Michael - Thanks for the clear rundown! Given the various "weight-conscious" extruder systems you've tried (Orbiter, OmniaDrop, Zesty, etc), it would be great to see a recap video where you explain the relative advantages of each, and whether one sticks out for everyday use (or, conversely, whether any have a specific printing niche you think they fill best)!
One of the things that I think people forget about with the Nimble is that while you are reducing the mass at the hot-end, you're also ADDING a ton of floppy mass with that drive cable. That cable flopping around can add all sorts of weird resonance issues. I had a Nimble 2 installed on my E3 for a while and getting it tuned was a challenge. Eventually I did away with the cable entirely and designed my own adapter to directly mount a Nema 8 stepper motor onto it. That worked but even that was a compromise. The high gear ratio meant that the torque from the small motor was just enough to work but, of course, I ran into a lot of the same acceleration limitations that you ran into. In the end, I gave up on it and replaced the whole thing with a Hemera. (Yeah, I know it's not really cost effective but I like to tinker and I can afford it so what the hell, I thought I'd play with it.) It weighs a lot more but it's also a lot more reliable and it will push just about any filament I throw at it. If I were to do it today, I might go with one of the smaller Hemera inspired direct drive systems but all in all I'm quite happy with what I've got now.
Just what I suspected then! i've been thinking about that cable flapping in the wind, and i think it may be even worse than just the mass of a heavy motor. the motor is predictable, the cable keeps swinging.
Thanks for sharing your insights. I went for the shorter of the two cables I ordered to try and minimise this. I'm just above the minimum 20cm bend radius and so far it doesn't seem to be a factor.
@@TeachingTech Great to hear. I love the engineering behind these things. I look forward to seeing how it works out for you in the long run. Great video as always. Keep up the great work.
I wonder if you could solve this by having the drive cable rigidly mounted to one of the delta arms. Then it only flexes outside of the kinematics and just for a bit between the arm and the extruder.
@@caramelzappa I was wondering why the new motor wasn't mounted where the old bowden system was. Seemed like a less messy/ floppy design and you could, like you said, "cable manage it" to one of the arms.
I think the reason for this extraordinarily large gear ratio is that those Bowdens don't take the torque well since it causes torsion. That's probably also why the extruders acceleration isn't great.
I read a bit and found that reducing the number of microsteps might also help with the performance, since we don't need them with that gear ratio and it gives the driver some headroom for faster steps.
All flexible shafts run better at high speed with a low load. The high gear ratio at the output end achieves exactly that. At high speed with a large reduction at the output the cable is under very light load, I.E minimal torsional flexion. The poor acceleration in the video is due to the very poor choice to reuse a large boat anchor stepper from his bowden with high rotor inertia, at the high speeds the cable wants that makes a very large difference, large mass at high speed = lots of angular momentum. He effectively added a large flywheel, then wondered why it wouldn't slow down and speed up quickly..... What's more, this is something that is explicitly covered in the documentation.
I'd be very interested to see an actual engineering analysis showing that there really are "on-paper" paper benefits to this kind of setup. There are two main benefits being claimed: reduced weight and faster/less laggy filament response to stepper inputs. On the first, moving the stepper off of the extruder definitely reduces some mass, but gearing way down (as the Nimble V2 does) should allow a much lighter stepper than standard which would reduce the relative benefit, though potentially with the same speed related issues. And in addition, there's now an added mass in the steel flex cable that is lighter than a stepper, but harder to account for with input shaping since it's only lightly constrained and has its own hard to model dynamics. On the 2nd claim about faster filament response, I think it's easy to forget than the remote motor isn't magically turning the extruder with no lag in this setup. The flex cable itself tends to have a fair amount of springiness in torsion, and depending on the orientation of the cable at any given moment can have varying amounts of friction to overcome. This is exacerbated as the length of the cable increases. So, if the cable gets an input to extrude a half a mm, there will be some lag before the actual extrusion happens as the flex cable takes up that torsion, overcomes the friction in the guide tube, and then equalizes and the actual extrusion happens. I wonder how much actual benefit there is to this setup compared to a quality Bowden setup of similar length. If anyone knows of an analysis like this that's been done, please post it!
To the second point: i am not so sure whether the torsion is or is not a problem in that case. The gear ratio make it so that basically all torsion doesn't matter because for a half mm you'd basically need 10 turns, so that the quarter turn lag by torsion basically doesn't matter, especially since the torque on the flex cable itself is so little. I think this kind of solved this issue but i also think a regular stepper/-driver isn't made for this kind of gearing.
After calculating torsion, I would compare that to the "travel" (I can't think of a word that completely describes) of the filament within the bowden tube -- filament width / bowden internal circumference. Would it be a wash? This cable system adds a second gear system when mounting it to an existing stepper motor instead of transmitting the motion to gear once in the current setup. Also -- as we move to faster print speeds, does extrusion "clamping" distance [where a screw actually grabs the filament to push it forward] really play a role in pushing filament precision? That is, (I hope I'm explaining it right) does the distance of a direct drive to hotend (15-20mm) actually deliver better, nominal pressure as compared to a bowden distance (around 300mm) when printing with higher speeds? Oh, sorry. One more thing: weight on the hotend unit is only a concern with CoreXY and Deltas, yes? Bed slingers, especially the ones with dual Zed screws don't care about the weight of the hotend unit.
The gearing doesn't actually save you much, if anything, on the stepper. As speed goes up the available torque of the stepper goes down, so unless the limitation is the hot-end melt rate, you will need the similar size stepper. Now unless you are on a volcano, the melt rate likely is the limiting factor. I did test this on a volcano however and found that the gearing actually made the stepper the limiting factor for print speed. Unfortunately the only way to fix this is to increase voltage. A bigger stepper does not help as it just put you somewhere else on the same torque/speed curve. (more or less)
Good video As far as I know extruder acceleration is only for E-only moves. During printing it don't do anything.. Decell_to _accel is a speed limiter and not in any way or form altering the acceleration of any moves. A bad name for it.. An interesting alternative would be a re-enforced bowden with optimal placement( maybe even with a moving joint)
I love watching your videos because they always humble me when it comes to my 3d printing skills. I read the video title.... "Oh maybe I'll try that upgrade" after watching the video "wow that's way more than I am currently capable of... Back to printing flexi toys for the kids"
$130 product that has a significant design flaw that needs a thingiverse fix? Sorry zesty but you need to try harder than that. Also seems incredibly poor value to me, just my opinions of course.
Nice work by the way. I appreciate the idea of displacing the weight from the hot end, and being able to shorten the path between extruder and hot end. I find this design rather interesting. However, at that price I'd expect the gears to be made with a longer lasting material like aluminum, and not nylon plastic. Considering the size of these gears, the weight would still be relatively light. Instead of 22g with plastic gears, the whole thing might be around 24-30g, which in my opinion is still very reasonable.
The gears, either POM or nylon, are not under high load by any stretch of the imagination. A far better choice than metal for the task. Mine have been running for years with almost daily use, there is no discernible wear. Both POMs and nylons lubricity beats aluminium hands down, theyd require regular oiling for no benefit and extra weight. Brass is the only metal that comes close for this sort of application, both POM and nylon are much better choices.
I tried to switch to zesty in order to improve printing speed, I ended up having to reduce acceleration so much that at the end of the day it prints slower than heavy direct drive :(. I also had to print without pressure advance. Input shaping I managed to get to work somehow but pressure advance I could not make work. Compared to bowden setup, you are replacing spring in the bowden tube with the spring in the cable, only, on bowden tube if your connectors are good and you use stiff filament like PLA you have way less springiness to counter than with the cable. Flex3Drive switched to much thicker cables to fix this issue. The best setup I managed to get between regular direct drive and these hybrids is flex3drive+nema8, should be possible with zesty too, since you don't need a lot of torque from the motor with 30:1 (or 40:1) you can use miniature nema8 motor and run G8 or NimbleV3 as direct drive with motor mounted directly, nema8 is so light that it makes sense. With the nema8+g8 setup I have great acceleration (5000) with high speeds (200+) and both pressure advance and input shaping work perfectly. It should be simple to add nema8 to the zesty (I don't have the v3 so can't try myself)
Thank you for those valuable infos concerning Klipper. I've converted the 2nd extruder of my T-Rex 3 to 3mm filament with a Nimble V1. It save me a lot of space. It prints quite well, but it can do better.
check out the G5 also. I have both. the G5 and the Nimble V3. do your research and go check out the communities for each and talk to people that have your machine and a setup with a system like this.
