@@OERYV I know how to drill, but what screws go into - carbon directly? I should get carbon and then fill it with plastic again (preferably with inserts) just to have something to screw in?
Weight is the enemy! Thanks for sharing this, looking forward to the results. I'd be interested to see the accelerometer stuff, for sure! I've considered drilling weight reduction holes in the sliding blocks on my delta's linear rails, but I couldn't find anybody who'd already tried it to give me enough confidence to go for it. It seems like there's way more metal in those blocks than the stresses of a printer could ever warrant, and pretty soon I'll have cut enough weight elsewhere that switching to titanium fasteners would actually mean something...
Hi, just for curiosity, why this will change somenthing on a delta? In delta the rail are fixed to the structure and doesn't move like in corexy or cartesian where one of the axis and the relative rail have to move (on corexy on x normally you have the x with related rail and head are moving connected to the y rail, while on cartesia is even worst since also the motor move with the gantry). On delta only the 3 rail block move on the rail but the rail is not moving.
@@SP_99999 The rail blocks are the things I'm talking about reducing in weight, not the rails they ride on. They contribute to the moving mass and at just over 50g apiece they're one of the biggest single-point sources of moving mass on my delta. 150g (the weight of all 3 together) might not seem like much on a cartesian system, but 150g is just about the same as my delta's entire effector with V6 and Nimble mounted. It's a different ballgame on the weight side for deltas. All that said, I'm not sure how feasibly the linear bearing blocks could be cut down. They are VERY hard steel and I don't think I could put a dent in one with any tools I have at home.
Finally someone figured out there are numbers OTHER than 12. To be honest I'm surprised you didn't go straight to mgn7. As far as load capability even 7 is still over-kill but that's the smallest affordable size.
@@VladOnEarth mgn7c has dynamic load capacity of 0.98kN, static 1.24kN, mgn7h is 1.37kN and 1.96kN respectively. So basically you can stand on it an it'll be fine. A 300g extruder and gantry? It doesn't even begin to feel it. If you feel adventurous you can order some higher grade bearing balls and find a better fitting size and fill with those for even less noise and higher precision/loading capacity.
Disclaimer first, I have absolutely no knowledge on klipper and Input_shaper. When comparing those diagrams I noticed that the first y axis (power spectral density) is scaled differently. If you compare the two measurements for the (printers) y axis, the value of the first peak stays roughly the same while the value of the second peak nearly tripels. So while the bigger peak is now the second, higher frequent one, the first isn't really altered. Or do I understand it wrong and higher values are better in this case?
@@Vez3D was going to make this comment as well. The graphs are actually showing that you've *increased* the resonance by lowering the weight. This makes intuitive sense as well - when there is less mass, the resonance will increase in amplitude faster. On the contrary, the data actually show here that the mass of the gantry was *not* the source of the resonance peaks you originally had (they are still there after the changes). This implies that the resonance is coming from another source. Probably: belts, frame, motors, or play in the rails. Since it looks like a corexy configuration, my guess is the long belts would be the first culprit to look into (as you've already stated). Try tightening, and see if there is any difference in frequency peak. If so, this could be further improved by upgrading to thicker belts. Seems like you're using 6mm belts? If so, try out 9mm? If the peaks don't change from that, another possibility is that you're hitting the resonant frequency of the motors. If this is the case, larger motors or higher currents should help. Great video! All the best.
Excellent explanation! I think you are spot on. I found that tightening belt more will just give way more amplitude and not change the frequency. So i think the motors are whats left to change here. Going with 60mm 0.9 degree
@@alexandereckhardt7153 biggest downside is that they may be too strong - can cause motor to stall if over tightened, or pull things out of alignment. But if you're careful with them, no not really :)
Great Video. I switched to the lighter rails due to mass. Rail types have a max. speed and acceleration rating that so far has not been limiting factor.
I’m really impressed of the input shaper system.. I think I’ll switch to Klipper just to try it. I wonder if the carbon tube is needed in general, maybe it’s discussed somewhere, but isn’t the steel rail stiffer than the carbon? Nice channel by the way, just subbed:)
The carbon tube increases the rigidity by increasing the moment of inertia. Think of a ruler. On the thin axis it is flexible but on the larger axis it a lot more rigid even if the length and weight doesn't change. In the same way, the tube increases virtually the thickness of the rail, making it stiffer.
