MeLt-N-AtoR Overview Metal 3D printer

[@jamieclarke321]

The more in depth explanations and dialogue is really great. Thanks rock salt

[@lolthekidison]

Awesome work, lots of really cool ideas used in gantry from alignment and tensioning. The new frame and Z axis look rock solid.

[@skimachco4198]

Keep these vids coming please. Cool work 👌

[@jamieclarke321]

Let me know if you want a set of parts made on a CNC machine once your finally settled on the design 👍

[@FrodeBergetonNilsen]

Thanks for sharing. Given the kind of tolerance this thing needs, with just one cross gantry, it is sort of highlighting the issue with tolerance and accuracy. It is funny how quickly binding becomes an issue, when things have to be aligned properly. Even more so, when I look at the specs for them rails, they are way more sloppy than proper rods, yet rods tend to bind less. If we are to make every joint and rail or whatever actually aligned, the joints will be crazy, with adjustments in all sorts of directions. Not to mention the accuracy of aluminum profiles, which is simply a big unknown, and I cant even remember seeing anyone actually compensating for that intolerance. Getting even the x-y plane flat is hard. Getting it actually square, anywhere near like in the 10s of microns, is hard. Not seen anyone work on that at all. I guess pushing for speed is an easy fruit to pick and to flash off, but getting things tight on genuine tolerance, well, that is an entire different ballgame. Since I am moving in the direction of cross gantry, it seems to me that the tighter accuracy of rods is desirable, and the more forgiving bearings for the rods is also desirable, at least until the tolerance issue might be dialed in. Also, when needing a more solid rod, I can just order a bigger one, and the rod itself, even when "reinforced" will still be super tight. Also, I just don't comprehend the trade-offs yet, and this praise for accuracy for rails are completely lost on me. The rails themselves are made at way looser tolerance than a tight rod. Often times, the rails are mounted onto parts with seriously sloppy or even unknown tolerance, such as aluminum profiles. So, if the rails are way soppier than a rod, why is the rail system supposedly of tighter tolerance? How sloppy are these bearings really? So again, thanks for sharing. My takeaway is still that in general, moving into cross gantry, chasing speed simply do not cut it, chasing accuracy and tight tolerance, is the mindset I actually need. So that is what I do. When I then look around, hardly anyone has any useful info on tolerance at all. People have no real interest in it, and useful info is like nowhere to be found. For instance, you use cartesian kinematics, with dual individual mechanisms for X and Y. For that to be anywhere near tight, you need to know that the tolerance of displacement on each of them mechanism runs within the the tolerance of the kinematics, as in need to make sure the displacement is within that tolerance. If not, the kinematics need to absorb the intolerance. Only when the actual, as in real world tolerance is dialed in, only then will rods and rails with super tight tolerance run without absorbing or binding. I don't hear anyone focusing on this. But you actually partly did here. Thanks for sharing. I will not use aluminum profiles, nor rail, as I need tighter tolerance and do not care for formula 1 speed. Any joint is a complete nightmare, when targeting tolerance. Cannot be overstated. Also, this chase for speed and heated chamber is even more baffling. That sort of throws tolerance out the window, again, and if using a gantry dependent on tight tolerance, this quickly becomes a nightmare. It is funny, how this industry do not realize that their industry depend on sloppy and seriously intolerant designs to even function, with this speed and high temp requirement. For instance, du you actually move that bed perpendicular to the x-y plane? How do you know that? Is that bed actually moving straight on the Z axis, and, how do you know that? How? And when you heat that chamber, how to you know that the tolerance is not messed up? Guys, I am not talking about what you think or presume, I am talking about what you actually know. What is the fact, and how to get the best understanding of the actual facts as possible. Because when you look at beds like this, and starts to seriously question actual accuracy, it kind of falls apart, and there is no way known to me, to actually measure it. Remember, for a heated chamber, you need to measure it while hot. Who does that?

[@MrRocksalt]

Hello, My understanding of the rails is .05mm (.0019")tolerance. I have not witnessed rods being tighter on 3DP, but... the parts they use are probably binned parts that didn't make the grade. The most problem with crossed gantry is over constraint. Plastic deposition tolerance seems to be more than this value. Making a 3d printer, in crossed gantry, with tighter tolerance seems less desirable. From my experience, which is not a decade, it is better to allow some unconstrained areas. Let opposing sides 'float'. To me, the point of crossed gantry is the shorter belts as the elasticity of the belts is the weak link. Also, the cross does a very good job of countering the toolhead from pivoting/rotation. And lastly, the ability to keep belt tension more consistent across the build plate xy positions. The way to make the frame is to use the same material as the rails/rods.. Steel, just for CTE sake. And not just the halo. The frame will swell an displace vertically. To answer your question about skew: The bed is trammed to the gantry via 3 point leveling. Yes, there is the issue of the rear member (Z system 1PZ) not traveling a true 0° vertically. But this can be easily calibrated with this system. Even if there is issue mechanically, the controls can compensate for the skew. You just have to do the required tests to obtain the values and enter them. If you have some test blocks tall enough, it can also be measured on the machine. But guess what.. toothpaste through a tube... it will always be that way with FDM. Choose your battles carefully with what you print. I will do a video, tramming the Z vertically, it's not that difficult to do with some jigs for the indicator and this 1PZ system is not entirely dependent on the front lift arms as they are only there to support the bed. They can have some influence, but their effect is much less than a printer with a different Z kinematic.

