When i clicked on this video i was thinking that this was just another printer trying to share a spot on the market with a bunch of other printers. I now see that this machine has its own place in the market and can fullfill a few different valuable roles.
I bought this a month ago after watching your vids. Some problems with my interpretation of the instructions, but all working, and the support from Dan and the community on Discord is second to none. - This thing is also a beast for ABS printing.
I run a Vcore 3.1 500 mm IDEX beta machine for 8 months now. It works great, software and mechanics. This printer takes another spin at the motion system which will limit speed and acceleration versus the Vcore quite a lot. Toolchanges on a Vcore are only a few tenths of a second (50k accel at 1000 mm/s) eliminating the need for a waste tower (at least with regular filaments). Nevertheless this is an interesting machine, especially the X-kinematic.
I'm concerned the response to worries about deflection being "he ran a simulation" rather than "he took measurements of the actual printer in motion". Removing the gantry around the rail while simultaneously adding the weight of the motor to the tool head is significant.
The gantry doesn’t move on the Z direction, so vertical wobble shouldn’t be an issue. Static sag is trivially calculated, and that’s taken into account in the mesh levelling process, unless you’re doing two identical prints.
I would love to see an IDEX Toolchanging printer, having the best of both worlds would be most welcome! Im getting the feeling it is about time for Bambu Lab to announce a new printer, hopefully a larger X1C and hopefully a toolchanging option.
I have pondered a IDEX with Tool changer and possible then one head going to an MMU. The trick is to set it up to cause the fastest changes to be the most common ones. So Primary on one head. Tool changers for support and other colors and if you need more colors beyond that use an MMU to feed one of the tool changer heads. It gets towards the I don't need it but I want to explore what is possible. I also an pondering a 3D printer with either a very long X or high Z for a unique use case.
I think if Bambu Labs adapted their Lidar and Load cell tech to an easy set up\calibration IDEX Printer combined with their AMS for general filament support management it would provide a very affordable\interesting option to a full tool changer.
I liked this concept for a printer so much I just put an order in for one. I’d love to see one more follow up video detailing how you’ve managed to leverage the printer to make unique dual material parts and maybe recap any new issues or ongoing fixes that formbot/dan have addressed.
Great video! You can also use the CXC to do manual camera-assisted calibration for XY offsets rather than setting up kTAMV entirely and it'll save you both time and improve precision. kTAMV is cool, but doing it manually is still lightyears ahead of the traditional, iterative printing method.
Interesting review. You did not encounter similar issues as the people at 3D Printing Canada ? Which all things summed up boiled down to : "firmware is highly unstable, on the brink of unusable even, and it forgets calibrations after power cycle, and assembling the machine took them close to 15 hours". If not, do you know if that's caused by possible improvements to the kit and firmware compard to the situation of the Marathon 3 months ago ?
@@kaween1000 there were some issues earlier on that have been resolved. That “review” was pretty misinformed in my opinion. 3 hours is a more accurate assembly time. As I mention in the video, the current state of the firmware is among the best I’ve seen on any Klipper printer.
Very nice, have had it on shopping card but resumed because of custom and delivery costs from China to Germany. Now in progress to klipperize old Creator 3 buyed cheap with fried mainboard. Good for learning handling of IDEX and reduce spareparts stock. May in future Marathon will be the successor.
Thank you for this clear & concise review & well done to the creator, I look forward to more of your content. With regards to the unsupported X gantry design. In my own personal experience when creating a similar but single tool head design (MGN12H)the assembly worked well on day one but over time the load curvature increased. In my case this became a bit of a vicious circle with belt tension being increased due to offset the increasing curvature and the increased belt tension adding to the rail curvature. The deflection was slight but enough to create layer shifting due the nozzle catching the filament at the most low point. I am not at all suggesting this is the case with this product\design and if there has been some very long term testing\analysis regards this feature, it is all good, just found this personally not to be the case. Could easily have been my design not the concept. In this design the belt tensioning is horizontal rather than vertical as per my more traditional Core XY 3D printed part design so may not be the case and this comment should be totally ignored has been not relevant with respect to this review. I just thought it was worth mentioning regarding a long term observation.
Thank you for the "follow up" video. Your initial video caused me to order the Marathon but I have not had time for really getting into it. Would love to see what is possible quality wise as I value quality over speed.
It’s a very nice looking printer, but as I wouldn’t use the higher throughput, I’d rather have a toolchanger with more than two options. I can only hope that the improvements made by this sort of printer and the XL’s shortcomings will be applied to a modern toolchanging printer.
