UPDATE:

It turns out that my mill is not perfectly perpendicular to the board. The error comes out of the fact that the end of the tool for the engraver mill and the drill is much different and when a small rotation is added to the mill head a linear difference is observed.
I adjusted the head rotation and now both the milled and drill points line up perfectly.

Tools>Options Menu, Display>Navigation tab
“Preselect in geometry View”
and
“Prehighlight faces for preselection navigator”

Easiest way to set this backup up is to run the client software from a windows machine. I have a Win7 VM running 24/7 anyway, so that is a perfect place to setup the CrashPlan software.

To get the CrashPlan software to read and backup a NAS drive, read the following links:

http://virtualj.net/20120126/use-crashplan-automatically-backup-network-drive-nas
http://lime-technology.com/forum/index.php?topic=17672.0
http://www.midnightspark.com/blog/?p=116
http://support.crashplan.com/doku.php/recipe/adopting_another_computer

They all talk about the same technique.

http://www.engineering.com/CATIACommunity/CATIAForum/view/topic/postid/91/forumid/25/tpage/1.aspx

Unfortunately there is no straight forward way to do text withing the part modeler directly. Oh well, but this got the job done.

My first concern was how thin a wall the printer could print at different slice thicknesses, so I made a test part that integrated short pieces of different thickness wall sections.

The same test part was printed at both 0.15mm z slice thickness as well as 0.4mm slice thickness. The top row consists of wall thickness of 0.2mm, 0.4mm, 0.6mm, 0.8mm and 1.0mm thickness. On the 0.15mm z slice thickness model, the first two (0.2 and 0.4mm) walls got ignored by the software. At the 0.4mm Z slicestep, the first 3 walls (0.2, 0.4, 0.6mm) got ignored.

Just because the walls got printed, it doesn’t mean that they got printed at the correct thickness. On the 0.15 z slice thickness part, the 0.6 and 0.8mm walls were measured to actually be around 0.8mm thick. The 1mm wall was measured to be about 1.1mm or so. On the bottom row, the 1.2, 1.4, 1.6, 1.8, 2.0mm walls all measured correctly within +/- 0.07mm which I’d say is pretty damn good.

On the 0.4mm s slice thickness part, the thicknesses had more variation (around +/- 0.12mm) in the wall thicknesses, and the wall thickness of the walls up to the 1.2 and 1.4mm walls, was not accurate at all. I have not measured all the parameters on this part.

The spacing between walls of the same thickness was designed to 5mm , and on the 0.15mm z slice thickness part, with the exception of 0.6 and 0.8mm walls (which were not the designed thickness), all the other walls came out pretty close to 5mm when measured; around +/- 0.05 or so.

The other thing to note was that at the 0.15 z slice thickness, the walls on the 1.2 and 1.4 mm walls was only made up of only an outer perimeter bead of plastic, with no fill between them. You can kind of see what I mean on the picture above on one of the 1.4mm wall. The 1.6 and 1.8mm walls seemed like they had a fill in between the outer perimeter skins, however the fill didn’t seem like it was connected very well to the skins. What this produces is 2 skins that are not connected together and loose I would think the 1mm wall would be stiffer (since the 2 sides are fused together) then the 1.2 or 1.4mm walls. I’d avoid designing with these thicknesses.

Once you get at the a 2.0mm wall thickness, the infill looks like it connects the sides together. These observations were made purely by looking at the final top layer of the print. I don’t know how well connected everything is inside the wall.

For those that want to print their own part, here’s the stl file : Thickness_Calibration_D01.zip

UPDATE:
Just found a post on the PP3DP forums talking about the same thing (HERE). In the same thread there is also a handy thingiverse generator to generate thickness calibration parts similar to mine, but parametric (HERE) and a chart of designed vs printed/measured wall thicknesses.

