Speed Limiter circuit

In theory if it is a 0-5 k ohm, you could use the resistors in parallel equation to find the appropriate resistor to put in parallel with it to limit the speed. The the resulting max resistance divided by 5 k ohms would be the percentage of top speed.

For example, if you had a 5000 ohm resistor than:

1/x = 1/5,000 + 1/5000 => x = 2,500 /5k ohms = 50% top speed

That being said, often controllers don't go the full range of 0-5k, but might start at something like 1k, and be full on at 4k ohms or so. Also, the throttle response may not be linear.

So, the most practical way to do this I would think is a 2nd 0-5k ohm pot in parallel with the existing throttle pot. That way you could adjust the top speed precisely.

It's important to note that it needs to be in parallel if it is a 0-5 k ohm pot input to the controller. If it's a 5-0k ohm then it would need to be in series.

It would limit the maximum voltage to the motor. But, you are right, this would not limit the max speed of the go cart directly and it will go faster down hill and slower up hill. But, how much is important.

This is not limiting the power per se unless you have no current limit. It's limiting the speed that the motor will provide torque up to. If you get a PM pancake motor that has a fairly flat torque curve, then it will provide a lot of torque until the motor EMF (or the voltage that the motor generates) approaches the applied voltage. It appears that the Mars "brushless etek" might have a flatter torque curve than the brushed etek.

For the brushed etek with a .032 ohm terminal resistance, and 7 amps no-load current:
Assuming you are running at 48v, and the maximum torque needed is at 100 motor amps, then the speed variation from no torque to max torque would be about 6.2%. If the speed was limited to half (or running at 24v), then the speed variation would be about 12.4% (or double). If geared for a 35 mph top speed, and assuming it takes 30 motor amps to cruise on the flat, then the motor would draw 100 amps at 33.4 mph (up a hill), and produce no torque (if going down hill) at 35.5 mph.

If limiting to half motor voltage, and assuming this corresponds to half speed (it might not) and it takes half the torque at this speed or half motor current, and cruises at 17.5 mph at 15 amps, then the speed variation would be 15.5 to 17.7 mph from full torque to no torque.

In theory, what you would want is a perfectly flat torque curve where the motor provides full torque then drops to no torque at a given speed. That's why I would recommend the brushless etek.

Anyway, maybe a cycle analyst could do that? It can vary throttle control for certain parameters like current. Maybe it could also do it with speed as input? It would just be more expensive than an added pot.

Hope that helps some.

BTW, sorry for the long calculation and difficult explanation. What I really mean to say, is try the extra pot first. See how that works for you. A cycle analyst might be a better solution if the pot doesn't work well enough. It appears that it can limit speed:
http://www.ebikes.ca/drainbrain.shtml

And again, the pot will work better for a motor with a flat torque curve like a brushless mars motor, where it may not work so well for a series wound motor to accomplish your goal of a speed limit without limiting power.