Meanwhile the Propeller Clock of Bob Blick is known quiet well in the DIY communities http://www.bobblick.com/techref/projects/propclock/propclock.html .
Some people do crazy things http://www.ledcontrol.narod.ru/eng/index.htm .

One problem building such a clock is to find a motor. The motor has to be silent, it has to make more than 25 turns per second and the power for the clock circuit has to be transfered somehow to the pov pcb that is mounted on the turning spindle of the motor.

Bob used a motor from an old floppy disk drive in his project. He modified the motor to use an extra coil as a generator for powering of the circuit. The motor that bob used was a non regulated one from a very old floppy disc drive. These are hard to find nowadays and the speed is not very stable.
Daryl bender http://www3.sympatico.ca/surfin.dude/creative/clocks/propclk/blick.html also modified a motor. Other people transferred the power using brushes. Brushes make noise, give bad contact and are ugly ...
One can also mount a battery onto the moving part http://www.atomicfireballs.com/pivot/entry.php?id=53#body , but this battery need to be replaced now and then.
Phillip used a fan as a motor and build his own generator http://www.activevb.de/rubriken/kolumne/kol_27/propclock.html .
Here are many links to different solutions http://www.luberth.com/analog.htm and pictures of clocks that use brushes to transfer the power to the pcb.

I decided to solve most of the problems by using motors of old 5 1/4 inches floppy disc drives. These motors are brushless dc motors that are very silent while running. They are cheap. The build rather big mounting platforms. The speed of the motors is rather constant because it is regulated. The motors should be found rather easy in surplus stores for a small amount of money.

Step 1: What You Need

Material
- two motors with working circuit boards
- one voltage regulator 7805
- one variable resistor (values depend on the circuit board. mine was 50k Ohms)
- one capacitor (value depends on the circuit board, mine was 10nF)
- glue, best use is epoxy based
- some material to fill "space! (I used broken experimental pcb boards)

Tools
- a soldering iron
- pliers
- solder
- screwdriver
- dremel tool
- ... the standard stuff ... ;-)

Step 2: Working Principle

The idea is to use one of the motors to make the pov circuit spin around and the parts of the other motor to work as a generator for electricity.

The driving motor part is basically untouched. The speed of a floppy disc drive is to slow. Therefore the motor driver need to be modified to make the motor turn faster.

Parts of the other motor can be used to build a generator. Using a dc brushless motor we get three phase current, see http://en.wikipedia.org/wiki/Three-phase_electric_power . The three phases are rectified with three diodes and the resulting voltage can be used to power the pov circuit.

The coils of the generator need to spin with the pov pcb. Therefore the magnets need to be mounted onto the static pcb of the motor and the star with the coils onto the spindle of the motor.

Step 3: Preparation

One disc need not to be torn apart, put it aside for later.
Take the other and remove the motor disc. There is a star build of coils that we are interested in. This star is the heart of the brushless dc motor. It contains 3 windings that form a 3-pase motor. The three windings are connected together in single point, that is the common connection.
See how it works on wikipedia http://en.wikipedia.org/wiki/Brushless_dc_motor . We have the star version here.
The moving part of the motor need to be mounted fix on the chassis of the other motor. The spindle of the other motor needs to go through this part. We need to make the mounting hole bigger. Use a dremel tool to cut a really big hole into the disc of the motor.
The other motor that we try to use as the basis chassis has some mounting holes that we do not need anymore. Just cut them of with the dremel tool.

Step 4: Mounting

Mounting the two motors is not an easy job. There need to be taken care of the small gaps between the moving star and the outer magnets. While putting the star into the magnet disc the magnets pull the star to the wall ... not good.
The base chassis and the disc cannot be mounted flat to each other, there is some space that need to be filled.
Take the broken pcb parts and cut them so that they fit into the space.
To prevent the star form touching the magnets, cut some pieces of paper and wrap them around the star. Try to fill the gap completely on all sites. Be exact here!
Now glue all these things together using the epoxy glue.
Also here you need to very exact! Try to exactly the middle. The star need to turn as exactly in the center as possible.

Step 5: Modify the Basis Motor Circuit

The floppy disc drives I used need two different voltages to work. The drives need 12V and 5V. Luckily the voltages on the pcb are connected in a way that makes it possible to place a voltage regulator on it and get the 5V out of the 12V. Just solder the regulator to the holes. There is a motor enable pin that need to be shortened to ground to make the motor turn, do so.

Old floppy disc drives spin the storage medium with 300 turns per minute. This is too slow to build a POV. The repeat rate is only 5 turns per second and the display will flicker. The motor needs to be tuned a little bit to turn faster. An other reason for faster turning is that using the original speed we get only about 2,5V for the circuit. To drive a small micro processor we need a minimum of 3V, 8V would be fine to get regulated 5V.
To make the motor turn faster remove the capacitor that is responsible for the frequency and put a smaller in. I changed the 22nF to a 10nF. To have a greater band of speed adjustment I removed the 22k Ohm variable resistor and put in a 50k Ohm instead.

Step 6: Testing

Mount a small pcb on top of the platform for testing Solder some rectifier diodes onto the pcb, add a capacitor a white led and a resistor. The video shows the test run and demonstrates the adjustment of the motor speed to get a good, non flickering display.

Tests showed that the construction generates about 7V easily. That should be enough to feed the pov circuit.

The motor is rather loud during testing. The reason for that is that the parts on the pcb are not mounted exactly and that the whole construction is out of balance and vibrating. After mounting the pov circuit this need to be fine tuned. An other source for unwanted noise is that I did not glue the motor exact and that some insulating plastics are sliding on the wall of the disc. I need to remove the plate ones more and cut the plastic of.