It's bed size is too small for my needs. My Troodon 2 has the best bang for space with it's given bed size, really wish it had convenient two color capability though. I know, have my cake or eat it... :(
You can do simulations all you like but practical experience and measuring is always better. Using a linear guideway alone as the X axis is a bad idea and is some of the source of the bad bed height map. On my Voron 2.4 I had one of the super light X axis installed and it slowly sagged which was visible on the bed mesh as a hump in the bed. It eventually got up to 0.4mm range or more than a whole layer thickness. I swapped the X axis to a solid machined version and the bed mesh instantly went to less than 0.05mm positive and negative. The system can of course cope with the variation but it completely defeats the point of using a cast aluminium bed if the gantry is warping. Especially in an IDEX machine. You may as well use a cheap PCB bed and make the machine cheaper.. There is a lot to like about this machine otherwise so I hope they rethink it and make some changes.
I am thinking about IDEX for quite some time but i suspect there are some reasons behind IDEX not being mainstream. It exists for quite some time already...
I've been interested in the Marathon for a while now, and this video has helped solidify my plan to have this be my next printer. Until the Marathon released, the overall state of development of IDEX printers had largely stalled since Sovol released the SV04, which is surprising given the increasing attention paid to multi-color and -material printing as well as filament waste. Dr. Dan's attention to detail throughout the design is obvious, even right on down to the instructions on how to upgrade the screen. And the biggest thing for me that puts the Marathon ahead of the RatRig V-Core 4 is the relative ease of assembly...the size options RatRig has to offer are nice to have but I don't like having to do that much assembly, especially the wiring. One thing that wasn't obvious either from this video or the Vivedino website: it very much looks like the nozzle and heat break assembly uses a standard V6 nozzle but then you mentioned that it was like a Revo in its overall construction. Are the nozzle and heat break a single part that is proprietary to Vivedino or is it possible to use off-the-shelf V6 nozzles? I'd love to use either a CHT nozzle to increase the volumetric flow rate (perhaps even with one of the huge 1.8 mm CHT nozzles for extra-strong infill) or a Diamondback for worry-free printing of abrasives and high-temp materials alike.
@@justinchamberlin4195 nothing about the nozzle / hot end is proprietary. I haven’t removed the nozzle to confirm but I think I was mistaken in saying that the nozzle and heat break are a one piece assembly. This hot end is compatible with those but I believe the stock hardware is just as pictured on the AliExpress page with a smooth top side heat break and threaded lower side, with a standard hardened steel v6 nozzle below it. So given that, you could replace it with any other v6 nozzle, including a CHT.
Very interesting printer and configuration, though I feel like some of the x-axis gantry stuff is done in a weird way for the sake of being different, motors on the sides of the Y carriage would put them more in-line with where the forces are applied by the belts at the edges that drive the Y axis beam, also sharing a belt between the two toolheads.. that one seems like a maybe bad idea, they are constantly tweaking at each other's tension levels, subtle belt stretch etc.. klipper tunes individual toolhead resonance via shaping, but what about external forces.. like the adjacent toolhead? I can see that they turned 4x of possibly needed belt(2 loops), into a 1x length both motors now run along, but belts were never that expensive anyways. Also surprised it's not a CAN setup with motor drivers on the toolhead for both X motion and extruder, instead all discreet harnesses.
Quite a nice printer. One point I do not quite understand: The bed has 3 axis like the vCore printers, so it can tilt. So actually hardwarewise it could compensate the hight of two different points. If tool 1 needs maybe +0.2 mm correction and tool 2 needs -0.1 mm, the bed could be tilted to archive exactly this to be done. Just some calculating to be done. The problem at the moment is Klipper does not have this function. However, as Klipper is contantly developed further and with the new vCore 4 Idex another great printer would profit from such a function, we might see it in future. This is unlike other Idex-printers, were you neither can tilt the bed nor the x-axis.