I've used Titan's, Bondtech's, the new Orbiter and going to try the Sherpa mini soon, along with the Zesty nimble. While I love mines it did have the wavy issue and I got a free upgrade to it to V2. I have three Zesty's, two V1 and one V2. I've noticed one thing. I believe length makes a difference. My V1 had the 1400mm long cable, and I've noticed there was a delay from it to when the motor spins to when the extruder itself actually turns due to the immense length. That time delay makes the difference in retraction. Also I don't like how the gears are plastic too, the non adjustable tension is also a big no no for me. So since then, it's been the following for me: For bowden, Titan or Bondtech (glad to see Bondtech working on lighter stuff now). for direct, Orbiter or very likely the Sherpa mini that I'm getting soon.
so, what you are basically saying that it is still less precise extrusion than DD, and that like a bowden setup the length of the tube/shaft makes a difference in the extrusion accuracy correct? i think in that case it is basically only better for printing flexibles because the filament path is shorter, and stringing would be the same or close since the extruison accuracy is reduced. It is also not good for very fast printing since it is still added weight and the resistance of pulling the shaft along. IMO it's a niche product and is still not a universal solution since it does not have all the pros of dd or bowden but does carry over the cons of both. what a shame. The only use case that i can think of is printing flexibles and rigid materials semi-fast (200-300mm/s) but it is still not a perfect use case because you could make that happen with a good DD setup. I do however think that when the tech gets cheaper it will be very worth it since if it s much cheaper than a good DD setup it will just be a cheaper alternative to bowden with the ability to print flexibles reliably.
This was the biggest thing that I wanted to know about. When you make a long member transmit torsion, it twists. The longer the member, the bigger the angle it will twist before transmitting everything. This shouldn't be a problem if you are only twisting in one direction, but if you try adding retraction, I imagine that it would become very difficult to tune since you have to compensate for the bending twist in both directions.
@@alexanderunguez9633 Frankly for me retract was not the issue, issue was that pesky piece that puts tention and cannot be adjusted. It did not woirk for me and extruder made skips even on PLA
Completely the wrong stepper to have used, high rotor inertia = slow acceleration at the speeds required. The breech takes a little time to become flexible from new, a few prints, that's normal. The flex cable should be doing 90 degrees or less, unless you want to optimise for high friction, you have yours at 180 input to output. All of this is obvious stuff and is also all in the documentation. As for what you did to it to get those TPU results, well that's anyone's guess. It takes some tuning, it's not like other extruders. Mine runs fast and produces incredible results, 300mm/s is my default speed for standard prints and I'm still using a V1
Your experience is your experience and mine is mine. If Zesty sold an 'ideal' extruder stepper as part of the kit I would have got it. The flex cable position matches the images on their website and adheres to the minimum diameter of 20cm. I went through the documentation, that's for sure. I also highlighted in the video where the website claimed that printing soft flexibles was as simple as closing the breech and adjusting temperature and speed. The temperature matched what I used in the previous test shown, and the speed was lowered to 10mm/sec. I'm glad yours is working well but you don't need to try and discredit me when it should be clear I've followed the website and docs closely.
Yes, and while the drive cable IS "flexible", it is still much stiffer than something like a bowden tube and can put extra force on the carriage, so you have to make sure your motion system is stiff enough to handle it.
A while ago, I bought and installed the Nimble V2 on an old CR-10 I have and although the extruded works just fine, the print quality had decreased dramatically and I’ve been chasing improvements ever since :/
I installed V2 on ender 3 and it was a disaster - missing steps for any harder work. That tensioner was purely designed and they do not seem to fix it. Terrible product.
The v2 contained a flaw in the main gear which created artifacts in the pri ts. This was a well documented issue but they failed to send out replacements as they said they would.
Thank you for the well thought out review and tuning guide Michael! Now that we have a light weight extruder revolution with the LGX Light, Orbiter, HEVort, etc we may not need the remote DD anymore. I can't help but feel the compromises far outweigh the lowered mass. You proved in your tuning that it lacks the acceleration for faster printing, which is what these new direct drive extruders excel at. Lastly I wonder if you experienced any of the dreaded zebra banding from these remote DD like the Flex3Drive and Zesty.
I have recently replaced the extruder on my FLSun Q5 with a Orbiter v1.5. It took at few revisions to get right but I'm sure the quality hasn't been affected (if not improved) and the extra functionality will make it worth it. I've been asked why I didn't go flex tube and to be honest it was simply cost, Orbiter was £40 vs £140 for the Flex. I'm admittedly using that £100 for a Linear Rail install so its still going on the printer and maybe not needed if I had the flex.
As always your reviews are honest and extremely well done. You did miss the opportunity to question your use of a friggin' Delta in the first place. Signed, A former Rostock owner.
Hello. I do not see much advantage in transferring torque from a stepper motor through a rigid flexible hose with an iron cable, so that 3D printing is of high quality and fast, including on a delta printer. The hose is an added weight and resonant design. In my opinion, there is nothing better than a direct drive based on a Nema 14 light motor with a volcano, as well as Klipper firmware. This gives speed, accuracy, quality and high productivity, as well as reduced retracts and no snot. I designed such a print head on a delta printer, with a total weight of 300g, where it is easy and quick to refill plastic. The price of which is quite low. I print with a 0.6 nozzle at speeds up to 140mm/s, and accelerations up to 15000mm/s2, and plastic extrusion up to 30mm3/s. I am very pleased with the quality of the printed result. I also made a video about this.
The breech is definitely a weak point. Mine broke after using it only three times. I requested some help from Zesty but, contrary to your statement in the video, product support has not been very responsive. Three weeks and I'm still waiting for some feedback that they are going to replace this faulty part. The least they could do is let me know what they intend to do about it.
Honestly, this seems to be the best of both worlds, as stated. Still needs some work, but I do see it as the best alternative to bouden and direct drive.
I had originally designed a cartesian printer around the Nimble, then waited patiently for the product to arrive. Once it did I was disappointed in the fit and finish of the parts, and had severe problems with the end of the flex cable fraying and becoming unusable. I never got it to work, as there were no instructions regarding lubrication (which I did NOT do). It had severe stick-slip friction, which I later suspected could be due to lack of lube, but frankly I doubt it. As assembled it had no chance at operating. I also found the squeeze-to-engage pressure wheel to be quite unsure, and not adjustable. A failure to be sure, partly my own doing. I changed over to a more conventional direct drive (e3d titan), and never looked back. The Nimble parts are still in a bin somewhere, never to be used.
What type of top speed were you able to achieve with this extruder? Also what volumetric flows are achievable with this extruder? Light tool heads are nice, but if you are not able to get past 100mm/s, and or run at high acceleration (5k or higher) then what is really gained with this extruder?
He hasn't spent enough time with it yet, it's nowhere near tuned. The first issue is that boat anchor stepper he strapped to it. Mine can easily feed the high temp mosquito it's attached to, the hot end is the bottleneck.
I have a Nimble v1. I used it for a while and just don't see the benefit over bowden. It took dialing in the extruder steps/mm very precisely and lots of pulling the thing apart checking for alignment problems and re-lubricating it to get rid of diagonal banding (very common problem at the time, not sure about the newer versions). The print quality on the Nimble was never as good as a well tuned bowden (in my case, BMG + V6 on a tool changer and Hypercube Evolution). I have somewhat high standards for print quality, so I chucked the thing in a drawer and it hasn't seen any life since. Also, on the HEVO, I noticed that the stiffness of the drive cable would put enough force on the carriage that it would create a small amount of deflection. This is something to pay attention to if you have a not-so-stiff motion system. For filaments that need direct drive, I just use an Orbiter.
