Fringe Robotics is officially two years old, this last year has been quite busy. Lots on new projects started.

CD industrial group has kept me busy with assorted odd jobs, I end up doing a fair bit of engraving and electronics, a typical example:

I designed the panel layout, Engraved the text in the black powder coat and milled out holes to mount everything. There was quite a bit of wiring in each of these as well.

Early in 2013 I built a set of digitigrade legs:

I am working with a fantastic costume/mascot builder in Vancouver to put these into a suit. Currently finishing up the design of an improved pair.

Back in May I started working on the gearing project with a couple artists in Calgary:

more info can be found at http://gearingproject.com/

I built the display base, milled out the carbon fiber accessories and got the laser cut stainless steel parts made.

Mid August was liquid nitrogen day, the best way to flash freeze peaches for smoothies later.

Currently I am designing a solar powered light trailer prototype for a new company called Lamplighter Solar, a partnership between my brother and I. Pictures of that will be on the Lamplighter site once I have that up and running.

The first year was all about setting up, the second year was meeting potential clients and starting all sorts of projects. This year will see further development of said projects and less time setting up and more time building. Should be an interesting year.

TTFN,

GrantM.

Been running the mill for over a year now so I have a pretty good idea of the performance and capability.

Mill Stats:

Travel Area(Work Envelope): 620mm * 620mm * 250mm (XYZ) (24.4in*24.4in*11in)

Table Size 26in X 26in Table is 1.75in thick. Table alone weighs over 300lbs

Spindle Motor: 2.2Kw 5000RPM to 24000RPM

Collet: ER20 (comes with 1/8, 4mm and 6mm collets) Also comes with a few endmills (1/8, 4mm and 6mm and a couple engraving bits)

Drive System: 750W Panasonic brushless AC servo motors

Hiwin Linear Ways – 4700lbs dynamic load, High precision class 0.01mm variation max. Running parallelism is 13u meter (0.5 thou of an inch)

Min step: 0.0005mm (10,000 steps/turn) (ballscrew 5mm/turn) (can be re-configured, mine set to 0.001)

Theoretical max speed: (servo: 3000 RPM) 15meter/min, 590IPM (mine set to 236IPM)

Crates arrived in good condition, smaller box has the servo controller, spindle driver and computer. Took about two months from initial payment to arrival.

Unpacked and moved into place. Considering this weighs as much as my car, we needed to use a forklift.

I opened up the covers and took a look at the internals. Pretty clean actually, no casting sand, decent quality casting and well laid out. Easy to service and disassemble if needed. Linear ways may be a bit undersized. Ball screws are metric 5mm/turn not the usual 5 turns per inch.

Panasonic servo motors, 1HP output. Better than I expected, high speed and ample torque.

Also matched with Panasonic servo drivers. They were not set up correctly and the wiring needed a bit of work, but it was well labeled.

Assembled and up and running, ran a few test parts then built the acrylic tie down table. Better to cut into that then the steel underneath. Nice shiny and clean, not for long…

I use the mill to create my own PCB’s, this is a breadboard adapter for a TQFP44. I may be able to go a bit finer with some effort. Pretty reliable accuracy down to 2 thou.

A still image from a mill depth test. Maximum depth of cut was about 1.5mm before it started to chatter. Need to upload the video of this.

So review of the mill so far:

The Good:

– The mechanical hardware is good. German ball screws, Panasonic servos, Panasonic drivers, good casting, solid steel.

– The mill runs smooth and is very quiet, all axis and especially the spindle.

– The work envelope is larger than most comparable mills.

– Decent accuracy

– Low maintenance linear ways

The Bad:

– Linear ways are a bit under sized, machine not as rigid as it could be.

– The coolant tank leaked and the pump couldn’t actually pump coolant up to the spindle. I ended up building a new coolant tank and used a new pump.

– The PC interface is not great, NC studio card didn’t work, supplied PC software is horrible. I replaced the whole works with Mach3 and a new breakout board (C11). Things are working well now.

– Cabling is labeled well and cabling organization is ok, could be better but it’s not a mess. I can see that they didn’t use a proper crimp tool for the connectors, someone just used a pair of side cutters, a small tug and the wire came out. I re-crimped them with a proper crimp tool.

– The Y-axis way covers needed to be moved to get the full movement of the axis.

– Seems a bit odd that the inductive limit switch for the y-axis is at the front and not the back. I like to “home” at the bottom left corner not the top left.

