Deltas: Kossel versus Rostock MAX versus Cerberus versus Kossel

First of all: Hello!

I recently got back into the topic of 3D printing after following it a couple of years but having to concentrate on other things. As I never learned too much about electronics it had been quite difficult to dive into this topic for me.

Now is the time.

I am specifically intrigued by the family of Delta RepRaps originating from the Rostock. As the objects I'd like to print are rather larger than smaller, I do not plan to go for anything the size of the Rostock Mini (for now), but intend to build either one of the following four. Unfortunately, even though I have looked through tons of material, I am stuck right here as there are several things I dont understand, I have opinions about, etc. And now is my time to ask.

Kossel: used to be my favored choice, until i realised that the three types available are either the mini (handy, but no), openbeam kossel pro (in kickstart, so no) and something called "legacy kossel" branded "deprecated" in reprap wiki.
Hasnt got the heated base, as it became custom with the cartesian ones. (I know you can change around later on, but for now I'd like to keep it a little more simple than redesigning the whole thing.)
What I liked about Kossel was the idea of the fishing line. it seemed not to have any backlash at all. => But is that such a big problem compared to others?
One more thing I liked is, that the function of the carriages (along the vertical beams) seemed logical to me. Liked them more than the Cerberus design.

Looks like Kossel is less for the beginner and rather for the tinkerers that have significant experience. So maybe in half a year or so.

Rostock MAX: Was intriguing for a while. Unfortunately it seems to have some heat issues (can provide the link if I find it). I would prefer a less closed of system. + I dont want to order too many parts from the US

Cerberus: Also no heated bed. At least its not in the design goals. I still wonder why.
The carriages on the vertical beam have a rather strange setup, but I would do with them. The base construction seems far away from the stability of MAX or Kossel, but I may be mistaken. I simply think the compromise in stability for using selfmade parts shouldnt be made there.
The one think especially hard to understand is the use of the motors packed tightly together. Why? I always heard of skipping and heatissues. But why would you pack them together? (No rhetorics, I would really like an answer to that. Honestly.)(Maybe for heating the glass?)
Besides all those points, and as this posting is of inquisitive nature I cant say "negative points", I love the implementation of the fishing line in this delta.

...Well... now the last idea is already my first project.

Still. I hope to get some answeres to thos questions posed above and hope to be able to make a final decision soon.
Keep up the good work! Its inspiring!

What size prints do you need? If you need tall prints all of the above would be awesome. I haven't really looked in detail at all the specs of the above printers but I can assure you that they all will need some TLC from you to make them sing. You will be able to get amazing and crappy prints on all of them. It is all about calibration and getting to know your printer.

That said...to be self promoting...if you need super large print volumes a RepRap Simpson can be scaled very large (My undersided prototype has a 10L print volume.) with standard electronics. Its print volume is much wider than it is tall. Of course, I am probably months off from even offering a beta kit. However, all the arms are very simple and can be made quickly by anyone with a jigsaw and a drill. A table saw and a drill press would be better. Simpson will also need a lot of TLC to make it sing for now.

Side note: The autobed leveling that Johann has put into the kickstarter Kossel is awesome and highly advised. If you want a printer that is ready to rock with the least amount of tweaking on your part then that is it.

The buildvolume for all options is about the same. But why ditch the Kossel/Rostock? I'm building the Rostock Mini but it can easily be scaled to as big as you want. Some have made it 150cm tall. Even the XY size kan be adjusted. You just need to alter some parameters in the firmware. The mechanics stay the same. I think there even is a OpenSCAD file which calculates all the required measurements.

On the auto bed-leveling. Yes, this is aweseome and the sliced bread of 3D printing as far as i'm concerned but besides the video's and the sourcefiles there is no info on it on how to implement it. I'm waiting for that to.

The rostock bed is entirely static, so what in the design do you really need to be able to put one on? Just sandwich a heater of your choise and call it a day.

I built my mini Kossel to 300mm x 250mm (although I'm only printing 200x200x180 ), using Misumi linear rail and Aerostock style mag ends, and the only part that I had to modify/design myself was the carriage (because I wanted perfect screw match to my rails and integrated belt attachment that the Aerostock carriages didn't include) Everything else just dropped together completely unmodified. IMO it was far, far easier to build than my first printer (original printrbot)

I haven't built a Cerb, but for any of my personal friends that may want to get printers of their own I would not hesitate to reccomend a Kossel for them as their first build.