@@etle113 the steel alone is not stiff enough at high accel. It will "spring" a bit. It needs a stiff support. Either a 2020 extrusion (heavy) or a carbon square tube...latter is way better :)
@@Vez3D I´m sure you booth are right! I´ve never worked with carbon, so I probably underestimating it´s strength. Would be interesting to see a input shaper graph without it, to see the introduced resonance frequency.
I would suggest using inserts made for this purpose. You used inserts for injection molding.how do one get a setup like yours? Love the bracket without a 2020 profile.
@@Vez3D CNC kitchen has a video showing of some types. I prefer ruthex which also seem to be the strongest of them all. It has started to pop up some China clones of the ruthex. Voron use them for example.
What size square tubing are you using? 30x30x28mm? Have you tried molding the x/y/x carriage out of hollowed carbon fiber powered by servo motors? It's my plan for the next 3D printer/laser/cnc project. The cheapest servo motor I can find cost about $60-70 bucks a pop. OUCH.
I added m3 nuts and lock washer inside. I printed a tool to reach the middle and used a pair of pliers really long (like surgon)..and on each end i printed small blocks with m3 nuts inserted ..so i could tight them to the Y carriage
@@Vez3D im interested to see how those lock washers do in terms of holding the fastener torque and whether they'll beat up the CF. my experience with them has always been terrible, wouldn't blue locktite be better for a low torque application?
@@Keechization i only had red locktite with me so I didnt use it. I dont know how washers will hold .. we will see. If they dont..then i will buy blue locktite :)
Its input_shaper and I think it came from cnc world.. though im not sure. Its Klipper firmware measuring vibration with a gyro/accelerometer sensor and countering vibration with some voodoos
I'm an idiot, don't be surprised but: "Could you drill/grind parts of the rail and carbone tube to reduce its weight without affecting its rigidity too much?"
how big is the build volume :O do you have any files anywhere? for the printer! v-slots are even lighter if I'm right :o also where did you get that carbon fiber block it's so awesome
I think you are doing really interesting work here. Input shaping is a way that the 3dp community can make some big improvements, both to improving performance, but also in measuring performance objectively. I have a suggestion upon looking at your results: Please plot the before and after on top of each other. @6:33, you claim that you moved your resonance frequency from 45Hz to 60 Hz.The plot I see has two major resonances in both the before and after. Before 45 Hz, 2.1 amplitude and 70 Hz, 1.2 amplitue. After has 45 Hz, 2.4 amplitude, and 70 Hz, 3.5 amplitude. This seems worse to me. It's super easy to focus on what you want in frequency response plots. Looking forward to you further videos on this. Despite my quibbling with you data, I would have expected the larger carbon fiber tube to help, even if I can't see it in your data.
Hey Vez, I see you have some nice cable sleeves in you cable chains. Would you mind sending me the link wheee to buy them. Im currently trying a similar thing but my sleeves are to thick and are not that bendable. Thanks nice vid by the way.
Could you explain why are you using a carbon tube to stiffen up the X axis? Carbon has a negative thermal expansion coefficient in comparison to steel.
the X rail needs a support. Eventhough some builders will mount it directly in the air, its not a good idea if you plan to run at high speed/accel. It needs to be supported or it will spring a bit. So you have to use 2020 extrusion. OR, if you want to save weight, using carbon fiber. As for themal expansion, its not an issue here. No direct heat on the carbon tube and the enclosure will only get to 40C max when printing ABS. Im not sure why you compare that to steel. It should be compared to aluminum since it is what it is replacing.
@@Vez3D I am comparing it to steel as you are supporting steel with carbon fiber tube. The difference between 400mm of steel and alu is about 0,1mm when the temperature rises by about 20C. The difference between steel and carbon fiber may even be 0,4mm which is starting to be significant. This was visible with the E3D Tool changer, as they had problems with calibration when using CF beam as support under the rail.
@@Boogie3D see thats the kind of comment that I like on youtube :) we can share knowledge. I am far from being a mechanical engineer and I dont have all that knowledge and/or dont pretend to know it all.. Im just a hobbying who is trying to improve 3d printing, and having fun sharing with others. Thanks for your constructive feedback, really appreciate it.
