@3dprinting One of the first things I did was repurpose the metal reinforcing brackets that used to support that crossbeam into supporting the bottom where the electronics housing used to be. These needed two M5x12s and two M5 hammer nuts each. I couldn't reuse their existing M5s, as they were designed to go into tapped holes in the frame and were thus too long.
3/n
@3dprinting Oh, almost forgot. Anyone who's interested in what an Endorphin is can check out the project page here: https://endorphin3d.com/docs
The tl;dr of the Endorphin is that it's a minimal set of Ender 5 upgrades designed to fix many of its issues, by making use of an unusual choice of kinematics (hybrid CoreXY). Unlike something like a Mercury One, which is a near-total rebuild, this is a small set of upgrades.
4/n
@3dprinting I won't limit myself to the Endorphin's upgrades though: at minimum, I will have to rebuild the entire Z axis, and I'll have to do plenty more besides. However, I always liked the Endorphin for how small and simple it was, and thus, decided it would be a workable project that can use up a bunch of workshop junk.
5/n
@3dprinting One of the first _printed_ upgrades will be some brackets to reinforce the frame. I found one that I liked here: https://www.printables.com/model/82832-90-degree-bracket-for-2020-extrusion-for-ender-5-p
However, the intended print orientation is somewhat suboptimal (thanks, Slant3D!), so I modified it a bit to add a sacrificial layer, which allows it to be printed in a more optimal way. I'll need quite a few of these: I'll start with the ones reinforcing the two Z columns at the front.
6/n
@3dprinting For material, I'm sticking with my 're-use' theme and using KiwiFil rPETG, which comes from 100% recycled feedstock. It helps that I have about 1.5kg of it lying around too. As I have no plans to enclose this printer, this will work fine, and will look amazing, as it's Race Car Red.
For those interested, this is the filament: https://www.kiwifil.shop/products/race-car-red-rpetg-100-recycled-1-75mm-1kg
I _will_ make an exception for the toolhead fan duct, which will be made of ASA, but that's a future Koz problem.
7/n
@3dprinting I'll also need to find some way of covering various bare extrusion ends (four in total) that are left by removing the top-front crossbeam and the original screen mount. I'm thinking of using these: https://www.printables.com/model/642540-2020-end-cap-with-chamfer
If anyone can suggest a better one, I'm all ears (eyes?).
8/n
@3dprinting One more-than-mildly annoying problem with the Ender 5 (in any variant) is that two of its motors are not really reusable. The original E motor has a round shaft with a weird press-fit gear on it, which makes it hard to use for anything. Additionally, its weird dual-shaft Y motor isn't too great for anything either. The X and Z at least can be re-used (and match AFAICT).
Not a huge issue for me though, I have two LDO Speedys gathering dust.
9/n
@3dprinting Stage 1 of the Endorphin needs some idlers of a type I don't have, so unfortunately I had to order some. Luckily, Powge had them, and they weren't expensive. Still waiting for them to arrive though.
10/n
@3dprinting One down, _many_ to go. I'll publish the design as a remix of the original. This also shows how vibrant the Race Car Red PETG looks next to the black extrusions.
11/n
@3dprinting Decided to publish my remix while I wait: https://www.printables.com/model/1413551-ender-5-4-bolt-corner-bracket-slant3d-optimized
#3DPrinting #FDM #3DDesign #FreeCAD
12/n
@3dprinting So, time for some absurdism. Remember that destroyed Z axis I posted pics of earlier? Seems like the Z motor was also a victim: the shaft is _so_ bent it has _visible_ runout.
However, not all is lost! I thought I had two LDO Speedys in my workshop junk pile, but it turns out I actually have _four_.
Yes, I am going to run _Speedys_ on my _Z axis_. Like an absolute maniac.
13/n
@3dprinting A little bit later, progress! Unfortunately, I ran out of the M5x10 bolts I need to make more brackets, but even these four make the frame _feel_ stiffer.
I also printed some covers for the extrusion ends exposed by some of the stuff I removed, which looks much better.
Next step would be using some 2020 extrusion covers to hide the rather ugly holes drilled in the extrusion in several places. Going to try these: https://www.printables.com/model/208845-2020-aluminum-extrusion-cover
14/n
@3dprinting Predictably, that design didn't work out. However, I found a design that I could remix, and now, the ugly holes can be covered!
