K3 filters
Locked
 
|
Nobody seems to talk about Group Delay Time.
The steeper the LC or Xtal filter is, the higher are Group Delay variations. One filter side up, the other side down. In narrow filters these are very close. This is at least one reason for "ringing". 8 Xtal filters are clearly steeper than 5 Xtal types, and have higher variation of Group Delay. DSP makes very steep filters WIHOUT group delays. K3 concentrates best selectivity here. Correct! Roofing filters just protect the rest of RX from too big signals. We tend to think that roofing means final selectivity. Wrong! It would be interseting to measure Group Delays between different Xtal roofing filters with and without DSP. Let us imagine an ideal 9MHz SSB-filter plus/minus 1,5kHz. 10 Xtal filter has shape factor of 1,33. Now -60dB points are plus/minus 2kHz. Now, the Group Delay difficulties are on both side from 1,3 to 2,2kHz. Now the best way of using DSP is to delete these variations by choosing (IF) DSP bandpass of plus/minus 1,2kHz. Still OK for contest, but you do not agree. A 3kHz good Xtal filter became only 2,2kHz final selectivity. I would rather use 5 xtal roofing and all bandpass DSP. Group Delay was importand in analog TV-transmitters (ancient times). It was found the hard way. Once understood, it was controlled and measured every time. Final solution was to use transmitters with normalized IF signal processing. Centralized block for pass-band filtering with standard Group Delay was the answer. Benny OH9NB _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
Benny,
Group Delay is I agree a very important consideration, but I would suggest that another very important factor which determines how well a receiver will perform when the DSP is used as the main 'selectivity block', is the odd order dynamic range of the first IF filter and *everything* else which follows it, which includes the DSP. If you widen up the first IF filter or its skirts and do NOT have a 'strong' system, in a strong or busy signal environment there is a real possibility that some of the IMD products generated within the IF filter and / or following stages will appear within the DSP's passband. An AGC loop after the filter can help if its component parts or action do not worsen IMD, but if activated will cause weak signals just above the noise floor to become 'weaker', or even disappear - i.e. the DX you are chasing. These comments are not only about the K3 but receivers in general. I suppose that another problem is the cost of a 'strong' system, which includes that of high performance Crystal filters. 73, Geoff GM4ESD Benny Aumala <[hidden email]> wrote on Saturday, December 01, 2007 9:31 AM > Nobody seems to talk about Group Delay Time. <snip> > I would rather use 5 xtal roofing and all bandpass DSP. _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
In reply to this post by Benny Aumala
Benny Aumala wrote:
> > DSP makes very steep filters WIHOUT group delays. > K3 concentrates best selectivity here. Correct! > I think you meant that they don't have group delay variations. They certainly have group delays; they work by adding together delayed versions of the signal. Actually I have a feeling that constant group delay is only true of finite impulse response filters. -- David Woolley Emails are not formal business letters, whatever businesses may want. RFC1855 says there should be an address here, but, in a world of spam, that is no longer good advice, as archive address hiding may not work. _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
|
|
In reply to this post by Benny Aumala
> I think you meant that they don't have group delay variations.
That's right, even FIR filters have group delay. It's typically equal to half the length (impulse response) of the filter. > Actually I have a feeling that constant group delay is only true of > finite impulse response filters. Right, FIR filters with symmetrical coefficients (i.e. almost all of them) are "linear phase" which means flat group delay. That is, the delay through the filter is independent of the frequency. That is very important for digital modulation and I believe it also makes SSB sound more natural. Al N1AL > Benny Aumala wrote: >> >> DSP makes very steep filters WIHOUT group delays. >> K3 concentrates best selectivity here. Correct! >> > I think you meant that they don't have group delay variations. They > certainly have group delays; they work by adding together delayed > versions of the signal. > > Actually I have a feeling that constant group delay is only true of > finite impulse response filters. > > -- > David Woolley _______________________________________________ Elecraft mailing list Post to: [hidden email] You must be a subscriber to post to the list. Subscriber Info (Addr. Change, sub, unsub etc.): http://mailman.qth.net/mailman/listinfo/elecraft Help: http://mailman.qth.net/subscribers.htm Elecraft web page: http://www.elecraft.com |
«
Return to Elecraft
|
1 view|%1 views
| Free forum by Nabble | Edit this page |