The mill  works well for what I use it for; prototyping work and small batch production. It can mill aluminum and brass without much problem but steel causes more vibration than I would like. If a smaller work envelope was acceptable I would look at the Tormach PCNC1100 as something in this price range, better rigidity and product support.

While milling some stainless steel I noticed that the spindle was vibrating a fair bit. After poking around with a dial indicator I found I could move the plate the spindle was mounted on by about 10 thou of an inch just by pushing and pulling on it by hand. The other axises I could only move about 1 thou.

I found the holes drilled into the plate to be awful, the holes were very over sized and the bottoms were not flat. I think several of the bolts were not actually fastening the plate to the linear ways. The cast iron was also a bit porous and the screw heads were squishing the metal underneath. Here they also had to grind the head on the socket head screw so that it would fit, some of the holes were out by about 1mm.

So began the process of building a new Z axis plate. This is the raw hot rolled steel plate, 1.25 inch thick. This plate weighed about 60 lbs.

Cutting the plate close to size on the bandsaw.

After cutting the plate I used a face mill to finish the top and bottom. The face mill is a bit beat up, it left deeper machining marks than I would prefer but the surface was flat according to the dial indicator.

After surfacing the top and bottom I squared and cleaned up the sides. Did this with a single full depth pass of a 5/8 endmill, the chips it left were like thousands of razor sharp needles.

The surface finish on the sides was quite nice.

Next step was to set the plate up in my milling machine and pocket the holes.

After a bit, all the holes are milled. Besides the two small holes in the bottom left corner everything was done with a 4mm endmill. I had to take this slow to prevent vibrations due to still using the old Z axis plate at this point.

Manually tapped the M8 holes for the Spindle mount.

Here is the back of the plate, I lightly sanded the surface.

Here is the new plate, installed. Fit perfectly. Now I can only move the dial indicator about 1 thou of an inch.

So far so good, there is less vibration and when it does vibrate they are lower frequency. Will be interesting to see how deep I can cut in a single pass now.

TTFN,

GrantM.

Today marks one year since Fringe Robotics started. It’s been a busy year, the vast majority of the time has been spent setting up everything from the paperwork to the cnc mill.

Besides a trip to Japan in Spring and a recent trip to Washington DC I have got the lab functionally up and running.

The cnc milling machine has been running beautifully:

The CNC mill, soon to never be this clean again.

For the hot days in August I picked up a 20L nitrogen Dewar (not pictured) and did liquid nitrogen day. We froze peaches and berries for later use in smoothies and I experimented with the nitrogen, submersing running motors and LED’s. I was thinking I would try milling otherwise difficult to machine materials while frozen, like rubber.

Some liquid science for liquid nitrogen day at Fringe Robotics. Seen here freezing some fresh peaches for smoothies later.

The latest project has been building shelves for the office:

Jig to hold the plates in alignment will being milled. Lots of aluminum cuttings.

Drill jig for lining up the holes, I used a steel plate for durability.

A dozen nice shiny brackets ready to be installed.

Here is what they look like all finished.

I could have just bought some cheap shelf brackets at wall mart but sometimes I like to do things the hard way and learn some new techniques.

So now with the lab up and running I can finally start on some new projects, fun stuff.

TTFN,

GrantM.

Thought I should post some pics of the lab so far:

The electronics lab so far and if you look closely you can see the lab cat.

This is the laptop I use to do most of the CAD work for now, if you look closely you can see the cat.

Setting up the milling machine, it’s all up and running now, as a bonus if you look closely you can see the cat.

The neural net board that everyone who sees my office asks about and if you look closely you can see the cat.

Picked out a CAM package, went for SprutCAM. I tried out a bunch of CAM software but SprutCAM “thinks” the most like me, so that’s what I went for. I also prefer the way the simulation mode is handled. It’s very easy to toolpath PCB’s and other complex foam shapes. One of the first things I milled out once the full version of Sprutcam arrived was a few tie downs:

Brass, Aluminium and Copper. I typically use the brass ones, I will try out the aluminium ones and the copper was just for fun. The copper one was a pain to tap, very sticky metal.

I really need to do more updates. Much has happened, the lab is built, the mill is up and running, my monster computer is the fastest I have ever used. I am just in the process of picking what CAM software to use and I can start creating parts. I have finished the CNC table, this is V3 and the best so far.

Table