So last week I scanned a quarter. I did it twice, once at a low resolution and once at 0.002 inch. The high resolution scan took about 16 hours to complete, or at least that is what it told me it would take before I left.
Scanning is very easy. You have to have the scanner tool on the mill, and remove the protective cover off the sensor, (a metal plate that is screwed over it). Make sure the cable is on the USB2serial connector and run the Dr. Picza software. You set the place to scan in the software, start it off and then wait for a LONG time.

Milling is done by running the Modela player software. The help feature will guide you through all the steps and is fairly easy to follow. The last step is a bit more tricky, you need to go into the layout and set the locations you will be milling, as well as show the height of the material to be milled. You can do a surface mill at this stage as well, which will make the surface of the area you are milling level with the bed. Setting the height of the material is done by moving the bit down with the up/down buttons on the mill itself. You want to just touch the tip to the material.

The following picture was done by first doing a surface, then doing a draft print, then doing the fine print. The surface and draft print took a very long time to finish (a couple of hours in total), however the fine print was done raster fashion and took only 20 minutes. The software views the mill like a printer, so you’ve got all the wonderfulness of print queues etc. If you make a mistake, deleting the item from the print queue will not be deleted until the “printer” asks for more data. It takes data in 4K chunks and depending on how long the lines are a chunk can take 15 minutes to complete.

There are 4 or five bits available for the mill, however they are all 3MM so your choice is actually pretty limited, but they do have different head shapes. But even with 3mm you can get some pretty good results.

Made with 3MM bit.

All in all, I think this mill is a perfect addition to Kwartzlab as it is easy to use, and has a number of nice features that lend themselves to a hackerspace. One of the features that I liked is that when setting up the job, the material you are milling is available through a drop down list, (if it’s not on the list, then don’t use the mill for that material).

My goal in this is to learn about motor drivers, distance sensors, and have a platform that I can use for experimentation.

To that end, I preformed a frontal lobotomy on a remote control hummer. The car was originally around $20, so even if it went up in flames I wouldn’t be out that much (initialy).

The first step was to take the outer body off. Then snip snip the brains are out. The mechanics of the car seem to be really well designed, the electronics not so much.
Next is upgrading the battery to a lithium-ion battery instead of the double A’s. I’ve done this with other projects. I use the VPX’s which are a 7 volt battery used in home power tools. They are no longer in production, so I snarfed up the last available from Home Depot. They are great because I’ve got a charger made for them, they are relativily small, and they don’t put out enough voltage to make me want to wear rubber gloves.

The next step was put in a motor controller, as well as an arduino. Both were bought from solarbotics. The motor contoller (a fun little project in itself) is mounted above the battery, and the arduino above that. The motor controller has LEDs that show the direction of the motor and when the motor is turning, and the motor controller can control two motors, so 4 lights in total. These got attached to the outer body.

The next thing added is a distance sensor. This is the first sensor added and where the software starts. My first project with my new platform was to make it stay a certain distance from whatever is in front of it. Through experimentation, I’ve found that I need to have a delay in the loop, otherwise the poor thing just sits there nervously shaking, decisions happening faster then it can react. I then tried from memory adding PID code, which gave it a rather agressive stance. It would run at an object, slow down, then back away from the object, over and over. I believe that I’m getting closer. wikipedia has some psudocode for PID controller

previous_error = setpoint - process_feedback
integral = 0
start:
  wait(dt)
  error = setpoint - process_feedback
  integral = integral + (error*dt)
  derivative = (error - previous_error)/dt
  output = (Kp*error) + (Ki*integral) + (Kd*derivative)
  previous_error = error
  goto start

That i’m going to try soon.

One problem that I noticed was that the usb connected to the Arduino was affecting the readings, so I could get things nicely tuned on a stand with the Arduino attached, but as soon as I disconnected the PC and ran strictly off battery power, the settings changed.

To that end I’ve added a bluetooth addapter. I’m excited about this, because I realized that I can do some major cpu programming on my laptop and the little Arduino only has to handle the I/O. I’d like the whole thing to be autonomous, but for now this is good. Maybe my next one will lug around a laptop.

The blue tooth adapter works fine on the Arduino, but getting the laptop to talk to it ended up being rather difficult. Finally got a device to work at /dev/rfcomm0 and if I cat’d that device I’d get all the data that the Arduino was spewing out. (Distance measurments and so on).

A trick I discovered the hard way is programming the Arduino can’t be done while the Bluetooth is online. Another switch to come for easy Bluetooth disconnect.

At this point, the RC Car has:

I’ll do more updates as I progress.

First the pi is very picky about it’s usb. I don’t know what the differences are, but it apparently only like high quality keyboards and mice.
Second the pi is very picky about power. I had read on the forums about this so I actually went out to get a power supply for it. (USB charger that does 5V, exactly).

The PI’s video can either go through composite or hdmi. I opted for hdmi (the family tv) since I haven’t had composite around for at least a decade.