I didnt weight them. But I dont use them anymore. On the square tube i have used m3 nuts. But not used ALL of the mounting holes. I dont think its needed to use them all.
@@Vez3D would a 430mm tube works? there is one on sale at this size for a pretty cheap price and free shipping for my country, raising to 440mm would make the price twice as big
@@cardosodouglas probably.. but one of the screw might be too much on the edge. But honestly..it sits so tight in the gantry that it could hold there with no screw. I paid my 500mm about 20$
@@Vez3D i'll give the 430mm a shot, you have have a few days ahead until i receive it and start to spam your videos asking for the files release. congrats again, what you did achieve is a piece of art.
@@Vez3D Cool and do you have any reference or the name of the seler ??? In this jungle of chinese copy it's very difficult without loose many money to found good copy. I've bought a real one in a shop and it cost a lot. Thanks for your help.
To get that any lighter you would probably have to skeletonize the mgn9 rail. Should be pretty easy to do on a drill press. Thats one very impressive gantry though, my WIP design is probably going to be horrible in comparison with all the stratasys baffles on it.
Drilling holes to reduce the weight was my first thought as well I have read in other posts and replys that the 9 and 12 have the same bearing size so you gain a bit by going down a size without and real cost but drilling holes would still be on my list
why not drill more holes in the 12mm rail? or simply larger holes. Also, wouldn't simply using larger motors be smarter? Enough beefy motors, and the weight would no longer really be an issue.
I have tried with 60mm 1.8. I didnt see improvements. These are e3d high torque 0.9 and it really improved quality. They are powerful enough. Even with high power motor, you still need to be light to remove vibration
Bigger engine doesn´t make you any faster around corners! It´s the reduction of weight which stops your tires from getting pushed out of the corner. The same goes for any acceleration.
Yeah the keyword is inertia. Yes with a beefier motor you can accelerate more mass in the same time. But this also means higher forces. Which means higher (reactive) forces against the rigid structure, which means more deflection and vibration. Just to recap the chain of consequences :)
why not do 2x of these rails, one for x axis and one for y axis, 2x belts and 2x motors per belt, and have the x+y axis move in an intersection of 2 rails like a + sign ? then you can prob drop the carbon fiber rod, and get smaller belt sizes aswell.
@@Vez3D mabye, I was just thinking of stiffening the setup and drop the carbon fiber rod and reduce the amount of belt, also it will give support x+y movement so you can increase acceleration in both x+y.
This is very cool printer you have built. Can you provide the stl files for version with carbon tubes? I have those lying around and would prefer using them instead of bying square tube. Thanks
Lmao i just did the opposite an dwent fromt eh single mgn9 ona extrusion to a floating mgn 12 but i am on a 400mm printer. I felt with dd their was more comlpiance on the 9 guides even high end ones
please review some vibration information in regards to modes before making these videos. you are misleading 1000s of people. i will break this down for you. ANYTIME you reduce stiffness you will reduce the peak in terms of hz. this is why you can put rubber dampeners on things and remove a vibration. ANYTIME you reduce weight and leave stiffness the same, you will increase the peak in terms of hz. Your y axis frequency went UP because you didnt just change 1 thing. you increased the stiffness in the Y AND decreased weight. Those 2 tubes were likely really stiff in the x but where NOT stiff in the Y. the bigger the diameter of a tube the stiffer it will be and those two small diameter tubes were not stiff. remember, stiffness of a round tube is proportionally to its radius to the 4th power. a tube that is twice the diameter will have 16x more stiffness. this is what you did with the square tubing. that square tubing is easily 50% larger in width. Your x axis went down likely due to the decrease in stiffness of the rail plus maybe a slight change from the two tubes to the square tube. those two round tubes were likely just as stiff in the x direction as the square tube. modes (the peaks you see) are tricky things to work with. i deal with them daily and work to actively design them out of machinery we make. the end all be all solution is stiffer and less weight. sometimes the best solution CAN be to increase weight and decrease stiffness to get the peak below were you are running at or to go faster so you are not operating in that area. for your case this would cause ghosting so its not an option. a helpful thing to review is the formulas for second area moment of inertia (stiffness of an object). en.wikipedia.org/wiki/List_of_second_moments_of_area Here is a rundown on stiffness and deflection. www.fictiv.com/articles/how-to-design-for-stiffness-using-a-geometric-approach Another thing you can look into is finding something with as high of a modulus of elasticity as you can. that is a materials springyness. the higher the better for you. your rails being carbon fiber means those are good but the plastics you use elsewhere are up for replacement. here is a list of materials you can sort by specific modulus( modulus per mass, higher the better) en.wikipedia.org/wiki/Specific_modulus
Very helpful comment 👍 thanks. I'm far from an expert and it's always good to have experts responding to my stuff. Glad you did. Thanks a lot. I'm just a hobbyists but with people surrounding me like my team and people like you...the learning is going on :) have a terrific my friend
Oh also.. when I started this yt channel..I was alone..but now I have a team with great knowledge on stuff I don't have.. which is awesome now. So we are stepping up our game a bit with time :) people are so awesome with me.
