[Elecraft] Filter alignment once again

[Bill Coleman]

On Mar 14, 2006, at 5:47 PM, Michael Jensen wrote:

> The result is found in a word in the following location:
>
> http://aprs.the-jensen-dk.net/OZ1BZJK2filter.doc

OK, here are some comments.

The first thing I notice is that the FL1 positions for LSB and USB  
don't look right to me. The shape of the USB filter ought to be  
exactly backwards of LSB, but yours looks the same, heavy on the low  
frequencies and tapering off toward the highs. I wonder if your noise  
generator has a bias toward low frequencies, or if there's some  
frequency-limiting component in your measurement. Do you have the  
KDSP2 installed? Did you disable it when you were taking your  
measurements?

The second thing is that the FL1 positions for LSB and USB don't have  
(about) the same base frequency. The "knee" of the USB curve is  
around 800 Hz, but the same point in the LSB curve is about 600 Hz.  
If you listen to LSB and then press CW REV to go to USB, you'd  
probably notice a large change in the background noise. This is  
because your sidebands don't have sufficient frequency symmetry. (It  
is possible to get the LSB and USB background noise to sound ALMOST  
exactly the same when properly adjusted. There's a slight shift in  
"color" because of asymmetry in the bandpass fiilter)

The third thing is that both LSB and USB seem to be adjusted a bit  
high -- the bandpass is a bit far from the carrier. You're going to  
lose a lot of low-frequency energy in your voice -- which may be OK  
if you are a very high-pitched voice. The bandpass should start to  
rise around 300 Hz, with the "knee" about 500 Hz. Your FL2 settings  
seems closer to being correct, and are a lot more symmetrical (see  
the first two points)

The last thing I'll talk about is the other filter settings for LSB  
and USB. I have 1.8 and 1.6 kHz widths programmed in to FL2 and 3 on  
my K2, but frankly, I never use them. I don't like the way the  
variable-bandwidth filter sounds at these wide settings, because of  
it's wild asymmetry. (In fact, I have FL1 programmed for 1.0 kHz in  
CW, which makes tuning the band a pretty quiet experience, even in a  
contest. 1.0 kHz is about as wide as the variable-bandwidth filter  
gets while staying symmetrical) Instead, I use the filtering  
capability of the KDSP2 to clean up the sides of the crystal filter.

I also do RTTY and PSK31, but I've programmed the RTTY mode filters  
for this mode. I don't recommend using a low frequency for either  
PSK31 or RTTY, as you have programmed FL4 to pass 850 Hz. The reason  
is that if your AFSK source has any harmonic energy, they'll pass  
through the sideband filter and cause adjacent-frequency QRM.

Classic RTTY uses tones of 2125 and 2295 Hz. The "low tones" are 1275  
and 1445 Hz. The second harmonic of 1275 is 2550 Hz, which may be  
outside the K2 bandpass, depending on your filter.

The K2 makes it tough to use the classic RTTY high tones because of  
limited BFO range. You may not be able to get the BFO to shift far  
enough for both sidebands. (And as the BFO range increases, the BFO  
step accuracy decreases) Even if you get LSB set up correctly, the K2  
actually uses the USB settings on 21 MHz and above, due to the  
inversion at the first mixer.

I ended up using 1500 Hz as the center frequency for RTTY (I use  
tones of 1415 and 1585 Hz). This also made it easy to program the  
KDSP2 for RTTY as well. I still had to modify the K2 BFO to get  
sufficient range.

Hope this helps


Bill Coleman, AA4LR, PP-ASEL        Mail: aa4lr at arrl.net
Quote: "Not within a thousand years will man ever fly!"
             -- Wilbur Wright, 1901