[Elecraft] KDSP2

[Lyle Johnson]

Hello Earl!

> 1) I assume that the purpose of the NR and filter gain menus is to adjust
> their gains so that there is no difference in audio output when switching
> NR on/off and when switching between DSP on/bypass.  Is this correct?

The purpose of the gain menus is to balance the audio levels, as you
indicate.  However, due to very significant differences in how the LMS
(noise reduction) algorithms treat different signals, this balancing is only
approximate.

For example, you'll find the noise reduction reduces the amplitude of SSB
voice, but not a steady tone (like a CW signal).  So the nr gain would be
different if you are mostly using CW versus mostly using SSB.

What they are *NOT* intended for, is to add more overall gain to the K2.
Some people are using them this way, and it is possible to start to get
internal overload problems (math over-range) in the DSP if they are all set
to high levels.

> 2) In CW mode with the DSP filter in LPASS there is an audio hiss.  The
> hiss disappears when the NR is turned on.  It is only noticeable with the
> filter set to LPASS.  Is this normal?

Yes.  LPASS is wideband, so it will pass the IF hiss.  If you previously had
a KAF2 installed, it always had a roll off to attenuate the hiss.  Note that
the hiss is not coming from the KDSP2, but from the radio itself.

> If this is normal, the easy
> solution is for the user to configure all four filters as bandpass
> filters.

Yes.  There was discussion about this during Field Test, and in the end we
decided to go with LPASS as the default.  As you note, this is easily
alterable by the user.

> 3) I assume that the NR1, NR2, NR3 and NR4 levels are set by firmware and
> are not user adjustable.  Is this correct?  If so, I would hope that a
> firmware revision (my DSP CPU is rev 3.00) will be forthcoming to make
> the NR levels user adjustable.

HHmmm.  There are three parameters involved in noise reduction: filter
length, beta (attack) and decay (er,, decay :-)

The nr1..4 settings adjust the length (number of taps) of the filter.  There
is a lot that goes on to change these, and I doubt that there will be more
than 4 offered.  But I have noted the request for possible future action.

> 4) If the NR levels are indeed set by the firmware, I was hoping that the
> beta and decay would be adjustable for each individual DSP filter,
> however beta and decay remain the same for all four DSP filters.  Is this
> normal?  If so, how about a firmware revision to enable saving different
> beta and decay for each individual filter?

Yes, the beta and decay values are applied globally.

Your suggestion is an interesting one.  To implement it will require an
additional command to be issued to the DSPx module by the MCU at every mode
or filter selection.  This is not a problem on a system level.

Unfortunately, the MCU code currently uses every available byte in its Flash
memory.  If I find a way to free up the required code space, I will look
into implementing this suggestion.

> I'm still evaluating the KDSP2, mostly by using my infamous weak-signal
> A/B tests, comparing with my FT-1000MP.  On SSB, the NR is phenomenal - a
> bit overagressive IMO, but the agressiveness seems taken care of by
> reducing the decay.  On CW, NR helps only slightly (and only on noisy
> bands), however The 50 Hz DSP filter is definitely an improvement on weak
> CW signal copy compared to just the xtal filter.  Contrary to another's
> (I forgot who) opinion that LPASS was best for weak signals in the noise,
> I find LPASS useless for that purpose.

The availability of so many adjustments makes it practical to tailor the
KDSP2 to your personal preferences (within limits :-)

> I'll post the results of my A/B tests tomorrow.

Thank you for taking the time to comment, Earl.

73,

Lyle KK7P