I under-printed by about 6mm, so I'll probably just fix that up later.
The design I worked from is here: https://www.printables.com/model/1169141-slot-cover-for-2020-extrusion . I'll post my remix later.
15/n
@3dprinting Pictured: an absolutely normal amount of reinforcement for a frame. For reference, together with the four already printed and installed brackets, this adds over _half a kilogram_ of weight to the frame. PETG be thicc like that.
And yes, that is one of the Speedys going into the Z axis.
16/n
@3dprinting This is the plate for the Endorphin Stage 1. All but one of these parts ended up needing tweaks by me, mostly to add chamfers, remove captive nuts in favour of heatset inserts, and in the case of the idler mount, to fix some _serious_ weakness issues.
I also managed to find bugs in _both_ the AMF _and_ STEP exporters in #FreeCAD in the process. I truly am a bug magnet.
18/n
@3dprinting More positively, I think I've decided to go for the EVA3 as my toolhead. However, since the _only_ hotends I have as workshop junk are E3D V6s, finding a hotend mount was a bit of a challenge. However, this one here works great: https://www.printables.com/model/201093-eva-30-e3dv6-revo-revo-micro-hotends-mount . The fact the groove mount is still free to rotate is not ideal, but I don't really know of a good solution for that anyway.
19/n
@3dprinting "But Koz, won't this have absolutely garbage flow?". Normally, dear reader, you'd be right: however, through the magic of the ZS Plus nozzle collection I've acquired, this should be far less of an issue.
20/n
@3dprinting The upcycling continues! For the extruder on the Endorphin, I was originally planning to use a NEMA17 pancake I've had lying around for ages. However, I quickly found absolutely _zero_ extruder designs fitting that. So I did some more digging, and found:
* A TL all-metal K1 Max extruder replacement: https://trianglelab.net/products/k1-series-extruder ; and
* This 0.9 degree NEMA14 pancake: https://www.omc-stepperonline.com/round-nema-14-bipolar-0-9deg-9-ncm-12-75-oz-in-1-0a-%CF%8636x17-5mm-4-wires-14hr07-1004vrn
They fit together like it was meant to be.
21/n
@3dprinting "But Koz, how on $DEITY's green earth are you going to mount that to the EVA you're going to use?"
Good question, dear reader. However, I am in luck. This is meant to be a drop-in replacement for the K1 extruder, and some kind soul has done the work already: https://www.printables.com/model/770682-k1-extruder-for-eva-3
Now, the existence of this model implies someone wanted to put a K1 extruder on an EVA. This is tantamount to fuelling your racecar with moonshine, but I won't look a gift horse in the mouth.
22/n
@3dprinting Is this what everyone posts on OnlyFans these days? 😹
More seriously, I modified these from the Mercury One's feet. The changes were to use M3 instead of M5 fasteners to the frame, and replace the captive M5 nut with an M4 heatset so I could upcycle the feet from the Ender 5. Looks really fantastic, and adds yet more frame mass. Needed two M3x10s, plus one M4x12, together with two M3 hammer nuts.
23/n
@3dprinting With that, I can also install the Speedys that will run the Z axis. The brackets are of mismatched colours, but that's just how they came, and I'm not going to let metal motor brackets go to waste.
24/n
@3dprinting So, slight rethink. I decided that I'd rather use the Z axis Speedys on the Y with this dual-Y mod: https://www.printables.com/model/520947-ender-5-plus-dual-y-axis-motor-mount/files . In the process, I discovered that the _only_ way to tension the Y belts on the Ender 5 involves _throwing your frame out of square_. Yes really. You have to remove a _structural_ component to _tension the belts_. Wtf Creality.
So yeah, more frame stripdown!
25/n
@3dprinting In the meantime, I've been making progress on the EVA3 I'm going to use. The next step is to make mounts for the dual 5015 fans for part cooling, and also the duct, which will _likely_ be the Tri-Horn. The duct will be made of ASA, as PETG isn't really up to that task. I also need to mount a Biqu Microprobe I have lying around, which will involve both a BLTouch mount _and_ an adapter, because the jank will never cease.
