Custom Delta/Quantum Delta Hybrid Kinematics

turtlegod42

@o_lampe A quantum delta is like a delta, but the bed moves.
The bed is very light, lighter than the toolhead actually, as it’s only 65mm in diameter. It’s a fairly small proof of concept printer.
Currently I am planning on using endstops for the gantries.
The benefit here is that theoretically, each side only needs to go half the target speed, so overall high speed printing will be much easier to accomplish.
That’s what I’m building this for - to test it.
It also just looks cool.

o_lampe

@turtlegod42 said in Custom Delta/Quantum Delta Hybrid Kinematics:

The benefit here is that theoretically, each side only needs to go half the target speed, so overall high speed printing will be much easier to accomplish.

It'll sure look cool, but you can reach the same double speed goal by replacing the idlers of a conventional Delta with steppers. Two steppers on each belt will give you twice the torque, hence higher acceleration and top speed. (Although the bottleneck lies in the extruder IMHO)
Plus, you could have a full size bed and no hassle with the kinematics.

Speaking of kinematics: in your case, when all the rods are equal length and the bed and effector are same size, you can probably specify the extra motors (eg. M584 X0:1 Y2:3 Z4:5 E6) and just invert them

M569 P0 S0 
M569 P1 S1 
M569 P2 S0 
M569 P3 S1 
etc

turtlegod42

@o_lampe thank you for your response!
I mean, there’s no way to know for sure until I test it. But, I didn’t want to create the same delta as everyone else. It’s also cooler in my opinion even if it performs worse.
Wouldn’t I also need to figure out a method to split all the movements between the two sides? I mean, I was kinda thinking to just halve the rotation distance, but that feels like a bit of a cheating solution.

o_lampe

@turtlegod42 said in Custom Delta/Quantum Delta Hybrid Kinematics:

I was kinda thinking to just halve the rotation distance

You could halve M92 values (steps/mm)

turtlegod42

@o_lampe Yeah, that’s what I meant by rotation distance. Sorry, I get all the different names between firmwares mixed up.

dc42

@turtlegod42 if you want each delta mechanism to move half of the commanded amount, I can see a couple of ways of doing that:

Tell RRF that everything is twice the size that it actually is, i.e. in M665 set the rod length and delta radius to double the actual values.

Or:

Use M579 to set the scale to 50%.

RRF supports more than 3 delta towers (up to 6), however this isn't documented n the GCodes dictionary. To use extra delta towers, create the extra axes using M584, then in the M665 command provide as many L values as you have towers.

turtlegod42

@dc42 Very interesting, thank you. Can I treat 3 extra towers as an additional gantry?

dc42

@turtlegod42 not without some firmware changes.

turtlegod42

@dc42 Good to know!
If I just take the approach of specifying the additional motors and inverting them, along with using m579 to set the scale to 50%, would there be issues with delta calibration/homing of the two gantries? Or since they’re identical would I only need to calibrate one side.
Thanks!

dc42

@turtlegod42 that depends on how accurate your build geometry is. RRF normally calibrates the delta radius, endstop corrections and tower positions. It won't be able to calibrate all 6 endstop corrections or all the tower positions if you pretend it is a normal delta with 3 towers.

shaunppl

@turtlegod42 The only problem I can foresee would be that with the kinematics producing high speed directional changes in the 2 opposite directions therefore introducing momentum instability would mean that without counter contra movements of the same mass and speed the vibrations would be extremely high, Unless I am mistaken. plus the changes in mass would not be a constant due to the process of actually producing a part on the build plate would alter the mass and therefore the kinetic changes in acceleration and deceleration would mean that the counter contra masses would also have to change at the same rate ??.

I am not trying to pour boiling water on your project but at hight speed I do not think the kinematics would stand up, and you would be subject to very low speeds well below the most slowest of printers now available, but I do admire the concept.

If I am wrong then I apologise and would be thankful for any insight on where.

have a look at this PDF on Kinematics.

https://khazna.ku.ac.ae/ws/portalfiles/portal/16917665/20177.pdf

turtlegod42

@shaunppl There is definitely a possibility of issues like that. The reason I’m building this at such a small scale is because it’s purely for proof of concept purposes - to test if that will be an issue present.

o_lampe

@shaunppl The table always moves in the opposite direction from the effector. To me it seems almost balanced? Better than a classic delta at least.
There might be a little problem when print height increases.
My deltas always had the lowest print adhesion compared to other printers, because I couldn't squish the first layer onto the bed.
A moving bed might cause the part to tip off, plus there's a bit more air draft.

turtlegod42

@o_lampe Yeah that’s kinda what I was thinking. My only concern is that it should theoretically create a torque around the center point, so it might try really hard to fall over.

dc42

@o_lampe said in Custom Delta/Quantum Delta Hybrid Kinematics:

My deltas always had the lowest print adhesion compared to other printers, because I couldn't squish the first layer onto the bed.

Why not?

shaunppl

@o_lampe It will only be balanced if the mass, speed and direction of both parts are of equal properties with a counter mass to stop the tumbling effect, that's why 2 cylinder engines have balance shafts to counter the torque difference, the kinematics act in the same way and dictate the masses momentum of acceleration / deceleration you are still going to get to the point where it will be greater and greater the faster the movement and whether this be at a centre point or not the mass has to be a constant for it to work at a higher speed that's why nearly all delta printers are fixed bed designs but they will still create high vibrations from the heads acceleration and decelerations and DC42 has done incredible work on the software development side to help absorb a lot of these effects when setting up delta printers.

I don't have a problem with getting a good first layer with any of my delta printers in fact it used to be bloody difficult to get the print off of bed but since using a glue based medium they just pop the bed now when it get below 50 degrees but will hold steadfast until it cools below that.

I wish turtlegod42 all the best and hope that his prototype proves me wrong and I really mean that.

o_lampe

@dc42 said in Custom Delta/Quantum Delta Hybrid Kinematics:

@o_lampe said in Custom Delta/Quantum Delta Hybrid Kinematics:

My deltas always had the lowest print adhesion compared to other printers, because I couldn't squish the first layer onto the bed.

Why not?

I blame it on my own effector with an 8mm offset to fit the sherpa mini. I admit it wasn't the most rigid and accurate design. Although the calibration is spot on, I have to make compromises regarding Z height. There is a slight effector tilt I can't get rid off.
Unfortunately, I had a big (~200GB) data loss last year and all my earlier designs and selfwritten programs are lost. I'd have to start from scratch.