Ender 5 Plus "Medusa" Z system (In Progress)

[amber comet]

Just like the mythological Medusa turned people into rocks, the Medusa modification for your Ender 5 Plus aims to make your Z system rock solid

What is this modification?
The Medusa system is a quad motor system for the Ender 5 Plus based loosely on the Hydra system. The Medusa will be reusing the 10mm rods and linear bearings that the Ender 5 Plus ships with, requring a minimal amount of parts

What hardware do I need for the Medusa?
The BOM is currently unfinished as the project is still not ready to be mounted on a printer, although the necessary parts will include:

• 4x TR8 lead screws
• 4x 10mm kossel balls
• 4x 12mm magnets (for the kossel balls)
• 4x/8x 6mm rods
• 4x/8x LMK6UU bearings

Why not just build Hydra?
The Medusa aims to be a much more affordable yet still solid system for the Ender 5 Plus. The whole cost of this upgrade will be around 1/5th the price of required parts for the Hydra system (not counting the new bed). If you have the money for a Hydra instead, I would advise to go with Hydra, as that's a more thoroughly tested and maintained system, while the Medusa is (currently) maintained only by me, and it's future might be uncertain.

Do I need the MIC6/ATP5 Bed?
While advised due to the default Ender 5 Plus bed often being very uneven and unreliable, the Medusa reuses the mounting points that were occupied by springs on the stock system - so if your stock bed is flat enough, you absolutely do not need to get a new bed.

When's the release?
At this time the release date can't be pin pointed, but the system is quickly being developed. Currently it's in the phase of designing multiple mounting points and seeing what will theoretically be the most reliable in the long run. Once the files are tested by me, the project will be released as an open beta for anyone to try and give input on.

[crystal tangle]

real quick thing i learned while designing something similar:

• 4 points constraints the bed in some cases
• z using linear rods instead of rails will amplify your leadscrews artifacts due to cylinder connections

[amber comet]

Thanks for the insight!

I am designing better mounts right now, and once they are done I believe I will be able to test this system

I will come around to fixing these issues you brought up once I have it built, because right now I can't see what could be done to improve these, besides of course changing it to a 3 point system, and using linear rails instead of rods, which would in turn create just another Hydra

But again, thanks a lot for bringing it up, it's great to be ready for such issues before actually noticing them myself

[amber comet]

Changed the mounts to include 2 rods instead of one, to remove the possibility of the linear bearing skewing

[crystal tangle]

the wobble I was talking about comes from the z.
I would highly recommend switching to linear rails 🤔

[amber comet]

I know what you were talking about, the second 6mm rod was to fix another issue that I found myself

i also suppose in extreme scenarios the lead screw artifacts could be fixed by oldham coupling?

[crystal tangle]

hmm. maybe. it might be underconstrained then

[amber comet]

I will have to see how it all behaves in a real scenario, right now although with valid points and theories, it's all speculations, and I don't want to to overcompensate for something that might not even come up as an issue later on.

[bitter jewel]

also wobble x could be used. im running them right now. on the lead screws. what kind of pivot do you have for the bed connection to the arm, similar to hydra?

[bitter jewel]

also another consideration for room would be to remove the top extrusion that the bed rod mounts to and run a printed bed rod holder. currently have this setup. makes it easier to do auxilary fans.

[crystal tangle]

essentially the only two constraints I had was the z linear rod, and the leadscrew, meaning the linear bearing could twist according to the leadscrew which brought a whole other level to leadscrew aryifacts

[bitter jewel]

but the wobble x and oldhams should decouple the lead screw which even if they are bent no artifact should show correct? or am i just missing something

[amber comet]

exactly the same as hydra, kossel balls

[amber comet]

well that certainly could create artifacts, in my case though, at least in the assembly, there can be no movement of the mounts because they are all constrained by eachother.
Essentially each corner is on double linear rods, and these linear rods allow movement if one motor is moved independently, but due to the fact that the two motors that are in a perpendicular line to the motor that is being moved, it cannot create any side to side or front to back movement in the bed

I will draw it because I don't think my explanation is great

see, each of these motors can move independently

to move the blue mounts, the green mounts would need to be able to move in that same direction as the green mounts want to move

but they cannot

so these mounts will remain static as long as there is no difference in Z height of any of the brackets

see, the z motors are offset from eachother, the mounts that tilt the bed in these two corners have moved in the direction of the bed

[amber comet]

an extreme example that will not be possible due to the bed hitting the mounts, but I hope it kind of shows how the system operates

[amber comet]

okay, will look into that. I have ordered the required parts - without a MIC6 bed, it all cost about 104 dollars

[amber comet]

once I mount all these parts I will check the viability of this setup and think about creating more space on the top

also something that I have to address because these pictures will be confusing, my assembly has 400mm TR8s, while I ordered 500mm screws, so that's why my motors are offset by 60mm, otherwise the leadscrews would be in the path of the XY axis. That's also why it looks like the build volume would decrease.
Currently the hypothetical height the Z axis could reach would be around 430mm's - will see how it all works out once I have it mounted, but as long as it will retain 400mm, I think I will be happy.