I've used the nimble v1 and flex3drive G5. Not impressed with either. The main advantage to reduced mass is faster printing at the same quality. The gearing of the remote direct drives makes it almost impossible to get the speed benefit. I experimented with a 15:1 overdrive gearbox on the stepper side and saw some improvement on that front, but eventually went on to the new wave of extremely light direct drive units like Orbiter and Sherpa Mini. I have an IDEX converted geeetech A10 with a pair of Sherpa minis, a voron v0.1 with a Sherpa Mini, and a custom corexy with an orbiter. The round nema14 direct drives are my new go-to.
I have both the v1.2 and v3. Both are a 30:1 gear ratio. I have found that for Travel speeds, Acceleration, etc. that I set my value between 120-250mm/s. I find anything like 1000 and up is just to fast. Also, at least, I don't recall you mentioning Jerk. This is another big issue. Zestytech States that one needs to work with the Jerk and Acceleration numbers plus vref to get the right smoothness of operation. Zesty recommends a Jerk setting of, I think, .6mm/s. I think a 3-6mm/s value for Jerk works better. They also recommend .3v for ref. I like .5-.533v. I ran my Zesty Nimble v1.2 on my Tevo Tornado so fast that I broke the cable. The cable is braided not solid. On my Tevo I could never get rid of a layer shift issue. other wise my ZN worked pretty well at 150mm/s. I now bought a Creality Ender 3 Max and I am trying to fine Tune my ZN and getting my part to stick to the bed. I am also printing a custom mount which is an integrated ZN mount on top of a copy off the stock fan shroud. I think that trying to force to much filament through the hot end for the first layer, trying to get my part to stick to the bed makes the ZN/hotend grind. Changing layer Width seems ok it is if I increase layer height that my Ender 3 Max starts to grind. And only grinds on the first bottom layer. Maybe by turning up the layer height the sets nimble/ extruder motor is just trying to push to hard. Not smooth enough. Over all for speed of a printer it is not just adjusting the Extruder that will allow for both higher yet smoother movement/speed. Zesty is right by saying first in tuning the Nimble get the nimble to run as smooth as possible by adjusting the Jerk, Acceleration, and Vref until you get the perfect combination. It like a dance. Adjust each value a little until finally everything comes together. I found my Tevo works so much differently than the Ender 3. So with the Ender 3 Max it is like learning all over again. So far I have been able to print one version of my fan shroud using CURA 4.12.0. It looked pretty good to me. I just could not get the walls in cure nor the first/bottom layer to prints solid. I get Walls in CURA to print as individual walls connected at the edges but no infill in the walls. Were my only issues. I am trying Superslicer now. Again I use 120-250 for my travel, acceleration moves for everything X,Y Z and E settings. I figure is X and Y are to fast then of course E is going to get under extruded. so I set these values in marlin before adjusting the Global setting for Speed.
Why not have 2 motors? One beffy thats ni inte om the chassi and do the grunt work. Then a super small one in the head to help the big one with the fine movements.
i think given the price point it's a bit too expensive but it could be a great somewhat mid tier extruder setup where you don't quite have a super solid coreXY machine so you can't achieve super high acceleration in the first place but also want somewhat faster than average printing speed
Seems to me that a BLDC would be more suitable at those gear ratios. Also the lack of grip indeed seems like an issue. Not an unfixable one; but not the kind of basics you are likely to be keen on reinventing after having paid for a premium product. The extruder id like to see put on the market is a hybrid one; one bowden-style nice and massive motor and gear putting pressure on the filament; and then a small and light motor and gear near the hotend. Ideally the first motor would be putting in most of the watts; and the second one would merely there to regulate the stick-slip and flex in the system. Better yet, if that second motor isnt a rotary motor, but a linear piezo-actuator deforming a flexible melt-chamber. Not sure if the engineering of that would work out; but I suspect it might, and that would give you lightning-fast retractions and super precise control.
I used a sherpa mini direct drive extruder on my delta 3d printer that I designed which also uses klipper. The sherpa mini is heavier but much cheaper. You can see my build video and playlist here czcams.com/video/cypxnTIcbQ8/video.html
The Bowden arrangement suffers from slop in the tube. I like the flex idea, but it also suffers from slop in the tube, especially during retraction. Anyone who's played with a flex shaft on their dremel can anticipate the problem this introduces. It would seem to make precision in extrusion and retraction impossible.
How many problems are caused by mediocre Bowden tubes that could be minimized by a tighter fitting more rigid tube and better fittings that prevent any end movement movement or slop? Do 2.85/3mm Bowden's have as many issues as 1.75mm setups?
Perhaps a NEMA 14 would be better suited in this application? These tend to perform better at the higher rotational speeds and acceleration required for higher gear ratio transmissions.
While I love your videos, the term you are saying as tolerances is clearences. While the tolerances of you may be high because you will only allow for so much slop the clearences between two parts is too tight. I'm working on getting the correct terminology for what I'm doing.
@Teaching Tech Nice review. I have a Nimble on my FLSun SR, but I replaced the whole original mount with a printed one to allow for a Nova hotend. The quality of the prints is fantastic in my case. There are only two problems with this kind of extruder in my opinion: 1) The single gear is not doing a good job when I use silk filaments (or other special filaments that are too hard or with different plastic properties) - they slip and I have air printing sooner or later. The breech can make a bit of difference (I tried the original and some printed ones) but inevitably I had air printing, eventually. Did you try silk filaments as well? I bet you would have the same issue. 2) With such a gargantuan gear ratio the retraction speed (and overall motor speed) is "limited", as the stepper revs required with a 30:1 ratio are much higher than with "standard" extruders (using the same printing speeds). This second problem can be easily fixed with a "speedbox". I am using one that multiplies the speed by 8, connected to the stepper, so I can still have decent speeds (>=200 mm/s and 90mm/s retractions) without issues and with any stepper, no matter the specs. As for the cable flopping, I did not have problems with it, but I do not go above 7K acceleration as with my setup that is the limit the input shaper is recommending. Maybe at >10k that floppy mass could be an issue, not sure.
I feel like it could be improved by having a gearbox of about 1:10, so you effectively print at a 3:1 ratio at the stepper itself, so it doesn't have to rotate as fast
Hmm ... I was actually considering the Nimble a while ago to gain benefits from Bowden and Direct Feed extruders. You demonstrated brilliantly something that made me hesistate, and I'm glad I chose Orbiter instead. The initial idea behind the Nimble is great and pretty logical, yet there are quite a few improvements to be done before it becomes fully capable. There are too much compromises for now. NB : the definition you give about "Direct Drive" is not really correct ( 1:35 ). What you show is Direct FEED combined with Direct Drive extruder, not just "Direct Drive". "Full Direct" could be a correct naming. Remember that both Direct Feed and Bowden extruder styles can be either Direct Drive (gears mounted next to the motor) or Remote Motor (using a flexible shaft between motor and extruder).
I might be wrong, but it seems to me that there is still a wide space to improove the surrounding hardware to benefit more of flex shaft designs. The lack of adjustable bridge is a defect of this specific design and not of flex shaft design, and inability to run fast is more a problem of some stepper motors than of the extruder itself. For adjusting pressure there is already the community mod. What about using some different stepper optimized for fast movments (or a brushless odrive) ?
Not that it’s likely to help, but can you turn down your micro steps with such a high gear ratio? It mainly seems to boost holding torque, and kipper probably isn’t limiting your step generation speeds, but having to reduce your acceleration is rough.
I like the idea, in theory, but in practice, it does not seem to provide top notch print quality or faster speed. So if it is neither faster nor better quality, what was gained here? it is quite expensive, significantly more complex than basic extruders, way more uncommon, so less support, help, mods, etc. So all in all, what is the point? Is it simply that the tech is promising, but not there yet? or is it a configuration/tweaking problem?
I always consider videos like this as an inside look into a product so viewers can make an informed decision on whether they want to buy. For me, convenience is one of the most important factors in using a 3D printer, so I'm happy enough to leave it on. You have evaluated that the proposition doesn't look good for your needs, which is a perfectly valid position.
