Self Balancing Robot. PID terms?

[spitso]

Hey guys
Im trying to tune my PID loop to make my robot balance but i dont know if im using the right terms for my robot?
Im implementing a contemporary filter for my Gyro(98%) and Acc(2%).
Is the following right?

output=(Kp*error)+(Ki*(error*dt))+(Kd*((error-prev_error)/dt))

error=robot angle --Balance point at 0deg
prev_error=--Previous error
dt=0.01 --10ms delay

thankyou

[megafiddle]

Based on my somewhat limited knowledge of PID temperature controllers, it basically looks correct.

You may need to negate the PID sum, depending on the action direction of the servo mechanism.
It needs to be degenerative (negative feedback).

-

[megafiddle]

I should add something -

I don't know exactly how you are implimenting things, but normally the integral value isn't simply summed in
with the P and D values. The integral action produces a persistant shift or offset in the control output. This offset
remains even after error*dt has become zero. The I action removes steady state errors.

For example, the proportional action will not necessarily regulate the robot angle to zero. The loop gain will not be
high enough to do that. Integral action will remove that type of error.

The P, I and D values are summed in the sense that they all contribute to the control output. The contribution of the
I value remains though after equalibriun is restored. At least until another steady state error occurs.

-