Novel 5DOF printer design (Helios)

[grizzled dew]

gotta lock in on college apps soon but ill get it built by october for sure

[grizzled dew]

update

[umbral trail]

I would try to make the cantilever as rigid as possible and add way more reinforcment to the Z axis base to keep it from bending as much as possible

[grizzled dew]

yeah i agree that tiny bracket is not gonna do anything

[blissful narwhal]

i feel like using a combined extruder+hotend is a lot of weight to be putting on the pivot at the end of cantilever like this. Personally i'd be considering bowden

Either actual bowden with the extruder stationary on the top of the Z-Beam, or a N14-based Microbowden toolhead where the extruder remains stationary and just the hotend + fans tilt

[half lotus]

what keeps you from making this into a full-framed bedslinger?

[tardy fox]

looks like this https://youtu.be/VEgwnhLHy3g with a 5th bedslinger axis to make it non polar

pretty cool

[desert river]

dual shaft y motor like that is a fun idea

you should add double shear

but

it’s neat

[grizzled dew]

The extruder must be able to rotate +90 -180 degrees so a bar going across would collide with the print. Has to be supported on one side only

[half lotus]

nope. still not getting it

[grizzled dew]

Yeah i love that design but tried to make mine have bigger accessible nozzle tilt area so hence longer beam

[grizzled dew]

Like imagine prusa but the nozzle can rotate. If it rotates above horizontal then the gantry is gonna be below the nozzle and get in the way

Cut off half the gantry and it works

[half lotus]

are you talking about the top of the frame?

or the x gantry?

[grizzled dew]

I couldn’t find what you are referring to online? Link pls?

[grizzled dew]

X gantry

[desert river]

The X beam would hit the print

Because the hotend doesn’t stay perfectly up and down

With this setup the hotend is fixed to the end of the cantilever

So the X beam won’t get in the way

[half lotus]

ah, so you'd need to drop the hotend down somehow so it could be able to point upward

anyway, you might want to consider cpap or shop air to reduce weight as opposed to a fan

[desert river]

well yeah but that’s not what I’m saying

[grizzled dew]

I was thinking this but the fan is like +50g max

[desert river]

here

[half lotus]

I wasn't really considering printing downward yeah

[half lotus]

at the end of a cantilever, and on your image, also at the end of a second cantilever perpendicular to the first

[desert river]

you need to minimize weight as much as possible or you will have issues

[half lotus]

lightweight hotend, lightweight extruder, whatever the meta is now

[desert river]

ideally if possible you should have dual vertical towers and have the horizontal axis between the two

top down

blue is R red are the uprights

[umbral trail]

I would not use that heavy ass biqu thing at all to be honest. Put the extruder on top of the Z extrusion and feed the filament to the hotend with a bowden tube. That way you don't have to worry as much about weight on the end of the cantilever

[vocal walrus]

Your Y stepper mount will bend causing the belt to wear down quickly. For Z you'd want to have 2 carriages on a single rail instead of 2 rails. Your Y assembly could benefit from an extra carriage as well to resist the nod.

[grizzled dew]

Novel 5DOF printer design (Helios)

helios is a chatgpt name and its perfect in every way myth lore wise

[livid sequoia]

very cool

[grizzled dew]

CAD done

my school is letting me do independent block for this but theyre debating on money

[tardy fox]

independent block?

[half lotus]

Does it need to be that long?

[umbral trail]

adding a balanced counterweight would probably be more space efficient but I guess model for the render just uses a standard extrusion lenght

adding a "counterplate" to the wheels and the pulleys could easily increase the stability of the arm without much work I imagine

Here

[vocal walrus]

But counterweight can't be balanced here unless you add another sliding assembly

[umbral trail]

True, didn't think about that

[grizzled dew]

Shenanigans during assembly

[ionic thorn]

Post the portal turrets song ❤️

[grizzled dew]

Gotta get back to work lmao

[ionic thorn]

Okay okay last one. Green hill zone

You have to, everybody does it

[grizzled dew]

Almost done building

Figured out IK algorithm yesterday

[half lotus]

how does it determine if moving the bed side to side is better than doing normal polar bed stuff?

or visa versa. you could treat it as a standard cantilever bedslinger, except sometimes it might be optimal to rotate the bed, but when?