@@TeachingTech Hello! i wasn't criticizing your video, quite the contrary, i had seen this product, and was interested, but now i am much less. It is not that i don't want this product at all, but i fail to see what it is that it brings compared to other products, what is its strong point in terms of printing performances. Or again, perhaps that is is simply not there YET. Hemera or mosquito, i get why one would buy it. the new 3 holes nozzle, i get it. The crazy expensive ruby nozzle, i get it. But here, i have a hard time understanding for who is this product. Anyways, thanks for the videos in general, you are one of the most informative chanels in 3DP for medium to advanced users out there! (Also probably good for beginers, but i watch your content like the current video more :) )
Thanks for sharing. I’ve been curious about this product since your previous video on it. Mainly interested to see how it reduces ghosting and ringing when printing at high speeds on a delta printer, but it looks like it couldnt even make it that far. Looks like I’ll stay away :-/
i have the v2 on my delta...it works, though i will say the prints are not as "clean" as they were on bowden...i have some very slight ripples between the layers. But it is light weight, and its direct drive. I dont think ill trust this product enough to buy the updated v3.
Hello, are you still using it? Is it good? You didn’t tell us what length of the shaft you have used and also the length of the Bowden pipe from the extruder to the hot end.. I’m thinking to buy it, I don’t have klipper so probably it would be easier to adapt it since I don’t have so many configuration and changes.. thanks in advance ☺️
You could probably use two different motors. Have one motor operating on the bowden tube end operating as normal, and have the gear on the extruder/hotend driven by a separate motor. Also, you should check out MirageC's content. He has one video in particular regarding servo motors, and another about fixing Z-wobble.
I tried this on my super racer and I hate it. It might be lighter than some DD setups, but what is the point if you have to slow the speed and acceleration to a crawl? I tried to mitigate using the speedbox setup that some others have used, but even with 1:3.5 pulleybox, none of the stepper motors I have can maintain a reasonable extrusion rate. Wish I had seen this video first because it highlights a lot of the downsides that in my opinion completely nullify any of the supposed benefits.
interesting that you're having flexible issues, i remember when the first nimble came out they showed it printing 60A TPU at a fair clip, weird that the gen3 is struggling
Not sure when this will be but I have a project that I can only make using it due to the size. I plan to print some upgrade parts for that video to supplement the project. Don't know if you saw but I used it in the first episode of the RatRig build.
All of the reasons I’d want to try and keep things lighter and have the added force of a DD would be for faster printing with higher volume and acceleration. The two things this system literally did not do.
I want to get a FLSUN SR but I have some interest in doing flexible materials. Would this addition allow this printer to be capable of printing even ninja?
It looks interesting, I think based on this review I'd wait and see if there's a "v2" version coming out. I'm activiely looking at options for a direct drive extruder for my Ender 3 Max but my budget is very limited; I'd like to print TPU and I'm not sure that the bowden tube arrangement I have will work reliably enough. I'm probably going to try the Madau3D direct drive adaptor first, as I don't think the extra mass will make a difference with a bed-slinger printer with such a large bed.
hi my friend, which 3d printer recommends me. I want to buy one that can handle regular jobs well. I had Anycobic mega zero and it was good sometimes and sometimes not good. so I want to buy a new one, which one do you recommend me?
I will be honest for $140 that is very expensive for 3D printed parts that you have to assemble yourself if it was $60 without shipping I would say its reasonable and if it were $80-$90 assembled with out shipping that's not to bad either but when you get to $140 and above I would start expecting the parts to be made of metal or aluminium
That would mean pulling off all of the parts from this video which is not on the cards. I'm considering a Sherpa Mini for another upcoming build though.
A very nice and thorough review as usual! The product itself has made me curious for the very principle of it; but in practise, it looks like a lot of trouble for sub-par performance. My Monoprice Delta Pro with bowden setup, can print a benchy faster and nicer looking. And I believe I could still improve that by some margin. It would be nice to hear how fsat your FLSUN usually spits out a benchy, and if you force it to run at the same speed as the Zesty Nimble setup, how would it look in comparison.
i use the Consolidated Tabs plugin, in addition to UI Customizer. I still don't get quite enough on my screen at once, thanks in part to the minimum sizes of some of the OctoPrint controls, but it's very usable.
Guess i'm sticking with bmg+dragon for now. Hope klipper has done wonders on your FLSun , i'm still working around my Anycubic Predator to get it setup but having issues understand documentation. Either way good review. Would be great to get more delta printer content as it feels lacking online compared to CoreXY and cartesian. PS : Also saw a project with a Dual stepper motor per axis on klipper (one of top of printer and the other on bottom for 6 steppers total). Any ideas on whether the improvements that would bring are worth the cost?
The next mod for this printer is pretty exciting for me and I hope for viewers too. With the multi steppers, are you talking delta specifically? So a pair of steppers for each corner running in parallel? I guess the extra grunt might be good for those doing speed Benchys, but it might be easier to use a higher voltage board and drivers like the Octopus Pro in that case.
@@TeachingTech Yep. Saw a mod with a pair of stepper on each corner running together. Think it had something to do with improving the speed and accuracy while also reducing the strain on the belt, yeah. I can search it and send u the source later, if you want to have a look at it. PS: the guy was using dual skr1.4T but with the boards we got now the Octopus Pro and similar boards save alot of hassle. Just think it should be an interesting mod with the klipper possibilities
Considering how you had to downtune the extruder in Klipper, I'm kind of wondering how this would go on your ratrig, considering how the whole idea about CoreXY is to not have a heavy X/Y motor on the gantry/bed to increase speed; which is contradictary to how everyone goes direct drive on them (though with pancakes with massive gearing), pressing absurd amounts of plastic through stuff like mosquito magnums with volcano's strapped to them. I highly doubt it'll be faster/lighter/etc; but doubt and curiousity are often a good combination.
Well I guess they are not going to send out replacements to people who bought v2 with the defective gears.... they swore they would last year on their discord but nothing...
Michael - Thanks for the clear rundown! Given the various "weight-conscious" extruder systems you've tried (Orbiter, OmniaDrop, Zesty, etc), it would be great to see a recap video where you explain the relative advantages of each, and whether one sticks out for everyday use (or, conversely, whether any have a specific printing niche you think they fill best)!
i think this is a great request as these are the real competitors for a nimble.
One of the things that I think people forget about with the Nimble is that while you are reducing the mass at the hot-end, you're also ADDING a ton of floppy mass with that drive cable. That cable flopping around can add all sorts of weird resonance issues.
I had a Nimble 2 installed on my E3 for a while and getting it tuned was a challenge. Eventually I did away with the cable entirely and designed my own adapter to directly mount a Nema 8 stepper motor onto it. That worked but even that was a compromise. The high gear ratio meant that the torque from the small motor was just enough to work but, of course, I ran into a lot of the same acceleration limitations that you ran into.
In the end, I gave up on it and replaced the whole thing with a Hemera. (Yeah, I know it's not really cost effective but I like to tinker and I can afford it so what the hell, I thought I'd play with it.) It weighs a lot more but it's also a lot more reliable and it will push just about any filament I throw at it. If I were to do it today, I might go with one of the smaller Hemera inspired direct drive systems but all in all I'm quite happy with what I've got now.
Just what I suspected then! i've been thinking about that cable flapping in the wind, and i think it may be even worse than just the mass of a heavy motor. the motor is predictable, the cable keeps swinging.
Thanks for sharing your insights. I went for the shorter of the two cables I ordered to try and minimise this. I'm just above the minimum 20cm bend radius and so far it doesn't seem to be a factor.
@@TeachingTech Great to hear. I love the engineering behind these things. I look forward to seeing how it works out for you in the long run. Great video as always. Keep up the great work.
I wonder if you could solve this by having the drive cable rigidly mounted to one of the delta arms. Then it only flexes outside of the kinematics and just for a bit between the arm and the extruder.
@@caramelzappa I was wondering why the new motor wasn't mounted where the old bowden system was. Seemed like a less messy/ floppy design and you could, like you said, "cable manage it" to one of the arms.
I think the reason for this extraordinarily large gear ratio is that those Bowdens don't take the torque well since it causes torsion. That's probably also why the extruders acceleration isn't great.
