[Elecraft] Analysis of FT-1000 APF

[Bill W4ZV]

Al Lorona wrote:
> 
>  
> I assumed that the front panel control was a 1k potentiometer. Probably a
> reasonable guess.
>  
> As the pot is tuned from max to min resistance, the audio peak moves from
> 340 Hz to about 1000 Hz (as I reported earlier). The gain at the peak is
> like this:
>  
> 340 Hz :  7.5 dB ,   Q = 11
> 485 Hz :  8.1 dB
> 600 Hz :  8.6 dB
> 740 Hz :  9.1 dB
> 850 Hz :  9.6 dB
> 1000 Hz : 10.5 dB  ,  Q = 35
>  
> So it has the typical characteristic of an active filter as it's center
> frequency is tuned higher (it gets 'peakier'). Maybe this is more info
> than you all care about, but there it is. 
>  
> Should the front panel control be larger than 1k, then all it means is
> that it will tune lower in frequency than I have shown here.
>  
> 'APF' is just a marketing term for a narrow audio band pass filter, at
> least in the FT-1000.
>  
> I don't have any place I can post a picture of these plots, so my word
> description will have to do.
> 

Al if you could provide the 6 dB BW and 6/60 dB shape factors this would
give Lyle an idea of what's needed in the DSP.  If you send all 6 APF plots
directly to me I'll post them for all to see (attachments must be sent
directly since they will not go through the list filters).

>From what I recall of the DSP filter characteristics, they all have a
constant ~150 Hz slope from -6 to -60 dB on the skirts.  This means a
nominal 150 Hz BW filter has a 3:1 shape factor and a *true* 50 Hz filter
would have a 7:1 shape factor ([50+150+150]/50).  So there are several
issues for the DSP:

1.  Can a true 50 Hz (or better yet 30 Hz) filter be constructed?
2.  Is a shape factor of 7:1 sufficient for the peaking function to work as
in the FT1000's APF?
3.  Can the above be done on-the-fly without introducing latency or other
problems?

Thanks again for your work on this!

73,  Bill


-- 
View this message in context: http://n2.nabble.com/Analysis-of-FT-1000-APF-tp3455867p3457407.html
Sent from the Elecraft mailing list archive at Nabble.com.