26/n
@3dprinting Today has been a day of... interesting progress. The dual-Y mod is mounted, but is currently quite loose, as I cannot attach it from the top and bottom. The reason? The same M5x10s I need to finish reinforcing the frame.
I think this is _the_ first time I've had a build bottlenecked by _fasteners_. It's kind of absurd. More are on order, but Aliexpress takes a hot minute.
As an aside, this build is starting to look like something Ivan Miranda would make.
27/n
@3dprinting I'm kind of freestyling here, as I don't (yet) have the idlers I need to do the mod properly. I also have to re-belt the Y, but that can wait until I can do the proper Stage 1 work.
One 'fun' discovery is that the Y assembly is weirdly structural on the Ender 5. This is most notable when looking closely at the dual Y. That bearing there is supporting _nothing_, but I can't remove that part, because it ensures the corner doesn't come off the pillar!
28/n
@3dprinting The eagle-eyed among you might have noticed that the Speedys are gone from the Z axis. This is indeed what has happened, as the dual-Y mod is _definitely_ a better use of Speedys. I managed to find some random garbage motors I could use there instead. Their shaft lengths aren't _quite_ the same, which I suspect _will_ come to bite me later, but that's a problem for future Koz to worry about.
29/n
@3dprinting The wise among you might be scratching your heads right now. "How the hell is he planning to run dual Y _and_ dual Z off a Cheetah?"
Yes, that could be a problem for sure. However, I already figured out a solution to this, in probably _the_ stupidest way possible. I'll explain that next post.
Meanwhile, let's skip even _more_ steps and print the mirrored Y tensioner from Stage 2. Otherwise, I have a _very_ loose corner, and it bothers me.
30/n
@3dprinting To explain how on earth I plan to run 6 motors, we need to consider power. The Ender 5 comes with a single LRS-350-24, which is _definitely_ going to fall short with all my upgrades. I _do_ have an HS-250-24 lying around, which would give me the power I need, but the Cheetah only has one input for power.
The solution? Double up with the original board! I can run one with the LRS, the other with the HS.
Sounds crazy? Yes, but let me explain.
31/n
@3dprinting Given that the heated bed wants 220W, I can drive it from the original board with the HS-250-24. In addition to that, I can drive the E motor from that board, as the poor little 0.9 degree pancake I'm running there is 1A peak, which means something like 0.7A RMS. That's pushing my luck a tad, but _should_ fit into the HS.
32/n
@3dprinting _Everything else_ can come off the Cheetah, run using the LRS-350-24. This gives me four steppers, one of which has two motor connections. That means I have a dedicated driver for _each_ Speedy, and then I can share the garbage Zs.
I am a _touch_ bottlenecked given such mighty motors here. The Cheetah has soldered-on TMC2209s, which have 2A peak current, while the Speedys can tolerate 2.5A peak. This will limit their performance some, but hey, this whole build is jank anyway.
33/n
@3dprinting Now, my original plan was to have a swingout electronics box with a 7 inch touchscreen I had lying around. Obviously with all these boards, this isn't really viable anymore.
However, with those feet, we now have an actual electronics well, which _should_ fit both the PSUs and boards, just. Then, the swingout electronics box can just house the BPi running Klipper and the screen.
Easier wiring, less designing, perfect!
34/n
@3dprinting With that in mind, I need DIN rails, as well as a way to cut them. Luckily, we can crib lengths from the Mercury One project, and I do have a local source for DIN rail.
Guess who is _also_ the proud owner of a new reciprocating saw?
35/n
@3dprinting One rather _annoying_ aspect of the Mercury One is that their electronics enclosure (or well) has _zero_ documentation attached to it. No BOM, no instructions. While you _can_ get the CAD through their base Mercury One repository, this took me some searching to find.
I'll have to pull the CAD model to find out both what mounts they're using for their DIN rails, and how long said rails have to be. Nice to see this project _also_ apes awful Voron documentation practices.
36/n
@3dprinting While I wait for filament, I finally got the idlers I needed to finish Stage 1. I haven't bothered putting the pulley on the X motor, or belting anything, because I'd have to un-belt everything for later Stages anyway.