Bonjour de Sherbrooke, What a nice project! I was thinking of the next step in weight reduction. There are some carbon fibre linear rails out there. For example: www.igus.com/product/753. The down side is probably the fact that since these aren't ball bearing, the slider aren't likely to be preloaded and you could get some play that would result in a bit of rotation around the rail axis. But so long as the print head center of mass is centered on the rail.
If you would like to see a video on how to tune Input_shaper with the accelerometer please comment below. --- yes please!!!
ill try to make one sometimes soon
Question: How did you screw the rail onto the square carbon tube - how the screws are fixed inside?
I'd also want to know it
Print jig with hole in the middle for drilling.
@@OERYV I know how to drill, but what screws go into - carbon directly? I should get carbon and then fill it with plastic again (preferably with inserts) just to have something to screw in?
It is always good woking on projects when other people (e.g.GF) are sleeping. Most relaxing time ever when people are not bothering you. :D
Love your work and attention to detail! You are inspiring me to lighten up my gantry and start changing those acceleration values!!
Weight is the enemy! Thanks for sharing this, looking forward to the results. I'd be interested to see the accelerometer stuff, for sure!
I've considered drilling weight reduction holes in the sliding blocks on my delta's linear rails, but I couldn't find anybody who'd already tried it to give me enough confidence to go for it. It seems like there's way more metal in those blocks than the stresses of a printer could ever warrant, and pretty soon I'll have cut enough weight elsewhere that switching to titanium fasteners would actually mean something...
Hi, just for curiosity, why this will change somenthing on a delta? In delta the rail are fixed to the structure and doesn't move like in corexy or cartesian where one of the axis and the relative rail have to move (on corexy on x normally you have the x with related rail and head are moving connected to the y rail, while on cartesia is even worst since also the motor move with the gantry). On delta only the 3 rail block move on the rail but the rail is not moving.
@@SP_99999 The rail blocks are the things I'm talking about reducing in weight, not the rails they ride on. They contribute to the moving mass and at just over 50g apiece they're one of the biggest single-point sources of moving mass on my delta. 150g (the weight of all 3 together) might not seem like much on a cartesian system, but 150g is just about the same as my delta's entire effector with V6 and Nimble mounted. It's a different ballgame on the weight side for deltas.
All that said, I'm not sure how feasibly the linear bearing blocks could be cut down. They are VERY hard steel and I don't think I could put a dent in one with any tools I have at home.
@@jamespray yes I’ve misunderstood what you write and was thinking you refer to the rail :-)
@@SP_99999 No worries =D
Finally someone figured out there are numbers OTHER than 12. To be honest I'm surprised you didn't go straight to mgn7. As far as load capability even 7 is still over-kill but that's the smallest affordable size.
maybe I should have gone with 7. Maybe one day :) thanks for your comment
The possible problem with MGN7 is balls size, it can get noisy at high speed and wear faster. Afaik MGN9 is using the same balls as MGN12
@@VladOnEarth mgn7c has dynamic load capacity of 0.98kN, static 1.24kN, mgn7h is 1.37kN and 1.96kN respectively. So basically you can stand on it an it'll be fine. A 300g extruder and gantry? It doesn't even begin to feel it.
