[Elecraft] [K3] Elecraft roofing filters

[Joe Subich, W4TV]

> "The difference between the 250 and 400 Hz filters is not 
> enough to be worthwhile (approximately 375 Hz vs. 430 Hz). "
> 
> Joe, what is your source for these numbers?  I've seen numbers 
> like this mentioned previously on the reflector.

My source is Elecraft: www.elecraft.com/K3/K3_8_pole_plots.htm

Look specifically at the last two plots which show: 

           KFL3A-400      KFL3A-250 
  -6 dB      435 Hz         370 Hz 
 -60 dB      935 Hz         785 Hz 
   SF        2.15:1         2.12:1  

INRAD do not publish the curves for the Elecraft (or Yaesu 
8215 KHz) 250 Hz filter.  However, they do show the curve 
for the 8830 KHz (Kenwood) version of the same filter and 
it shows a similar percentage bandwidth (367.5/845 Hz at 
-6/60dB). 

> Here's what I came up with, with AGC off and with the DSP 
> bandwidth set at least as wide as 900 Hz for all tests (to 
> isolate the effects of just the crystal roofing filter):
> 
> 200 Hz, 5-pole: -6 dB BW = 210 Hz; -30 dB BW = 430 Hz
> 
> 250 Hz, 8-pole: -6 dB BW = 260 Hz; -30 dB BW = 500 Hz
> 
> 400 Hz, 8-pole: -6 dB BW = 380 Hz; -30 dB BW = 580 Hz

Did you run those tests in a Data mode or with the CW pitch 
set to 800 Hz?  If not, the low frequency skirts will be 
sharpened significantly and the measured bandwidth will 
less than the roofing filter.  With a CW pitch of 560 Hz 
I measure 10% less bandwidth than in DATA (CF = 1500 Hz or 
2210 Hz).  If I reduce the pitch to 300 Hz, the measured 
bandwidth is further reduced (due to the DSP filter's low 
frequency cut-off). 

73, 

   ... Joe, W4TV 
 


> -----Original Message-----
> From: elecraft-bounces at mailman.qth.net 
> [mailto:elecraft-bounces at mailman.qth.net] On Behalf Of ni0c
> Sent: Thursday, February 05, 2009 3:47 PM
> To: elecraft at mailman.qth.net
> Subject: [Elecraft] [K3] Elecraft roofing filters
> 
> 
> W4TV wrote:
> "The difference between the 250 and 400 Hz filters is not 
> enough to be worthwhile (approximately 375 Hz vs. 430 Hz). "
> 
> Joe, what is your source for these numbers?  I've seen numbers 
> like this mentioned previously on the reflector.
> 
> I recently tested my K3 CW roofing filters using the XG-2 generator 
> at 7040 KHz (50 microvolts input) and KS7D's nice software package, 
> "K3 Filter Tools."  
> 
> Here's what I came up with, with AGC off and with the DSP 
> bandwidth set at least as wide as 900 Hz for all tests (to isolate 
> the effects of just the crystal roofing filter):
> 
> 200 Hz, 5-pole: -6 dB BW = 210 Hz; -30 dB BW = 430 Hz
> 
> 250 Hz, 8-pole: -6 dB BW = 260 Hz; -30 dB BW = 500 Hz
> 
> 400 Hz, 8-pole: -6 dB BW = 380 Hz; -30 dB BW = 580 Hz
> 
> Note the uncertainty in each of the bandwidths above is 
> plus or minus 20 Hz, because I ran these sweeps in 10 Hz 
> increments to save time. (I'll repeat these tests using a 
> 1 or 2 Hz increment, when I have some spare time.) 
> 
> These tests on my filters (as well as my experience by ear) 
> indicate there is an appreciable difference in the 250 and 
> 400 Hz 8-pole filters.  I've always wondered, too, about 
> manufacturing tolerances on narrow crystal filters.  
> 
> 73, 
> Chuck Guenther  NI0C
> 
> 
>    
>