1) you are demodulating a signal between 100 kHz and 800 kHz bandwidth.
2) fidelity of the demodulated baseband signal is important.
Then the group delay (phase linearity) of the filter is of primary importance.
A ceramic filter can give nearly flat group delay and good bandpass charactaristics. The cost varies from a few pennies to 0.40 cents in production quantities. Here is a good catalog with application info:
formatting link
A loosly coupled IF transformer can give nearly flat group delay and good bandpass charactaristics. Depending upon the coupling between the primary and secondary, and the loaded Q, you can alter the bandwidth and group delay charactaristics. Fancier passive L.C. I.F. transformers / filters have multiple mutual couplings. Try searching on "helical filter", for instance.
It is challenging to do this as well with an active analog filter, even at this low of a frequency.
Yes, a passive filter or IF transformer will give the same or better performance at 1/10 the price.
If you want near perfect group delay and passband charactaristics, use a wideband flat group delay ceramic filter, digitizer and DSP, assuming cost is no object.
I have seen active filters used for digital video IF's
You need to switch your thinking. Just because you see active filters at audio does not mean you should use them at RF. In all but a handful of cases, you'd be using LC circuits at
10MHz, and given that you have a standard IF, the norm would be a ceramic filter. The latter may not be the solution, given your bandwidth, but since we don't know your actual application one can only guess.
Note that since 10.7MHz is a standard IF frequency, you can get IF transformers for the frequency, so the solution may be to build up something from them.
As for active elements, they'd be there to compensate for the loss in the LC circuits, and to provide the gain you need overall, as opposed to synthesizing a filter. Common would be simple transistor gain stages, or an IC intended for IF amplifier use.
Why ? I agree a ceramic filter would be good if you need a few to a few 10's of kHz bandwidth. But why should you use LC circuits? Because everybody used to do it that way or is an LC filter still better than a modern 1GHz GBW opamp with a few cap's and R's? Just curious..
I remember many many moons ago, tuning up an IF strip using a sweep generator and scope. I got a huge kick out of watching the humps on the response curve move while I tweaked the cores. :-)
Thanks for the explanation Frank. Deep down, I knew all this, but not working with analog stuf much, makes it really stay "deep down" and then you need someone to stir those dormant grey cells..... ;-)
Because LC is a whole lot cheaper and you can achieve a much better filter performance and group delay flatness. With your wide BW spec it would boil down to two filters in series. One high pass and another one low pass.
Today it would be possible to do this with opamps but that doesn't mean it would make economical sense. Actually, it doesn't ;-)
Besides cost and size there is another drawback of opamp filters at high frequencies. They consume a whole lot of power.
Good points, Joerg. I recently designed an RF generator board with 2 DDS chips and at that time I was very tempted to use a 1GHz GBW opamp as a low pass filter after the DDS at 400MHz. Eventually I decided to go for the "good ol' " LC filters....
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.