I read a bit and found that reducing the number of microsteps might also help with the performance, since we don't need them with that gear ratio and it gives the driver some headroom for faster steps.
I wonder if it would’ve been better to gear up the stepper before gearing it down at the extruder.
@@xenontesla122 I guess the loss to friction becomes an issue there but would be curious as well. 🤔
All flexible shafts run better at high speed with a low load. The high gear ratio at the output end achieves exactly that. At high speed with a large reduction at the output the cable is under very light load, I.E minimal torsional flexion. The poor acceleration in the video is due to the very poor choice to reuse a large boat anchor stepper from his bowden with high rotor inertia, at the high speeds the cable wants that makes a very large difference, large mass at high speed = lots of angular momentum. He effectively added a large flywheel, then wondered why it wouldn't slow down and speed up quickly..... What's more, this is something that is explicitly covered in the documentation.
I'd be very interested to see an actual engineering analysis showing that there really are "on-paper" paper benefits to this kind of setup. There are two main benefits being claimed: reduced weight and faster/less laggy filament response to stepper inputs.
On the first, moving the stepper off of the extruder definitely reduces some mass, but gearing way down (as the Nimble V2 does) should allow a much lighter stepper than standard which would reduce the relative benefit, though potentially with the same speed related issues. And in addition, there's now an added mass in the steel flex cable that is lighter than a stepper, but harder to account for with input shaping since it's only lightly constrained and has its own hard to model dynamics.
On the 2nd claim about faster filament response, I think it's easy to forget than the remote motor isn't magically turning the extruder with no lag in this setup. The flex cable itself tends to have a fair amount of springiness in torsion, and depending on the orientation of the cable at any given moment can have varying amounts of friction to overcome. This is exacerbated as the length of the cable increases. So, if the cable gets an input to extrude a half a mm, there will be some lag before the actual extrusion happens as the flex cable takes up that torsion, overcomes the friction in the guide tube, and then equalizes and the actual extrusion happens. I wonder how much actual benefit there is to this setup compared to a quality Bowden setup of similar length.
If anyone knows of an analysis like this that's been done, please post it!
To the second point: i am not so sure whether the torsion is or is not a problem in that case. The gear ratio make it so that basically all torsion doesn't matter because for a half mm you'd basically need 10 turns, so that the quarter turn lag by torsion basically doesn't matter, especially since the torque on the flex cable itself is so little. I think this kind of solved this issue but i also think a regular stepper/-driver isn't made for this kind of gearing.
After calculating torsion, I would compare that to the "travel" (I can't think of a word that completely describes) of the filament within the bowden tube -- filament width / bowden internal circumference. Would it be a wash? This cable system adds a second gear system when mounting it to an existing stepper motor instead of transmitting the motion to gear once in the current setup.
Also -- as we move to faster print speeds, does extrusion "clamping" distance [where a screw actually grabs the filament to push it forward] really play a role in pushing filament precision? That is, (I hope I'm explaining it right) does the distance of a direct drive to hotend (15-20mm) actually deliver better, nominal pressure as compared to a bowden distance (around 300mm) when printing with higher speeds?
Oh, sorry. One more thing: weight on the hotend unit is only a concern with CoreXY and Deltas, yes? Bed slingers, especially the ones with dual Zed screws don't care about the weight of the hotend unit.
The gearing doesn't actually save you much, if anything, on the stepper. As speed goes up the available torque of the stepper goes down, so unless the limitation is the hot-end melt rate, you will need the similar size stepper. Now unless you are on a volcano, the melt rate likely is the limiting factor. I did test this on a volcano however and found that the gearing actually made the stepper the limiting factor for print speed. Unfortunately the only way to fix this is to increase voltage. A bigger stepper does not help as it just put you somewhere else on the same torque/speed curve. (more or less)
Haven't tried the v3, but been using 2 v1s and 1 v2 for quite a while now. Very satisfied!
I'm so glad you're doing this on the Flsun SR. I just got mine and I'm learning it through experimentation and your videos.
Good video
As far as I know extruder acceleration is only for E-only moves. During printing it don't do anything..
Decell_to _accel is a speed limiter and not in any way or form altering the acceleration of any moves. A bad name for it..
An interesting alternative would be a re-enforced bowden with optimal placement( maybe even with a moving joint)
I love watching your videos because they always humble me when it comes to my 3d printing skills. I read the video title.... "Oh maybe I'll try that upgrade" after watching the video "wow that's way more than I am currently capable of... Back to printing flexi toys for the kids"
$130 product that has a significant design flaw that needs a thingiverse fix? Sorry zesty but you need to try harder than that. Also seems incredibly poor value to me, just my opinions of course.
Being all plastic it's doomed from the start.
I do like the concept of the Zesty, but it seems kind of pointless if you have to lower speeds..
Especially on a delta lol
Nice work by the way. I appreciate the idea of displacing the weight from the hot end, and being able to shorten the path between extruder and hot end. I find this design rather interesting. However, at that price I'd expect the gears to be made with a longer lasting material like aluminum, and not nylon plastic. Considering the size of these gears, the weight would still be relatively light. Instead of 22g with plastic gears, the whole thing might be around 24-30g, which in my opinion is still very reasonable.
The gears, either POM or nylon, are not under high load by any stretch of the imagination. A far better choice than metal for the task. Mine have been running for years with almost daily use, there is no discernible wear. Both POMs and nylons lubricity beats aluminium hands down, theyd require regular oiling for no benefit and extra weight. Brass is the only metal that comes close for this sort of application, both POM and nylon are much better choices.
I tried to switch to zesty in order to improve printing speed, I ended up having to reduce acceleration so much that at the end of the day it prints slower than heavy direct drive :(.
I also had to print without pressure advance. Input shaping I managed to get to work somehow but pressure advance I could not make work. Compared to bowden setup, you are replacing spring in the bowden tube with the spring in the cable, only, on bowden tube if your connectors are good and you use stiff filament like PLA you have way less springiness to counter than with the cable. Flex3Drive switched to much thicker cables to fix this issue. The best setup I managed to get between regular direct drive and these hybrids is flex3drive+nema8, should be possible with zesty too, since you don't need a lot of torque from the motor with 30:1 (or 40:1) you can use miniature nema8 motor and run G8 or NimbleV3 as direct drive with motor mounted directly, nema8 is so light that it makes sense. With the nema8+g8 setup I have great acceleration (5000) with high speeds (200+) and both pressure advance and input shaping work perfectly. It should be simple to add nema8 to the zesty (I don't have the v3 so can't try myself)
I've never had that issue with the breach on the older one, but it was red plastic and looked better made.
Thank you for those valuable infos concerning Klipper. I've converted the 2nd extruder of my T-Rex 3 to 3mm filament with a Nimble V1. It save me a lot of space. It prints quite well, but it can do better.
Perfect timing man! Was just going yo look for a direct drive tomorrow even though I hated the weight factor. I now know what I’m getting!!
check out the G5 also. I have both. the G5 and the Nimble V3. do your research and go check out the communities for each and talk to people that have your machine and a setup with a system like this.
I've used Titan's, Bondtech's, the new Orbiter and going to try the Sherpa mini soon, along with the Zesty nimble. While I love mines it did have the wavy issue and I got a free upgrade to it to V2. I have three Zesty's, two V1 and one V2. I've noticed one thing. I believe length makes a difference. My V1 had the 1400mm long cable, and I've noticed there was a delay from it to when the motor spins to when the extruder itself actually turns due to the immense length. That time delay makes the difference in retraction. Also I don't like how the gears are plastic too, the non adjustable tension is also a big no no for me. So since then, it's been the following for me:
For bowden, Titan or Bondtech (glad to see Bondtech working on lighter stuff now). for direct, Orbiter or very likely the Sherpa mini that I'm getting soon.
so, what you are basically saying that it is still less precise extrusion than DD, and that like a bowden setup the length of the tube/shaft makes a difference in the extrusion accuracy correct? i think in that case it is basically only better for printing flexibles because the filament path is shorter, and stringing would be the same or close since the extruison accuracy is reduced. It is also not good for very fast printing since it is still added weight and the resistance of pulling the shaft along. IMO it's a niche product and is still not a universal solution since it does not have all the pros of dd or bowden but does carry over the cons of both. what a shame. The only use case that i can think of is printing flexibles and rigid materials semi-fast (200-300mm/s) but it is still not a perfect use case because you could make that happen with a good DD setup. I do however think that when the tech gets cheaper it will be very worth it since if it s much cheaper than a good DD setup it will just be a cheaper alternative to bowden with the ability to print flexibles reliably.
