[Collins] testing 51J4 mech filters

[Dr. Gerald N. Johnson]

On Sat, 2009-04-25 at 23:42 -0400, W2HX wrote:
> Hello all, 
> 
> About a year and a half ago I put out a query on this list to find out how I
> might sweep/test a 51J4 filter to see if its ok or not.
> 
> Well, I finally got around to trying it (boy does the time fly). I am using
> a sweep generator and a spectrum analyzer.
> 
> My question to the group is: When one of these mech filters are rated as
> BW=3.1 kHz is this the -3dB bandwidth or -6dB bandwidth? I ask because my
> measurements seem to be very close to the stated bandwidth using -6dB, but
> further off using -3dB.
> 
> My set up is rather crude. I am not resonating the input or output of the
> filter, or attempting to match impedances, just in and out. 50 ohm sweeper
> in and 50 ohm SA out.
> 
> In case anyone is interested, here is a plot of a 1.4 kHz BW filter 
> 
> <http://www.hertzmail.com/51j4/collins%201.4%20kHz.gif>
> 
> As you can see, this -6dB plot is still a little narrower than 1.4 kHz and
> the -3dB is narrower still.
> 
> Perhaps if I used the proper input and output impedances, would my results
> differ?
> 
> Thanks
> Eugene W2HX
> 
> 
Both impedance matching and resonating affect the pass band ripple, and
likely the location of the band edge. In Collins bulletin 1031, a 1962
vintage mechanical filter catalog, the filter bandpass is rated at 6 and
60 dB down.

For 50 ohm source and load, you'd series resonate the coils. Probably
about 100 to 150 pf. Whatever parallel value +5 pf is in the 51J-4.

-- 
73, Jerry, K0CQ, Technical Advisor to the CRA
All content copyright Dr. Gerald N. Johnson, electrical engineer