If you feel adventurous you can order some higher grade bearing balls and find a better fitting size and fill with those for even less noise and higher precision/loading capacity.
@@predragbalorda so if you are only using 2 rails for z would 12 or 15 be optimal. and for y would 12 or 9 be optimal. seems like 9 for x is good.
@@alexandereckhardt7153 the smaller - the lighter - the better - so mgn7 would be optimal actually.
Disclaimer first, I have absolutely no knowledge on klipper and Input_shaper.
When comparing those diagrams I noticed that the first y axis (power spectral density) is scaled differently. If you compare the two measurements for the (printers) y axis, the value of the first peak stays roughly the same while the value of the second peak nearly tripels. So while the bigger peak is now the second, higher frequent one, the first isn't really altered. Or do I understand it wrong and higher values are better in this case?
I noticed it too. I think its due to loosen belts. Ill do more testing
@@Vez3D was going to make this comment as well. The graphs are actually showing that you've *increased* the resonance by lowering the weight. This makes intuitive sense as well - when there is less mass, the resonance will increase in amplitude faster.
On the contrary, the data actually show here that the mass of the gantry was *not* the source of the resonance peaks you originally had (they are still there after the changes). This implies that the resonance is coming from another source. Probably: belts, frame, motors, or play in the rails. Since it looks like a corexy configuration, my guess is the long belts would be the first culprit to look into (as you've already stated). Try tightening, and see if there is any difference in frequency peak. If so, this could be further improved by upgrading to thicker belts. Seems like you're using 6mm belts? If so, try out 9mm?
If the peaks don't change from that, another possibility is that you're hitting the resonant frequency of the motors. If this is the case, larger motors or higher currents should help.
Great video! All the best.
Excellent explanation! I think you are spot on. I found that tightening belt more will just give way more amplitude and not change the frequency. So i think the motors are whats left to change here. Going with 60mm 0.9 degree
@@megamac125 would there be any downside to going up to 12mm belts or would there be marginal results that are not worth the extra cost?
@@alexandereckhardt7153 biggest downside is that they may be too strong - can cause motor to stall if over tightened, or pull things out of alignment. But if you're careful with them, no not really :)
“Shaved 3g.” Man, you’re like the Formula 1 driver of 3D printing 🚗 💨
“The devil is in the details.” Good stuff man. 💪🏻
Great Video. I switched to the lighter rails due to mass. Rail types have a max. speed and acceleration rating that so far has not been limiting factor.
I’m really impressed of the input shaper system.. I think I’ll switch to Klipper just to try it. I wonder if the carbon tube is needed in general, maybe it’s discussed somewhere, but isn’t the steel rail stiffer than the carbon? Nice channel by the way, just subbed:)
The carbon tube increases the rigidity by increasing the moment of inertia. Think of a ruler. On the thin axis it is flexible but on the larger axis it a lot more rigid even if the length and weight doesn't change. In the same way, the tube increases virtually the thickness of the rail, making it stiffer.
@@etle113 the steel alone is not stiff enough at high accel. It will "spring" a bit. It needs a stiff support. Either a 2020 extrusion (heavy) or a carbon square tube...latter is way better :)
@@Vez3D I´m sure you booth are right! I´ve never worked with carbon, so I probably underestimating it´s strength. Would be interesting to see a input shaper graph without it, to see the introduced resonance frequency.
Great work, well done. What software/hardware are you using to measure the resonant frequency? Can you record a dedicated video please?
I would suggest using inserts made for this purpose. You used inserts for injection molding.how do one get a setup like yours? Love the bracket without a 2020 profile.
you just ask me :) send me an email. As for inserts... interesting. I didnt know about that. what inserts do you suggest ?
@@Vez3D CNC kitchen has a video showing of some types. I prefer ruthex which also seem to be the strongest of them all. It has started to pop up some China clones of the ruthex. Voron use them for example.
What size square tubing are you using? 30x30x28mm? Have you tried molding the x/y/x carriage out of hollowed carbon fiber powered by servo motors? It's my plan for the next 3D printer/laser/cnc project. The cheapest servo motor I can find cost about $60-70 bucks a pop. OUCH.
Great video! how did you fix the rail to the carbon fiber? Did you tap the carbon fiber tube?
