Solving 90 Degree Phase Shift Problem with Acclerometer Signal

[morten84]

Hi.

I get a signal from an acclereometer that I integrate using an op-amp to get the velocity. But the problem is that this signal is phase shiftet 90 degree.
Because I'm supposed to compare this signal with a reference signal that is 180 degree phase shiftet, and now I'm getting wrong values.

Anyone know an easy way to phase shift this 90 degrees more, so that the total phase shift is 180?

 

[antoker]

OK, the answer is simple. Do you need lag/leading network? There are four typical topologies that are used to lag/lead a signal: T,Pi networks with passive components.

T-network, Lag:
Here is spice code for the network

R1  1  2  100
C   2  0   1u
R2  2 3   100

T-network, Lead:

C1  1 2   1u
R1   2 0  100
C2  2  3  1u

Pi-network, Lead

C1  1  0   1u
R1  1   2   100
C2  2   0  1u

Pi-network, Lag

R1  1  0   100
C1  1   2   1u
R2  2   0  100

NB. values in the spice code are just arbitrary.

You just need to derive the transfer function for the network of interest, in
$$j\omega$$ terms and find the equation that defines phase shift (something with $$tan^{-1}()$$), so you'll basically need to work your way back. Just remember that those networks will also transform input/output impedance of your circuit and be sure to identify the frequency bandwidth of your accelerometer, so your phase lead/lag network won't interfere with the sensor.

BTW, that's the first time I've heard of need to shift the signal in order to get the accelerometer to function properly, I've worked with them before and have not experienced any problems with lagging. Probably because of the integrator.


P.S Are you Norwegian?

 

[berkeman]

Hi.

I get a signal from an acclereometer that I integrate using an op-amp to get the velocity. But the problem is that this signal is phase shiftet 90 degree.
Because I'm supposed to compare this signal with a reference signal that is 180 degree phase shiftet, and now I'm getting wrong values.

Anyone know an easy way to phase shift this 90 degrees more, so that the total phase shift is 180?

Why would a reference signal be shifted by 180 degrees? If you have sinusoidal acceleration, you also have sinusoidal velocity variations, phase shifted by the 90 degrees. A 180 degree phase shift between a sinusoidal acceleration and the associated sinusoidal velocity plot would be non-physical.

 

[light_bulb]

you should be able to do it with an lcr circuit in the feedback loop of an op-amp.