[bitter jewel]

I have a mic 6 if it matters

[amber comet]

My MIC6 was also shipped recently

I designed the brackets to use the default mounting points of the bed

so despite the stock bed being awful, it can also be used

[opal yacht]

One thing I learned during my mechanical design job is how to adjust tolerances, thermal expansions and the related solution:
I would

1. always go with 1 „fixed“ point (locking x,y and maybe even z) - to make sure you have an origin or reference point of the bed positioning - which will be always the same and you can depend on. The other mounts should have the chance to adjust for movements, expansion and tolerances. Your concept concerns me are moving on the bed plane freely during print and could lead to layer shifts by self-adjustment.
2. a plane is perfectly defined by 3 points - if you „over-constrain“ it with 1 additional points it’s best to make it not same strong as the other 3 to allow some flexibility.

[amber comet]

I will first build what I have designed now and then will go to troubleshooting but to address your points:

1. Theoretically what I have right now is stable so the bed should not shift layers. Will see in practice though.
   As for the fixed point, I do not think Hydra has a fixed point, does it?

2. I do see how the 4th motor could cause problems, if that turns out to be the case I will scrap the Quad Z concept and transform everything into a Triple Z concept just like hydra has, but rather with the linear rails and bearings (IF leadscrews won't produce artifacts - if that turns out to be an unfixable, or a rather expensive to fix problem, I will move onto a Belt Driven Z system)

[amber comet]

aliexpress is sadly very slow with shipment recently so I do not even have half of the parts that I require for this build, I suppose I won't be done with it earlier than early December

[bitter jewel]

so when can others expect testing lol

[amber comet]

Well some parts still haven't arrived (love aliexpress lol) but I'd say do not expect beta files before Christmas! i will first need to print the files that I have right now and then find issues, issues that can be managed/fixed easily or issues that can only be managed by swapping to a 3 axis system - if that turns out to be the case, the system needs to be reworked, and it might take some time.

I refuse giving you guys a beta project that barely works, so that's why it's taking so long, I hope you understand.

[rich star]

Are the little slides rigidly bolted to the bed? If that joint can rotate from vibration I could see the arms rotating around the smooth rods. I see these 4 point beds and I wonder if it could in Pringle a stock bed.

[amber comet]

the slides utilize Kossel Balls

so they are not rigid

[hexed hawk]

I would love to try this out, do you have files for testing yet?

[rich star]

This thing has been has been stuck in my brain for two weeks. So I drew it out to figure out what my brain doesn't like. I hate when people come in and crap on my designs but i think this is a problem. I think the build plate can shift.

[bitter jewel]

wouldnt the lead screws prevent rotation? especially with top bearings.

[bitter jewel]

when you anxious for new things.

[crystal tangle]

We never intend to hurt when giving advice, most of us just have made the same mistakes 😭
I did already talk abt this, but this design does have some things that could affect it 😅
the fact that the bed tilts on 4 different points could over constraint it
the biggest thing though is the choice of linear rods as the linear guide. the cylindrical mate of this guide forces the leadscrew to be the second guide. due to the way that leadscrews are designed, they often have inaccuracies/bends, and when this is coupled to our z, it transfers not only vertical z motion, but also x/y motion. This is the reason people decouple their leadscrew nuts, use wobblex, or use oldhams coupler to avoid transferring x/y movements to the bed.
In this case, since the linear rod allows for rotational movement, we will have issues in every scenario:

In the first scenario, our leadscrew nuts are coupled to our z. alrhough this may be somewhat stable, this also means all the errors in our leadscrew are transfered to our bed, and our print. multiply this by 4, and you get the most extreme case of z wobble.

In our second scenario, we decouple our nuts. in this case, our bed will constantly wobble with the lack of constraints and will be less stable than my mental health.

