开云体育

I use LTSpice for "low frequency" (below 100 MHz) work with filters and simple linear designs.? The standard models for capacitors and resistors are quite accurate enough, especially if you are building using surface mount components (which have virtually zero lead inductance -- no leads!) and modern surface mount multilayer ceramic capacitors have very low series inductance and resistance.

Inductors are a bit tougher, but if you can find manufacturers Spice models (Coilcraft has these) the sim results can be quite accurate.? Accuracy is still fairly good even if you just use the published values for "Q", series resistance, and self-resonant frequency and plug those into the Spice inductor parameters.? Toroids are a bit trickier since the published values for the parameters are harder to come by, and Ferrite cores are actually pretty complex.? Iron powder cores are more predictable.

The LTSpice "Monte Carlo" process lets me assign tolerances to my filter components so I can see how the filters are likely behave under real-world component variations.? I usually start a filter design using one of my familiar favorites, reference material, ELSIE, or on-line filter design tools, then play with the filter using LTSpice.? I do try to make my filters as parameter-tolerant as practical, avoiding the need for heroic "high loaded Q" behavior.

But eventually you need to build the circuit and measure the actual performance.? By using the simulation tools I'm usually not terribly surprised by the final results.?

Recently I've been skipping the "breadboard on a blank PC board" step and going directly to having a circuit board made.? Using KiCad and JLCPCB is extremely quick and easy, and cheap enough that it makes more sense for me than hand-carving traces on a blank board.? Last night I sent this filter design off to JLCPCB for fab and assembly, specifying components in the JLCPCB/LCSC inventory.? I will get five assembled boards (minus the connectors) for under $12 each, including shipping.? It would be still cheaper if I ordered more, which I will once I get the design dialed-in.? Yeah, it will take about two weeks before the boards arrive here, but I've got so many irons in the fire that I will be plenty busy in the meantime.? This board combines a two-stage 10 MHz shelving filter, and a steep four-section 30 MHz elliptic low-pass filter.? This is intended to go between an antenna and a SDR having a sample rate of 66 MHz.? The shelving filter reduces the low-frequency signal level (where SDR overload is typically a problem), and the LPF is for additional anti-aliasing.? The "pure" design gives 70dB stopband attenuation, simulations show better than 60dB, and if I get at least 50dB then I will consider it a success.

--
Paul Elliott - WB6CXC