Something that a LOT of people forget when printing with a remote direct drive extruder as opposed to a bowden on a coreXY setup is that even though the weight increases lightly, the print speed actually also increases, especially with detailed objects. This is because of the inherent elasticity that bowden suffers from. at 200 mm/s corners for example will always turn a bit ugly due to the extruder lowering nozzle pressure at the wrong time (due to elasticity of the filament in the bowden tube). Slicers can compensate for it a little bit, but its a shitload of work to make it perfect (wasting also a lot of filament) and each time you switch to new materials you have to redo it. It takes so much time and effort that a remote direct drive is just far superior and elegant of a solution.
Just when I get my SecKit SK-Go to an acceptable standard something new comes along. But that is a good thing. The never ending quest for perfection continues.
Michael you made a great video here, covering all the right topics in just the right amount of detail. It was a pleasure to assist where you needed some help and to be fair that was pretty minimal. 20k views and counting! Thank you for having a look and great review.
@@michael90810 One of the key design objectives was to make it fully printable for all processes, for the purpose of allowing printerists everywhere to printer it themselves :) It holds up well if the parts are correctly printed and fully cured with most resins.
I only bought an MSLA printer when the resins became structural. Some of the parts have better mechanical properties than FDM parts because resin parts are isotropic and don't suffer from layer delamination.
THere are plenty of functional resin prints. In fact the cheapest way to print true engineering parts as a hobbyist is with resin at the moment (30% less strong than PEEK (But also 30% less expensive), same thermal properties). Reason? you dont need a 100k stratos printer but can just use any crappy SLA LCD printer you can find .
@Thu Nell Ⓥ - I've seen the salt remelting process and it's intriguing and a genuinely useful trick for FDM printing. It's a bit tedious and resin prints will always produce nicer surface finish and better resolution. I love that resin prints are now VERY structural.
Bowden still for the win. I've never had ringing ever. No artifacting ever, stringing is kina bad, but very easy to clean in post. Retraction took about a month and a kilo of pla to finally dial in, but once I did I'm getting near SLA quality out of a $260 bone stock ender3 pro. I learned from this very channel what to do.. so don't ask me
This video actually talked me out of getting a flex drive system. Not because I wasn't impressed, but like yourself it wouldn't make a huge improvement for my system. That said, your tip about putting a nut on a bolt before cutting it, made the video worth watching for me. Mind blown! ;-) Good stuff, and thanks for all the great content. One of the best channels on UA-cam.
If you consider an STL opensource sure. Stp files available for the mounts only but oddly no stp files for the actual extruder parts, with a license that allows for personal derivatives but no means to actually create them apart from laborious redraw from the ground up using the STL as a template. About as far from being open as you can get while still claiming opensource. little more than feel-good wankology.
Wow that is again a good review. There are many parameters coming together in a 3D printer. It is a good thing that you emphasize that again in this video, and yet still you are not afraid to try this very promising extruder out. Thank you for giving us a more and more complete understanding of 3D printer parameters and technology in general. You are truly teaching Tech and a true Tech teacher!
This looks like a game changing technology. I'm glad to see innovation like this, I'm grateful to those developing and popularizing the new technology and I look forward to a flex shaft drive on commercially available printers.
Just discovered your channel. I think it is the most informative 3D printing channel I have found. I am learning heaps. Thank you for spending the time to make these videos.
Another great detailed video. The real benefit I see is increased print speed while maintaining quality. For mass production this is certainly a plus. The limiting factors become how fast you can melt and cool the filament and of course the mechanics of the printer. For hobbyists, printing slower compensates for quality but we usually have the luxury of time (or a second (or third) printer)). Would like to see a revisit where you push the G5 and the printer to its limits using a volcano hot end and extreme part cooling.
One way it could be better, would be by using dual feed gears instead of a idler arm. I really feel that it helps prevent slipping, and having a more consistent extrusion.
You mentioned in passing Cetus ghosting. This drove me crazy until I made a very simple steel bracket for the Z axis stepper. It constrains the stepper to the printer base plate using the two lower threaded stepper motor mounting holes and a right angle bracket. This did required a redesign of the Z axis clamp/brake. Mine and several other brakes are available on thingiverse. Now I don't see any ghosting, however I usually print 0.1mm layers on fine print settings. Cheers -Peter
I've since gifted it to a family starting out in 3D printing who will not care in the slightest about the ringing, but thanks very much for sharing your solution.
Thank you for finally disclosing how you get your products, please keep this up in every video, with this sort of honesty I can finally get behind you! You really worried me with your last few videos.
For the record, the last few videos: Hemera on CR-10 Max: I purchased the Hemera myself and mentioned that in the original Hemera video. TFT35: Purchased one, paid for one. Originally omitted from description as an oversight. Creality LD002R - Disclosed free from Creality.
@@TeachingTech I understand, and Im not trying to start any trouble, but you really should say in every video how you get everything you feature, it only takes an extra 30 seconds to do it and there is no reason not to do it in the video. The description or past videos is no substitute for total transparency in the actual video, its the moral thing to do. Once again thank you for this! I know you probably hate me by now, but as I said previously I am saying these things because I think you make good content and I care about this issue. Look at all the bigger tech channels, they do exactly what I am trying to explain, and as you get bigger( and you will if you keep doing good work), you will only be put under the microscope more. More than once while watching your videos I have been left feeling uneasy, and I hate to say that, and it could have been avoided with an extra 30 seconds of video.
@Neil Siebenthal I think the guy who does it best is Lon TV, but he is far from the only one, even LTT does this and Linus is the biggest sell out on UA-cam. Im not knocking Linus by calling him a sell out, he makes fantastic content and I am pretty sure even he would call himself a sellout. Im just saying Dyson lol.
@Neil Siebenthal Once again Im not Knocking Linus, I love LTT. My whole point was he has 50 million subscribers, does a minimum of 2 commercials in pretty much every video, and well once again the Dyson video or the even roast. The man built an empire, you have to sell out a bit to do that, but he sold out in the best possible way. He makes some of the best tech content on UA-cam. If you have ever watched the Wan show or some of the more odd ball content he has done, you would know that he wouldnt have any problems with anything I have said about him, he would probably even back me up. The one thing I have never seen Linus do though is fail to disclose something in the video, even when he is being paid to do a video he states it right up front, he tells you in every video where he gets his content, he is 100% transparent. ua-cam.com/video/0AQFQMeOAig/v-deo.html
On an I3 style printer the weight of the direct drive motor won’t limit the speed, the moving bed will. No matter if you have the lightest possible hot end, you still won’t be able to go any faster than the bed can move without ringing
Cool idea. The thing about direct drive is that most systems are still lighter than what the Y axis has to throw around unless you have an Ender 5 design. You'll have ringing from the Y as speed/accel/jerk increase long before the X axis. At least this is the argument going around Facebook constantly.
Yep, my ender 3 is still bowden, and I use the glass bed, and honestly that glass bed is making my bowden setup kind of pointless. I'm limited by bed weight far more than hot-end weight. I plan to switch to DD at some point, although bowden is working out fine for me as long as I use linear advance, which I have to use to kill stringing on the micro-swiss.
Mike, you are awesome for doing all of this for your supporters. I think you do a great job. I hope you are ok. This no longer seems like a labor of love. I feel like you are suffering through these vids. Either way, thanks you and I hope that I'm a crazy person and don't know what I'm talking about.
