Okay, I *think* I've found the problem - only, unfortunately, it lies in the part of code that completely threw me during the WDL tutorials. It's this bit:

code:

---

	temp = max((1-time*friction),0);
	my._SPEED_X = (time * my._force_x) + (temp * my._SPEED_X);    // vx = ax*dt + max(1-f*dt,0) * vx
	my._SPEED_Y = (time * my._force_y) + (temp * my._SPEED_Y);    // vy = ay*dt + max(1-f*dt,0) * vy
	my._SPEED_Z = (time * absforce.z) + (temp * my._SPEED_Z);
	// calculate relative distances to move
	dist.x = my._SPEED_X * time;  	// dx = vx * dt
	dist.y = my._SPEED_Y * time;     // dy = vy * dt
	dist.z = 0;                      // dz = 0  (only gravity and jumping)

---

ie the bit that gives you a constant speed on all machines. I don't seem to be having the problem if I use a less accurate way of calculating dist... Anybody good at physics who might know why this is?

Thanks,
Keith