I'm suggesting the following changes to the way renewable energy works.
1. a. In order to make the drawbacks consistent, wind generators should continue to break their blades during storms, and there should be a small chance of this happening at any altitude.
b. Solar panels should also wear out in some way. I suggest adding a GUI with a slot, and a visual indication of some type that a panel has gone bad. If a panel has gone bad, you open up the solar gui and install a replacement electronic circuit into the slot. The chance of a solar going bad should be higher in the harsh sunlight of a desert. The exact chance per hour of play should be very carefully tuned to be a drawback but not ridiculous.
c. For the same reason, water generators should lose their blades like wind generators. Since water gens are cheap, their replacement blades only would cost wood.
2. I've heard that Al intends to implement a central "controller block" for solars and then add on panel blocks you click down to form an array. I suggest extending this idea :
The controller block should apply to ALL renewable energy. The tin wire should no longer be part of the normal wiring system, but ONLY suitable for connecting a renewable energy source to a single controller block. You should not be able to place tin wire anywhere but against a controller block or a piece of tin wire attached to a controller block. Cross connecting 2 controller blocks with tin would vaporize the wire. The tin wire would have associated with it the ID number assigned to the controller block (a data field set on initialization) Connecting a renewable energy source to the tin wire would register it with the associated controller block permanently.
Power would NOT flow in packets down the tin wires, instead, the controller block would simply keep track of how many wind/solar/water are attached to it, and simply supply the right amount of energy to the energy net every tick. So you could have 1000 solars and 1 controller block, and the only active tile entity is the controller block (the solars would be inert and not use OnTick at all in their code) which simply supplies one packet of 1000 EU every tick. The controller block would need to have in a slot like a slot for an overclocker enough transformer upgrades to handle the attached load, of course.
So a 1000 solar farm would cause as much tick lag in SMP as a single solar panel. For solars, this would be the same simulation speed advantage you get with compact solars, without the compact part, which is unrealistic and dumb. (because it removes the only drawback to solars at all, the low density)