Damn it Michael, yesterday evening I was up to deciding whether to get a Hemera or the G5 flex3Drive for my Ender 3. So I opted for the Hemera as TT has such a great video of it online (and because you DOSed the flex3drive website :). But this morning the race is open again. I hate making decisions... Thx!
Nice video and timely too because I was just looking at the Flex3drive today. You started by showing a truly awesome benchy and wondering whether the Flex3drive would make an improvement. That question never go answered. I hope that you can answer it in a follow up video snippet.
If you have the room and parts, you can put 2-3 nuts on the bolt, outer to snugged against each other at the cut end, then make your cut through the outer nut. You get a much cleaner cut, and less chance of messing up the shalf thread on the to-be-kept section. By snugging up two nuts against each other, you cN clamp the screw in a vise without the screw turning or needing to clamp on the threads or screw head. Don't overtighten the two nuts though, you'll stretch and weaken the screw/threaded rod.
Great little circuit! Great channel too! I'll try it out for sure, but would place it next to the main board, so I don't have to cut the wires and loose the product warranty for sure. Also asking me if it is possible to adapt the circuit so it ramps the fan speed instead of just turn it on and off. Additionally thinking if I even can hook it up to the already existing thermistor for temperature control to save additional wiring and a thermistor. No idea if that will work, but I'll give it a shot.
Thank you for this video and great hardware review! It was great to see this hardware in action, too bad the results shown were neglected by the cooling fan issue thus is hard to see any benefits by comparing it with the previous hardware.
It would be nice to have an evaluation of the flex in that cable. In the end if that is not minimized enough, it has the same inconsistent effect as a bowden tube. And also an evaluation of the top extrusion speed and its relationship to the gear ratio driving the final gear.
I researched numerous cables before buying my G5 kit for my E5+ (got the 800mm flex shaft for mounting motor outside the frame on the back right top corner and it is the perfect length). Any cheap cables are just that, and the seemingly good ones were outrageously expensive. For the money spent, the G5 kit is hands down the best bang for buck so far and this guy has been making and selling this style extruder for 8 years so it is a more evolved product. I initially printed my parts in structural PLA as well as regular SLA resin printed on an el cheapo Creality LDR-002, but I plan to reprint the case parts after the initial build is running and my FDM precision increases with carbon fiber ABS for strength and temperature handling (for possible enclosure heat). Trying to piece-meal source the bearings, flex shaft, hob gear, and screws would cost more and take much more time than the G5 kit, the full G5 ready unit even faster.
@Arpad Toth you just have to change the reverse E0 direction=true to false in the firmware for that. reversing the wiring can cause other problems that fry control boards & steppers.
@@teamlieblong I'm pretty sure he's referring to the flexible drive shaft being composed of steel wires arranged in 2 layers twisted in opposite directions, not electrical wires being crossed. flex3drive.com/product/driveshaft/
Could you do a review on a printer that prints very fine detail work and the best filament for fine printing. Not every person whats to print large objects. Most hobbies require small scale objects with high detail and strength. Speed of printing is not a issue its the quality that is important. Thanks
Will fix problems in the X axis but not so much the Y axis . What i can't understand (barring cost) is why the X and Y axis movement isn't done by the bed ? That way you could run bigger more powerful steppers and hot end on the z axis as it only moves up and down . Another issue is that with the X axis movement on the carriage the effect of inertia is increased as the Z axis moves upwards as the printer frame will flex more when the Z axis is higher . With the X and Y axis movement done by the bed any flexing can be controlled if not eliminated as it is down low and encased in a rigid frame . This is the basic principle of cnc milling machines which can machine parts at high speeds to tolerances that are better than any 3D printer ever will. Down side is the printer would be as wide as it is long or have a square base rather than rectangular as the typically do now and a bit heavier .
now you only need a remote cooling fan, to decrease the weight evenb further and increase the cooling.^^ I guess a long tube and a nice strong blower fan should be enough?
Or get a sturdy printer so the weight doesn't matter, industrial fdm printers have hotend's alone that already weight 2.5kg and the complete carriage over 10+ kilo's and prints over 120mm. the weight doesn't matter if your printer casing / base / frame is sturdy enough. And a lot of materials want slower print speeds so a light gantry so you can print "fast" doesnt even matter.
Are you going to do a proper review once you have your fan fixed? I am curious about this setup at high acceleration but the introduced flex seems like it could bring up different issues
hey, I'm ElWiz, the one mentioned in the very end. you can check my yt channel, I have some short vids with high(er) speed printing. I can push the print speed to 220mm/s with certain materials. PETG 140 - 180 with very minor ringing. it's prefectly fine for functional parts imo.
@@TeachingTech as many indicated to me when I was first pushing the limits one should always do the maths for his calibration models to figure out whether you even reach your top speed at a given acceleration. I also found acceleration having a much larger impact on print time, while the preset values are more than conservative. I'm curious to see your results!
Thanks for the very useful info. Helped me decide to try one. I will intend to modify it to use a dual-drive gearset once I've used it for a while. At 4:52, you mean SLA parts, right?
There is a water cooler available for the print head cooling, with the part fan, I don't know that a fan would be much heavier then the tubing and air spreader required for remote cooling.
I put a Zesty Nimble on my CR10 S5 and it did reduce the ringing, but finding a good mount point for the stepper motor was problematic, as was sorting out the mounts for the hot end and zesty (I had an older e3dV6) Zesty were quite helpful during the process. I'm not sure how much weight the flex cable adds to the head though, and resistance to movement vs a bowden or direct drive, it is hard to take into account as the flex cable flexes differently across the build surface and if you don't mount the stepper on the z-rail, then it also changes over the height of the print job, which means the forces on the print head also change.
If you see how a Flex3Drive E motor is mounted on CR10/Ender3 style machines you may see how simple elegant and correct it is. Don't use original E motor location, mount on the front of gantry instead of behind.
hey Michael, hope this finds you well! could you do a follow up on this please regarding print speeds? with the reduced weight its one of the benefits you might expect, but i've heard that due to the gear ratio it isn't possible. would love to know what the limits are in that regard.
After seeing your reviews of the SK-Go I ordered a SK-Go² got e 2 weeks ago and got it running. While assembly I was taking your idea with the PTFE tube to support the cables, I designed a bracket as part of a SK-Go² starter kit (www.thingiverse.com/thing:4567427) which can be mounted on the extruder carriage and on the frame at the back. This bracket has holes included to hold the PTFE firmly in place which makes it much easier to apply the cable spiral. It also works fine with cable sleeves (see pictures on thingiverse). I just ordered the Flex3drive G5 and when I will have it installed, I will share the experience. I am also looking in getting a Zesty Nimble since it is even lighter in design, but they have stopped accepting orders due to Covid 19 for the time being. Thanks again for your great tutorials and let me know if you like some of my own designs in the starter kit. Also just contact me if you need the step files for remodelling.
You neglected to weigh the drive cable in the moving parts. This is a serious oversight most people do when spec-ing the moving mass. I maintain the best quality will be achieved with a series of two standard extruders, the first with a large motor, and the second with a vastly smaller motor for fine control. This system will work with the MMU2 very well and is being prototyped by E3D
I get far less ringing on the Nimble setup than with the bond tech extruder and the ringing is no worse than the stock bowden setup. The cable weight should be taken into account, I agree, but if you are mounting on the arm that moves up, then the weight is constant (nearly) and can be reduced by mounting it correctly at the top of the arc. I would be interested in working out how much weight it adds, but it doesn't seem to cause any ringing vs a bowden setup, so the weight doesn't appear to be much of an issue, at least on i3 style printers I have tried the Nimble with.