@@WhiteG60 "Basically, before you order," or better do not order it
This was the biggest thing that I wanted to know about. When you make a long member transmit torsion, it twists. The longer the member, the bigger the angle it will twist before transmitting everything. This shouldn't be a problem if you are only twisting in one direction, but if you try adding retraction, I imagine that it would become very difficult to tune since you have to compensate for the bending twist in both directions.
@@alexanderunguez9633 Frankly for me retract was not the issue, issue was that pesky piece that puts tention and cannot be adjusted. It did not woirk for me and extruder made skips even on PLA
@@bumbarabun That stinks. Which cable length did you go with?
side note, wearing a Williams racing shirt is a nice touch. RIP to the legend
Hi Michael, I find that your honest and professional reviews are mandatory for anyone interested in up dating their 3D Printers . Thanks once again
Completely the wrong stepper to have used, high rotor inertia = slow acceleration at the speeds required. The breech takes a little time to become flexible from new, a few prints, that's normal. The flex cable should be doing 90 degrees or less, unless you want to optimise for high friction, you have yours at 180 input to output. All of this is obvious stuff and is also all in the documentation. As for what you did to it to get those TPU results, well that's anyone's guess. It takes some tuning, it's not like other extruders. Mine runs fast and produces incredible results, 300mm/s is my default speed for standard prints and I'm still using a V1
Your experience is your experience and mine is mine. If Zesty sold an 'ideal' extruder stepper as part of the kit I would have got it. The flex cable position matches the images on their website and adheres to the minimum diameter of 20cm. I went through the documentation, that's for sure. I also highlighted in the video where the website claimed that printing soft flexibles was as simple as closing the breech and adjusting temperature and speed. The temperature matched what I used in the previous test shown, and the speed was lowered to 10mm/sec. I'm glad yours is working well but you don't need to try and discredit me when it should be clear I've followed the website and docs closely.
Thanks, another nice review , haven't seen you vlogs for a long-time.
I feel like at least half of the weight of the drive cable needs to be included in the weight of the extruder. It's still having a move that cable.
Yes, and while the drive cable IS "flexible", it is still much stiffer than something like a bowden tube and can put extra force on the carriage, so you have to make sure your motion system is stiff enough to handle it.
A while ago, I bought and installed the Nimble V2 on an old CR-10 I have and although the extruded works just fine, the print quality had decreased dramatically and I’ve been chasing improvements ever since :/
I installed V2 on ender 3 and it was a disaster - missing steps for any harder work. That tensioner was purely designed and they do not seem to fix it. Terrible product.
The v2 contained a flaw in the main gear which created artifacts in the pri ts. This was a well documented issue but they failed to send out replacements as they said they would.
Thank you for trying! Really much appreciate it
But so many headaches for a possible failure
Ooh, speed test is live on the page? Perfect, I can tune the newly mounted volcano hotend on my X5SA :D
Neat! I had wondered why no one had built such a thing - I guess they had!
Thank you for the well thought out review and tuning guide Michael! Now that we have a light weight extruder revolution with the LGX Light, Orbiter, HEVort, etc we may not need the remote DD anymore. I can't help but feel the compromises far outweigh the lowered mass. You proved in your tuning that it lacks the acceleration for faster printing, which is what these new direct drive extruders excel at. Lastly I wonder if you experienced any of the dreaded zebra banding from these remote DD like the Flex3Drive and Zesty.
I have recently replaced the extruder on my FLSun Q5 with a Orbiter v1.5. It took at few revisions to get right but I'm sure the quality hasn't been affected (if not improved) and the extra functionality will make it worth it. I've been asked why I didn't go flex tube and to be honest it was simply cost, Orbiter was £40 vs £140 for the Flex. I'm admittedly using that £100 for a Linear Rail install so its still going on the printer and maybe not needed if I had the flex.
As always your reviews are honest and extremely well done. You did miss the opportunity to question your use of a friggin' Delta in the first place.
Signed,
A former Rostock owner.
Hello. I do not see much advantage in transferring torque from a stepper motor through a rigid flexible hose with an iron cable, so that 3D printing is of high quality and fast, including on a delta printer. The hose is an added weight and resonant design. In my opinion, there is nothing better than a direct drive based on a Nema 14 light motor with a volcano, as well as Klipper firmware. This gives speed, accuracy, quality and high productivity, as well as reduced retracts and no snot. I designed such a print head on a delta printer, with a total weight of 300g, where it is easy and quick to refill plastic. The price of which is quite low. I print with a 0.6 nozzle at speeds up to 140mm/s, and accelerations up to 15000mm/s2, and plastic extrusion up to 30mm3/s. I am very pleased with the quality of the printed result. I also made a video about this.
The breech is definitely a weak point. Mine broke after using it only three times. I requested some help from Zesty but, contrary to your statement in the video, product support has not been very responsive. Three weeks and I'm still waiting for some feedback that they are going to replace this faulty part. The least they could do is let me know what they intend to do about it.
Sorry to hear that. My statement was made based on lurking in the Discord. Hopefully they do the right thing.
Honestly, this seems to be the best of both worlds, as stated. Still needs some work, but I do see it as the best alternative to bouden and direct drive.
I had originally designed a cartesian printer around the Nimble, then waited patiently for the product to arrive. Once it did I was disappointed in the fit and finish of the parts, and had severe problems with the end of the flex cable fraying and becoming unusable. I never got it to work, as there were no instructions regarding lubrication (which I did NOT do). It had severe stick-slip friction, which I later suspected could be due to lack of lube, but frankly I doubt it. As assembled it had no chance at operating. I also found the squeeze-to-engage pressure wheel to be quite unsure, and not adjustable. A failure to be sure, partly my own doing. I changed over to a more conventional direct drive (e3d titan), and never looked back. The Nimble parts are still in a bin somewhere, never to be used.
Great review, with the good, bad and the ugly notes
Thanks for sharing :-)
What type of top speed were you able to achieve with this extruder? Also what volumetric flows are achievable with this extruder?
Light tool heads are nice, but if you are not able to get past 100mm/s, and or run at high acceleration (5k or higher) then what is really gained with this extruder?
He hasn't spent enough time with it yet, it's nowhere near tuned. The first issue is that boat anchor stepper he strapped to it. Mine can easily feed the high temp mosquito it's attached to, the hot end is the bottleneck.
Thanks for the video. For all the trouble you went to, and with questionable results, I'd say it's not worth hacking your machine for this.
"...so this added lightness can improve print quality." I see what you did there.
Thanks for the effort and the nice vid. Seems like a very _zest(y)less_ (effort+cost)/gain ratio - not so _nimble_ at all.
Reducing microsteps from 1/16 to 1/4 or even 1/2 should help with speed and acceleration
I'd love to see a video comparing the stock extruder design and the zesty against a flying extruder design for the delta printer.
I have a Nimble v1. I used it for a while and just don't see the benefit over bowden. It took dialing in the extruder steps/mm very precisely and lots of pulling the thing apart checking for alignment problems and re-lubricating it to get rid of diagonal banding (very common problem at the time, not sure about the newer versions). The print quality on the Nimble was never as good as a well tuned bowden (in my case, BMG + V6 on a tool changer and Hypercube Evolution). I have somewhat high standards for print quality, so I chucked the thing in a drawer and it hasn't seen any life since. Also, on the HEVO, I noticed that the stiffness of the drive cable would put enough force on the carriage that it would create a small amount of deflection. This is something to pay attention to if you have a not-so-stiff motion system. For filaments that need direct drive, I just use an Orbiter.
Dude nice shirt!