I added m3 nuts and lock washer inside. I printed a tool to reach the middle and used a pair of pliers really long (like surgon)..and on each end i printed small blocks with m3 nuts inserted ..so i could tight them to the Y carriage
@@Vez3D im interested to see how those lock washers do in terms of holding the fastener torque and whether they'll beat up the CF. my experience with them has always been terrible, wouldn't blue locktite be better for a low torque application?
@@Keechization i only had red locktite with me so I didnt use it. I dont know how washers will hold .. we will see. If they dont..then i will buy blue locktite :)
The video has the input shaper config lines commented out, did that get fixed?
Sure
What is that resonance analysis you did? Is that something I could apply to a CNC router table that I built?
Its input_shaper and I think it came from cnc world.. though im not sure. Its Klipper firmware measuring vibration with a gyro/accelerometer sensor and countering vibration with some voodoos
Where did you buy your rails from? are they cheapies from aliexpress or more expensive ones?
I would also love to know.
Hello nice video and explanation. What is your carbone square tube used please?
I'm an idiot, don't be surprised but:
"Could you drill/grind parts of the rail and carbone tube to reduce its weight without affecting its rigidity too much?"
This is why Ultimaker has the best core XY design. Not using a ton of belts and rails to move around.
huh may i ask how you realized you have too high of a tension? i'm suspicous i have that issue too
how big is the build volume :O do you have any files anywhere? for the printer!
v-slots are even lighter if I'm right :o
also where did you get that carbon fiber block it's so awesome
Is there a reason to use a rod and then mount a rail to it? Wouldn't using only a rail be even better?
May go carbon fiber or aluminium screws?
'tonight, while my girlfriend is sleeping' usually covers a large percentage of the time i spend researching something for projects... like now XD
Would ball screws be too much weight
have you noticed any downsides from switching to mgn 9. Like it the print head being able to deflect?
No its good :) same size ball bearing between mgn12 and mgn9
@@Vez3D The stiffness decrease with the mgn9 seems not to be a problem? i.e, The mgn12's were probably overkill to begin with?
@@ryanleslie7425 Im gong to MGN7 now ;)
I think you are doing really interesting work here. Input shaping is a way that the 3dp community can make some big improvements, both to improving performance, but also in measuring performance objectively. I have a suggestion upon looking at your results: Please plot the before and after on top of each other. @6:33, you claim that you moved your resonance frequency from 45Hz to 60 Hz.The plot I see has two major resonances in both the before and after. Before 45 Hz, 2.1 amplitude and 70 Hz, 1.2 amplitue. After has 45 Hz, 2.4 amplitude, and 70 Hz, 3.5 amplitude. This seems worse to me. It's super easy to focus on what you want in frequency response plots. Looking forward to you further videos on this. Despite my quibbling with you data, I would have expected the larger carbon fiber tube to help, even if I can't see it in your data.
I have played so much with belt tension that i cant we cant compare the 2 graphs here. My mistake. I should measure with same tension
Hey Vez, I see you have some nice cable sleeves in you cable chains. Would you mind sending me the link wheee to buy them. Im currently trying a similar thing but my sleeves are to thick and are not that bendable. Thanks nice vid by the way.
Where did you buy them? I had bad luck with cheap ones
What if you use the rail alone without any other gantry structure?
Could you explain why are you using a carbon tube to stiffen up the X axis? Carbon has a negative thermal expansion coefficient in comparison to steel.
the X rail needs a support. Eventhough some builders will mount it directly in the air, its not a good idea if you plan to run at high speed/accel. It needs to be supported or it will spring a bit. So you have to use 2020 extrusion. OR, if you want to save weight, using carbon fiber. As for themal expansion, its not an issue here. No direct heat on the carbon tube and the enclosure will only get to 40C max when printing ABS. Im not sure why you compare that to steel. It should be compared to aluminum since it is what it is replacing.
@@Vez3D I am comparing it to steel as you are supporting steel with carbon fiber tube. The difference between 400mm of steel and alu is about 0,1mm when the temperature rises by about 20C. The difference between steel and carbon fiber may even be 0,4mm which is starting to be significant. This was visible with the E3D Tool changer, as they had problems with calibration when using CF beam as support under the rail.