You can't just weigh the cable because its weight is not solely supported by the moving print head. It's a factor, but no where near enough to overcome the 161 grams saved from this modification.
I'm aiming at a Flex3Drive G5, but unfortunately the Coronapocalypse has other plans :( And I'm not yet confident enough to design the mounting brackets. And I need to figure out how to mount the V6 on my ender5 also! If only the coronapocalypse actually afforded me more free time at home, but it seems to have done the opposite.
Informative Video thank you. Gizmo seems to work. Although at 16:15 those two Grey prints are rife with ringing ..despite claims. Wonder how long before Clones appear
@1:25 is the printer really jumpy like the vid shows, make sure the accel and jerk settings are all the exact same for x/y/z, it seems that maybe the z could be set to a different value, thats common with DIY marlin config for delta printers.
If I remember correctly, the Gear down ratio of the worm gear is 40:1, so while there is a little backslash in the drive cable, it is not measurable at the actual drive gear any more.
Potentially the worm drive could have lash but I agree with Ole in that the 40:1 ration will sort that out. The tolerances are also good which helps matters.
@@TeachingTech I have to admit, that I never checked the dimension of backlash. But only looking at the retraction parameters for direct drive (1-2mm) and bowden (4+mm), it looks like I can expect at least 1mm of extra backlash for bowden. I can see that the (potential) torsion spring behaviour is tackled by the 40:1 ration to reduce it to a tiny fraction of the 1mm. I haven't even thought about the worm drive lash. Technically the 40:1 doesn't apply to it (e.g. if the worm gear itself jumps up and down). Although I always have to think about lego worm drives, which hat serious backlash, I see that it can potentially be 0.5mm and below. Even if the flex drive type does not destroy bowden drive in case of PLA, I see the huge advantage for TPU. The question: Why isn't there a hemera like version offering an ultra short path between the extruder and nozzle, while keeping everything even lighter? Might it be the cost? Bowden is often a cost reduction, too. In that case the flex drive would add so much cost, that many companys probably would use direct drive, as it is less parts than flex drive?
You are all generally on the right track here and scratch the surface. Delve deeper and there is a lot one can do during design to avoid the issues, like using lower gearing to minimise that load, use a thicker shaft, and drive it in the right direction. Core features of all Flex3Drives and only found in Flex3Drives. I will be sharing all detailed info about this on the website to share so people can create their own designs and avoid the pitfalls.
I have an fl-sun QQ and have been chasing nozzle jams. the idea of remote drive extruder intrigues me I would be tempted to leave the drive where the extruder currently is (on the top) with just a small 20mm piece of PTFE to guide it through the arms to the extruder/printer head. I am not sure if this is a practical idea or not, IDEAS PLEASE.....
Oh, forgot, that recommended retraction is a bit high if linear advance is even close. I'm running k0.12 with retraction of 0.35mm. I was running 1.25mm with a bowden and L.A. tuned marginally.
Please remember I'm showing the recommended settings from Flex3drive. They don't recommend linear advance after setup. My retraction is lower like yours.
There are some weak aspects to the design of this thing. First, they should give you a proper metal coupler for the motor, that printed one just seems like it's asking to creep and come loose. Also, I don't like the idea of having toleranced tension arms instead of having a spring with a thumbscrew to adjust. Overall it seems like a design more well suited to a non coreXY printer, although on an i3, you'll run into the issue where your bed is gonna be way heavier than the hot end no matter what. I feel like on an i3 design you end up getting a better print quality increase just by ditching glass beds for PEI spring steel sheets.
Adjustable tension arms impact your effective e-steps with every filament change unless you get exact same position. Using a spring can also introduce variance in the same way.
@@3dprintingguru378 Why not just use a bigger motor? Since it isnt on the print head, you could use a much longer motor than a weak pancake stepper. In fact for every bowden printer I have, the motor I use on the extruder is way overpowered. On my ender with a bowden BMG and a 40-42 motor, I feel sure I could lift the printer by the extruder if the filament didn't snap first. Torque should only be an issue when you're trying to reduce weight.
@@Mutley3D That is a fair point I guess, but in practice the only time I've encountered that is on direct drive printers with underpowered pancake steppers. A bigger motor should be able to overcome any inconsistencies by its sheer torque alone. I tend to keep a spreadsheet with steps/mm and linear advance K for every filament/printer combo anyway.
There should be also said that reduced ringing due to bowden/flexdrive setup has most advantages on coreXY or delta printer. On i3 style printer the Y axis with movable bed is probably as heavy as direct drive setup on X axis. Fitting bowden/flexdrive on such printer will only eliminate ringing in X axis.
Even in core xy I sometime wonder if a remote direct drive will help a lot, If there is a linear rail supported by a steel traverse moved (like on Saphire S). I think a crossing rods design (ultimaker) maybe with Carbon rods would profit even more than most core xy. But definitly: For Delta (esp. big ones like Predator) it is most interesting. As there is a massive gear down ratio by the worm gear, the motor can be very smal anyway (moving fast, only 1/4 Microstepping, but more than 1000microsteps/mm). So how much would be the weight of a "normal" direct drive with such a smal Motor with massive gear down? I think on anything but maybe an Delta you would not have a difference; at least printing PLA and PET-G where cooling beiing the limiting factor.
How could you use this on Artillery printers? it would mean having the stepper cable running along the flex tube? Also we'd have to do a mount for the PCB but that's not hard. I also searched their website and could not find anything concerning Artillery printers.
Correct, no z wobble, as the colour changes as the speed increases. The actual geometry of the face has no wobble, if it did the period would be a lot smaller.
I would love to see how this compares to a lightweight direct extruder like the EZR with a pancake motor on something like the Ender 3. I have a Tronxy xy-2 pro and it is need of some upgrades.
The flex cable runs really high RPM to filament motion, this basically gears out any of the slack in the flex cable making it so minor as to not be an issue.
Forgot to mention that a U shaped mounting geometry for the drive cable creates a twisting force as the extruder moves. Typically less than 90 degrees but the twist causes a reduction in extrusion accuracy. Cheers -Peter
I think it's one of those things that is true in theory but whether the effect can be seen in the results is another matter. It will also depend a lot on the quality of the cable.
Yes this is correct and depending on mounting location and range of movement this can result in a difference that is lower than an equivalent of a singled skipped micro-step.
@@edwardpaulsen1074 the g5 is the 5th iteration of this extruder, the slow burn is in reference to the 5th iteration being released, not the product as a line. The flex3drive has been around since 2012
this seems gimmicky, the point of a direct drive is to get rid of the springiness of the bowden tube, the drive cable will have some springiness to it. a steel braided bowden tube would probably be on par with this and still be lighter.
The bowden tube adds a lot of friction to push that molten fillament over quite a large distance in a narrow tube and requires bigger motors to overcome that friction. They can't print a lot of flexible filaments for this reason too as it gums up the inside and wears out the tubes. This seems like a far more clever solution, get the weight off the direct drive for faster print speeds and rapid direction changes or stops like a boden but at the same time cut out the issues of pushing a viscous molten filament through a long tube. It just intuitively makes sense as a concept.
Stupid question perhaps, but why so many prints with a bad part cooling fan duct? Surely the very next print should be a new fan duct? Or do you mean the duct in the fan itself?
