I found this Imac 17″ 2.1Ghz dualcore with a defective panel. I thought that such a nice machine should not die like this and that is why i decided to rebuild me a Mac with the guts from this one.

Next step ,disassemble the imac 17″, -you will find plenty of videos on how to do that on youtube.

Every part inside the Imac is connected with nice long cables so you can rearrange everything very easy.

I did that and mounted everything on a 9mm white plywood 250*600mm with use of plastic spacers and plywoodscrews.

Here you see the Wlan antenna and the Bluetooth antenna, they also have long cables and can be placed away from big metalparts.

There is a buildin webcamera with mic which happens to be an USB camera. I have not yet desided what to do with it.

The only 3 cables i had to extend was the power ON cable and the 2 cables for the small integrated loudspeakers.

Next part was to make the white plastic cover . It’s mounted with only 4 screws , 2 in each side.

Here is some pics to se the difference.

Finally I added some perforated sheetmetal like the original Imac had in the bottom, but with larger holes.

The final result. It took aprox 8 hours from start to finish to make me a brand new Mac !

The robot is based on the free and opensource robot project Leaf.

The primary programming language is Lisp and C.

Usefull links:

Leaf project

Leaf in WIKI

Yahoo group

This Vacuum forming is a very simple setup that could be improved in several ways.

I made everything in wood because its fast and easy to construct something in wood.

The oven is made of plywood and 6 pcs. ceramic IR heaters as used in professional vaccum forming machines. I have seen people using many different heating sources that is cheaper and could be as good as the heaters i have.

Each ceramic IR element is 200W /230V

The oven is mesuring 31*47cm, so its pretty small and easy handle.

The frame for holding the plastic is also made in wood.

I had one frame made in aluminium that was my first choice of material for the frame holding the plastic. It had the problem that it cooled the plastic to much at the sides, so i made one in wood instead with simple clamps to hold the plastic.

The frame take sheets of 33*50cm which is very practical mesures because the big sheets you get from the reseller is 100*200cm. Then i can easily cut the sheets up with minimum waste of material.

The vacuum table is also made in wood. Its a sandwich contruction made of to pieces plywood with a 3mm aluminium spacer profile inbetween to have a small champer to “distribute” the vacuum if you can say so.

Bottom view shows the plumbing that is glued together with PVC glue and silicone.

The tube fits together with a standard vacuum cleaner tube for easy connection.

The final setup on the floor vacuum cleaner at left, vacuum table in the middle and oven at right.

This is the power panel with Fuses and switches for Motorcontroller, Mini ITX board and the rest of the power consuming boards. I used white LEDs instead of green or red because i want to keep everything in white and black colors where possible.

Im really excited about this new robot feature : SMS control.

www.bladox.com sells this unit called Turbo lite 2 for 49 euro.

I had one for a while without knowing what to do with it, but when Marvin began to take shape I thought why not implement the SMS controller in Marvin?

In short the features are : 2+2 opto-isolated I/O, 13 digital I/O (4 can be used as ADC inputs).

This unit can send you a SMS about levels on ADC input. Send Alarm SMS at a preset ADC level (nice for Battery-low warning). Send status SMS for all in-outputs.

And the really cool feature is that you can send SMS to the unit and control all ouputs and ask for status on all inputs.

I have hooked the unit directly to my motherboard by connecting the Opto-isolated output to the on/off input on the motherboard. The Turbo lite 2 has a command called BTN for Button on all outputs, that means it will simulate a momentary button. Thats just perfect for the power on/off of the motherboard.

Picture this : you leave your robot at home alone and after some hours you get a SMS telling you that your kid has arrived from school (the PIR sensor is activated).

Later you get a message saying that the battery is low. You send a message back that sets the motherboard in standby to save power.

If you have some good ideas then please send them to me.

I bought this Mini ITX motherboard with the new Atom CPU from Intel D945GCLF.

I choose this instead of buying a laptop because it is really cheap: you can buy it for 85$ in Denmark ! Add a 2Gb DDR Ram and a SATA disc and you are running !

As you can read in the PDF it also has a real serial and parallel com port, and where can you find these nowadays?

I couldn’t use my USB to serial converter to upload the Microcontroller testing program with my laptop. When i got the D945GCLF and hooked the microcontroller to the serial port there was no problems at all uploading the test SW.

How do i connect this motherboard with a 12v battery ?

You buy a picoPSU-60-WI (find it on ebay at 40$). It will supply the motherboard and you can supply it from 6-26V ! thats really impressive. The picoPSU-60-WI will deliver 60W – I only use 24W so i guess i can use the remaining power for other power needing boards onboard the robot.

If you want to use the picoPSU then remember that you can NOT use 3.5″ HDD or 5.25″ CDROMS because of the high power consumption.

I had an old slimline slot-in CD drive from a mac-mini that is is perfect for the robot. I got anIDE converter from dealextreme for 5$ which fits at the back of an slot-in drive so that it can be connected to IDE.

Then I added a wifi/bluetooth PCI board from dealexteme for 25$ (worldwide shipping is free!).

In total with a 2.5″ 80Gb SATA HDD, 2Gb Ram, CDrom, Wifi/BT PCI and a logitech webcam it draws 2A which is around 24 W. I think its okay for now, but i have to see when i try in real life conditions.

I got these Pittman motors GM9234S022-R1 from ebay (4 pcs for aprox. 120$) from a shop called raysnow. The gear ratio is 38.3:1 which gives 127 RPM.

My plan is to use standard differential steering with driving wheels from a kickboard which have excellent high friction and good stability with double ball bearings and they are cheap too. Standard diameter off kickboard wheels are 12cm.

The speed of the driving mechanism will be around 80cm/sec with no load and i guess when i put a 25-30kg robot on top of it the speed will be reduced.

I had very good experience with omniwheels as Idle wheels in a differential drive robot. I have a 10kg robot with some smaller omniwheels and it just works perfect. Normally these wheels are used for transport systems in the industry, but they are perfect for robots too.

This type of wheel takes no extra space when the robot is turning and they look really cool too. I got 8 for 46$ at Kornylak . Order RW10 OMNI 2.580 the diameter is 8cm and remember that you have get 2 pcs. to form 1 wheel. They fit together as you can see on the photo.

he Pittman motors i have is unfortunately not prepared for Encoder mounting.

On the ungeared side of the motor there is only about 2mm shaft free.

My solution to this was turning a small aluminium shaft extender and presfitting it on the small shaft and fastening with some lock-tight. The shaft is now about 12 mm and is sufficient for mounting the encoder wheel.

I got 2 pcs. HEDS-5500 F06 encoders from Avego Technologies. They are 2 channel and with 256 counts per revolution i will have 256*38 (gear ratio)= 9728 CPR 9728/37.6 (wheel circumference)= 258 counts per 1cm equals ~ 25.8 counts per 1 mm.

They are very easy to mount , mount the bottomplate with 2 2.5*5mm bolts, click the top part containing the wheel and encoder on and tighten the umbraco and you are ready to test!

The shaft you see below is the homemade 12mm shaft extender the diameter is

6 1/4″ . The bottom plate from the encoder is mounted too.

Just finished mounting the motors on an angle aluminium. The shaft is 8m an is running trough 2 ball bearings and a cardan joint. I had a lot of trouble getting the shaft straight and in center of the motor shaft, but managed to get all alligned at last. Tomorrow i will put the wheels on. Everything has to be custombuild and therefor takes some time to get finished.