Next, An Input Sine Wave Of 1850hz Was Applied (the Midpoint Of The Pass

Next, an input sine wave of 1850Hz was applied (the midpoint of the pass band), and the amplitude of the input voltage was increased until clipping of the output signal occurred. The op-amp cannot provide a voltage swing greater than its supply voltage, so theoretically this should occur when:
(8)
This equates to a 15mV pk-pk input signal. In practice it was found that this clipping occurred at around 13mV pk-pk. The datasheet states 13V as typical swing at 15V supply voltage if load resistance is 2k? or greater [4]. Note that the minimum load resistance in the specification given for the pre-amplifier is 2k?. To solve this problem the supply voltage could be increased to 20V, giving a minimum swing of 15V according to the datasheet. For higher input voltages, the op-amps could be replaced with higher voltage devices allowing still greater swing.
Although noise was not a particular problem, a low-noise op-amp would perhaps be more suitable for general production. The overall specification for the pre-amplifier is given below:
Table 2 - Pre-amplifier Specification
Input resistance: 50K?
Input pass band range:300Hz - 3.4kHz
Gain: 61.8 64.7 dB in pass band
Attenuation:-20db/decade of frequencies outside pass band
Min load:2k?
Typ. max. input voltage:13mV pk-pk at ±15V supply

4. Conclusion
Overall the circuit performed as expected, and as such was a success. If greater attenuation was required outside the pass band it might be interesting to investigate higher order filtering techniques, at the expense of complexity.
It would also be interesting to experiment with a transistor output stage following some form of op-amp pre-amplification in order to drive lower resistance loads, but since this was not in the original design specification this was not attempted due to time constraints.
There are also many programs which simplify the design of high order filters, it would also be interesting to experiment with these also, but this was not done as the emphasis of this report is on design calculations as well as the resultant circuit [5].
Ultimately the design presented here proved simple and effective for the purposes stated. For higher fidelity signals, it is likely that more complex filtering would be required.
As with all engineering problems, the task is one of balancing opposing factors, in this case attenuation performance, cost, and complexity. It is felt that this has been achieved successfully in this case.
5. References
[1]Although common in the literature, this transfer function is stated in Frequency Response and Filters available at:
http://cc.ee.ntu.edu.tw/~thc/course_ckt/note/chap11.pdf
[2] The equation referenced in [1] may also be verified from pp81 'The Non-inverting Configuration of Microelectronic Circuits 4th Edition, Sedra & Smith, Oxford University Press, 1998. ISBN: 0195116631
[3] Resistor values obtained from the standard E24 list, available in many places.