On 6 Oct 2012 , at 6:08 PM, hellphish wrote:
> I'm not sure if the accuracy of the commercial printers is due to mechanics
> or due to them controlling both the plastic and the slicing engine. Could
> be they have modeled the way their plastic behaves particularly well and
> can slice better because of it.
It's likely two things
1. An plastic formulation which doesn't shrink as much after cooling, and
2. A factory calibration process
Indeed, 2 can compensate for 1 when you constrain the operation to a small
set of known plastics (as in the case of the Mojo).
Folks who assembled their own bots are familiar (or should be) with the
process of calibrating the bot's steps per mm in the x, y, z dimensions.
There's what it theoretically should be given the stepper motor, timing
belt, and pulley characteristics and then what you actually get in practice.
You make a few prints and tweak the steps per mm in the machine definition
file and then you get a very reasonable dimensional accuracy.
But with the Replicators, MBI has not seemed to call attention to this
particular process which users might do themselves. (Indeed, when the
accelerated firmware was first introduced, the axes lengths and steps per mm
were hard coded in the firmware. That's recently changed and from
looking at github, RepG can now transmit that info from the machine
definition to the bot's firmware. Hmmmm. Really good idea. Why didn't
we think of that for the Jetty Firmware ;) ;)
Anyhow, circling back: a Replicator 1 or 2 or even a ToM or a Cupcake
can produce reasonable dimensional accuracy. However, you have to take
the time to do some calibration prints to get their. I know for a fact
that doing that is part of routine maintenance on some of the big, high end
commerical 3D printers -- I've seen it in the manuals. They may do an
initial in-factory calibration on the Mojo's.
Dan