MarkForged Kinematics vs Hbot vs CoreXY

Hi everyone,

It's time to upgrade my faithful Prusa Mendel with something more rigid and capable of producing larger objects; something in the order of 200 x 400 x 300mm (xyz). After some research, I was settling on designing and building a CoreXY machine - until I saw the configuration of the xy axes in the MarkForged Mark One, in this promo video by Onshape.

Here's a screen capture from that video that's got me excited by the minimalism of their configuration:


Below is a diagram I created of the Mark One, (accuracy T.B.C.).
It's mirrored in the horizontal plane compared to the image above,
to make it easier to compare it with the diagrams at corexy.com and fabtotum.com.


I believe it's a simplified Hbot configuration, and that it likely overcomes any potential racking issues by virtue of the stiffness of the attachment of the x-axis to the driven side of the y-axis; on the y-axis driven side there appears to be two slides supporting the x-axis, vs one slide on the other side.

So far I can't find anything more about the history or merits of this configuration, so I'm really interested in discussing those.

Here's some initial thoughts:

Pros:
- simple!
- belts for each axis don't cross.
- belts remain coplanar.

Cons:
- requires super-stiff (expensive) THK rails or similar, but these can be bought relatively cheaply on eBay (second hand from South Korea).

Edited 1 time(s). Last edit at 05/01/2015 01:43PM by hairykiwi.

---

Cheers,
Hamish

Checkout RichRap's Sli3dr. I don't know where his layout came from but it is also coplanar but has the added benefit of being antiracking.

[richrap.blogspot.com]

---

ConceptFORGE
Wally, GUS Simpson, LISA Simpson, THOR Simpson, Sextupteron, CoreXZ

Thanks Nicholas.

I follow RichRap pretty closely and even had a good look over his Sli3DR design at the TCT show last year - very nice - and super quality output.

I won't rule out building a Sli3DR, but I'm still interested in better understanding the pros and cons of the MarkForged belting configuration.

Speaking of RichRap's machines - does anyone know the name of this (pure Cartesian) one spotted on his blog recently:

---

Cheers,
Hamish

that "MarkForged" configuration is not so good.
why? becouse lets name it y carriage is moved by belt on only one side.
so you need very expensive hiwin rails, and machined bridge to avoid flexing when moving carriage.

in reprap style that will not work, it is hard to say if its ok on this demo machine with high speeds and acceleration settings.


i suggest core xy as better alternative.

wheel is allready invented.....

and one more thing, when i see those thin Z axis linear rails i can not understand that manufacturer seriously.
i hate to see when people are removing part from bed, and then everything flexes 10 mm....

Edited 3 time(s). Last edit at 08/09/2015 05:43PM by woo.

After dealing with a standard hbot configuration for a pick-and-place application, I see one thing that's really nice about the MarkOne's hbot variation. Since X/Y map directly to M1/M2, you can home the gantry by homing one motor at a time. The issue that comes up with the standard hbot config (in my application) is that you have to move both motors simultaneously. The timing between the moves isn't much of an issue, but if you have a motor / controller failure during homing that goes undetected, your end effector then starts to move diagonally. Depending upon your system layout, now you could potentially run into other object in your workspace.

Thanks for your thoughts guys.

Sorry for not seeing your replies earlier - I was distracted from my original 'next build' plan by hacking the Mendel90 design for the last few months for myself and a friend; I've lost count of the number of times I've fallen in and out of love with OpenSCAD.

I little off-topic but related nonetheless, I found out from a UK sales engineer that THK now have a cost effective option in the HSR rail range called 'Ct':
Although their 317E HSR-Ct brochure says "...we offer an LM Guide series for reasonable prices that meet the price demand in the market for general conveyance systems and the like.", looking at the included graph of tolerances, I think it's likely to be plenty accurate enough for RepRap machines - and possibly offer an improvement in stiffness over 'traditional' RepRap linear shaft configurations when paired with a suitable frame.

Here's the quote I was given for purchasing this product through THK UK distributor OEM Automatic:
1 off HSR15C1SSCT+500LCT - £53.67 ea
2 off HSR15C2SSCT+500LCT - £71.62 ea
1 off HSR15C1SSCT+270LCT - £41.84 ea
(First bold number indicates number of sliding blocks, second bold number is rail length)
All ex-stock 1 to 2 days
£13.95 delivery
Ct rail can be cut to length locally, compared to the 4 - 6 weeks waiting for rails to be made to order in Japan + airfreight + onward ground freight charge.

Wait a second, the MarkForged's config still requires both motors to move purely in Y (vertical in the diagram). The belts and motors themselves aren't coupled, but the movement of the carriage is, so throwing extra stiffness at the gantry isn't necessary. It should actually be easier than an Hbot since it has only one side of the torquing effect. I'd say it's somewhere between hbot and corexy for functionality, without any additional length of belt over hbot.

Edited 1 time(s). Last edit at 01/25/2016 11:33PM by brazenrain.

Quote
brazenrain
Wait a second, the MarkForged's config still requires both motors to move purely in Y

Off course, but the "X" motor only turns to clear space to allow the Y movement to happen. I don´t like the design either. woo already said enough. the prototype has very expensive linear rails to compensate for this "off centric"
momentum - ladden movements.