Setup:

TOWER: this is the file archive NAS. It’s running on  UnRAID software and is running inside its own separate case. It currently has 5TB of storage

NAS4FREE: is a zfs raid NAS that’s running as a virtual machine inside the ESXi 5 server. It has 2 TB of dedicated storage in raid 0 configuration. These two 2TB drives are dedicated to this VM exclusively, and are being accessed as raw devices.

Results:

The results I get are 66MB/sec write and 52MB/sec read

The results I get are 52MB/sec write and 70MB/sec read

The results I get are 25MB/sec write and 66MB/sec read

I ran the first set of tests from another VM on the ESXi5 server inside a Win7 VM using a software called “LAN_SpeedTest 1.2 (Free)“, to use as a benchmark to compare against. The test ran on the Win7 local drive. This local drive is actually a file on the ESXi5 server, and not a dedicated drive, so speeds are not reflective of raw disk access but are close.

The second test was run to the NAS4FREE server. Here I was getting excellent read and write speeds of 52MB/s write and 70MB/s reads. I can certainly live with that, as it appears to be faster then the local drive. This high speed data transfer is due to the fact that both VMs (both the Win7 and NAS4FREE VMs) are running on the same physical computer, so there’s no “network slowdowns” to speak of as all communication is internal to the same physical computer.

Third I ran the test to the TOWER NAS machine. I know this machine has a slow write speed due to the nature of the UnRaid software used, so I’m not surprised with the 25MB/s write speed, however I was surprised with the read speed of 66MB/s. This was taking place over Gigabit LAN, and the physical computers are right besides each other. At lease I know that over short distances I can nearly saturate the gigabit LAN to give me near native drive speeds. (I would imagine that native drive speeds would be about 80 or 90MB/s)

I don’t remember what transfer speeds I was getting from my workstation to both the TOWER and NAS4FREE NAS machines, but it was significantly lower then those numbers.

UPDATE:

Running the test from a wired laptop on the other side of the house, through CAT5e cable I get:
(Test is done using 2000MB file)

For comparison, to local HD, I get Write: 89.2 MB/sec and Read: 76 MB/sec
To the NAS4Free server, I get Write: 61.1 MB/sec and Read: 47.2 MB/sec
To the TOWER server, I get Write: 24.7 MB/sec and Read: 75.8 MB/sec

These speeds seem very reasonable to me. Nearly as fast as a local hard drive.

UPDATE2:

I thought I’d test the WiFi speeds from a couple of computers on the network with this tool.From the old T41p thinkpad, with a dLink DWA-121 (wireless N) adapter I get about 1.6Mbps both up and down. Pathetic. Can’t even watch videos properly as it needs to rebuffer every couple of minutes.
Put in an older DWL-G650 card, and tested again. With the laptop on my lap, I get Write:3.5Mbps and Read:4.9Mbps. If the laptop is 12″ away from me on the table, I get Write: 9.5Mbps and Read: 13Mbps. I am about 50ft away and one floor up from where the router is currently located.
For reference, the best speeds I have achieved on the G WiFi network at home is from a T61p laptop sitting right besides the wireless router. Those speeds were 16Mbps write and read.

Then last evening it came to me. The light bulb came on.
Use the milling machine as the printer! Just get a print head, and control it from the existing software (Mach3). The work area is huge, and the cost should be quite reasonable since I only need the extruding head. Between a good quality plastruder, and a temperature control board, it should be about $250 ($170 for the plastruder, stepper motor driver and $60 or so for the temp control board)

Feed speeds of a makerbot are around 40mm/s (2400mm/min). Currently I can do about 1000mm/min on my milling machine (which is 1/2 of the makerbot speeds) but with a ballscrew upgrade I should be able to do about 5000mm/min feed speeds (83mm/sec).
I have read about some kit 3D printers (like the Ultimaker) that can do 200-300 mm/sec feeds (12,000-18,000mm/min). I know my mill will never go that fast.

Resources:

Mach3 as 3dprinter software (here)

Good info on whats required in order to use Mach3 as the river software (here)

Everything is printed on demand, shipped to the customer, and I get the royalties. Easy.

Talk about inDesign templates (here)