescalon541]

What’s your opinion on “OpenRail linear rails” by openbuilds. Essentially an inverse of v slots and wheels where the wheels have the v slot. Should it be considered a decent budget upgrade from normal wheels or would spending 4-5x researching and sourcing decent linear rails?

[@riba2233]

Nice vide0, but you need to fix your audio, it is extremely quiet.

[@SpencersDesk]

Audio has plagued this channel for so long… I’m getting closer to a solid workflow though! Thanks for the feedback!

[@riba2233]

@@SpencersDesk np, good luck :)

[@MisterkeTube]

POM V-wheels are forgiving whereas linear rails are so hungry for attention and out for revenge when you don't give enough of it.

[@DynoRC]

You made so many errors it completely took out the "scientific" aspect of the experiment, which doesn't leave you with much of anything to be honest. Basic alignment of the rails that requires a dial indicator? Out the window. Rigid mounts for the bed? Well you had aluminum but decided to use plastic to go with the rest of it, again losing the whole point, not to mention painting the aluminum ruining the precise machined surface. Using quality rails? Appears you passed on that one too. For printed structural parts the perimeter thickness has way more impact on the strength and rigidity than infill so the resonance results you got reflect the mass of each part more than anything, rendering your experiment meaningless. The list goes on... You measured wrong things with wrong tools and got wrong results, it's almost impressive. So how about not making content on topics you're not qualified to talk about?

[@SpencersDesk]

If we stopped making videos we weren't fully qualified to talk about/didn't make mistakes in, KZbin would dry up real quick. Science is based around making mistakes and learning from them. Take a walk down the history of Physics and see how much we got wrong. I did my best, presented my results, and discussed them without hiding anything. That's how progress is made. If you'd like to help the community, try adding to it. As for the alignment, there is very little gain for setting things up perfecltly. I moved the bed and repeatedly until there was no binding left. For a 3D printer, that is good enough. Maybe you didn't watch the video so closely as I did use the aluminum mounts for the rest of the video, not plastic. The precise machined surface of the aluminum truly means nothing. Again, it's good enough. Quality rails are again of diminishing returns in the world of 3D printing. Most of the community is using cheap rails, especially if upgrading their printer. So, this is applicable. not every can afford to drop 100's of dollars on rails. Have some perspective. While the perimeter thickness has "Way More" effect, the infill also affects it. So the experiment is still valid. The difference in mass is negligible between the different infill parts. Again, 3D printers are not precise machines. Their motion system could be perfect and there are still going to be issues because of extrusion. There is so little gain from throwing money at tools and parts to setup your printer perfectly. Big picture.

[@DynoRC]

@@SpencersDesk I'd agree with you but then we'd be both wrong Making mistakes okay, doing it confidently in a "informative/guide" format is not As for the "KZbin drying up, no it'd be less diluted with misinformation I'm sure there are plenty of topics you dominate, try sharing those with people instead

[@SpencersDesk]

@DynoRC I’m sure I never presented this as a guide or a how to. Simply a bunch of tests and my interpretation of the results. Again, science is full of misinfo. Nature of the beast. We would be set back so far if we tried to avoid it at the cost of progress

[@flagman3116]

Free product in expectation of a review is 100% sponsorship. You handled it well, but US law is very clear: "Assume now that the consumer joins a marketing program under which participants agree to periodically receive free products from various manufacturers and write reviews of them. If the consumer receives a free bag of the new dog food through this program, their positive review would be considered an endorsement under the Guides because of their connection to the manufacturer through the marketing program."

[@SpencersDesk]

At no point did PCBWay request or make me feel obligated to review the item or say anything about them at all! This is good to know though for the future so I can be extra careful. Thank you!

[@KeithZim]

STOP upgrading.... Things are moving too fast. Save your money and buy the new flavor of the month when you can. Upgrading is just throwing money away in denial. Your 1 year old printer is "old junk" already. Move on...

[@Kevin.L_]

I just ordered a Qidi Q1 PRO rather than upgrade my ender 3 pro. Couldn't see wasting money on something so old.

[@SpencersDesk]

It's all about why you have your printer. I have mine because I enjoy the process of testing, designing, experimenting. If I bought a printer that just, worked, I may get bored of it. I also find that I'm increasingly attached to this printer. It's like a little pet that I can make better and better. No one size fits all :)

[@46449771]

Your video has some really good testing, but the clickbait thumbnail is pretty bad... Your testing is done on your cheap bed slinger machine, but when you get into higher performance machines I suspect you will have a different opinion. I guess being dramatic gets views, but many of us in the 3D printer community prefer a more humble approach.

[@SpencersDesk]

I can see how it could be seen as click bait but I hold the same view. Thumbnails and titles are meants to make you click and watch. I don't feel that my title and thumbnail were a misrepresentation of my video. This video was on whether upgrading to linear rails was a waste of money, not on whether linear rails are better than V-Slot wheels. Most of us in the community can only afford a cheap bed slinger which is my target audience.

[@46449771]

​@@SpencersDeskHey, that's fair enough, thanks for your reply, and I did enjoy the video, so thanks for sharing.