I decided to use a shield because it makes my life a lot easier and I still have enough pins left for the rest of my sensors.
The toy came with a, "TYCO R/C 6.0V Jet Turbo" rechargeable NiCd battery pack (rated at 700 MaH... wiki says it would charge for 8 hours and then run for 20 minutes.) So I'm guessing the motors run close to 6v, if not a little under that.
I don't know what the current needs of the motors are, but the battery only had .7A so that has to be the max, right?
Not necessarily. So far, we have an idea (based on the wiki entry) that the car (as an R/C toy) will run for 20 minutes on a .700 Ah battery; 20 minutes is 1/3 of an hour, so it is likely necessarily pulling 3 times as many amps as the battery will supply - or 2.1 amps. If it was only pulling .7 amps, it would run (ideally - in reality, probably not) for 1 hour. If it was pulling only .35 amps, it would run for 2 hours. See how this works?
You can almost guarantee that most of that current consumption will be from the motors (little will go toward the receiver in the vehicle). Since there are two, that means the motors are each consuming about 1 amp each (and who knows how much they consume when stalled).
What you need to do, though - to know for certain - is to measure the current draw from the motors while being run from a 6 volt (or 7.2 volt - whichever you choose) battery, while in "free running" (that is, just driving the wheels with no load on them other than the gearbox), "loaded running" (normal weight of vehicle and other components you'll use), and "stalled" (super heavy load - wheels barely turning or stopped - caution, this can strip gears!). Note: Stalled current consumption is best done using the motor alone (no gearbox), with a pony brake or substitute (a pony brake is easy to build - get a piece of wood, drill a hole in the end the size of the shaft, then cut a notch from the end to the hole along the centerline of the hole. Install a screw through the cut such that when turned, you can tighten or loosen the pressure on the shaft). You could also try to grip the shaft while applying power, but some motors can be impossible to stop (they have very high torque) - and can give you friction burns (do not do this if the pinion gear on the motor is not removable - you will be injured).
The L293D delivers .6A according to their data sheet which is .1A under the max the old battery delivered. If that .1A is a problem I have read that you can piggyback L293Ds to double the current if you need to. I'm not sure what that does as far as restricting heat dissipation though. I'm guessing it's nothing good.
Seeing as how these motors are likely going to exceed the capabilities of an L293, what I am going to write won't matter much. The L293 will not likely support these motors under running load, let alone stalled (but double check the current consumption of the motors with a meter first), which means that shield won't either. If the shield supports paralleling the L293s on board, you might be able to use two shields, but even then there may be overheating under a stalled or highly-loaded condition. Without knowing the actual current consumption (and only making an educated guess as I did above from the wiki info), I can't say for certain. So once again, measure it first.
You will probably have to go with using a better h-bridge (an L298 in bridged mode will allow you almost 4 amps - there are other chips and driver boards out there as well - check out pololu.com, for example - but get your measurements first before you purchase - no sense in buying something bigger than you need).
Another option is to replace the L293D with SN754410 motor driver ICs which is rated for 1A which will work for sure but it doesn't have the kick-back protection the L293D does (I have no idea what that means or why it is bad).
Kick-back protection are simply diodes wired such that they shunt the inductive kick of the collapsing field of a coil (such as inside a motor, relay, or solenoid) to ground. The voltage of such "kickback" can be much, much higher than that of the voltage driving the coil, which can easily damage (instantly, or quickly over time) the driving transistors of the h-bridge chip. Some h-bridges have these built in, some do not. I'd advise always adding your own, using high-voltage schmitt trigger (fast acting) diodes - even if the h-bridge has them internally. Note that the L298 may or may not have such diodes in place internally (I believe the 298D does).
Sorry about the ambiguity of "all day." I hope to run the robot 8 hours a day and recharge it over night for the next day. The motors would probably be running about half that time. So, its duty cycle would roughly be 1/6. I'm hoping 6v at 2450MaH would do that?
Well, if the motors are taking 2.1 amps, and the motors are only running for about 4 hours overall, you would need at least 8.4 aH (8400 mAh) of battery supply to meet your goal; that's discounting stall conditions, and in an ideal situation. The truth is, a battery is only rated in the mAh it will give for a particular current draw (see the battery's spec sheet for details). If you go above (or below) this number, the actual amount of time you can get will change (ie - above, it is de-rated, below, pro-rated). Things will also change depending on certain environmental factors (heat being the main one - and yes, batteries get hot and change this as well!). Its also likely you will need even more than the 8400 mAh above, especially if the motors are being worked hard (under extremely loaded or close to stall conditions). Then again, you might need less - if for instance you only use one motor more than the other (but on average, probably not).
-- had to break up the post - too many characters! --