There was no point making a new duct that might not have fit over the new extruder, and I couldn't design one just for the new extruder without ruining the back to back test. I was really stuck for options.
At 16:18 what causes the vertical, repeating pattern visible on the inner walls? I definitely want to try this speed test on my E3 w/ titan clone. Seems the walls are long enough for 500mm/s2 acceleration to hit speed. Will be a good test of the microswiss' ability to melt the filament volume at 205C.
I honestly wonder if reducing weight on the SecKit SK-GO is worth it? Have you weighed those Aluminium mounting plates? Those are also part of the extruder/hotened assembly... And the entire linear rail is technically also part of the moving parts, only in one axis mind you but stil....
I wasn't willing to disassemble the lot to make that comparison, but I have net change of 161 grams with this setup. The stock setup is very good, however.
Perhaps if you worked a little bit more on your tuning and parameters you could eliminate that ringing from your X1 and the Cetus... Angus doesn't seem to have those problems as much as you do.
The X1 sure, but the Cetus is proprietary slicer with premade profiles so that one is on them. Plenty of others agree ringing is a weakness for that machine.
I know a core-XY (or similar) stands to gain the most from a low weight hotend, but for printers with an X axis hotend and Y axis bed (cr-10), will the lighter hotend mass have any significant affect on the Y motion induced ghosting? It seems gains would be limited by the bed mass.
The hot end doesn’t move in the Y direction (except high up as the gantry swings back and forth, unintentionally), so any Y direction ghosting is caused by bed mass and not hot end mass.
How many operations do you predict before the retaining arm stops locking in properly to hold the filament? I'd almost rather have it spring-loaded like a titan.
The tension arm design has been extensively tested both in-house and in users hands before release, and its operation has been solid. Its a clever little geometric arrangement in the way it locks down and prevents any reduction in tension over time.
The vibration from these speeds shook it loose, so I was going to add a fair bit of heft in fixing it properly. Now this extruder is installed I can design a nice new duct.
Is it possible to adapt it to an existing extruder assembly instead of using the provided extruder? I have a Titan Aero, I'd like to keep that as my extruder.
I'm new to this but is this better than the MDD direct drive? I just ordered a ender 3 pro and for mostly printing flexible prints. Should I upgrade to this or the MDD? Any help would be appreciated.
Doesn't he mean SLA parts not SLS parts? SLS stands for selective laser sintering. Where a high power laser melts fine powdered nylon (usually), but also sometimes other plastics and even metals.
I paid on flex3drive website a year ago. They never shipped it to me. Initiated a dispute on PayPal. A guy immediately called me and told me the package was going to be on the way very soon. I canceled the dispute. After a few months. Still no package. I found that one can only initiate a payment dispute once. A total hoax.
wish it was dual drive and had a nicer drive gear. those brass ones wear very badly with abrasives and in general. i'd like at least a metal one! edit: 15mm/s retractions?! ;-;
can you even still buy from this company? started a project building a variant of this using some off the shelf parts but i cant find a flex shaft cable for my life; ive tried so many times too purchase from their site but there just isnt any way too add an item too a cart or purchase & their contact page doesn't work xD
Zesty is the copy that stole from Flex3Drive. And the G5 is mechanically better as the zesty team made changes without the full engineering understanding
What are the speeds we are talking about when ringing generally shows up? (say ender 3 with direct drive mod) I don't usually care about print duration so I'm wondering if these print speeds are ones I'd never use.
Yes this is the initial release for a large body of work that is being made available. Everything still in early stages of release and licensing may change, crucial thing was to get files out there and freely available. With CC you cannot go backwards so its a top down approach on the licensing.
Exactly, not to mention no actual source available, just output files for the extruder itself. You can modify and innovate with the mounting brackets though! Woo hoo hooo smell the freedom! using the term "open source" as a nothing but a marketing ploy. With the old 'we will release the source later" bollocks that you see all the time, as if exporting a STP file as well as the STL is so vastly resource and time-consuming. Complete BS.
I would rather just strengthen the Y axis support (or get an over sized one), maybe get stronger stepper motors, top end bearings and be done with it vs going through this over engineered process, but thats just me...;)
Stronger steppers can make the ringing worse, and making the system more rigid can also make it worse, the kinetic energy has to go somewhere. Reducing the mass is the best way to reduce ringing.
Something that a LOT of people forget when printing with a remote direct drive extruder as opposed to a bowden on a coreXY setup is that even though the weight increases lightly, the print speed actually also increases, especially with detailed objects.
This is because of the inherent elasticity that bowden suffers from. at 200 mm/s corners for example will always turn a bit ugly due to the extruder lowering nozzle pressure at the wrong time (due to elasticity of the filament in the bowden tube). Slicers can compensate for it a little bit, but its a shitload of work to make it perfect (wasting also a lot of filament) and each time you switch to new materials you have to redo it. It takes so much time and effort that a remote direct drive is just far superior and elegant of a solution.
Just when I get my SecKit SK-Go to an acceptable standard something new comes along. But that is a good thing. The never ending quest for perfection continues.
It's a fun printer to work on, part of the reason I enjoy it so much.
Michael you made a great video here, covering all the right topics in just the right amount of detail. It was a pleasure to assist where you needed some help and to be fair that was pretty minimal. 20k views and counting! Thank you for having a look and great review.
Interesting to see a resin print being used for something functional for once.
Please keep us updated as to how it holds up.
@@michael90810 One of the key design objectives was to make it fully printable for all processes, for the purpose of allowing printerists everywhere to printer it themselves :) It holds up well if the parts are correctly printed and fully cured with most resins.
I only bought an MSLA printer when the resins became structural. Some of the parts have better mechanical properties than FDM parts because resin parts are isotropic and don't suffer from layer delamination.
Hi, do u have an external lcd? I’m revillete, I put it in my collection.
THere are plenty of functional resin prints. In fact the cheapest way to print true engineering parts as a hobbyist is with resin at the moment (30% less strong than PEEK (But also 30% less expensive), same thermal properties). Reason? you dont need a 100k stratos printer but can just use any crappy SLA LCD printer you can find .
@Thu Nell Ⓥ - I've seen the salt remelting process and it's intriguing and a genuinely useful trick for FDM printing. It's a bit tedious and resin prints will always produce nicer surface finish and better resolution. I love that resin prints are now VERY structural.
Bowden still for the win. I've never had ringing ever. No artifacting ever, stringing is kina bad, but very easy to clean in post. Retraction took about a month and a kilo of pla to finally dial in, but once I did I'm getting near SLA quality out of a $260 bone stock ender3 pro. I learned from this very channel what to do.. so don't ask me
This video actually talked me out of getting a flex drive system. Not because I wasn't impressed, but like yourself it wouldn't make a huge improvement for my system. That said, your tip about putting a nut on a bolt before cutting it, made the video worth watching for me. Mind blown! ;-) Good stuff, and thanks for all the great content. One of the best channels on UA-cam.
Flex3drive G5 is the best on the market and open source to, it is the original flexible drive extruder
@npgoalkeeper _ what drive cable do you mean?
If you consider an STL opensource sure. Stp files available for the mounts only but oddly no stp files for the actual extruder parts, with a license that allows for personal derivatives but no means to actually create them apart from laborious redraw from the ground up using the STL as a template. About as far from being open as you can get while still claiming opensource. little more than feel-good wankology.
