Software Integration

Since the hardware is deterministically approaching a usable state, its time to begin the Software integration...

To that end I have setup 5 ROS machines to begin creating and testing the first phases of getting a CubeSpawn machine into a virtualized State

 The 5 Physical machines have remote desktop sessions to cut/paste commands from the tutorials

The integration will focus on getting ROS and Rviz setup, then building a URDF for one of the Cartesian machines.

The integration consists of ROS, Machinekit, MTConnect, and possibly STEP-NC

I'll be focusing on getting some State machine info out of ROS and Machinekit via a very simplified MTConnect configuration, Since SwRI has built a MTConnect instance for ROS Industrial, there is a fairly solid foundation for that half of the channel...

More info here
http://rosindustrial.org/news/2013/11/9/ros-industrial-mtconnect-integration-program-completes

On the Machinekit side, it appears this has been discussed, but not implemented, however, I'll be dissecting the MTConnect/Makerbot repo
https://github.com/mtconnect/makerbot_agent

For while I have no love for Makerbot, a MTConnect to serial Cartesian device code base of some sort should provide a basis for retrieving state info, which should make hacking together a Beaglebone/Machinekit 3D printer example within the scope of this preliminary effort.


Thanks to John Morris, (A Machinekit Developer) I have this nifty System Diagram:

The "CubeSpawn Coordinator" is probably beyond my skills at this point, but if I can hack some basic communication FROM the machines with the available pieces, then writing a simple polling loop to push info to them should at least make some rudimentary control possible - then we'll bring in a professional.. ;-)

An' inch at a time...

The End of Prototyping is The Beginning!

Closing in on the end of trying out ideas and starting the control software integration efforts











The Electrical Module houses the control electronics for a range of Cartesian machines.

Although much module evolution is expected, this gets the idea out of our minds and into yours ;-)

Buildout

As with the last few posts, I am finishing off the part runs for 8 Cubes,

This leads to 16 modules: 8 electrical modules and 8 mechanical modules..

Five of these are 3D printers, one for the array and 4 for various kinds of testing

    

All the driveshafts

The four pieces of stock, and additional parts to make the "Z" axis

 All in all, quite a lot of grunt work to get all the setups together, cut the stock, prep the fixturing, write, then validate the Gcode, run the parts, deburr, drill and tap, deburr... I REALLY look forward to the automation part... ;-)







But grinding through this portion is helping to shape how the modular fixturing for the milling machine will need to look if a robot is going to have to set it up

4 sets of  modules and  frames

A lot of work, for an automated, form factor based, self extending, recursive, on-demand network of inter-operable manufacturing machines, but probably worth it... ;-)

Marching onward...

Since the design work is substantially over, ( for the first-cut, tested design) nothing to do but make the parts for the first batch of enclosures and modules...

Blocks for the backplane

Front and Rear side panels test-fitted

Cube module frames - stacked to conserve bench space.

A stack of module side panels

This amounts to (essential) grunt work, but is starting to bring the modular fixturing together and helping to define process steps and other less tangible elements of the part production process together.

I am excited about the near future, when the modular fixturing can start to come together for the mill, and the 2nd gen lathe design prototype can be built.

Much to do!

Chrono-insectile Tau depletion

Time Flies! but its not running out...

During the course of the last month several important steps forward have been taken:

the 3D printer has been running with all the minor issues one would expect cropping up

A nozzle jam, leading to the hobbed bolt eating the filament, some issues with bonding to the currently cold bed and all the usual growing pains for a new printer.

            

Meanwhile, another engineer has embarked on building the lathe module out in a preliminary form, having a partially complete mechanical presence on the bench has begun to inspire a whole list of issues to consider and better methods of handling a turning machine in this small space even before the controls and electrical system has begun...

Alongside that effort, the majority of the parts for the milling machine module are collected with a similar high rate of improvements being added to the design even before implementation has begun

Needless to say, the days have been full and long, and I have a lot of semi-completed models to polish up and add to the 3DWarehouse in the coming weeks..

NB #5 updated Git

A little afterthought on post #4

Here are some more specific details about the events and their consequences...

I added Git-LFS and added the model files back to the Repository last week.

The current state of the projects consists of this list of tasks:

Finalize all the details of the "Ultimaker compatible" 3D printer:

-Bowden Tube - since this printer design does not have an open space above it - the free form Bowden tube concept does not apply, I  will modify the Ultimaker printhead design to accommodate a connector oriented on the side of the head at the lowest angle that will still provide clearance, a shelf or thin gauge shield above the front-Y drive rod and the Right X drive rod may be necessary to prevent the tube and wiring  from interfering with the drives

X and Y
The drives, limits and all other motion features are functional and tested

Z
The pallet arms have been rebuilt, clearance for the pallet loader will need to be machined in the forward facing ends

Custom Pallet,
The removable pallet that carries the print-bed has been designed, but needs to be machined

I have allowed for a self leveling mechanism and for an electrical socket to engage power for a heated bed option, the mechanical lock-down latches are built and installed.

NewBlog #4 Ultimaker Module ++

Much toil this week in finishing out the "Ultimaker 2" Module.
I have encountered a few minor problems due to the constrained space to fit in the bowden tube - a misunderstanding on my part about direct drive extruders and 3mm filament resulting in a switch to 1.75mm filament for this printer and the attendant conversion of the hot-end and extruder feeder

The real intent here is to make it easy to transition an Ultimaker printer's original parts into an automated machine. Anyone who has operated a production machine knows the tedium associated with setting up the job on a machine, then baby sitting it as it performs some short run task.

So ultimately, getting the machine and its support services to:

• Set up the job (from a section in the digital template) Calibrate, Align, Level, check for supplies on hand etc etc

• Run the job (next section ;-) )

• Cope with failed runs, errors,  (?? are we there yet ??) and re-run or correct the problem if necessary... 

These 3 little bullets represent a lot of bundled concepts, none of which are beyond the realm of the possible so down the garden path we go...


While this represents an approach to the solution we will also have to accumulate these case specific  solutions into a database relevant to the model/version/type of machine and process, once a significant number of these type of solutions are accumulated, then training an AI to attempt to extrapolate new solutions should become possible as well.

But back to the real and current plans:

The next machine will be a better incarnation of the small CNC mill. The milling machine is kind of a swamp of extended detail, since a robotic mill needs a lot of support services to operate:

Stock prep, probably a few cutoff saws, a small bandsaw and possibly a cold saw

Fixturing, which leads to a requirement for a modular fixturing solution, which then leads to a pick and place fixturing cell to set up jobs and that means a machine-command-translation-template to get a fixtures requirements encapsulated in a meaningful way to translate the fixture "recipe" into placement motion. not to mention fastening the pieces down so a screw gun with torque checking/strip prevention/and the fastener handling - then there will be fixture component storage and transport.....

Tooling, this one is big, since a toolchanger is only the frontend to a tool setter, possibly a tool sharpener/custom tool grinder/profiler and a tool magazine for several different cells (Tapping, anyone?)

Needless to say staying focused can be difficult!!

And tomorrow, we shall do it again!! (Solve some small portion of these problems, I mean)