[GreenKeys] [External] HF radio relating to DSP TU Update

[Harold Hallikainen]

On Wed, July 6, 2022 8:03 pm, Jones, Douglas W wrote:
> From: Jim Haynes [jhhaynes at earthlink.net] -- Wednesday, July 6, 2022 8:46
> PM
>
>> I think the most successful attack on the multipath problem was a system
>> developed at MIT called RAKE, which put the signal through a delay line
>> and then tried to take signals from several taps on the line and combine
>> them constructively.
>
> See:  Clarence W. Hansell, "Method and Means for Reducing Multiple
> Signals," U.S. Patent 2,310,692, granted Feb. 9, 1943, applied for in
> 1939.  Yep, patent does exactly what you describe, using short hunks of
> coax as delay lines.  The problem they were working on at RCA at the time
> was multipath interference in TV signals, a common source of ghost images.
>
> I found that patent when I was working on the analog delay line Wikipedia
> page:
> -- https://en.wikipedia.org/wiki/Analog_delay_line
>
> That page is still over 50% my work, despite the passage of time.
>
>          Doug Jones
>          jones at cs.uiowa.edu

That's great! The combining signals from several taps on a line reminds me
of a DSP Finite Impulse Response filter which, has memory for a series of
samples, starting from the latest, then one earlier, then one earlier,
etc., then multiplying the contents of each of those samples by a
coefficient and adding the results. So, if you know the transfer function
of the path (with reflections), you can do the inverse of it in the FIR
filter. With HF, it seems tricky since the path is always changing. But I
guess now and then you could transmit a known "sync" signal to determine
the transfer function of the path at that time and adjust the filter as
appropriate.

More fun stuff I don't remember from school!

Harold
https://w6iwi.org




-- 
https://w6iwi.org