Update 12/12/2012 : PCB received

I've finally received my new circuits boards. They have been realised by Eurocircuits, it was the first time for me. I used to have them manufactured by futurlec previsously. Eurocircuits is a bit more expensive but realisation and transport delay is far more interesting.

I've started to solder power supplies, and already found some mistakes which will have to be corrected in next revision. Soldering needs to be patient because of the full SMD realisation and the fact that I used 0805 chips for place constraints. Furthermore, the 50 bicolor leds soldering needs to be as perfect as possible. A small misalignment will be visible while running.I'm satisfied with the "Power SO8" soldering as discussed previously, the 3mm hole under the chip easily allows power pad soldering.

I'm currently working on Propeller software writing. The display code is almost done. Now the big part is to implement the serial communication protocol between Base and Propeller...

Update 13/12/2012 : Photos updated

As promised, here are a few pictures of the new propeller and transformer.

The transformer is winded around the motor's rotor. Turns are glued with cyanoacrylate during winding process, and will then be covered with a thin layer of epoxy glue. The secondary is made of 0.56mm copper wire. 2 layers are winded, with 10 turns on the first layer and 9 turns on the second one.

The primary is then winded around the secondary, using a small circle of cardboard. This cardboard should ideally be the thinnest as possible, but keep in mind it must provide sufficient gap when removed to allow rotation of the secondary within this primary.

There are to primaries. The first one is made in the same maneer as the secondary : 2 layers of 10 + 10 turns, the transformer center point is then taken out, and winding process continues with the second primary, with exactly the same characteristics. While the winding is completed, a thin layer of epoxy is applied on the outer layer to add strengh.

I agree that's the tricky operation of this project, but the rotating transformer and the contactless energy transmission is far more interesting than a pair of brushes...

Theorically, with this construction, each primary should be around 18µH. I'll try to measure it at work... 

Update 13/12/2012, a bit later :

I found an old network analyser at work to measure my transformer. At first I was quite happy, each primary gave 23µH... but this was primary only. I then tried to measure with the secondary + the motor inside the primary, and found only 9µH !

With U = Ldi/dt, we find that a 9µH powered by 9v gives a current ramp of 1A/µs. The mosfets driving the transformer are given for 2A max, so I need to switch at a frequency of about 500kHz... way too much !

I'll have to wind it again, I'll try with a 0.4mm wire with more turns...

Update 03/01/2013, A few issues encountered, on the way to improve transformer :

As I told previously, I tried to wind a new transformer with 0.4mm wire instead of 0.56. I managed to wind 2x27turns on primary and 33 turns on secondary. I then measured a primary inductance around 25µH which is acceptable with a switching frequency of 100k to 200kHz.
 However, first runs with this transformer weren't as good as I expected. I only had a few volts on secondary (around 3.5V). Simulation told me it should be around 7.5V with load connected and 18V without.

After spending christmas holidays thinking to this problem, I see 3 possible explanations to this :

1 -The aluminium casing of the motor captures a big part of magnetical field, acting like a short circuited turn.

2 - Although mosfets (FDC6561N) are specified for a 3V Vgs, their gate driving is to weak (driven directly by a microcontroller's output). Simulation confirms this hypothesis

3 - Magnetic coupling of the transformer is too weak. My previous propeller clock used more turns (around 130 on primary and 180 on secondary)

I'm on the way to experiment a new design :

- Add a totem pole driving between the microcontroller's output and mosfet's gate

- Change mosfet to an IRLML0060 which can handle a higher Vds voltage (oscilloscope captures shows a peak voltage of around 80V on transistor opening without any snubber, and 30V with snubber connected)

- Wind a new transformer. This time, it'll be wound on a 32mm PVC pipe to add some distance with the motor's aluminium core, and allow a few more turns

I'll keep you in touch...

Update 08/01/2013, Motor issue:

I've assembled the propeller on the motor's hub for the first time, but it's a disappointing experience. With no load, motor runs very smoothly without any noise, but with the propeller, it does not manage to start. Sensorless brushless motors have a particuliar startup procedure which fails with the MTD6501. I guess the "locked rotor protection" activates because of the high inertia moment of the rotating assembly. I tried to fool the controller by adding a few ohms on motor's wires but it doesn't solve the problem.

I wanted to use this HDD brushless motor because of its quality and it's no noise running, but I have to reconsider the problem: I am going to replace it by a fan motor, driven with a DC voltage. Hopefully, I've found a good one at work, on ball bearings, which tends to be rare nowadays. Furthermore, it seems to have a hub diameter compatible with the transformer I wound, will check this evening...

Finally, using a fan motor will be easier to find than a HDD one for other peaople who wants to try this project !

Update 09/01/2013, Transformer works!

Yeah ! After 5 evenings of windings many transformers, this one is the good one !

The driver circuit has been ruggedized, aluminium hdd motor has been thrown out and replaced by a 92mm fan motor. I improved the snubber circuit to attenuate some high voltage spikes.

The transformer now works at 50kHz, due to its increased inductance, but this is easier to avoid ringing issues.

I measured an output voltage aroung 12V even when heavily loaded (20 leds lit at maximum power). I did not tried to measure output power, it's just sufficient for the needs I have.

This is the first 'magical' step of the project. With base powered, it so magical to just approach the propeller and see it illuminating without any wire !

Furthermore, infrared transmission perfectly works too, which is helpful to debug. Next step : new motor mounting...

Update 03/03/2013, Demonstration video online!

The hardware is now 100% validated, the big part now is completing the software.

I am currently working on propeller's code, serial frames processing, different types of display, color management, numerical/analog display format.

Have a look at the small demonstration video : http://youtu.be/CVzaX0vt5qg