[@FrodeBergetonNilsen]

@@MrRocksalt Not sure where you get 50micron from? There is no 50micron listing in the tolerance table. As for rods, even H9 is tighter than 50 microns, even for Ø18mm. A 10mm rod at H6 is +0 to -9microns. An H4 rod is listed at 4 microns a Ø10mm. The tightest rail listing I have seen is 15microns. As for skew, using a Califlower or something like that, well, once you start running down the list of assumptions for such a test, you're not left with much accuracy. Not to mention, accuracy that is applicable to what? Take this claim: _"traveling a true 0° vertically"_ Parallel to what? What is the assumption here? There is no parallel to any gantry, because by definition, it is not a flat plane. Even removing flat, it not even a plane, given the tolerances at play, so what could be "parallel" in one spot of the "plane", is not "parallel" in another. Give it some thought, because these rods or rails for the Z need to run in parallel. Even the most tight rods, which I have never seen but clearly exists, are simply not perfectly straight. Also, I simply fail to see what on earth that is supposed to guide any moving bed to move straight, other than the rails or the rods. Please explain that to me. What supposedly guides the bed to travel factual straight, or rather, within what tolerance can you actually claim the bed to be moving, and more importantly, why? Just because you say so, simply don't cut it: That is gaslighting. It is kind of funny when people measure a printed part, with all the issues that comes with it, using a caliper, then claiming that their measurement is universal accurate beyond any reason. The level of noise is simply insane. This is destroying the industry. It needs to stop. Take ABL. No matter how anyone tries to distort the truth, in its current form, ABL does not level the bed at all. No ups and owns are corrected, other than maybe the bedpalte itself is tilted, but not to be flat, nor to be in parallel with the rods or rails guiding it. Even worse, ABL propagates through the entire print, unless you actively deactivate that. Most people do not even print along the gantry "plane", whatever that is.

[@MrRocksalt]

@@FrodeBergetonNilsen Look at the assembly tolerance section for the HIWIN rails. www.hiwin.com/wp-content/uploads/Linear-Guideway-Assembly-Manual-HIWIN-Corp.pdf If the gantry plane is the master, then the probe leveling of the bed translates the bed to be on plane with the gantry, using a 3 point bed. The Z position at max to min can be calibrated mechanically or in software to achieve a true, square box, within the limits of FDM (toothpaste through a tube).

[@Tinkerer5888]

@@FrodeBergetonNilsenyou do realize aluminum expands with just the given off by hand… so you can only make a printer so tight of tolerances that a 60c temperature change would make literally everything binding up just due to thermal expansion and rods are not a good option hence no good cnc or mill runs on them, they have no support and cause a lot of unwanted vibrations

[@FrodeBergetonNilsen]

@@Tinkerer5888 CNC mills are a totally different beast, and once again, you argue about the need for support for rails that are simply not that strong, and will probably bend by their support. Once you get the point, you realize that the height of the rail thus varies by the supporting structure, not the rail. The supporting structure influence the accuracy. But sure, some use cases requires pretty hefty support for the moment, CNC is one of them. Even then, in the very example you offer, you seem to not understand the importance of the support structure accuracy and tolerance. A 3Dprinter is typically a completely different set of tradeoffs. Just like a bike differ from a motorbike.

[@chrisgray21]

a fantastic overview of your printer set up, please keep sharing.. could you share where i could find information on the 1PZ Z, i cannot seem to locate any information on it.. Keep up the great work.

[@MrRocksalt]

www.printables.com/model/595277-firstprincipals-z-motion-system facebook.com/groups/3dprinterdesign/

[@tokehgekko9455]

First 3min 50sec is 90deg wrong 😬just something to consider for the future :P

[@Tinkerer5888]

I believe thermal expansion is your issue. I feel a welded steel frame would have been much much better. If just one vertical extrusion gets more heat than the others I think it would still warp the gantry plane especially to the Z plane

[@MrRocksalt]

yes.. you are right. Mitsumi had a sale on 1st order of CNC job. But I put in a request and the cost was nearly 900.00US. So I passed. When I come across some steel, I will have it machined flat/surface ground. It will fit right on top of this printer. Can't help the legs though.. they are going to expand, but at least the gantry will be solid.

[@SirRobinII]

ye these projects takes a lot of time. Ive converted my ender 5 to a croxy and I bought parts to build a large croxy printer. But theres no energy in me left for the next couple of weeks.

[@jamieclarke321]

At the start of the video I need to hold my phone upside down 😅