Wow that is again a good review. There are many parameters coming together in a 3D printer. It is a good thing that you emphasize that again in this video, and yet still you are not afraid to try this very promising extruder out. Thank you for giving us a more and more complete understanding of 3D printer parameters and technology in general. You are truly teaching Tech and a true Tech teacher!
Thanks for the kind words. I would have liked to have done better with the testing but I might follow up in a later video.
This looks like a game changing technology. I'm glad to see innovation like this, I'm grateful to those developing and popularizing the new technology and I look forward to a flex shaft drive on commercially available printers.
That's quite a neat solution for CoreXY machines...
Just discovered your channel. I think it is the most informative 3D printing channel I have found. I am learning heaps. Thank you for spending the time to make these videos.
Another great detailed video. The real benefit I see is increased print speed while maintaining quality. For mass production this is certainly a plus. The limiting factors become how fast you can melt and cool the filament and of course the mechanics of the printer. For hobbyists, printing slower compensates for quality but we usually have the luxury of time (or a second (or third) printer)). Would like to see a revisit where you push the G5 and the printer to its limits using a volcano hot end and extreme part cooling.
Was looking for such a solution, thanks Michael!
One way it could be better, would be by using dual feed gears instead of a idler arm. I really feel that it helps prevent slipping, and having a more consistent extrusion.
Members have suggested this for the G6.
@@TeachingTech gonna be great. I'll get it then in that case possibly!
You mentioned in passing Cetus ghosting. This drove me crazy until I made a very simple steel bracket for the Z axis stepper. It constrains the stepper to the printer base plate using the two lower threaded stepper motor mounting holes and a right angle bracket. This did required a redesign of the Z axis clamp/brake. Mine and several other brakes are available on thingiverse. Now I don't see any ghosting, however I usually print 0.1mm layers on fine print settings.
Cheers
-Peter
I've since gifted it to a family starting out in 3D printing who will not care in the slightest about the ringing, but thanks very much for sharing your solution.
Thank you for finally disclosing how you get your products, please keep this up in every video, with this sort of honesty I can finally get behind you! You really worried me with your last few videos.
@Neil Siebenthal if you say so, but the last several videos have proven otherwise.
For the record, the last few videos:
Hemera on CR-10 Max: I purchased the Hemera myself and mentioned that in the original Hemera video.
TFT35: Purchased one, paid for one. Originally omitted from description as an oversight.
Creality LD002R - Disclosed free from Creality.
@@TeachingTech I understand, and Im not trying to start any trouble, but you really should say in every video how you get everything you feature, it only takes an extra 30 seconds to do it and there is no reason not to do it in the video. The description or past videos is no substitute for total transparency in the actual video, its the moral thing to do. Once again thank you for this!
I know you probably hate me by now, but as I said previously I am saying these things because I think you make good content and I care about this issue. Look at all the bigger tech channels, they do exactly what I am trying to explain, and as you get bigger( and you will if you keep doing good work), you will only be put under the microscope more.
More than once while watching your videos I have been left feeling uneasy, and I hate to say that, and it could have been avoided with an extra 30 seconds of video.
@Neil Siebenthal I think the guy who does it best is Lon TV, but he is far from the only one, even LTT does this and Linus is the biggest sell out on UA-cam. Im not knocking Linus by calling him a sell out, he makes fantastic content and I am pretty sure even he would call himself a sellout. Im just saying Dyson lol.
@Neil Siebenthal Once again Im not Knocking Linus, I love LTT. My whole point was he has 50 million subscribers, does a minimum of 2 commercials in pretty much every video, and well once again the Dyson video or the even roast.
The man built an empire, you have to sell out a bit to do that, but he sold out in the best possible way. He makes some of the best tech content on UA-cam. If you have ever watched the Wan show or some of the more odd ball content he has done, you would know that he wouldnt have any problems with anything I have said about him, he would probably even back me up.
The one thing I have never seen Linus do though is fail to disclose something in the video, even when he is being paid to do a video he states it right up front, he tells you in every video where he gets his content, he is 100% transparent.
ua-cam.com/video/0AQFQMeOAig/v-deo.html
On an I3 style printer the weight of the direct drive motor won’t limit the speed, the moving bed will. No matter if you have the lightest possible hot end, you still won’t be able to go any faster than the bed can move without ringing
Cool idea. The thing about direct drive is that most systems are still lighter than what the Y axis has to throw around unless you have an Ender 5 design. You'll have ringing from the Y as speed/accel/jerk increase long before the X axis. At least this is the argument going around Facebook constantly.
Yep, my ender 3 is still bowden, and I use the glass bed, and honestly that glass bed is making my bowden setup kind of pointless. I'm limited by bed weight far more than hot-end weight. I plan to switch to DD at some point, although bowden is working out fine for me as long as I use linear advance, which I have to use to kill stringing on the micro-swiss.
Mike, you are awesome for doing all of this for your supporters. I think you do a great job. I hope you are ok. This no longer seems like a labor of love. I feel like you are suffering through these vids. Either way, thanks you and I hope that I'm a crazy person and don't know what I'm talking about.
Thanks for the concern. I was really burner out at the end of 2019 but now only working one job has improved my quality of life massively.
@@TeachingTech I'm glad to hear it. Do what makes you happy amigo! I will be looking out for your next vid!
Damn it Michael, yesterday evening I was up to deciding whether to get a Hemera or the G5 flex3Drive for my Ender 3. So I opted for the Hemera as TT has such a great video of it online (and because you DOSed the flex3drive website :). But this morning the race is open again. I hate making decisions... Thx!
Given the low cost of the Flex3Drive you could probably get both, and then you'll have an excuse to build another printer.
Im so glad you re iewed this.not much about it online
Nice video and timely too because I was just looking at the Flex3drive today. You started by showing a truly awesome benchy and wondering whether the Flex3drive would make an improvement. That question never go answered. I hope that you can answer it in a follow up video snippet.
I'd love to know how to get a benchy looking that good!
Excellent video Michael. Good to see options in the marketplace. Cheers, JAYTEE
If you have the room and parts, you can put 2-3 nuts on the bolt, outer to snugged against each other at the cut end, then make your cut through the outer nut. You get a much cleaner cut, and less chance of messing up the shalf thread on the to-be-kept section. By snugging up two nuts against each other, you cN clamp the screw in a vise without the screw turning or needing to clamp on the threads or screw head. Don't overtighten the two nuts though, you'll stretch and weaken the screw/threaded rod.
You gave their server the hug of death!
Great little circuit! Great channel too! I'll try it out for sure, but would place it next to the main board, so I don't have to cut the wires and loose the product warranty for sure. Also asking me if it is possible to adapt the circuit so it ramps the fan speed instead of just turn it on and off. Additionally thinking if I even can hook it up to the already existing thermistor for temperature control to save additional wiring and a thermistor. No idea if that will work, but I'll give it a shot.
Thank you for this video and great hardware review! It was great to see this hardware in action, too bad the results shown were neglected by the cooling fan issue thus is hard to see any benefits by comparing it with the previous hardware.
I think your video got them a bit more traffic than they were prepared for :P site down resource limit reached!
Aye woops, working to resolve but its a good problem to have.
It would be nice to have an evaluation of the flex in that cable. In the end if that is not minimized enough, it has the same inconsistent effect as a bowden tube.
And also an evaluation of the top extrusion speed and its relationship to the gear ratio driving the final gear.
