Lab 5 - EE420L 

Authored by Rodolfo Gutierrez

gutie284@unlv.nevada.edu

3/4/2016

  

Op-amps III, the op-amp integrator



    (Vin - 0)/R1 = (0 - Vout)/C1
    C1/R1 = -Vout/Vin
    C1 = 1/jwC
    Vout/Vin = -1 *(-j)/wRC
   
    Vout/Vin = j/wRC = 1/ 2*pi*1k*1u = 159.15 Hz

    With the +j we can state that the gain leads by 90 degrees
   
   Since frequency response requires AC signals we can ignore the 2.5 V entering Vp. The 100k resistor can also be ignored due to it being large enough to block AC current flow.
In an ideal scenario the circuit would work, however in the real world you would have a DC offset, so the 100k resistor is there to compensate the offset.
No because the resistance is too large for AC current to flow through. In comparison the 1u capacitor is closer to a short circuit.

fig1_waveform.JPG

    We see that we get a triangular output for our square input at 159 Hz, confirming our calculations for unity gain frequency.We see that the output is leading the input by 90 degrees, which we should expect from the +j term in the frequency response.

Vout  = T*Vin / 2RC

4 -1 = 2.5*(1/10k) /2RC

RC = 2.5/10k /6 = 41.667 u

Next we can decide either the resistor value or capacitor values. Since it would be easier to acquire unique sized resistors, it would be simpiler to choose the capacitor size

R = 41.667 u / C

We see that if C = 1u, then we will have to use a 41.667 sized resistor. We would rather use a smaller capacitor to get a larger resistor. At C = 0.01u  R = 4166.7 Ohms

Simulation_Schematic.PNG

Next we simulate the circuit with our calculated values for R and C, this is to insure that we will get a triangular output that falls within the given constraints.

Simulation_Waveform.PNG

Here we see that our Vout meets the given requirements.

In the lab experiment we see a output waveform that nearly matches our experimental results

experiment_waveform.JPG

    Here at 10kHz we get a triangle output, centered at 2.5 volts. Now the reason why the tips of the triangle wave spike up is because we are using a small capacitor, if the capacitor was larger we would have smoother triangle edges, as seen with the intergrator output seen in the first experiment. Sofor the trade-off between the resistor and capacitor is that if the capacitor is too small we will have larger edge spikes in our intergrator. However if the capacitor is too big we would need to use small resistors to compensate the RC size. We would try improving the design by using a 1k resistor with a 41.667 nF capacitor, although obtaining that capacitor in the lab would require combining smaller capacitors together, we should see an improvment in our output signal.




Return