@@Boogie3D interesting. I will see how it goes.. but so far so good.
@@Boogie3D see thats the kind of comment that I like on youtube :) we can share knowledge. I am far from being a mechanical engineer and I dont have all that knowledge and/or dont pretend to know it all.. Im just a hobbying who is trying to improve 3d printing, and having fun sharing with others. Thanks for your constructive feedback, really appreciate it.
@@Vez3D I am glad to help, I have some education and practical knowledge in aerospace and robotics.
If you have any doubts, feel free to ask :)
Very nice and detailed video.
What is the total weight of the brass inserts? Maybe not much, but another linkage type could be lighter. what do you think?
I didnt weight them. But I dont use them anymore. On the square tube i have used m3 nuts. But not used ALL of the mounting holes. I dont think its needed to use them all.
@@Vez3D I was wondering this exact thing. Every other hole is fine. Are you missing ones in the middle? How many? That can make a huge difference.
How the screws are fixed inside carbon tube???
What did you print your parts out of, looks like polycarbonate carbon fiber. That is what I use for everything structural.
Good eyes !! Polycarbonate carbon
@@Vez3D what brand? im using priline carbon fiber pc, they sometimes have good prices on amazon.
@@Anonymousg64 priline
What's the material of the new support for CF tube?
There's lite weight pla might help for printer head
CNC Kitchen made a video about the effectiveness of brass inserts, threading into the plastic is actually almost as effective as using an insert.
True..but not if you unscrew them often
waiting anxiously for your files release. what's the square cf tube length? did you lose any print area?
440mm. Bought a 500mm and cut it. Not loosing anything in volume. Its the same lenght as the 2020 extrusion that I started with
@@Vez3D would a 430mm tube works? there is one on sale at this size for a pretty cheap price and free shipping for my country, raising to 440mm would make the price twice as big
@@cardosodouglas probably.. but one of the screw might be too much on the edge. But honestly..it sits so tight in the gantry that it could hold there with no screw. I paid my 500mm about 20$
@@Vez3D i'll give the 430mm a shot, you have have a few days ahead until i receive it and start to spam your videos asking for the files release.
congrats again, what you did achieve is a piece of art.
Just one question : do you use real mgn9 or chinese copy ???? and where you bought it ??? Thanks
Copy. Bought them from amazon
@@Vez3D Cool and do you have any reference or the name of the seler ??? In this jungle of chinese copy it's very difficult without loose many money to found good copy. I've bought a real one in a shop and it cost a lot. Thanks for your help.
@@bitosdelaplaya Usongshine is the seller
To get that any lighter you would probably have to skeletonize the mgn9 rail. Should be pretty easy to do on a drill press. Thats one very impressive gantry though, my WIP design is probably going to be horrible in comparison with all the stratasys baffles on it.
Drilling holes to reduce the weight was my first thought as well
I have read in other posts and replys that the 9 and 12 have the same bearing size so you gain a bit by going down a size without and real cost but drilling holes would still be on my list
The rails are hardened so not as easy as you think to lighten them
@@trp951 yeah I forget not everyone has access to a industrial drill press and carbide drill bits
why not drill more holes in the 12mm rail? or simply larger holes. Also, wouldn't simply using larger motors be smarter? Enough beefy motors, and the weight would no longer really be an issue.
I have tried with 60mm 1.8. I didnt see improvements. These are e3d high torque 0.9 and it really improved quality. They are powerful enough. Even with high power motor, you still need to be light to remove vibration
Bigger engine doesn´t make you any faster around corners! It´s the reduction of weight which stops your tires from getting pushed out of the corner. The same goes for any acceleration.
Yeah the keyword is inertia. Yes with a beefier motor you can accelerate more mass in the same time. But this also means higher forces. Which means higher (reactive) forces against the rigid structure, which means more deflection and vibration.
Just to recap the chain of consequences :)
Good idea!! I also wanted to know if you could send me the image on the sidebar of your mainsailos pls.
why not do 2x of these rails, one for x axis and one for y axis, 2x belts and 2x motors per belt, and have the x+y axis move in an intersection of 2 rails like a + sign ? then you can prob drop the carbon fiber rod, and get smaller belt sizes aswell.
