buildlog.net

How do these two approaches compare in print performance. I see a lot on how fast delta machines are. It doesn't seem to me that the could be as accurate as a conventional printer. Is this assumption likely correct?

The benefits of delta are overstated. Its primary advantage is looking cool.

Claimed advantages include no backlash (since gravity preloads the actuators) but this is no longer true at the high accelerations of a 3d printer. There is a relatively low mass being moved which contributes to its high speed but there is more to a 3d printer than just speed.

I am not a huge fan of the nonlinear kinematics since it means that many accuracy errors (for example steps/inch, backlash compensation) will vary over the working volume and are very difficult to measure and correct effectively.

Accuracy doesn't really matter for 3d printers since you are smooshing plastic out of a nozzle which is an inherently inaccurate process. The real contributor to surface finish/print quality is repeatability which essentially translates directly to rigidity. The arrangement of the various axes will have basically no effect on print quality when compared to the quality/rigidity of its implementation. A flimsy printer will make bad prints.

In essence, there are good printers and there are crappy printers. Being delta won't make a crappy printer good, or a good printer bad. IMO there are fewer bad delta printers because the designer at least had to know some basic geometry and they tend to be more expensive.


For what its worth, I do not believe delta has any inherent advantage over traditional systems. 99% of cnc, pick and place, laser, plasma, router, etc... are not delta for a reason. Attached are two prints from a commercial belt-driven gantry FDM system.

Very informative post, thanks. Nice prints too

r691175002 wrote:A flimsy printer will make bad prints.

In essence, there are good printers and there are crappy printers. Being delta won't make a crappy printer good, or a good printer bad.

This. Except applied to any printer construction and geometry currently out there (excluding the obviously unstable or otherwise bad designs, but those die out naturally anyways).
Deltas do have the advantage of a light actuator, but there are cartesian bots that do too (Ultimaker style designs).

Also, there's more advantages to a light actuator than speed. Acceleration and jerk being one of those (and not because of maximum attainable speed, but because of extruder physics).

Cartesian is the word you are looking for.

delta type machines are good for things that need to move quickly since they have reduced mass. They are usually used for things like sorting. The basic design is significantly less rigid than a cartesian system and thats why you dont see them much in the cnc world short of special machines that have all sorts of position capabilities.

You can make a lightweight actuator cartesian bot too.
Just not one that can do all three axes with that light actuator.

Which is largely irrelevant in case of 3D printers, because high Z speed, acceleration and jerk are not needed.

A Delta can have much lower BOM item count. They look complicated, but they really have very few parts and the axes are totally identical.

What Bart said - I can build a much more "accurate" delta for a lower cost that an Cartesian, and it will print faster. One of the current limitations of the delta firmware is the virtually nonexistent floating point....

The above said - the cost of increasing the Z build volume is lower on a delta as well.

I just want to know who's eating all those pancakes