I've used the nimble v1 and flex3drive G5. Not impressed with either. The main advantage to reduced mass is faster printing at the same quality. The gearing of the remote direct drives makes it almost impossible to get the speed benefit. I experimented with a 15:1 overdrive gearbox on the stepper side and saw some improvement on that front, but eventually went on to the new wave of extremely light direct drive units like Orbiter and Sherpa Mini. I have an IDEX converted geeetech A10 with a pair of Sherpa minis, a voron v0.1 with a Sherpa Mini, and a custom corexy with an orbiter. The round nema14 direct drives are my new go-to.
I have both the v1.2 and v3. Both are a 30:1 gear ratio. I have found that for Travel speeds, Acceleration, etc. that I set my value between 120-250mm/s. I find anything like 1000 and up is just to fast. Also, at least, I don't recall you mentioning Jerk. This is another big issue. Zestytech States that one needs to work with the Jerk and Acceleration numbers plus vref to get the right smoothness of operation. Zesty recommends a Jerk setting of, I think, .6mm/s. I think a 3-6mm/s value for Jerk works better. They also recommend .3v for ref. I like .5-.533v. I ran my Zesty Nimble v1.2 on my Tevo Tornado so fast that I broke the cable. The cable is braided not solid. On my Tevo I could never get rid of a layer shift issue. other wise my ZN worked pretty well at 150mm/s. I now bought a Creality Ender 3 Max and I am trying to fine Tune my ZN and getting my part to stick to the bed. I am also printing a custom mount which is an integrated ZN mount on top of a copy off the stock fan shroud. I think that trying to force to much filament through the hot end for the first layer, trying to get my part to stick to the bed makes the ZN/hotend grind. Changing layer Width seems ok it is if I increase layer height that my Ender 3 Max starts to grind. And only grinds on the first bottom layer. Maybe by turning up the layer height the sets nimble/ extruder motor is just trying to push to hard. Not smooth enough. Over all for speed of a printer it is not just adjusting the Extruder that will allow for both higher yet smoother movement/speed. Zesty is right by saying first in tuning the Nimble get the nimble to run as smooth as possible by adjusting the Jerk, Acceleration, and Vref until you get the perfect combination. It like a dance. Adjust each value a little until finally everything comes together. I found my Tevo works so much differently than the Ender 3. So with the Ender 3 Max it is like learning all over again. So far I have been able to print one version of my fan shroud using CURA 4.12.0. It looked pretty good to me. I just could not get the walls in cure nor the first/bottom layer to prints solid. I get Walls in CURA to print as individual walls connected at the edges but no infill in the walls. Were my only issues. I am trying Superslicer now. Again I use 120-250 for my travel, acceleration moves for everything X,Y Z and E settings. I figure is X and Y are to fast then of course E is going to get under extruded. so I set these values in marlin before adjusting the Global setting for Speed.
Why not have 2 motors? One beffy thats ni inte om the chassi and do the grunt work. Then a super small one in the head to help the big one with the fine movements.
i think given the price point it's a bit too expensive but
it could be a great somewhat mid tier extruder setup where you don't quite have a super solid coreXY machine so you can't achieve super high acceleration in the first place but also want somewhat faster than average printing speed
Think the parts are HP MJF Nylon with black dye. I wonder if you could use one of these with a BLDC motor with better results?
Nice..and I interested how to reduce print with lass support pole.
Seems to me that a BLDC would be more suitable at those gear ratios. Also the lack of grip indeed seems like an issue. Not an unfixable one; but not the kind of basics you are likely to be keen on reinventing after having paid for a premium product.
The extruder id like to see put on the market is a hybrid one; one bowden-style nice and massive motor and gear putting pressure on the filament; and then a small and light motor and gear near the hotend. Ideally the first motor would be putting in most of the watts; and the second one would merely there to regulate the stick-slip and flex in the system.
Better yet, if that second motor isnt a rotary motor, but a linear piezo-actuator deforming a flexible melt-chamber. Not sure if the engineering of that would work out; but I suspect it might, and that would give you lightning-fast retractions and super precise control.
I used a sherpa mini direct drive extruder on my delta 3d printer that I designed which also uses klipper. The sherpa mini is heavier but much cheaper. You can see my build video and playlist here czcams.com/video/cypxnTIcbQ8/video.html
Honestly it's just too expensive for me.
The Bowden arrangement suffers from slop in the tube. I like the flex idea, but it also suffers from slop in the tube, especially during retraction. Anyone who's played with a flex shaft on their dremel can anticipate the problem this introduces. It would seem to make precision in extrusion and retraction impossible.
How many problems are caused by mediocre Bowden tubes that could be minimized by a tighter fitting more rigid tube and better fittings that prevent any end movement movement or slop? Do 2.85/3mm Bowden's have as many issues as 1.75mm setups?
I've never had a printer than used larger filament than 1.75mm.
The zesty nimble is actually printed on MJF.
Perhaps a NEMA 14 would be better suited in this application? These tend to perform better at the higher rotational speeds and acceleration required for higher gear ratio transmissions.
Yes indeed, or any low rotor mass stepper for that matter. I.E the opposite of this clumsy effort.
While I love your videos, the term you are saying as tolerances is clearences. While the tolerances of you may be high because you will only allow for so much slop the clearences between two parts is too tight. I'm working on getting the correct terminology for what I'm doing.
I hear you, and I probably should have said clearances. In my defence, the clearances will also be affected by manufacturing tolerances.
@@TeachingTech I appreciate your positive vibes, and I fully agree with you on this.
@Teaching Tech Nice review. I have a Nimble on my FLSun SR, but I replaced the whole original mount with a printed one to allow for a Nova hotend. The quality of the prints is fantastic in my case. There are only two problems with this kind of extruder in my opinion: 1) The single gear is not doing a good job when I use silk filaments (or other special filaments that are too hard or with different plastic properties) - they slip and I have air printing sooner or later. The breech can make a bit of difference (I tried the original and some printed ones) but inevitably I had air printing, eventually. Did you try silk filaments as well? I bet you would have the same issue. 2) With such a gargantuan gear ratio the retraction speed (and overall motor speed) is "limited", as the stepper revs required with a 30:1 ratio are much higher than with "standard" extruders (using the same printing speeds). This second problem can be easily fixed with a "speedbox". I am using one that multiplies the speed by 8, connected to the stepper, so I can still have decent speeds (>=200 mm/s and 90mm/s retractions) without issues and with any stepper, no matter the specs. As for the cable flopping, I did not have problems with it, but I do not go above 7K acceleration as with my setup that is the limit the input shaper is recommending. Maybe at >10k that floppy mass could be an issue, not sure.
I feel like it could be improved by having a gearbox of about 1:10, so you effectively print at a 3:1 ratio at the stepper itself, so it doesn't have to rotate as fast
amendment: the extruder consists of a feeder and a hot end
Hmm ... I was actually considering the Nimble a while ago to gain benefits from Bowden and Direct Feed extruders. You demonstrated brilliantly something that made me hesistate, and I'm glad I chose Orbiter instead.
The initial idea behind the Nimble is great and pretty logical, yet there are quite a few improvements to be done before it becomes fully capable. There are too much compromises for now.
NB : the definition you give about "Direct Drive" is not really correct ( 1:35 ). What you show is Direct FEED combined with Direct Drive extruder, not just "Direct Drive". "Full Direct" could be a correct naming. Remember that both Direct Feed and Bowden extruder styles can be either Direct Drive (gears mounted next to the motor) or Remote Motor (using a flexible shaft between motor and extruder).
I might be wrong, but it seems to me that there is still a wide space to improove the surrounding hardware to benefit more of flex shaft designs. The lack of adjustable bridge is a defect of this specific design and not of flex shaft design, and inability to run fast is more a problem of some stepper motors than of the extruder itself. For adjusting pressure there is already the community mod. What about using some different stepper optimized for fast movments (or a brushless odrive) ?
Not that it’s likely to help, but can you turn down your micro steps with such a high gear ratio? It mainly seems to boost holding torque, and kipper probably isn’t limiting your step generation speeds, but having to reduce your acceleration is rough.
I like the idea, in theory, but in practice, it does not seem to provide top notch print quality or faster speed.
So if it is neither faster nor better quality, what was gained here? it is quite expensive, significantly more complex than basic extruders,
way more uncommon, so less support, help, mods, etc.
