I require a passband filter with

  1. bandwidth of 500 MHz.
  2. 4 GHz central frequency.
  3. minimize the attenuation of the signal (but there is no restriction on this as I am going to use power amplifiers after the filters to increase the signal strength).

Can someone guide me as to how to design such a filter? Are there commercial ICs available at such a high frequency?

Additional info: I plan to use this filter in a high bandwidth communication circuit. The input signal is a pulse of lower frequency 10 MHz and hence the frequency spectrum would be a sinc curve. I plan to use the lobes around 4 GHz for my transmission and hence would be using this filter on the transmission side. The receiver would also have the same filter after the antenna and the LNA. The detection of pulse is using RF energy detector. Note that there is no modulation happening here.

Edit-1: Input power level is around -50 dB, which I will amplify to around -30 dB using ADL5542 (RF amplifier). The output of the amplifier will be fed to the filter. Post-filtering, I require an output power level of around -35 to -40 dB.

Edit-2: I am not sure of the width of the transition band as I am still identifying the components for my communication circuit but I think I can allow a transition band of around 50 MHz. As far as the ripple is concerned, I can allow a ripple of 3 dB. I need an attenuation of about 10 dB in the rejection band. I am ready to compromise on the above parameters for input and output power constraints. Also, ripple and attenuation of rejection band is of higher importance to me than width of the transition band.

Reference for measurements in dB is 20 dBm.

  • \$\begingroup\$ We need power levels to have any bearing, there is more then that, but without general power levels filtering is really hard to define. \$\endgroup\$
    – Kortuk
    Commented Feb 16, 2012 at 7:42
  • \$\begingroup\$ Input power level is around -50 dB, which I will amplify to around -40 dB. The output power level would be around -35 to -40 dB \$\endgroup\$
    – Neel Mehta
    Commented Mar 13, 2012 at 12:07
  • \$\begingroup\$ edit it into your question. \$\endgroup\$
    – Kortuk
    Commented Mar 13, 2012 at 13:43
  • \$\begingroup\$ Other important parameters are the width of the transition band allowed and eventually the maximum ripple that you can accept in the pass band \$\endgroup\$
    – clabacchio
    Commented Mar 13, 2012 at 16:16
  • 3
    \$\begingroup\$ That's a convoluted way of saying your reference is 100 mW. \$\endgroup\$ Commented Mar 13, 2012 at 17:40

1 Answer 1


I think you need to talk to a filter vendor instead of stackexchange. For 4GHz you'll probably need some form of waveguide resonator-based bandpass filter. But I don't know if you can get that wide of bandwidth with such a steep 50Mhz transition band without paying a small fortune. Here's an IEEE article on a 4Ghz 500Mhz BP filter.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.