There are three distros.
The Debian squeeze
The Arch
and the QtonPI.

Several problems with rebooting, and messages about usb’s disconnecting. (The peripherals that I used are ones I use all the time, so I think it’s the pickyness showing up again).

Of the three distros I was able to get two of them working, the QtonI (which in spite of being for developing qt stuff didn’t include X.) and the Debian.
The Debian looks like it is the go to distro. It has X, some games and looks to be generally working.

I wasn’t able to get sound out of the hdmi, and alsamixer wasn’t present so I don’t know what the status of sound is.
Mr. Weber pointed out that this thread has some possible solutions to audio hdmi problems.

It seems that power is the main concern with this board, on the main page they say use 5v 700ma powersupply at least, but in the wiki they say 5v 1000ma at least. I’m using a 5v 1000ma.

He also has USB business cards as well as other projects.

Should be interesting.

In general the CES was a disappointment, and I probably won’t make a special effort to go again. There were a lot of 3d TV and it looks like Sony is going into it in a big way. I also saw 3d Software which will take two pictures and merge them into one 3d image, (presumably to use on your 3d tv).

The Parrot AV drone blew me away. The display was a cage mesh (which appeared to be for keeping the drones from flying off into the neither regions of the show, but were probably more for keeping people from putting their grubby hands over the drones). Two drones floated in the center of the cage, and would randomly jump up or down 2 feet or so, in sync. I asked the fellow there about it and he explained that the drones run Linux, and were running a pre-programmed routine. They go for $299. I’m seriously thinking of getting one.

There was one item that I thought would be useful in particular for Kwartzlab to sell as kits.

Ladybug robots. These little kits go for around $20, and are from Japan, although the creator’s family lives in Ottawa so there are possibilities of importing from there. The robots run on a toothbrush drive system, (vibrating motor, attached to a brush) and run surprising ly well, doing line following and music beeps. programming is done by running them over black and white bar code things, which can be printed, and rearranged as needed. Advanced programming can also be done via the USB port. A closer look can be seen at http://js-robotics.com/ and http://js-robotics.com/PDF/JX-TTM-Flyer-English.pdf Pretty impressive tech for $20. Looking on the site, the retail looks to be 4980 yen or about $60.

I’ll bring in the info I collected next time I come in, and leave it in the reading area.

I’ve written to the company explaining a bit about Kwartzlab and what we do.

They’ve written back and sound enthusiastic, and are asking about quantities.

I’ve replied to them that the project I would be interested in pursuing is a drop in replacement for the SG-5010, (a generic brand servo).
Specs available here

I gave him my thoughts on this project, which at this point are very blue sky, but I’m hoping he can quickly say that the project is feasible or not.

My thought is to stack multiple double sided stators to make a gearless servo, as well as have an encoder to determine the position. Your product is well suited for this, in that 1. It is gearless, 2. is forgiving to external forces, 3. additional electronics can easily be incorporated on the same board. 4. I has a low footprint, allowing it to be stackable, 5.once it has reached it’s position it can be turned off (assuming friction will hold it in place), thereby making it a much lower power unit then a typical hobby servo.

My hope is that this could become an useful product that kwartzlab could produce, either as a kit form or as a completed package. These would be of great interest to maker groups like ours, as well as hobbiests looking for a lowpower accurate quiet servo.(In other words everyone who has bought a hobby servo and wanted something better).

I’ve asked the company producing these PCBMotors if they think the project is feasible or not. I don’t want to spend time on something that is clearly unworkable. I suspect it is feasible but not cost effective, until large enough quantities are obtained.

• In (input from sensor, or program)

• Or (the maximum of all inputs)

• And (the minimum of all inputs)

• Xor (the difference between the maximum of all inputs and the Minimum of all inputs)

• Fuzzy Logic Block (a range of outputs mapped to a range of inputs)

• Out (the resulting value)

The program allows you to adjust values in the Fuzzy’s and the Inputs and will show the resulting values at each component’s output.

Where this becomes useful is in robotics, where you want to quickly determine output levels according to the inputs.

My FuzzyBuilder program saves it’s datafiles as include files (.h) which when linked with a fuzzy.c file produces the desired effect.

So what does this have to do with QT you ask? Well my first version was written using Embecardaro’s Rad Studio (c++ version) and while it works well enough it is a windows only solution. I want to be able to run this on Linux as well, so it made sense to recreate it using QT http://qt.nokia.com/products/developer-tools which is a cross platform solution.

My biggest problem so far has been understanding QT’s graphic system. I want to have the same functionality that I have on the windows system where I can move components around, and draw connections between them. I’ve been spending some time learning about scenes and view ports and such. Last night I made some progress and was able to finally make a blue box, which knows when it’s been clicked.

I know, not exciting. Just a programmer doing his thing, but to me it’s a wonderful breakthrough!