So all in all, what is the point? Is it simply that the tech is promising, but not there yet?
or is it a configuration/tweaking problem?
I always consider videos like this as an inside look into a product so viewers can make an informed decision on whether they want to buy. For me, convenience is one of the most important factors in using a 3D printer, so I'm happy enough to leave it on. You have evaluated that the proposition doesn't look good for your needs, which is a perfectly valid position.
@@TeachingTech Hello! i wasn't criticizing your video, quite the contrary, i had seen this product, and was interested, but now i am much less. It is not that i don't want this product at all, but i fail to see what it is that it brings compared to other products, what is its strong point in terms of printing performances. Or again, perhaps that is is simply not there YET. Hemera or mosquito, i get why one would buy it. the new 3 holes nozzle, i get it. The crazy expensive ruby nozzle, i get it. But here, i have a hard time understanding for who is this product.
Anyways, thanks for the videos in general, you are one of the most informative chanels in 3DP for medium to advanced users out there! (Also probably good for beginers, but i watch your content like the current video more :) )
Thanks for sharing. I’ve been curious about this product since your previous video on it. Mainly interested to see how it reduces ghosting and ringing when printing at high speeds on a delta printer, but it looks like it couldnt even make it that far. Looks like I’ll stay away :-/
i have the v2 on my delta...it works, though i will say the prints are not as "clean" as they were on bowden...i have some very slight ripples between the layers. But it is light weight, and its direct drive. I dont think ill trust this product enough to buy the updated v3.
to avoid klipper being funky, stealthchop threshold should only ever be 0 or 999
A Nextion display? Haven't heard anyone use those in quite some time. They are my go-to for ESP3w projects.
The price is just insane, and you don't even get an assembled product.
It should cost half the price of a direct drive
Good video, think I'll pass on this until some more engineering is done.
Hello, are you still using it? Is it good?
You didn’t tell us what length of the shaft you have used and also the length of the Bowden pipe from the extruder to the hot end.. I’m thinking to buy it, I don’t have klipper so probably it would be easier to adapt it since I don’t have so many configuration and changes.. thanks in advance ☺️
You could probably use two different motors. Have one motor operating on the bowden tube end operating as normal, and have the gear on the extruder/hotend driven by a separate motor.
Also, you should check out MirageC's content. He has one video in particular regarding servo motors, and another about fixing Z-wobble.
Sounds like a starwars laser gun.
I tried this on my super racer and I hate it. It might be lighter than some DD setups, but what is the point if you have to slow the speed and acceleration to a crawl? I tried to mitigate using the speedbox setup that some others have used, but even with 1:3.5 pulleybox, none of the stepper motors I have can maintain a reasonable extrusion rate. Wish I had seen this video first because it highlights a lot of the downsides that in my opinion completely nullify any of the supposed benefits.
interesting that you're having flexible issues, i remember when the first nimble came out they showed it printing 60A TPU at a fair clip, weird that the gen3 is struggling
When will you do another video about the lowrider 2 cnc?
Not sure when this will be but I have a project that I can only make using it due to the size. I plan to print some upgrade parts for that video to supplement the project. Don't know if you saw but I used it in the first episode of the RatRig build.
All of the reasons I’d want to try and keep things lighter and have the added force of a DD would be for faster printing with higher volume and acceleration. The two things this system literally did not do.
I want to get a FLSUN SR but I have some interest in doing flexible materials. Would this addition allow this printer to be capable of printing even ninja?
It looks interesting, I think based on this review I'd wait and see if there's a "v2" version coming out. I'm activiely looking at options for a direct drive extruder for my Ender 3 Max but my budget is very limited; I'd like to print TPU and I'm not sure that the bowden tube arrangement I have will work reliably enough. I'm probably going to try the Madau3D direct drive adaptor first, as I don't think the extra mass will make a difference with a bed-slinger printer with such a large bed.
There's a Chick Hellebuck video about the extruder where with a different arrangement it can do tpu on an ender 3 just fine
Очень толково рассказал. Благодарю!
This product doesn't click for me. The breach design is a problem waiting to happen. Seems like the engineering only went so far, then they stopped.
Why would they use 30:1? Is it just to say you have finer control could they not of went with a more reasonable ratio
hi my friend, which 3d printer recommends me. I want to buy one that can handle regular jobs well. I had Anycobic mega zero and it was good sometimes and sometimes not good. so I want to buy a new one, which one do you recommend me?
If gear ratio is such a problem, why not set microstepping to 1?
I will be honest for $140 that is very expensive for 3D printed parts that you have to assemble yourself if it was $60 without shipping I would say its reasonable and if it were $80-$90 assembled with out shipping that's not to bad either but when you get to $140 and above I would start expecting the parts to be made of metal or aluminium
I would love to see an update about the Flex3drive G5!
Have you used it since the video you made about it?
czcams.com/video/qvY_dChxj0E/video.html
Hello, for this delta printer, how many accelerometer do you need?
Thanks
Is there a link to the mount he uses to connect the accelerometer to the FLSUN super racer?
Can you make a Tutorial how to "pimp" the FLSUN SR with the Mini Sherpa ?
That would mean pulling off all of the parts from this video which is not on the cards. I'm considering a Sherpa Mini for another upcoming build though.
@@TeachingTech Ok got it
A very nice and thorough review as usual! The product itself has made me curious for the very principle of it; but in practise, it looks like a lot of trouble for sub-par performance. My Monoprice Delta Pro with bowden setup, can print a benchy faster and nicer looking. And I believe I could still improve that by some margin. It would be nice to hear how fsat your FLSUN usually spits out a benchy, and if you force it to run at the same speed as the Zesty Nimble setup, how would it look in comparison.
11:27 how do you have the terminal besides the temperature tab? Do you have a video on your octoprint plugins?
i use the Consolidated Tabs plugin, in addition to UI Customizer. I still don't get quite enough on my screen at once, thanks in part to the minimum sizes of some of the OctoPrint controls, but it's very usable.
Do you have a link to the Zesty Nimble Discord?
The link was on the receipt once the kit came. If it's not publicly available then it's not my place to post it in public.
Guess i'm sticking with bmg+dragon for now. Hope klipper has done wonders on your FLSun , i'm still working around my Anycubic Predator to get it setup but having issues understand documentation.
Either way good review. Would be great to get more delta printer content as it feels lacking online compared to CoreXY and cartesian.
PS :
Also saw a project with a Dual stepper motor per axis on klipper (one of top of printer and the other on bottom for 6 steppers total). Any ideas on whether the improvements that would bring are worth the cost?
The next mod for this printer is pretty exciting for me and I hope for viewers too.
With the multi steppers, are you talking delta specifically? So a pair of steppers for each corner running in parallel? I guess the extra grunt might be good for those doing speed Benchys, but it might be easier to use a higher voltage board and drivers like the Octopus Pro in that case.
@@TeachingTech Yep. Saw a mod with a pair of stepper on each corner running together. Think it had something to do with improving the speed and accuracy while also reducing the strain on the belt, yeah. I can search it and send u the source later, if you want to have a look at it.
PS: the guy was using dual skr1.4T but with the boards we got now the Octopus Pro and similar boards save alot of hassle. Just think it should be an interesting mod with the klipper possibilities
Mine never worked good, lots of grinding noise... Such a money waste
Probably heavier than the bowden setup with that cable weight added in..
Considering how you had to downtune the extruder in Klipper, I'm kind of wondering how this would go on your ratrig, considering how the whole idea about CoreXY is to not have a heavy X/Y motor on the gantry/bed to increase speed; which is contradictary to how everyone goes direct drive on them (though with pancakes with massive gearing), pressing absurd amounts of plastic through stuff like mosquito magnums with volcano's strapped to them.
I highly doubt it'll be faster/lighter/etc; but doubt and curiousity are often a good combination.
LGX lite all the way!
That noise looks like a very frown R2D2...
What length of flex did you use?
Well I guess they are not going to send out replacements to people who bought v2 with the defective gears.... they swore they would last year on their discord but nothing...
Who even sells this thing I've been looking and there isn't any available.
Amazing product
Amazing simrig