[desert river]

Because it’s not being treated as polar

It’s being treated as an additional rotary axis

It’s not replacing anything it’s adding on

To answer how you determine what’s best

Frankly there isn’t automatic slicing past 3 axis

So you manually tell it to do what you want

[ionic thorn]

There are some developed algorithms for 4 axis slicing that utilize deforming a 3 axis slice of the object to remove any overhangs. Not sure how you’d go past that though

This video explains it pretty well

https://youtu.be/VEgwnhLHy3g?si=EgXr5xVgeaISil9z

So technically not a slicer but an external program that uses g-code from a 3-axis slice and optimizes it so everything can be printed without supports

[desert river]

Oh i thought it was still fairly manual

[ionic thorn]

Nah, unless you consider dragging and dropping from one program to another to be manual

Effectively it’s no different from normal 3-axis nonplanar, but the 4th axis makes it so the nozzle is always held perpendicular for the best line adhesion

[half lotus]

but what do you do with an extra axis?

[grizzled dew]

Theres only 1 solution to any input of 5 position variables xyzcb

But yeah if the point is too close to the center of the bed it could sweep out a massive angle for a small translational change

Speed isnt an issue though

[grizzled dew]

That video is nice but i wonder if it can be generalized to 5 axis

Because his printer basically MUST point extruder towards center of bed while it can change elevation angle

[grizzled dew]

Adding fifth axis makes it so you can place the nozzle anywhere and angle it in any direction

4 axis gives you full translational freedom but you only can spin in a single angle

[half lotus]

I get that much

well, the 4th axis was obvious

[ionic thorn]

Yeah it is really something where you have to combat the conceptual “how and why”

Cause yes a 5th axis gives extra freedoms, but also WHAT extra freedoms is difficult to figure out

And how to actually use them properly to do something that can’t be done with something less complicated

[half lotus]

it is easier to conceptualize when all degrees of freedom are on one object

and makes it more obvious that a 6th degree of freedom is unnecessary when the nozzle is radially symmetrical and homogenous... or however you call something that doesn't change if you rotate it.

I dunno. I studied biology.

[ionic thorn]

Very fancy belt knot but damn that’s impressive

[uneven flower]

Looking forward to see the first print

[grizzled dew]

thanks!

[grizzled dew]

1220mm gt2 belt coming from china rn lol

[umbral trail]

Looks like a fun ride

Awesome work

[desert river]

damn you did this fast

where’s the toolhead 👀

[half lotus]

I think it's gonna be a biqu toolhead. relatively cheap and compact option as direct extruder designs go.

I think most designs won't work great, since they have fans to the sides, which in this configuration would potentially be rotated into the print at extreme angles. Seems it’s actually pretty good for purpose.

[ocean mirage]

Looks really cool, impressive progress so far!

[grizzled dew]

yeah that exactly

[grizzled dew]

school didnt want me to have "fire hazard" in my room 😔

so its a fake extruder

[ionic thorn]

Any concern about torque forces or do you expect the Y axis to not move very fast

Cause my worry with a toolhead that big hinging on a single point is the amount of wobble that could be caused by torque

[half lotus]

oh uh, you know, it might even create a non-negligible twist in the y gantry (he labeled the extendy arm the y. dunno if that's normal for polar designs).

[ionic thorn]

the extending arm is the y axis and the toolhead spins on a separate rotational axis

So my thought was when the arm goes in and out for the Y axis it would create torque

[half lotus]

torque with what as the center?