37/n
@3dprinting I also decided to try using my new reciprocating saw to cut some DIN rails. I wasn't particularly successful, as I simply could not workhold anything against this saw's might. Ended up using a hacksaw, some 3-in-1, and a lot of patience.
These are square and parallel, the camera just makes them look like they're not. Might change them later.
38/n
@3dprinting One tip I got taught when doing tricky idler stacks involving shims: use a hex key to hold everything in place, then bolt down from the opposite end, displacing the key. This proved _especially_ useful here, as the thin shims I was using would have _certainly_ shifted otherwise.
39/n
@3dprinting I also discovered that I did, indeed, have two flex couplers for lead screws, so I installed them too. The Z motors have _slightly_ different shaft lengths, so I will likely have to adjust them a bit. They look very cool if you give them a little nudge, as the springy couplers make them wave.
I can't do the rest of the Z assembly, due to missing 10mm rods. Until I get all the stuff I'm missing, I can't really continue.
40/n
@3dprinting Mounted the Y rails while I wait for parts. This is a bit of quite confusing documentation from the Endorphin project. According to the pictures, the written instructions and the CAD model, the right rail goes on the _outside_ of the frame. According to the instructional video, on the _inside_. So which is it?
Opened an issue about it: https://github.com/endorphin3d/endorphin/issues/31
For now I followed the video, so it looks a bit goofy. Can always remount once I have the printed parts.
41/n
@3dprinting The Stage 2 instructions also aren't particularly specific on _where_ the rails are supposed to go exactly. This is definitely not obvious, as each rail is non-trivially shorter than the extrusion it mounts on.
For anyone reading this as an #Endorphin guide, it's 15mm from the rear brackets.
I might just go sensorless on Y and design some 15mm spacers/stoppers. Spoppers? There should be more than enough space there to M3x8 and hammer nut something in place.
42/n
@3dprinting More filament arrived today, which means more work on the #Endorphin. Several things have happened today.
First, rail stuff. I moved one rail to the outside of the frame, then designed and added some rail stops. These are both spacers (ensuring the rails are symmetrically placed), but also sensorless homing targets. As per the meme of this build, they fasten with M5 hardware, because what else.
These are fun to bump with the carriages.
43/n
@3dprinting Next, we finally have a second Y tensioner! This is a redesign by me, because the original was held in place by _one_ screw, which meant it was neither holding the corner together (like the original tensioner) nor particularly well-constrained (as it could rotate around that screw). My design fixes both.
44/n
@3dprinting Third, as you probably noticed, I restored the top front extrusion. While I admit it gets in the way, I simply could not justify not having it when I realized that I couldn't reinforce the two side top extrusions with corner brackets. This is because the Y belts would have nowhere to go then, whoops.
I stuck some small corner brackets on it just to be sure though.
45/n
@3dprinting Last, but by no means least, I have replaced the toothless idlers with toothed ones for the Y idlers, and the idler opposite the X motor. These not only better match my colour scheme, but should also help with precision. Having to redo the shim stacks was a bit annoying, but ultimately felt worth it.
46/n
@3dprinting Next on the agenda is completing Stage 2, which means mounting the two parts holding the X axis on the MGN12 carriages. The original parts have... _suspect_ belt clamping, so I'll probably test-print the parts first to check if they actually can hold a belt under tension. If not, I'll have to modify them too.
I'll also have to finally thread and tension the Y belts too, since those mounts will stay with us into Stage 3. This is the _easier_ belting step at least.
47/n
@3dprinting I also need to finish mounting all the reinforcing brackets, but for _that_ I'm going to need a lot more M5 fasteners. They are arriving soon. I'll also need to properly mount the Y motors (those mounts have quite a bit of vertical play at the moment), and most likely will need a _different_ reinforcing solution for the rear top frame corners, as my original brackets collide with the Y motor mounts. I have some options for those, but they'll need test-printing too.
48/n
@3dprinting Also, somewhat tangentially, I'll need to finish the Z axis. While I can't do this yet (still waiting on some key parts), this will determine a lot of other things, including where the electronics will go. I'll have to print some parts for this as well, but I decided not to even think about that until I have the rods and extrusion in hand.
49/n
@3dprinting #Endorphin progress picture! I've (mostly) finished Stage 2 at this point: only the X tensioner remains. I've also added (and tensioned) both Y belts, which was a lot more annoying than I expected.
