Posted: 7/10/2017 5:34:00 PM EDT
|
I have a specific 6m FM frequency I want to receive using a cheap USB SDR receiver. I'd like to build a bandpass filter for that ~20khz of FM. I figured this would be an opportunity to improve my ham radio skills.
I don't want to use a cavity filter due to size and expense. I started with using an online filter calculator to come up with an LC circuit. Mind you the picture below is one of my test examples. Some of the capacitor values are large, but not unobtainable. 
Talking with an experienced RF engineer he advised me to look at building crystal filters for a sharper cutoff. They seem interesting. Just wondering if what I am doing is silly or theres a better way to do it. |
|
Wondering why you want to filter the 6 M band so sharply? Is there some horrendous adjacent channel interference you need to knock down?
Anyway, the LC filter is impractical. Once you add in and simulate the real-world components with their finite Q and series resistance and so forth, you will not be anywhere near the desired response, and have a ton of loss to boot 
That is after all why folks go to the trouble, size, and expense of cavities and the like. If small LC were possible, everyone would already do it. Re xtal filters - you may be able to do this, although loss is gong to be pretty high. And it would be pure luck if you were able to get xtals on the frequency that you wanted. You -may- be able to build a reasonably low loss filter of a MHz or there-abouts with a tweeky LC filter using good quality components, beefy inductors, etc.... Most often, one would want to filter the width of the 6 Meter band in order to eliminate out-of-band overloading signals - this would be a good practical design that you may be interested in trying 
This difficulty in get sharp narrow-band filters up in the MHz and 10's, 100's and higher of MHz range is exactly the reason that hetrodyne receivers were developed. And also why super high-dynamic range SDR was developed. It's easier to filter at the low IF frequency in the case of a hetrodyne RX, and narrow band filtering is done digitally at baseband in the SDR software program. |
|
Quoted:
Wondering why you want to filter the 6 M band so sharply? Is there some horrendous adjacent channel interference you need to knock down? Anyway, the LC filter is impractical. Once you add in and simulate the real-world components with their finite Q and series resistance and so forth, you will not be anywhere near the desired response, and have a ton of loss to boot
That is after all why folks go to the trouble, size, and expense of cavities and the like. If small LC were possible, everyone would already do it. You -may- be able to build a reasonably low loss filter of a MHz or there-abouts with a tweeky LC filter using good quality components, beefy inductors, etc.... This difficulty in get sharp narrow-band filters up in the MHz and 10's, 100's and higher of MHz range is exactly the reason that hetrodyne receivers were developed. And also why super high-dynamic range SDR was developed. It's easier to filter at the low IF frequency in the case of a hetrodyne RX, and narrow band filtering is done digitally at baseband in the SDR software program. But I knew frequency response was going to be the problem and just needed to hear it from someone experienced. |
True what you say regarding some raw SDR dongles and the like - they are fairly sustainable to overload and other nasties. I use one for VHF and UHF satellite work, and built various filters to address just the issue that you are talking about - so kudos for being on the right track 
If you want to pursue it more, I would suggest shooting for a few MHz width, or perhaps the width of the 6 Meter band. You'll probably do well with a 3-section filter (par LC input, series LC mid, par LC output). Just for reference, you should end up in the very rough ballpark of these values, and even this example is a bit extreme. If anything we want smaller input and output C's and larger L's. We want just the opossite for the middle section ===> 
Note that it's almost always the inductors that determine loss, so be sure to purchase or wind units of high-Q |
|
Quoted:
True what you say regarding some raw SDR dongles and the like - they are fairly sustainable to overload and other nasties. I use one for VHF and UHF satellite work, and built various filters to address just the issue that you are talking about - so kudos for being on the right track
If you want to pursue it more, I would suggest shooting for a few MHz width, or perhaps the width of the 6 Meter band. You'll probably do well with a 3-section filter (par LC input, series LC mid, par LC output). Just for reference, you should end up in the very rough ballpark of these values, and even this example is a bit extreme. If anything we want smaller input and output C's and larger L's. We want just the opossite for the middle section ===> http://oi67.tinypic.com/20gk7ly.jpg Note that it's almost always the inductors that determine loss, so be sure to purchase or wind units of high-Q |
|
Might give this a try. I am going to try one for HT as soon as I get my gear back up.
|
|
Quoted:
Have you tried using the receiver without a filter? Just curious. It may be easier to build a relatively effective notch filter if a station nearby overloads your front end. 
|
|
I am all about home brew, but here is a little kit that may help. Scroll to the bottom for a 6m mod for the kit. Will still take some testing and changes, but gives you a platform to go off of.
https://www.qrp-labs.com/bpfkit.html |