I researched numerous cables before buying my G5 kit for my E5+ (got the 800mm flex shaft for mounting motor outside the frame on the back right top corner and it is the perfect length). Any cheap cables are just that, and the seemingly good ones were outrageously expensive. For the money spent, the G5 kit is hands down the best bang for buck so far and this guy has been making and selling this style extruder for 8 years so it is a more evolved product. I initially printed my parts in structural PLA as well as regular SLA resin printed on an el cheapo Creality LDR-002, but I plan to reprint the case parts after the initial build is running and my FDM precision increases with carbon fiber ABS for strength and temperature handling (for possible enclosure heat). Trying to piece-meal source the bearings, flex shaft, hob gear, and screws would cost more and take much more time than the G5 kit, the full G5 ready unit even faster.
@Arpad Toth you just have to change the reverse E0 direction=true to false in the firmware for that. reversing the wiring can cause other problems that fry control boards & steppers.
@@teamlieblong I'm pretty sure he's referring to the flexible drive shaft being composed of steel wires arranged in 2 layers twisted in opposite directions, not electrical wires being crossed. flex3drive.com/product/driveshaft/
There have been a couple of flexible shaft drives, like Flex3drive.
the barebones kit is only $45! That's damn tempting!
Not $ but £ ! Quite a difference. And you have to add shipping.
@@oleurgast730 but bang for buck the value is undeniable given its performance and accessibility, which is one of the aims of the Flex3Drive project.
Could you do a review on a printer that prints very fine detail work and the best filament for fine printing. Not every person whats to print large objects. Most hobbies require small scale objects with high detail and strength. Speed of printing is not a issue its the quality that is important.
Thanks
Will fix problems in the X axis but not so much the Y axis . What i can't understand (barring cost) is why the X and Y axis movement isn't done by the bed ? That way you could run bigger more powerful steppers and hot end on the z axis as it only moves up and down .
Another issue is that with the X axis movement on the carriage the effect of inertia is increased as the Z axis moves upwards as the printer frame will flex more when the Z axis is higher . With the X and Y axis movement done by the bed any flexing can be controlled if not eliminated as it is down low and encased in a rigid frame . This is the basic principle of cnc milling machines which can machine parts at high speeds to tolerances that are better than any 3D printer ever will.
Down side is the printer would be as wide as it is long or have a square base rather than rectangular as the typically do now and a bit heavier .
now you only need a remote cooling fan, to decrease the weight evenb further and increase the cooling.^^ I guess a long tube and a nice strong blower fan should be enough?
Or get a sturdy printer so the weight doesn't matter, industrial fdm printers have hotend's alone that already weight 2.5kg and the complete carriage over 10+ kilo's and prints over 120mm. the weight doesn't matter if your printer casing / base / frame is sturdy enough. And a lot of materials want slower print speeds so a light gantry so you can print "fast" doesnt even matter.
Fair enough but this is a channel for enthusiasts who aren't in the market for the size and price of an industrial printer.
I might be nearly 50 years old, but I still giggled when you talked about wiping off the lubricant and slide the .... on to the shaft.
Great build
INteresting as ever
Thanks for sharing :-)
I'd like to see you give 3D optimizer a test when tuning slicer settings, im liking it quite a bit so far except for the startup code
I'd love to see this on a X1
Heh, website is overloaded ;) you killed their website! *grin*
Are you going to do a proper review once you have your fan fixed? I am curious about this setup at high acceleration but the introduced flex seems like it could bring up different issues
I'm thinking about a better test in future upping the acceleration rather than the speed on the tower. After the cooling is fixed of course.
hey, I'm ElWiz, the one mentioned in the very end. you can check my yt channel, I have some short vids with high(er) speed printing. I can push the print speed to 220mm/s with certain materials. PETG 140 - 180 with very minor ringing. it's prefectly fine for functional parts imo.
@@TeachingTech as many indicated to me when I was first pushing the limits one should always do the maths for his calibration models to figure out whether you even reach your top speed at a given acceleration. I also found acceleration having a much larger impact on print time, while the preset values are more than conservative.
I'm curious to see your results!
Thanks for the very useful info. Helped me decide to try one. I will intend to modify it to use a dual-drive gearset once I've used it for a while.
At 4:52, you mean SLA parts, right?
REMOTE COOLING would be very nice. with like an aquarium pump or such.
There is a water cooler available for the print head cooling, with the part fan, I don't know that a fan would be much heavier then the tubing and air spreader required for remote cooling.
@@originaltrilogy1 it wouldn't make much of a difference in weight but it cools betters apparently. would love to see it tested by him
I put a Zesty Nimble on my CR10 S5 and it did reduce the ringing, but finding a good mount point for the stepper motor was problematic, as was sorting out the mounts for the hot end and zesty (I had an older e3dV6) Zesty were quite helpful during the process. I'm not sure how much weight the flex cable adds to the head though, and resistance to movement vs a bowden or direct drive, it is hard to take into account as the flex cable flexes differently across the build surface and if you don't mount the stepper on the z-rail, then it also changes over the height of the print job, which means the forces on the print head also change.
Good points, thanks for sharing.
If you see how a Flex3Drive E motor is mounted on CR10/Ender3 style machines you may see how simple elegant and correct it is. Don't use original E motor location, mount on the front of gantry instead of behind.
@@Mutley3D On the S5 it doesn't have the same options for mounting as on the standard CR10
@@originaltrilogy1 send me an email or ill try to get a pic on the site if its no on there already. You really dont want motor on the back though.
hey Michael, hope this finds you well! could you do a follow up on this please regarding print speeds? with the reduced weight its one of the benefits you might expect, but i've heard that due to the gear ratio it isn't possible. would love to know what the limits are in that regard.
you don't like my printer? well you can SECKIT! 😄
I expect the crispness of retracts will be diminished by the backlash from the Flex3drive torsional drive system.
After seeing your reviews of the SK-Go I ordered a SK-Go² got e 2 weeks ago and got it running. While assembly I was taking your idea with the PTFE tube to support the cables, I designed a bracket as part of a SK-Go² starter kit (www.thingiverse.com/thing:4567427) which can be mounted on the extruder carriage and on the frame at the back. This bracket has holes included to hold the PTFE firmly in place which makes it much easier to apply the cable spiral. It also works fine with cable sleeves (see pictures on thingiverse).
I just ordered the Flex3drive G5 and when I will have it installed, I will share the experience.
I am also looking in getting a Zesty Nimble since it is even lighter in design, but they have stopped accepting orders due to Covid 19 for the time being.
Thanks again for your great tutorials and let me know if you like some of my own designs in the starter kit. Also just contact me if you need the step files for remodelling.
You neglected to weigh the drive cable in the moving parts. This is a serious oversight most people do when spec-ing the moving mass. I maintain the best quality will be achieved with a series of two standard extruders, the first with a large motor, and the second with a vastly smaller motor for fine control. This system will work with the MMU2 very well and is being prototyped by E3D
I get far less ringing on the Nimble setup than with the bond tech extruder and the ringing is no worse than the stock bowden setup. The cable weight should be taken into account, I agree, but if you are mounting on the arm that moves up, then the weight is constant (nearly) and can be reduced by mounting it correctly at the top of the arc. I would be interested in working out how much weight it adds, but it doesn't seem to cause any ringing vs a bowden setup, so the weight doesn't appear to be much of an issue, at least on i3 style printers I have tried the Nimble with.
You can't just weigh the cable because its weight is not solely supported by the moving print head. It's a factor, but no where near enough to overcome the 161 grams saved from this modification.