I did some other stuff besides, which I will explain in subsequent posts.
50/n
@3dprinting One of the additional things I did was mount both the top and bottom Z rod brackets. I had enough metal ones upcycled from the original to do the bottom, but the top ones had to be modified from the Mercury One. These aren't likely to be their final positions, as I won't be able to judge that until I get the rods and extrusion to build the axis.
Also, the clearance between the top brackets and the moving parts of the Y axis is _super_ close. Like, _millimetres_.
51/n
@3dprinting These belt clamps are simultaneously wondrous and terrible. They're wondrous because, in spite of _repeated_ attempts to pull them loose under tension, I couldn't. They are held in with nothing but friction against the belt teeth.
They're _terrible_ because there is no way in hell to thread belts into them once the axis is mounted to the rail carriages. Seriously, I tried, it can't be done. Had to disassemble the entire axis to get those damn belts in there.
52/n
@3dprinting So, after the X axis tensioner, what's left?
1. X linear rail (easy).
2. Z axis (fairly easy, blocked on parts).
3. LED mount (fairly easy).
... and then the suffering begins, because I have to build and wire the toolhead, set up (and wire) the electronics well, and probably figure out the skirt. All of these _will_ require custom parts and crimping.
Send help.
53/n
@3dprinting Been a few days without an update, but not because I haven't been working! I've since mounted the X rail, (most of) the X tensioner, and even started on the toolhead. However, I'm bottlenecked right now, because the belt clamps don't align with the belt path. The fix for this _should_ be easy enough, but I've been a bit stalled because of that.
54/n
@3dprinting The observant members of the audience are probably yelling something like 'WHERE ARE THE CAT EARS KOZ', which is a fair question. I'll definitely have some, but I'm not 100% sure if this cowl design is the one I'm going for. To save myself potentially re-doing a bunch of work, I'll first figure out how all the parts fit, and _then_ remix as appropriate.
55/n
@3dprinting So, good news bad news. The good news is that the belt alignment issues in _Z_ are fixed, using the stupidest possible method (chopping 3mm out of the part that mounts to the MGN9 block).
The bad news? With the stock belt clamps, I now have belt alignment issues in _Y_, which means at extremes of movement, belts now rub against the Y belt clamps.
Whee, time to redesign some belt clamps I guess!
56/n
@3dprinting This seems to fix it. All I did was add a 'foot' to the belt clamp out to about 10mm. It looks a little off-square in the picture, but that's not the part bending under tension, it's the belt itself interfering. This _should_ be fine I think.
I also decided to see if the Tank shroud looks better. It doesn't. Going to work from the original Small shroud to add the nekomimi we all crave.
57/n
@3dprinting The build continues - we finally have an X belt! This was a _real_ pain, as not only did I have to build another shim stack for the idler, I also had to thread the belts _without_ the hotend mount attached, as there's literally no way to do it otherwise.
Next up - proper X rail stops for sensorless homing, assembling the rest of the hotend, and of course, the Z axis.
59/n
@3dprinting Once the X axis rail stops are done, we will know _exactly_ how much X travel I'm allowed. I don't suspect I'll have any issues, given that the bed on an Ender 5 only has 220mm of usable X, but it would help for future mods to know _exactly_ how much slack I have in that axis.
I ended up going for the inductive probe instead of a Biqu Microprobe for two reasons. Firstly, this printer won't be enclosed anyway; secondly, the Microprobe mount I found didn't fit and I'm lazy.
60/n
@3dprinting And now the toolhead is assembled! Note that the wiring is 100% not done, but I figured I'd get it all in there to figure out how it all has to be arranged. _Definitely_ going to need some wire guides, many of which I will likely have to design myself.
X total travel with this (and test bump stops) is 280mm, plus or minus 1, which should be enough.
61/n
@3dprinting The _sweatiest_ front clearance ever. I swear that's like, a millimetre.
62/n
@3dprinting Figured I'd post a full photo. Mechanical assembly is almost done at this point: only the Z axis really remains. After this is installed, it's time for electronics and wiring, AKA 'the part Koz hates and cannot do properly'.
63/n
Koz Ross