What heat issues with the Max? I just bought my kit a few weeks ago and got it up and running. Extruder is great, but if you haven't used a Bowden before it'll take some time to figure it out. The bed takes me about 15 minutes to get to 120c, but there ways to speed that up fairly easily.

As for the print quality, it's fantastic. The seemecnc forum is also 2nd to none for support with deltas...

I agree with cope413 on the support with deltas, it seems to be the only delta with actual support truthfully. The Kossel is more of a bullet train of iteration and advancements.

I liked what the Cerberus Pup did with the carriages that I'm using it for my Kossel.

Laton Wrote:
-------------------------------------------------------
> Cerberus:
> The one think especially hard to understand is the
> use of the motors packed tightly together. Why? I
> always heard of skipping and heatissues. But why
> would you pack them together? (No rhetorics, I
> would really like an answer to that.
> Honestly.)(Maybe for heating the glass?)

I do not have Cerberus myself, but there is one obvious reason why to put motors as far away as possible from the pulley which turns the spectra line upward. It is to minimize nonlinearity between motor turn (in degrees) and carriage movement (in milimeters).

The point is that the line is being reeled on/off a drum which is on the corresponding motor shaft. This moves the point where the line touches on the drum vertically (up/down) from the center position on the drum. This movement changes the angle at wich the line bends around the pulley at the corresponding tower base. I do not know exactly how Cerberus is done but in a simplified situation lets mark the maximum angle between a horizontal plane and the line from the drum to the pulley as α, lets the distance from motor shaft to the pulley be marked as d, and lets the maximum (up/down) movement of the touch position (of the line on the drum) from the drum center be marked as x. Then x is a fixed constant (depends on the drum diameter, fishing line diameter and the cerberus height). The introduced error to the carriage movement is something like e = (d - d*cos(α)). And α can be computed as α = atan(x/d). Investigate function e(d) = d - d*cos(atan(x/d)). You want it to be zero. Clearly it happens when d→0 and when d→∞. Since d=0 does not make sense you want it to be as big as possible. Hence the motors are in the center.

> Hence the motors are in the center.

Why not have the motors at the next "leg" to maximize the length d?

TBog Wrote:
-------------------------------------------------------
> > Hence the motors are in the center.
>
> Why not have the motors at the next "leg" to
> maximize the length d?

They would cross in the center. Maybe you can make it so that the filaments do not touch each other in all valid cariage positions. If you mind the small non-linearity that much then just make the drum/spool diameter smaller at the sides so that the ciscumference decrease is just as big as the error term (d-d*cos(α) ... or slightly different expression if the cerberus geometry is not just the same as I assumed).

What size prints do I need?
Approximately the size of a A4 Sheet of paper possibly more. I tried to get away from the small printers as quickly as possible. Though that its not going to be plug and play was kinda clear. Calibration seems to be the main topic quite a lot of times.

10l? Sounds amazing!
I try to go as big as I dare for now, as I am not able to calculate eventual swinging of beams and such, thus creating inaccuracies which are not necessary.

@nicholas.seward
At this point, its gonna be a Delta. I didn't rule out Simpson, I was simply not aware of it and the progress it has made. You might be able to tell me though how far along your idea with the newly designed arms is? I went through the thread and felt that there is far more to come...


Worst case scenario includes mostly the same electronics in another projects, while I figure out what I did wrong with attempt one.

Edited 3 time(s). Last edit at 09/04/2013 08:32AM by Laton.

I wanted to go with the spectra line. I really appreciate the idea of using it. Besides it being a little harder to work with, are there any other drawbacks about it?

Thats also the reason, why I did try to avoid the Kossel. The Kossel Legacy seems to have those features, but as I said, its my first 3D printer which I plan to make and to make a decision a quote on github facilitated the decision: "Work in progress: most of these parts are unfinished and will change. If you like to live on the bleeding edge, feel free to print parts, start building and provide feedback."

Any more points against the Kossel Legacy besides building and working with it seems to be rather more complicated than what I had intended.

Considering that the Kossel can easily be scaled up, I think I will go with that one. Just had the opportunity to talk to an engineer from our factory and he supposes it to be a little more reliable. I think I will do with that for now.

[groups.google.com]

This is probably the best place for Kossel support at the moment.