DSP

[William Carver]

W7AAZ SAID:
> It's not true that once signals get through the xtal filter
> they cannot be removed.  A DSP filter CAN remove them...after
> all, the FLEX radio works without ANY roofing filter!

W4TV SAID

That's not true.  Even the Flex will not be able to remove a
-50 dBm signal (S9 +20) and allow you to copy a -135 dBm signal
(at the noise floor) 50 or 100 Hz away.  That's a blocking issue
not an IMD issue.  True IMD requires two or more interfering
signals with a specific relationship and with narrow filters
one of the two tones is almost never inside the filter.  

W7AAZ now sez:

I agree, separating two signals isn't an IMD issue per se: third order
intercept or two tone dynamic range is only an indicator of linearity.

The difference between -135 dBm and -50 dBm is 85 dB. That's within the
envelope of DSP filter capability: it CAN separate two signals 85 dB
apart in amplitude if there is reasonable frequency spacing between
them. Choosing 50-100 Hz is NOT a reasonable criterion (there's no
ham-practical xtal filter than will do it, either!).

I have an RX with 3.1 KHz roofing filter and it can separate two CW
signals a few hundred cycles apart, both inside that filters passband. I
can easily copy a 0.1 microvolt signal with an S9+20 dB signal in the
passband of the filter. Its DSP is several generations old and less
capable than the K3.

The bottom line is if you can preserve second i.f linearity and make the
DSP better you can reduce dependence upon the roofing filter.

Alternatively if you can supply a narrow roofing filter you can relax
dependence upon strong second mixer and DSP performance. In fact, if you
have a just-wide-enough "roofing" filter, you don't NEED the second
conversion or DSP at all.

It's an engineering/economic tradeoff. Given the linearity of the K3
second i.f. the only option is more xtal filters.

Bill



______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Joe Subich, W4TV-4]

> The difference between -135 dBm and -50 dBm is 85 dB. That's
> within the envelope of DSP filter capability: it CAN separate
> two signals 85 dB apart in amplitude if there is reasonable
> frequency spacing between them. Choosing 50-100 Hz is NOT a
> reasonable criterion (there's no ham-practical xtal filter
> than will do it, either!).

100 Hz is more than adequate separation ... the K3 with a 200 Hz
filter can do it quite easily, particularly if the desired signal
is set slightly to the side so the interfering signal is off the
side of the roofing filter and the DSP is narrowed down.

Unfortunately, the ADC would need another 40 dB of headroom to
handle the S9+60 dB signal without overflow.  You're not likely
to see that without some kind of hardware AGC and that puts
the "unfiltered" DSP receiver back into the blocking limited
class.

Of course if the interfering signal has clicks forget it ...

> -----Original Message-----
> From: [hidden email]
> [mailto:[hidden email]] On Behalf Of William Carver
> Sent: Friday, August 07, 2009 3:19 PM
> To: [hidden email]
> Subject: [Elecraft] DSP
>
>
> W7AAZ SAID:
> > It's not true that once signals get through the xtal filter
> > they cannot be removed.  A DSP filter CAN remove them...after
> > all, the FLEX radio works without ANY roofing filter!
>
> W4TV SAID
>
> That's not true.  Even the Flex will not be able to remove a
> -50 dBm signal (S9 +20) and allow you to copy a -135 dBm signal
> (at the noise floor) 50 or 100 Hz away.  That's a blocking issue
> not an IMD issue.  True IMD requires two or more interfering
> signals with a specific relationship and with narrow filters
> one of the two tones is almost never inside the filter.  
>
> W7AAZ now sez:
>
> I agree, separating two signals isn't an IMD issue per se:
> third order intercept or two tone dynamic range is only an
> indicator of linearity.
>
> The difference between -135 dBm and -50 dBm is 85 dB. That's
> within the envelope of DSP filter capability: it CAN separate
> two signals 85 dB apart in amplitude if there is reasonable
> frequency spacing between them. Choosing 50-100 Hz is NOT a
> reasonable criterion (there's no ham-practical xtal filter
> than will do it, either!).
>
> I have an RX with 3.1 KHz roofing filter and it can separate
> two CW signals a few hundred cycles apart, both inside that
> filters passband. I can easily copy a 0.1 microvolt signal
> with an S9+20 dB signal in the passband of the filter. Its
> DSP is several generations old and less capable than the K3.
>
> The bottom line is if you can preserve second i.f linearity
> and make the DSP better you can reduce dependence upon the
> roofing filter.
>
> Alternatively if you can supply a narrow roofing filter you
> can relax dependence upon strong second mixer and DSP
> performance. In fact, if you have a just-wide-enough
> "roofing" filter, you don't NEED the second conversion or DSP at all.
>
> It's an engineering/economic tradeoff. Given the linearity of
> the K3 second i.f. the only option is more xtal filters.
>
> Bill
>
>
>
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:[hidden email]
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html