[rich star]

My 5+ leadscrews are not nearly rigid enough to counter that load. Maybe if you upgraded to 10mm leadscrews with top and bottom bearings, but that's kinda against the spirit of this mod.

[rich star]

I think over over containing the bed would have the benefit of being able to remove some hyperbolic warp from the bed. Not sure that's worth it tho.

[crystal tangle]

It would likely bind the axis before removing warping of the bed by stretching it (assuming i understand what you mean)

[bitter jewel]

its mainly just resisting the rotation of the motor movement? since all 4 lead screws will be supporting that movement.

[crystal tangle]

Yesnt. 10mm screws doesnt mean the screws will be more straight, what it does mean is that it will be more rigid

with a low rigidity leadscrew (like 8mm), if you constraint it at the top and the screw has any twist or bend, it will be exaggerated and bend the screw more.
with a high rigidity leadscrew (like 12mm), instead of bending itself, it will likely bend the frame, which is arguably worse.
A compromise go this is Helical Flexible Shaft Coupling which will eliminate most of the innacuracies on a rigid leadscrew, but still requires a linear guide with no rotational movement

[rich star]

I believe if the bed is pushed horizontaly, all 4 arms will rotate and let the bed move. In that processes it will deflect the z screws.

[crystal tangle]

assuming they are PERFECTLY straight, yes, but in the real world, it will be the cause of rotation

[bitter jewel]

well yes your not wrong but the amount of deflection would probably be basically none unless you have nozzle strikes. obvioulsy i understand what you are convaying. its always the comprimise in engineering. i know my lead screws arent straight.

[rich star]

Yeah, I think that the leadscrews won't be enough, only a test will prove it one way or the other. We'll see when @amber comet gets parts.

[amber comet]

no! sadly not yet, the last things I need is to print the parts, but sadly my Ender 3 is occupied right now, but I suppose this week I could get to printing them finally! Then I will test everything

[amber comet]

it can't

it's constrained by linear rods

as you can see on this picture, there's no way for the bracket to rotate, so what you drew out is practically impossible I think

but of course I will have to test for that

so thank you for the question

I'd rather have someone bring up an issue themselves, it's better than me not realising that something could be an issue later on

[rich star]

I thought that the bed attachment could rotate. Maybe I misunderstood.

[amber comet]

the corners of the bed itself are attached in a way that does allow pivoting

but that pivoting is only affected by the other brackets

you guys do bring up valid issues that I could encounter so I really appreciate that

I can't wait to print the parts and test them myself

[rich star]

Ah, I understand. Cool

[amber comet]

currently there are two issues that I am concerned about, because they are quite unfixable:

1. Overconstrainment of the bed by the 4th point - unfixable in any way, the system would have to be reworked into a 3 axis system (basically a poor man's hydra)
2. lead screw wobble could impact the print quality and transfer to prints - although possibly fixable by something like oldham coupling or wobbleX, for the price of 4 of these you might aswell just get hydra parts lol, and let's say the system does get reworked into a 3 axis system, absolutely no point to fixing this by oldham coupling - just print hydra at that point

[crystal tangle]

Good luck 🫡
im excited to see

[amber comet]

one part prints 8 hours on my E3, damn, this will take a lot of time and filament

[crystal tangle]

@amber comet take a look at this message. if you decouple the nuts using wobblex or alternatives, the bed will be way too underconstrained

[amber comet]

even if I fail I still will be happy with this project as this is my first big 3D printer mod, that the concept of was enough for me to feel accomplished with

[amber comet]

yes that's why I believe that lead screw wobble will kill the project

becuase even if you do "waste" money for these devices

you will have a bed that can shift into any direction

[crystal tangle]

im planning on making hydra which uses the stock bed. sometime soon™

[amber comet]

I still do have some little hope that maybe 4 linear rods will counter linear bearings wobbling

but if not, eh! At least I already have a hydra bed, and I will have a reason to buy linear rails for actual hydra

and if this project does in fact get killed, I will leave this channel for future makers to see that in fact it was never worth it to cheap out on the Z system

[bitter jewel]

my doomicron is ready to print test parts.

[amber comet]

actually

sure

I could maybe release these?

Okay I will send the files here, at least I won't feel like I am forcing you guys to wait for something that might never come

I will zip em up and send them

absolutely no instructions though!! I haven't had the time to even think about these

so I will probably also zip up the assembly of this project for you guys to have some reference

[crystal tangle]

ye, i already attempted this w stock bed and 3 points. same exact idea. thats how ik lol

[amber comet]

BEFORE ATTEMPTING TO PRINT THESE ^
READ THE DISCLAIMER IN THE FILE!