[ionic thorn]

That

That's how cantilevers work

[half lotus]

the arm going in and out isn't going to make much torque there

but it may change the amount of twist in the gantry member

...we might be talking about the same thing

[ionic thorn]

Torque as in the rotational force created by an object at the end of a motion arm

The arm going in and out creates torque on the extended toolhead

Due to it sticking out away from the point where it connects

Meaning that back and forth motion on the Y axis will cause the toolhead to create a torque force

It's the same as how a toolhead on a CoreXY can cause torque if it isn't balanced since the back and forth motion on a Y axis can twist the gantry

[half lotus]

you can't do anything about there being a point mass at the end of the lever that's moving in and out. You'd need a reciprocal arm that would be useless weight and added complexity. You can mitigate the downward bend by stiffening it, which I guess you could do by using a HFSF5-2020, or if the pulleys at the ends are bigger than 20mm in diameter so the belts won't rub, you might be able to use a HFSC5-2020. I was concerned with it being off to one side, which would introduce a clockwise twist on top of the end drooping downward.

[ionic thorn]

My only comment was that it's a heavy extruder

I know perfectly well you can't not have it there

I was just saying going with such a big toolhead was more likely to be an issue

[half lotus]

I think he went with it because it's cheap for an all-in-one toolhead as well as slim on the sides.

what is the meta for a light toolhead here, assuming you have to go with a direct drive?

[vocal walrus]

I assume you mean direct extruder and not direct drive
But well, remote cooling for both the hotend and the parts cooling would be a step in the right direction, there were some lightweight hotends over the years as well

[half lotus]

Yes, a direct drive extruder. A 3010 isn’t a lot of weight to add. The extruder stepper is the heaviest thing that could be moved off of the cantilever easily. A couple of small fans is small change.

[vocal walrus]

Direct drive extruder like the one on stock Ender 3?
Remember that weight isn't everything here either

[half lotus]

Direct drive extruder as opposed to a bowden extruder.

[vocal walrus]

Ender 3 had a direct drive bowden extruder
Prusa Mk4 has a geared direct extruder
These are 2 different things and can coexist

[half lotus]

You know what? I think this is another “cartesian” thing.

[umbral trail]

Direct drive extruder: The motor directly drives the filament, no extra gearing etc. See Ender 3 extruder
Direct Extruder: Extruder sits on top of the hotend

[desert river]

a direct direct drive extruder directly directs directly directed filament in the right direction

[grizzled dew]

the h2 extruder is like 200g total with fan and ducts and its obviously not built for speed

also at this point stuff like slight bending is a second order concern behind actually making it print

[umbral trail]

I mean you could probably quarter that by using a hotend with a short bowden extruder but yeah, actually printing should be a higher priority

[ionic thorn]

I’m not saying it won’t work. And I know the goal is to just make it print. I guess I’m just looking at things on an accuracy and longevity standpoint. I’m simply looking at how it has the potential to make the Y axis a diagonal line simply because of how cantilevers work, and then the additional cantilever of the toolhead. I don’t at all think it wouldn’t work, and this is an incredible project, I’m just applying my own perspective. And yeah the H2D toolhead is massive but it also isn’t being hung from a cantilever on a cantilever

I’m trying to provide talking points or ideas because I do want to see this project succeed.

It might be beneficial in the long run to maybe design in a lighter extruder and cooling system, potentially via a Bowden and CPAP setup, but for the time being it definitely will print

[grizzled dew]

i think having a double gantry would be the next step. it would get rid of that second cantilever you're talking about

might cause more collision issues though

[ionic thorn]

That’s actually something that the addition of another axis can help with because you can keep the printer from running into the object on the bed even if you don’t have it as a cantilever

But I don’t have a possible clue how you’d implement that into the motion algorithms

Normally on a polar printer you can’t have a normal gantry and HAVE to make it a cantilever because of collision

But having an X and C axis makes it possible

BUT it would entirely change how the printer functions on a fundamental level and wouldn’t work with how your motion system works right now

The toolhead would have to be able to rotate a full 360 degrees

Or I’m just spouting nonsense bullshit I’m still figuring it out in my head but it wouldn’t be worth the sheer amount of complication it would add

Cantilever is just way easier to implement

[ocean mirage]

What do you use for slicing?

[half lotus]

A better question is if he’s using an existing slicer as a base or is doing it all from scratch

[grizzled dew]

using orca for getting toolpaths

but then those are deformed and the angle of approach for extruder is calculated

[grizzled dew]

yeah no way im making a slicer

[grizzled dew]

overhang correction rotation and diffused corrections

[umbral trail]

Just tell a company about your project and ask if they would like to sponsor you

[desert river]

wait, are you running with only the pulley hubs on the shaft?

