Another idea I'm currently thinking about implementing on the Tercel is an electric water pump. The Tercel is blessed to be a base model, so it doesn't have power steering or A/C. That leaves the alternator and the water pump left on the accessory belt. I'd ideally like to eliminate both of those accessories and not have any accessory loads on the engine, and no belt losses. The alternator delete is easy. I've done that already on the Paseo I had. The water pump is slightly more difficult. I've seen only a couple people do them here and they seem to be ridiculously expensive (hundreds of dollars). While not super inexpensive, my proposed setup is quite a bit cheaper than that. If you want to scrounge for parts, it'll definitely drop the price.
My idea is first off not to replace the OEM pump, but instead just drive it with an external electric motor. There are tons of small motors out there that will work. I was thinking of using one of those little permanent magnet scooter motors. They're rated for way more than enough power to spin a water pump around. You can pick these up for nearly free if you can score a dead scooter, or off ebay for ~$30. Here is a picture of one I've had kicking around for a couple years just waiting to be put to good use.
Of course, you'll likely have to derate them for 12V (vs 24-48V). I'd recommend ~100W derated for anything up to 3.0L. This fits inline with OEM electric water pumps. Oversized motors are fine as we'll be programming this with an arduino and we can slow it down to sip as much power as we choose. Mine happens to handle 16.5A. That gives me (16.5A * 12.5V = ) 206W of power. Its WAY oversized for this application where I think I'll probably only need about 50W max.
As mentioned, an arduino will be used to control the motor speed. We don't just want it running full blast as that would be quite wasteful. So, my current idea is to monitor coolant temperature and the MAP sensor and vary speeds accordingly. The warmer the engine is, the faster the pump runs. The higher the engine load, the faster the pump runs (with a few second delay after loads have decreased to take out any heat retained in the cylinder walls and head). This keeps the programming and wiring quite simple.
Now, the arduino can't handle providing anywhere close to 100W of power to the pump, so you need some other board with some transistors so we can PWM control the pump accordingly. I was thinking of making my own (and you could to save money), but then I found
this 10a motor controller shield that makes life a lot easier. Just plug in the motor wires and you're ready to go. It even handles reversing the motor direction programatically if you want.
The connection between the motor and pump isn't figure out yet. Chain or belt will be the way to go. There isn't enough room to direct drive it as the frame member will get in the way.
All in all, this setup, if purchasing all components new would cost ~$70 plus shipping of the components. If you scrounge parts you could put this together for $20.
I like this idea because it'll work great in combination with the alternator delete, block heater, and
using engine coolant to heat the transmission. The engine and drivetrain could be preheated before even starting the car just by plugging in to recharge overnight. It also allows you to get heat in winter when your engine is off.
So, that is the general idea. Its far from being implemented as the car isn't even running yet. But, I figured it would be fun to share, talk about, and refine the idea with you guys.