I'm aiming at a Flex3Drive G5, but unfortunately the Coronapocalypse has other plans :(
And I'm not yet confident enough to design the mounting brackets. And I need to figure out how to mount the V6 on my ender5 also!
If only the coronapocalypse actually afforded me more free time at home, but it seems to have done the opposite.
You should also try their nema 8 drive, instead of the flex drive. weighs a little more but should be quieter and be less fussy.
Teaching Tech
Informative Video thank you. Gizmo seems to work.
Although at 16:15 those two Grey prints are rife with ringing ..despite claims.
Wonder how long before Clones appear
That may be the infill showing because the walls are too thin.
@1:25 is the printer really jumpy like the vid shows, make sure the accel and jerk settings are all the exact same for x/y/z, it seems that maybe the z could be set to a different value, thats common with DIY marlin config for delta printers.
Doesn't that introduce the backlash by the flex drive instead of the bowden? Maybe I underestimate the flexdrive.
If I remember correctly, the Gear down ratio of the worm gear is 40:1, so while there is a little backslash in the drive cable, it is not measurable at the actual drive gear any more.
Potentially the worm drive could have lash but I agree with Ole in that the 40:1 ration will sort that out. The tolerances are also good which helps matters.
@@TeachingTech I have to admit, that I never checked the dimension of backlash. But only looking at the retraction parameters for direct drive (1-2mm) and bowden (4+mm), it looks like I can expect at least 1mm of extra backlash for bowden. I can see that the (potential) torsion spring behaviour is tackled by the 40:1 ration to reduce it to a tiny fraction of the 1mm. I haven't even thought about the worm drive lash. Technically the 40:1 doesn't apply to it (e.g. if the worm gear itself jumps up and down). Although I always have to think about lego worm drives, which hat serious backlash, I see that it can potentially be 0.5mm and below.
Even if the flex drive type does not destroy bowden drive in case of PLA, I see the huge advantage for TPU.
The question:
Why isn't there a hemera like version offering an ultra short path between the extruder and nozzle, while keeping everything even lighter?
Might it be the cost? Bowden is often a cost reduction, too. In that case the flex drive would add so much cost, that many companys probably would use direct drive, as it is less parts than flex drive?
You are all generally on the right track here and scratch the surface. Delve deeper and there is a lot one can do during design to avoid the issues, like using lower gearing to minimise that load, use a thicker shaft, and drive it in the right direction. Core features of all Flex3Drives and only found in Flex3Drives. I will be sharing all detailed info about this on the website to share so people can create their own designs and avoid the pitfalls.
I have an fl-sun QQ and have been chasing nozzle jams. the idea of remote drive extruder intrigues me I would be tempted to leave the drive where the extruder currently is (on the top) with just a small 20mm piece of PTFE to guide it through the arms to the extruder/printer head. I am not sure if this is a practical idea or not, IDEAS PLEASE.....
Oh, forgot, that recommended retraction is a bit high if linear advance is even close. I'm running k0.12 with retraction of 0.35mm. I was running 1.25mm with a bowden and L.A. tuned marginally.
Please remember I'm showing the recommended settings from Flex3drive. They don't recommend linear advance after setup. My retraction is lower like yours.
There are some weak aspects to the design of this thing. First, they should give you a proper metal coupler for the motor, that printed one just seems like it's asking to creep and come loose. Also, I don't like the idea of having toleranced tension arms instead of having a spring with a thumbscrew to adjust. Overall it seems like a design more well suited to a non coreXY printer, although on an i3, you'll run into the issue where your bed is gonna be way heavier than the hot end no matter what. I feel like on an i3 design you end up getting a better print quality increase just by ditching glass beds for PEI spring steel sheets.
Fair point on the non metal part, but personally I don't mind because I can alter the design or manufacture spares as I see fit.
Adjustable tension arms impact your effective e-steps with every filament change unless you get exact same position. Using a spring can also introduce variance in the same way.
the very low torque demand makes a metal coupler even worse. introducing rotational inertia is a bigger deal then rigidity at sub 2Ncm
@@3dprintingguru378 Why not just use a bigger motor? Since it isnt on the print head, you could use a much longer motor than a weak pancake stepper. In fact for every bowden printer I have, the motor I use on the extruder is way overpowered. On my ender with a bowden BMG and a 40-42 motor, I feel sure I could lift the printer by the extruder if the filament didn't snap first. Torque should only be an issue when you're trying to reduce weight.
@@Mutley3D That is a fair point I guess, but in practice the only time I've encountered that is on direct drive printers with underpowered pancake steppers. A bigger motor should be able to overcome any inconsistencies by its sheer torque alone. I tend to keep a spreadsheet with steps/mm and linear advance K for every filament/printer combo anyway.
There should be also said that reduced ringing due to bowden/flexdrive setup has most advantages on coreXY or delta printer. On i3 style printer the Y axis with movable bed is probably as heavy as direct drive setup on X axis. Fitting bowden/flexdrive on such printer will only eliminate ringing in X axis.
Even in core xy I sometime wonder if a remote direct drive will help a lot, If there is a linear rail supported by a steel traverse moved (like on Saphire S). I think a crossing rods design (ultimaker) maybe with Carbon rods would profit even more than most core xy.
But definitly: For Delta (esp. big ones like Predator) it is most interesting.
As there is a massive gear down ratio by the worm gear, the motor can be very smal anyway (moving fast, only 1/4 Microstepping, but more than 1000microsteps/mm). So how much would be the weight of a "normal" direct drive with such a smal Motor with massive gear down? I think on anything but maybe an Delta you would not have a difference; at least printing PLA and PET-G where cooling beiing the limiting factor.
Bondtech dual drive gear version?
How could you use this on Artillery printers? it would mean having the stepper cable running along the flex tube? Also we'd have to do a mount for the PCB but that's not hard. I also searched their website and could not find anything concerning Artillery printers.
Hm... I see a serious z-wobble strips on your printed parts.
That is just the speed ramping up. The print ramps the speed from 60-200 over X layers.
Correct, no z wobble, as the colour changes as the speed increases. The actual geometry of the face has no wobble, if it did the period would be a lot smaller.
Do mounting guide for Sidewinder X1 please!
I would love to see how this compares to a lightweight direct extruder like the EZR with a pancake motor on something like the Ender 3. I have a Tronxy xy-2 pro and it is need of some upgrades.
I "think" there will be some Flex #in the turning of the flex cable leading to some issues
The flex cable page on their website is worth a read to see the technical specs.
They run a stupid high gear reduction. Listen to his steps/mm
@@TeachingTech OK GREAT VIDEO added to my 3D Printer Playlist, keep up the good work
The flex cable runs really high RPM to filament motion, this basically gears out any of the slack in the flex cable making it so minor as to not be an issue.
@@legionof0ne441 I remember a high rev cable like these getting tied up in knots, dangerous
I did something similar with my Tlm. But you could have gotten something similar from NZ. Zesty nimble.
What fan shroud is TT using there? Is that a Hero Me? It looks similar but with small differences, but hard to tell with the white plastic?
Forgot to mention that a U shaped mounting geometry for the drive cable creates a twisting force as the extruder moves. Typically less than 90 degrees but the twist causes a reduction in extrusion accuracy.
Cheers
-Peter
I think it's one of those things that is true in theory but whether the effect can be seen in the results is another matter. It will also depend a lot on the quality of the cable.
Yes this is correct and depending on mounting location and range of movement this can result in a difference that is lower than an equivalent of a singled skipped micro-step.
how does this differ from the Zesty Nimble?
seems to be much cheaper.