[half lotus]

bro needs to get smooth idlers

[grizzled dew]

Wdym

[grizzled dew]

Lmao i would have had to buy another item

But had a bunch of the toothed idlers lying around

Can’t imagine it would be fatal

[desert river]

the hub of the pulley is where the setscrews are

does the motor shaft not extend through the whole pulley

[grizzled dew]

Oh

Yeah only the thick part with screw is on there

[desert river]

[grizzled dew]

Would have had to got long shaft nema or make mounting plate thinner

[desert river]

that’s not good for your motor shafts and significantly reduces possible belt tension

[grizzled dew]

Oh i bet

[desert river]

long shaft motors are not very expensive

[grizzled dew]

If it breaks I’ll get those

But tension is solid rn

No noticeable deflection

[half lotus]

https://www.amazon.com/RLECS-12x8-5mm-Aluminium-Suitable-Machines/dp/B0D7CCF9ZW/ literally less than $6 and it's a straight upgrade on prospective print quality.

[grizzled dew]

Yall are so accustomed to speed optimization and surface quality lol

Its a 5 axis printer the best thing i can do rn it make it work

[half lotus]

it's just an idler bro

I'm not telling you to double shear your motor shafts

[desert river]

It’s not just about optimization and surface quality
It’s about proper design practices and manufacturer specifications (and like, caring about doing it right)

It’s not really even more work to do it proper

[grizzled dew]

Can I see your machine working?

[livid sequoia]

[grizzled dew]

Pls don’t nitpick the smallest things on my printer and say I don’t care about doing it right

I built this from scratch in 35 days

[livid sequoia]

its just so easy to replace them and get better print quality, not destroy your belts, and be able to run spec tension

[grizzled dew]

Thanks but realize print quality is like 10th on my priority list

[livid sequoia]

if it was something like completely redesigning part of it, id understand, but this is literally replacing four idlers

[grizzled dew]

This is a 5 axis printing research

Not about the idlers thats more reasonable

[livid sequoia]

its so easy

[grizzled dew]

I know

Dude i had to cut bom in half because of denied funding so everything counts

I happened to have them lying around and didn’t want to wait another day and all that packaging waste

Sure you’d do the same

[livid sequoia]

smooth idlers are literally a dollar on aliexpress

maybe as a temporary solution but id fix it

[grizzled dew]

If you’re thinking about “print quality” at the conception stage you’d never get anything done

I don’t know why we’re arguing about it

[livid sequoia]

also belt health and spec tensions

[grizzled dew]

“Its easy” sure but its its even easier to use what i have

[livid sequoia]

gonna cost you more in belts than the price of two smooth idlers

[grizzled dew]

Id appreciate critique on my algorithm more this

[livid sequoia]

i do printer design not slicer design

[grizzled dew]

If you could find something wrong with the collision penalized arap optimizer that would be more helpful

Yeah i can tell

[livid sequoia]

ima stay with the stuff i know

[grizzled dew]

I’m sure i could go to your machine and find 10 things wrong

[livid sequoia]

my machine is meh and i know it

[grizzled dew]

Dude if the belt ever breaks ill let you know

This is like worrying about aerodynamics befote you build a car prototype

[livid sequoia]

i consider both in my designs

[grizzled dew]

Show me a printer you designed operating then

[livid sequoia]

currently working on unfucking klipper for my polar

[vocal walrus]

Should've used RRF

[livid sequoia]

thats the new plan

[desert river]

Show me this printer printing.

I’m done interacting with you. Good luck with the build.

[half lotus]

why is it fuzzy?

did you tune it?

definitely cool tho

[livid sequoia]

i think it’s time to consider print quality

[grizzled dew]

https://tenor.com/view/idfk-gif-3761179052566277881

gang the printer barely works

95% chance it crashes and instantly fails

[umbral trail]

Its a start

[unreal zealot]

So there is a 5% chance it works!

[ocean mirage]

hahaha

Getting there!

I think it is lovely, how did you slice the model? @grizzled dew