Post Go back to editing

Is LTC1060 the correct device for a 2nd IF stage receiver f_0BP=455 kHz BW=300 Hz?

Hi, I've got an Analog Devices LTSPICE VII simulation of a bandpass filter based upon "Typical Application" on page 1 of the LTC1060 (https://www.analog.com/en/products/ltc1060.html#product-overview) datasheet, and my simulation with the sample values from Example 1 - Design closely matches the performance shown in Figure 1 4th Order Butterworth BP Filter, f_0PB = 2 kHz shown in the corresponding Linear Technology Application Note AN27A-1.

My goal is to use the same  op-amp strategy dual-stage 4th (or higher) bandpass filter with a target design center frequency of 455 kHz, and the target desired bandpass region of the 2nd stage: 

  • Selectivity (-6/-60 dB) SSB/CW: 2.2 kHz/4.5 kHz
  • target filter is 300 Hz/1.0 kHz or 500 Hz/2.0 kHz

Is the LTC1060A device the suitable choice for this application?

ex. f_0BP = 455 kHz, BW = 300 Hz so f_0BP/BW = 1516 @ 300 Hz BP, or 910 @ 500 Hz BP. 

What could I do to design such a filter using ADI products? This is to replace an unobtanium mechanical filter originally made by Collins. The shape of the filter actually matches the AN27A-1 Figure 1, hence my interest in LTC1060, but if there is a more appropriate device, which one should I focus on? 

Thanks. 

  • Hello.

    A high selectivity BPF based on a LTC1060 switched capacitor IC is limited to 10kHz.

    The filter specifications:

    "f_0BP = 455 kHz, BW = 300 Hz so f_0BP/BW = 1516 @ 300 Hz BP, or 910 @ 500 Hz BP",

    can not be implemented with any ADI filter IC because the f_0BP/BW is extremely high for an analog BPF.

    Using two LT1568 filter ICs an 8th order f_0BP/BW=20 can be implemented.

  • thank you. I used LTSPICE XVII to try different filter arrangements, and can try LT1568 as suggested. Is "8th order" referring to 8 identical sections in the context of a filter chain? I ask as I just studied a sample schematic, and duplicated into a series of 4 identical stages of LTC6268/LTC6269 op-amp - and got the interesting filter curve (green) which shows very nice roll-offs with a narrow bandpass region -- and I wonder if I did increase the stages, would I get even sharper roll-offs? How would I then define the 'order', is it the number of op-amps, or 2x the # of op-amps? 4-stage LTC6268/LTC6269 ~300 Hz bandpass RF filter

    Just curious: How do we define the Q of the circuit in this topology? Do we worry about Q of each stage, or just the Q of the overall circuit? This is based upon a "Bandpass Deliyannis Friend" schematic provided with the filter wizard output, but I think it is also known as "2nd Order Bandpass Multiple Feedback" topology. 

  • In a bandpass filter the 2nd order Q2nd is equal to (fcenter)/(-3dB BW). 

    A reliable narrow-band active RC BPF circuit, requires very low tolerance passive components (0.1% resistors and 1% capacitors) and the BPF op amps must have very high open loop gain at the BPF center frequency, 60dB or greater. In addition there is a practical upper limit to the 2nd order Q to minimize the BPF magnitude response variation due the passive and active component variations. Lab tests of narrow-band BPF circuits indicate a maximum Q2nd<25.

    For identical cascaded BPF stages the total Q of the BPF, Qtotal, is 1.56xQ2nd, 2xQ2nd and 2.3xQ2nd for two, three and four cascaded stages respectively.

    For example is the Q of one 2nd order BPF is 10 then Qtotal is 15.6, 20 and 23 for two, three and four cascaded stages respectively.