compare voltages
I think your tracker needs more than just photodiode sensors. Your principle is correct but not practical.
You will need some sort of 'intelligence' from a microcontroller as well as sensors to make good decisions on the motor activation times. You also have to bear in mind that your motor should be using as little power as possible so you do not waste the gains from using a tracker and that overnight you need to return your panels back to face the rising sun. If you don't return them, the chances are that by morning your photodiodes will be in shadow of the panel and never work again!
I think what you really need is two photocells, facing the same direction as the panel but with a small raised divider between them so that unless facing in exactly the right direction, one or the other will have the shadow of the divider cast upon it. You can use a microcontroller to directly measure the voltage from the cells. You need to drive your motor so the voltage from the cells is equal. I don't think a third sensor is necessary. You only need to check the two voltages perhaps once every few minutes, if you do it more often you may find the panel 'hunts' side to side as it tries to correct for noise, clouds etc.
If all you are doing is trying to optimize the output from a PV panel, a much easier way is simply to rotate it at a fixed speed on a polar (angled pivot) mount and then reverse it back again after dark. That method eliminates all sensors and just uses timers. It may not be quite as efficient but the loss would be very small and the cost much less.
Brian.