No BOM, you have to guess it from the assembly, sorry!!!

If any questions rise up, just DM me or even ping me here - I don't mind!

[bitter jewel]

does the step have the smaller rod size and bearing? also are the balls the same as hydra

[amber comet]

yeah the balls are exactly the same as hydra

I am not sure I understand what you mean by smaller rod size and bearing?

I designed it for 6mm rods

[bitter jewel]

thats what i ment. the 2 rods that hold the tank part and there bearings. havent looke for a source yet

[amber comet]

oh well aliexpress should be good

but I suppose any kind of smooth 6mm rods will do

i think the ones in the STEP are 100mm

[bitter jewel]

probably going to due amazong cuz speed, lol have a reference for the linear bearing for the 6mm rods? these work? https://www.amazon.com/uxcell-LMK6UU-Square-Flange-Bearings/dp/B07H8X9MFM/ref=sr_1_15?dib=eyJ2IjoiMSJ9.6klncWo62krIVLlbhJiXC_WVKA8wdzwVT2g7NJ248NgYCV1zIam2akO1iTvvk8QUrbwpg6FNkJVa6koJeIyKGTlLRKaPsqieF_rA4eezeOJAAMxw0qfEzLugEUDofjmpZziIjp34rY57UYNKifhiVVA6FpW8xlNaMVTODMLf87ciUMBWi-_57UxaHNICaa1NHjVU51WZTLqaTkEkuyKVGZ-mcamtSnT34_xxyFiVnunjRVGmCJL95KYqjtm2weTtdYxiu_wj43jA7yPQnmvd6uv_GskFGYW3ZEp-0sf22gc-q-Z1M5avG4Ywc64Cx2iiCkjCcsN5aqBPdglSbQPX4JuPb671Uyaa0GslEiq_cs8.TDS0K90fsxGWVPvF28Nsz-f3zX25rL67ZOHrtTmWyl0&dib_tag=se&keywords=6mm%2Blinear%2Brod%2Bbearing&qid=1734558695&s=industrial&sr=1-15&th=1

[amber comet]

I think LMK6UU are the exact ones I have but let me check my aliexpress order history

Yep LMK6UU

[bitter jewel]

Sweet.

[bitter jewel]

just need the balls 6mm rods and the bearings right?

[amber comet]

shit sorry, I don't get notifications about this channel for whatever reason.
Yeah! kossel balls, magnets for these balls, rods and bearings! Also some countersunk M4's for the bed but I suppose you can just reuse the stock ones.

[bitter jewel]

Magnets what magnets lol. ill have to look

[amber comet]

kossel ball magnets

[rich star]

Was reading through the Kipper docs, I did not realize quad_gantry_level was a standard setup. Faith increased. 🙏

[crystal tangle]

well, in that case, you do need 4 z axis, which is not recommended with leadscrews

[bitter jewel]

getting parts in will start build after christmas

[amber comet]

hey just letting all the interested people here know that I am still alive and the project is in fact not dead

didn't find the time during holidays and now some tougher times came and I really couldn't find the peace of mind to work on this project, also my Ender 3 broke so I couldn't even print the parts lol.

I do plan on working on this again soon but I first need to be done with exams in university, so it might take around a month.

[spark carbon]

Ive just found out about your mod. So this makes it possible to reuse the old "bad" bed any get a similiar experience as with hydra ?

[spark carbon]

Ive just ordered some new filament, this might be a good chance to give this a spin

[spark carbon]

right, so i´ve been going this chat. Does that mean you stuck to the 4 axsis or got back to the drawing board and went with a 3 axsis system? I´ld love to give this a shot on my E5+ Mercury One.1

[crystal tangle]

i would advise against it personally

the stock z system on e5+ is already pretty good and solid

[amber comet]

Sorry, I kinda forgot about this chat. Due to issues with my old printer I kinda had to shift my focus to other things and my Merc has been sitting untouched for months now

[amber comet]

I would honestly advise going with Hydra, this system probably won't give you any more improvement over it and considering the fact that I haven't even tested it myself, I can't in good conscience recommend it

And the stock system on the E5+ is pretty awful

Especially when you have a warped bed

[amber comet]

hey have you actually finished the build? I am wondering because I am thinking of returning back to the design as I might've found some ways to improve it, but I haven't found the time to actually print and mount the stuff myself due to university craze