I thought you were on to something, but actually this idea may not be so easy. A height offset between the two hotends would require a different bed angle depending on the distance between then, with a steeper angle needed when they are closer. If we try to continually compensate, we'd be wiggling the bed angle left and right, which would cause an increasing offset the higher the print goes... Now having said that, I just realized that the compensation could likely be done over the first one or two layers to get a level print layer and then just continue normally. 👍
@@Chris-oj7ro On duplex mode you always have about half the bed width distance between the extruders. Even in mirror mode you need the width of one extruder in distance (about 8cm or more) between nozzle to avoid collision. So if the distance between left and right z-axis is maybe 40 cm, the difference between left and right z has to be at maximum 5 times the difference between z-mesh-offsets of extruder 0 and 1. A good bed has a maximum deviation of about 0.3 mm. So the maximum tilt to compensate would be a difference in z-hight left and right would be 1.5mm. As theoretical worst case, if the maximum deviation would occur on an 8cm x-distance (unlikely). 1.5mm on 40 cm distance is not realy an steep angle. So while the angle has to be adjusted due to the x0 and x1 positions, I see not a real problem there. Even going further up in the print without fading out the compensation in the first cm, it would still be just a theoretical maximum difference between left and right about 1.5 mm all until you reach the top. While you have 3 z-axis, for calculations you only need two, as the two z-axis on one side have same x-coordinate, so they need same correction (you tilt arround the y-axis). So if two z-coordinates on the left and one of the right, the two on the left need same correction. Lets calculate the offset of z-left and z-right for theoretical worst case, assuming following dimentions, defining middle of bed as x=0 for more easy calculation: Bed: 260mm (distance between left z and right z = 400 minus 2 times width of an extruder =2*80mm to allow full dual mode) x-position z0&1 axis: -200mm x-position z2 axis: +200mm x-position e0: 200-40 = -40mm x-position e1: 200+40 = +40mm max deviation of bed hight: 0.3mm. So this is the maximum to be compensated for. ze0_comp = -0,15mm @ x=-40 (due to mesh) ze1_comp = +0,15mm @ x =+40 (due to mesh) resulting in tilt = (0,15-(-0,15))/(40-(-40)) = 0.00375 left-z-compensation = -0.00375 * (-200)mm = -0.75 mm right-z-compensation = 0.00375 * (+200) mm = +0.75 mm Remember: It´s just the compensation for the bed mesh by intentional tilting, after z0,z1 and z2 were already adjusted to no tilt. It does not include a physical z-offset between e0 and e1. However, if you use two identical manufactured extruders with precise parts (like Revo nozzles), the physical z-offset between left and right extruder should be less than 0.2mm. If you add max diviation of the bed and physical difference of z-offset between extruders, you would still only need to compensate for 0.5mm over am minimum distance of 8 cm. So maximum tilt of 0.00625, resulting in +/- 1.25 mm in above example. The correction needed are quite smal. Maybe I was wrong - I think at least with mechancal compensated heads and a maximum tilt of 0.00375 this might inside the tolerances of a bed-slingers ability to tilt the x-axis using two z-motors without blocking. And while theoreticaly you should compensate for x-length if x-axis is tilted, with such low tilt this might not even matter... Edit: On this printer two z are in the front, one in the back. So the tilt is done by the left and right front z´s, the back one has no effect for tilting along the y. I had the positions of the vcore 3.1 in mind. But same principle.
It took a long time after release to have at least one video on RU-vid. As far as I remember the printer was available in January 2024. I considered to buy it but was not very impressed by the first reviews. I fear that development is lagging behind due to the low sales figures. At the moment I'm waiting to see what Bambulab launches and then I'll have another look to see which printer will be right for me.
I'm looking at printers for nylon carbon fiber and this one is interesting because it would reduce the cost of the support material. Have you reviewed this material considering that it doesn't have a pre-heated chamber? Tnks.
I did some testing with glass filled ABS which has similar properties. The passive chamber temperature rose to 55 degrees which should be suitable for CF Nylon, especially with a bed adhesive for added adhesion.
@@slomo4056 it uses input shaping it just doesn’t have accelerometers. The same is true of the Prusa XL. You can add your own accelerometer if you want to measure the resonances yourself, as I have shown in the video.
I love IDEX, and hope it is the future along with tool changers. The marathon is an appealing higher level printer. If you’re looking for an entry level IDEX I recommend the LNL Tenlog D3 Pro ~$500.
Just bought a troodon 2.0 pro so no need for this printer at the moment. When i need a idex printer in the future it would be a race between this or a voron phoenix Biggest difference the build volume 😅
Dont really understand the idea to have the motor on the tool head, there seem almost zero benfit to it, added weight to the tool head, limited size and for assembly i am sure it easier to install it on a mounted frame than on the tool head with the belts
I really hope Formbot/Vivedino has improved the firmware fixes. I own a Formbot T-Rex 3.0 (also an IDEX at a price point 2000+ €) and it never received any firmware updates and is a constant pain. 🙁
@@andreask.2675 watch the firmware section of the video. The marathon is a different beast. A completely different team working on it. It’s Klipper now and not Marlin like the T-Rex.
As a Marathon owner I can tell you @ygk3d tells it how it is in practice. Using the latest image the Marathon is fully updateable, including the Marathon-specific code, through the standard Klipper Update Manager. Just click the check for updates button, click update all components, click the standard checkbox that you understand the risks, then click the update button - a few minutes later everything is updated including the latest Marathon-specific updates and macro adjustments. It just works.
I'm not sure how that would work. A second gantry with 2 more heads? Could be interesting but I think you'd need a significantly larger X/Y footprint to take advantage of all 4 heads for quad-mirror/copy mode.
@@Kloberman I'm less interested in the copy mode (though I want 4 of them to copy YES!) and more interested in just four different materials printing the same, or similar parts, maybe even parts of the same item. I have a ton of multicolor prints that I would love to do in multi material that I can't do with my P1S with AMS.
I am in the market to buy a few idex of this size minimum. The price is cheap for what it might do, but I dont see very tidy results from this video. Prints look pretty messy in general. Not convincing as I do not have time to finetune and fiddle with single printers.
@@mannytheengineer4513 yes. Looks nice but considerably more expensive. A good option if you want to print PEEK or other similar materials that need a heated chamber and 500C capable hotends.
