The Ceramic Spurs (not a rock group)

Paul Taylor VK3HN's magnificent AM receiver was the inspiration for my Quarantine-31 Shortwave Broadcast receiver.  Like Paul I decided to make use of ceramic filters at 455 kHz for selectivity.  I started with the +/- 3 kHz filters that Paul used, but I found them kind of narrow for SW listening.  So I went with some wider ceramic filters that Bruce KK0S had sent me.  But I misread the specs that Bruce sent.  I thought they were 10 kHz wide filters. I realized later that they were +/- 10 kHz -- really twice as wide as I needed.  So I went back to Mini-Kits in Australia and got some +/- 6 kHz filters.  12 kc wide should be just about right, I thought. 

The bandwidth was right, but I started noticing a problem:  I could hear strong SW broadcast stations at two places on my dial.  This brought to mind an admonition from R.A Penfold, author of "Short Wave Superhet Receiver Construction" (1991 Babani Publications).   He advised keeping a few standard 455 kc IF cans in the circuit because, he warned, the ceramic filters have spurious responses, spurs that the IF cans can help knock down.  

Penfold was right.  Look at the filter response curve on the right (above).  There is a nasty spur at around 640 kHz.   This was the cause of my problem. Here is why: 

Suppose I was tuning Radio Marti's big signal on 9805 kc.  My VFO would be running at 9350 kc. 

9805-9350=455.  Great, but... 

With that spur at 640 kc, I could tune down to 9620 kc on my dial.  My VFO would be running at 9165 kc. 

9805-9165=640.  Bad.   That 640 kc difference product would make it through to my detector and AF amp.  I'd have Radio Marti showing up in two places.  I didn't like this. 

I thought about putting a series LC circuit tuned to 640 kc at the output of the ceramic filter.  This looked like a possible solution, but on the bench it looked like I would have trouble getting a circuit of sufficiently high Q.  

So rummaging around in my junk box I found an old Murata CFM455B filter.  This filter is quite broad, but it does not have the spur at 640 kc.  I could use it as a kind of roofing filter just ahead of the +/- 6 kHz filter.  Putting it there would allow me to avoid having to build additional matching circuits for the 455B filter.  

+/-6kc filter upper left,  455B wide filter to the lover right. 

I'm happy to report that this fix works.   The 6 kc filter provides the needed selectivity, and the broader 455B filter knocks down the 640 kc spur. 

Beware the Ceramic Spurs!  

Q-31 with can for first IF amps and filters open