What you describe looks like the K3 printer
@@Vez3D mabye, I was just thinking of stiffening the setup and drop the carbon fiber rod and reduce the amount of belt, also it will give support x+y movement so you can increase acceleration in both x+y.
Since your Resonance peak in X went down instead of up, probably your belt tension is a little lower this time
This is very cool printer you have built. Can you provide the stl files for version with carbon tubes? I have those lying around and would prefer using them instead of bying square tube. Thanks
Will do
what if you used the rail just buy itself without the carbon
Too springy
@@Vez3D have you noticed any downsides from switching to mgn 9. Like it the print head being able to diflect?
Now I wanna see some mega fast sub 9 minute benchy footage 😄
I think he can easily do sub 7 tho :)
No.1
making Y axis MGN9 too is also a better priority for you rather than playing watercooling
We have mgn9 Y option already
Have you thought about building a voron?
No
I have a voron already.. a 1.5
Where did you purchase the carbon fiber square tube?
Aliexpress
any rc hobby shop can help you with that
I want the results !!!
Patience :]
Aren't people going to mgn12 for a better printer now??
I dont see why mgn12 would be better. Its way heavier.
Lmao i just did the opposite an dwent fromt eh single mgn9 ona extrusion to a floating mgn 12 but i am on a 400mm printer. I felt with dd their was more comlpiance on the 9 guides even high end ones
How is your result with floating rail? I was finding it way too much springy
please review some vibration information in regards to modes before making these videos. you are misleading 1000s of people. i will break this down for you.
ANYTIME you reduce stiffness you will reduce the peak in terms of hz. this is why you can put rubber dampeners on things and remove a vibration.
ANYTIME you reduce weight and leave stiffness the same, you will increase the peak in terms of hz.
Your y axis frequency went UP because you didnt just change 1 thing. you increased the stiffness in the Y AND decreased weight. Those 2 tubes were likely really stiff in the x but where NOT stiff in the Y. the bigger the diameter of a tube the stiffer it will be and those two small diameter tubes were not stiff. remember, stiffness of a round tube is proportionally to its radius to the 4th power. a tube that is twice the diameter will have 16x more stiffness. this is what you did with the square tubing. that square tubing is easily 50% larger in width.
Your x axis went down likely due to the decrease in stiffness of the rail plus maybe a slight change from the two tubes to the square tube. those two round tubes were likely just as stiff in the x direction as the square tube.
modes (the peaks you see) are tricky things to work with. i deal with them daily and work to actively design them out of machinery we make. the end all be all solution is stiffer and less weight. sometimes the best solution CAN be to increase weight and decrease stiffness to get the peak below were you are running at or to go faster so you are not operating in that area. for your case this would cause ghosting so its not an option.
a helpful thing to review is the formulas for second area moment of inertia (stiffness of an object).
en.wikipedia.org/wiki/List_of_second_moments_of_area
Here is a rundown on stiffness and deflection.
www.fictiv.com/articles/how-to-design-for-stiffness-using-a-geometric-approach
Another thing you can look into is finding something with as high of a modulus of elasticity as you can. that is a materials springyness. the higher the better for you. your rails being carbon fiber means those are good but the plastics you use elsewhere are up for replacement. here is a list of materials you can sort by specific modulus( modulus per mass, higher the better)
en.wikipedia.org/wiki/Specific_modulus
Very helpful comment 👍 thanks. I'm far from an expert and it's always good to have experts responding to my stuff. Glad you did. Thanks a lot. I'm just a hobbyists but with people surrounding me like my team and people like you...the learning is going on :) have a terrific my friend
Oh also.. when I started this yt channel..I was alone..but now I have a team with great knowledge on stuff I don't have.. which is awesome now. So we are stepping up our game a bit with time :) people are so awesome with me.
Your girlfriend is asleep a lot. :)
Bonjour de Sherbrooke,
What a nice project! I was thinking of the next step in weight reduction. There are some carbon fibre linear rails out there. For example: www.igus.com/product/753. The down side is probably the fact that since these aren't ball bearing, the slider aren't likely to be preloaded and you could get some play that would result in a bit of rotation around the rail axis. But so long as the print head center of mass is centered on the rail.
Wonder what the cost is on those.