most of the time I use an L298D motor driver IC for controling motors. However i find that the 1+ voltage drop is anoying. Some motors just dont go fast enough.
Using a 7.2V power supply, leaves about 6V going to the motors and, as I understand it, the L298D is not designed to take in higher voltages. Although I think it can handle quite a bit if you cool it properly.
One option would be to use a mosfet to provide the current, as shown in the picaxe manual example circuits. Another pin can then control a relay to reverse the polarity to the motor. The idea is to use the full voltage of the batteries to the motors. I've tried to sketch the circuit here with 9.6V, but 12V could be an option too.
Can anyone tell me whether this will work OK and what the pitfalls and downsides are?
Thanks
Thanks guys. I will check out those fet based h-bridges.
Sorry for the confusion with the relay drawing. I realise i didn’t draw it correctly, but you obviously got the idea.
The reason for this settup is that in R2D2 I want to place a motor controler in each foot: close to the motor and i have mosfets and relays laying around somewhere.
L9110
I find the L9110 works great for small low power motors i.e. yellow and white gear motors and the like. I get @2v drop in my usage.
But…
If you have a L293D hole, you could fill it with a SN754410NE. It’s pin and drop in compatable with the L293D and has higher current and voltage ratings.
Also…
I have a base I put together that uses one of the cheap (5$US) opto isolated 4 relay boards from China for driving two gear motors. It works fine and the motors get all the happy electrons they can take 
It was very easy to set up with a few jumpers.
Not saying this is best for all applications though.
Test results
I set up a small motor and ran it through the L9110 to compare input and output voltages. I ran the test at 7.2V 10V and 12V letting the motor run free and also putting some load on it by stopping the spinning wheel with my hand.
The L9110 drops about 0.6V - 0.7V running free with 80-100mA current. Straining the motor the current goes up to 600mA (at7.2V) to 1A (at 12v) which results in the voltage dropping another 0.2V
I did the same test using a IRF530 Power mosfet and got results that are comparable. Off course at these small loads, temperature is not an issue.
I conclude that a setup with my MOSFET and a relay, will not beat the L9110 so I’ll go with that one. (thanks for the tip Bajdi!)
It was very easy to set up with a few jumpers.