This is the company zesty copied.
@Neil Siebenthal flex3drive has been around since 2012.... The nimble guys literally bought an early flex3drive to design their product
@@edwardpaulsen1074 the g5 is the 5th iteration of this extruder, the slow burn is in reference to the 5th iteration being released, not the product as a line. The flex3drive has been around since 2012
@Neil Siebenthal the 2017 website was selling the g3, still the 3rd iteration of the extruder.
this seems gimmicky, the point of a direct drive is to get rid of the springiness of the bowden tube, the drive cable will have some springiness to it. a steel braided bowden tube would probably be on par with this and still be lighter.
The bowden tube adds a lot of friction to push that molten fillament over quite a large distance in a narrow tube and requires bigger motors to overcome that friction. They can't print a lot of flexible filaments for this reason too as it gums up the inside and wears out the tubes. This seems like a far more clever solution, get the weight off the direct drive for faster print speeds and rapid direction changes or stops like a boden but at the same time cut out the issues of pushing a viscous molten filament through a long tube. It just intuitively makes sense as a concept.
Stupid question perhaps, but why so many prints with a bad part cooling fan duct? Surely the very next print should be a new fan duct? Or do you mean the duct in the fan itself?
There was no point making a new duct that might not have fit over the new extruder, and I couldn't design one just for the new extruder without ruining the back to back test. I was really stuck for options.
At 16:18 what causes the vertical, repeating pattern visible on the inner walls?
I definitely want to try this speed test on my E3 w/ titan clone. Seems the walls are long enough for 500mm/s2 acceleration to hit speed. Will be a good test of the microswiss' ability to melt the filament volume at 205C.
I honestly wonder if reducing weight on the SecKit SK-GO is worth it? Have you weighed those Aluminium mounting plates? Those are also part of the extruder/hotened assembly... And the entire linear rail is technically also part of the moving parts, only in one axis mind you but stil....
I wasn't willing to disassemble the lot to make that comparison, but I have net change of 161 grams with this setup. The stock setup is very good, however.
Idk how I feel about this. I print at 120mm/s and don't have any ringing on my direct drive and I doubt this will allow me to increase my speed.
Perhaps if you worked a little bit more on your tuning and parameters you could eliminate that ringing from your X1 and the Cetus... Angus doesn't seem to have those problems as much as you do.
The X1 sure, but the Cetus is proprietary slicer with premade profiles so that one is on them. Plenty of others agree ringing is a weakness for that machine.
flexshaft with shroud also weighs. Not a fair weight comparison :(
Do you feel this will hold up to using high volumetric output hotends such as super volcano or mosquito magnum?
I know a core-XY (or similar) stands to gain the most from a low weight hotend, but for printers with an X axis hotend and Y axis bed (cr-10), will the lighter hotend mass have any significant affect on the Y motion induced ghosting? It seems gains would be limited by the bed mass.
The hot end doesn’t move in the Y direction (except high up as the gantry swings back and forth, unintentionally), so any Y direction ghosting is caused by bed mass and not hot end mass.
Yes, a hot end having only one moving axis will not get the full benefit from this system, but my Delta certainly can.
Moving Thick Cable UP/down left and right mmm
Did you melt-in the brass inserts in the resin printed stuff too?
How many operations do you predict before the retaining arm stops locking in properly to hold the filament? I'd almost rather have it spring-loaded like a titan.
On the Zesty Nimble the breech is replaceable, for this reason.
Limited testing but so far zero deterioration. Since I made the parts myself I can replace if the need ever arises.
The tension arm design has been extensively tested both in-house and in users hands before release, and its operation has been solid. Its a clever little geometric arrangement in the way it locks down and prevents any reduction in tension over time.
there is no hook or latch that could wear down, the geometry of the idler bearing in relation to the hobbed gear makes it durable.
Why didn't you just patch your part cooling with some tape?
The vibration from these speeds shook it loose, so I was going to add a fair bit of heft in fixing it properly. Now this extruder is installed I can design a nice new duct.
Is it possible to adapt it to an existing extruder assembly instead of using the provided extruder? I have a Titan Aero, I'd like to keep that as my extruder.
It'll only benefit most cartesian style printers on the x axis as they're still throwing a heavy bed back and forth :\
Hi, Can you make a Video for this extruder kit but mounted in the Prusa Mini??
Also, can you test this remote direct extruder? zesty.tech/products/the-nimble-v2 this are very interesting options
I'm new to this but is this better than the MDD direct drive? I just ordered a ender 3 pro and for mostly printing flexible prints. Should I upgrade to this or the MDD? Any help would be appreciated.
Doesn't he mean SLA parts not SLS parts? SLS stands for selective laser sintering. Where a high power laser melts fine powdered nylon (usually), but also sometimes other plastics and even metals.
I wonder if this would be better on a non-coreXY printer.
I paid on flex3drive website a year ago. They never shipped it to me. Initiated a dispute on PayPal. A guy immediately called me and told me the package was going to be on the way very soon. I canceled the dispute. After a few months. Still no package. I found that one can only initiate a payment dispute once. A total hoax.
What printer would you reccomend for someone looking to get third 3d printer in the under 300 dollar range?
wish it was dual drive and had a nicer drive gear. those brass ones wear very badly with abrasives and in general. i'd like at least a metal one!
edit: 15mm/s retractions?! ;-;
can you even still buy from this company? started a project building a variant of this using some off the shelf parts but i cant find a flex shaft cable for my life; ive tried so many times too purchase from their site but there just isnt any way too add an item too a cart or purchase & their contact page doesn't work xD
waiting for dual gear + direct attached nema8/14 stepper
Whats your opinion?
G5vsZesty nimble and whats the difference?
Zesty is the copy that stole from Flex3Drive. And the G5 is mechanically better as the zesty team made changes without the full engineering understanding
What are the speeds we are talking about when ringing generally shows up? (say ender 3 with direct drive mod) I don't usually care about print duration so I'm wondering if these print speeds are ones I'd never use.
It says open source, but it is not. It's actually Creative Commons, Non Commercial, No Derivatives. Not very open if you can't share it.
Yes this is the initial release for a large body of work that is being made available. Everything still in early stages of release and licensing may change, crucial thing was to get files out there and freely available. With CC you cannot go backwards so its a top down approach on the licensing.
Like Mutley3d said, it's an initial effort and he's being very cautious. He's essentially a one man army.
Exactly, not to mention no actual source available, just output files for the extruder itself. You can modify and innovate with the mounting brackets though! Woo hoo hooo smell the freedom! using the term "open source" as a nothing but a marketing ploy. With the old 'we will release the source later" bollocks that you see all the time, as if exporting a STP file as well as the STL is so vastly resource and time-consuming. Complete BS.
How does this compare to the Zesty Nimble and would you recommend either over any other DD solution, if not what would be your solution of choice?
This is the 'father' of Zesty.. They use the same principle.
I would rather just strengthen the Y axis support (or get an over sized one), maybe get stronger stepper motors, top end bearings and be done with it vs going through this over engineered process, but thats just me...;)
Stronger steppers can make the ringing worse, and making the system more rigid can also make it worse, the kinetic energy has to go somewhere. Reducing the mass is the best way to reduce ringing.
Also this is a coreXY machine, not a moving y axis bed i3 style.
Good job. You crashed their website. XD
How does the software or printer know the total printtime ??
Please give this another shot with Ender 3