The AGC in the RX340 has a 120 db range. There is 80 db for analog
AGC and 40 db for the DSP AGC. As long as there is not more than
40 db difference in the signals passing through the16 khz roofing filter,
the DSP AGC controls
everything and the DSP filters will work with no AGC pumping whatsoever.
But when you get a difference in signals through the 16 khz roofing filter
more than 40 db, the analog AGC has to come into play to keep the
DSP from being overloaded. So this is a design limitation of the DSP
system only having a 40 DB range. I will say that the analog system
employed in the RX340 is the excellent and makes this limitation hardly
noticable except in the extreme signal strange range you noted.
Carl Moreschi N4PY
----- Original Message -----
From: "Chip Brown" <email@example.com>
To: "Carl Moreschi" <firstname.lastname@example.org>; "Jim Reid" <email@example.com>
Sent: Friday, February 15, 2002 2:47 PM
Subject: Re: [TenTec] Re: RX-340 and Noise Reduction
> Good Afternoon Jim-- Your adventures with the RX340 agc are very
> interesting. My RX340 arrived a couple of weeks ago. It is all what you
> and Carl have told me. From a post of yours a few days ago I guess you
> don't have big signals out there only weak ones!
> I have been doing some one on one or A/B comparisons with the Omni VI+
> here. Mainly on 80 and 40 CW. Can not say that I have done a real
> comparison with ssb signals other than to quickly realize that with the
> stock 2.4 filters in the Omni and the comparable BW in the 340 the 340
> "sounds" much better. I have written that off as to why ssb people do
> the irad audio mod and filter change out.
> However, with regard to CW comparisons on 80 or 40 during the evening,
> there is something happening that I had not anticipated with the rx340.
> Now this is with regard to strong signals. W1AW on a bad night is 70 dB
> over the noise floor and can be as much as 90 dB on a good night. There
> are several locals here (W1 or VE1 land) that can also blow your socks
> off as well. I have found that either receiver can handle the signals in
> that I find no intermods. However there is significant AGC pumping in
> the 340. I can move off the strong signal a few kc (DSP BW maybe 100 to
> 300 cycles) and still see the receiver quieted by 10 dB during the "on"
> time of the signal with fast AGC. If I slow down the AGC and or reduce
> the gain then there is no pumping but have now lost sensitivity. I have
> also tried out various programmed agc settings to try and "control" this
> pumping but to no avail, as Carl pointed out. As I played with this
> some more I eventually realized it was essentually pumping over the
> bandpass (bandwidth) of the second IF namely + or - 8kc.
> When I would try all this same tunning on the Omni again it was true
> but only over the band pass on the 9mc IF filter I had in (2.4, 1.8,
> 500). Thus after I moved off W1AW a Kc or so I again had full
> sensitivity and just did not hear a trace of the W1AW signal or any agc
> action. ((I like to use W1AW code practice because they just keep
> going, not start and stop as in a QSO))
> All of which has led me to try and understand how the agc works in the
> 340. What I am talking about is shown in Figure 10-18 (page 10-25) of
> the 340 book. Starting on the left hand side the second LO comes in #55
> and the 45 mc IF comes in #54. then you see the mixer diodes and the amp
> following it. Following that is the first of the two 16kc ceramic
> filters (FL1) , a couple of amps and the second 16kc filter (FL2). Now
> the 455kc signal out of FL2 is split with part going up to a "digital"
> mixer if you like and off the page as IF3 at 16 2/3 KHz. The other
> branch of the signal comes down (see the little note that says 200mv
> p-p) and goes into the agc detector which is Q14 and Q15. (and the IF
> monitor at #56) Switching back to the text on page 3-6 section 3-5
> second paragraph is a description of this circuit. It continues on in
> the third paragraph. Also see figure 10-1 on page 10-2 for a big picture
> of all of this. The question of what agc function happens in "DSP Land"
> is not clear from this picture or from the text so far. From Figure 10-2
> one can see that agc control data goes into the DPS computer and the
> computer inturn provides digital data to "force" control of the gain of
> the 455kc IF. Page 3-8 paragraph 3-10, middle of the second column
> begins to provide some clues. From the description here it appears that
> the DSP processing provides the shapping (that is the rise, hang and
> fall time) or contouring of the control voltage. It does not appear
> that the signal strength in the DSP is evaluated and used to control the
> level of the IF signal. There must be some sort of delicate balance set
> up in the agc detector circuits to ensure the signal going into the A/D
> at 16 2/3kc does not full scale the A/D. It looks like the DSP just
> does filtering and demodulation and the signal strength in the digital
> filter is not deturmined or used for any control. In other words all the
> agc derived control is based upon signals (and noise) in the 16 kc wide
> 455kc IF.
> Also it looks like there is no adaptive DSP IF noise reduction
> processing that will give an equilvant effect to the omni's post
> detection processing, I guess something we all knew. But there is some
> DSP filtering action in the 340 that helps to control the noise. There
> is the "Fast Filter" option which appears to provide a different shape
> DSP BPF type that does not have a sharp phase shifts and thus provides a
> better fidelity in the data modes. However its bandpass appears to be
> wider so there is a bit more high frequency hiss with it in. There is
> also the "noise blanker" which is to reduce impulse noise I guess. I
> have no impulse noise so it really does not do much for me. Maybe that
> is an amplitude clipper? So far to me anyway it looks like the 340 has
> basically an analog agc system with the agc leaving enough headroom to
> ensure the DSP system is not over loaded. Going to a digitally derived
> agc is the next step I guess for TT and Doug Smith.
> My guess is that the DSP-599's adaptive noise reduction feature is
> similar to the one in the Omni and is probably an adaptive low pass
> filter. The DSP-599 also has an agc feature that also appears to be
> all digital within the DSP processor. I do not know if that will cope
> with the agc pump action that one hears in the 340's audio, probably not
> much, or if it will help with the readability of ssb signals.
> So Jim, it seems to me in order to help with your ssb readability
> problem and my strong signal problem what is needed is to first narrow
> up the 455kc IF. We don't need the entire 16 kc for ham ssb or cw
> operation. I have gone so far as to identify the model filters used on
> the board, they are Murata, and identify a suitable Murata substitute
> with a narrower bandpass. Have not talked with TT about it nor the
> filter company to see if possable to get a couple filters. On the other
> hand I have only had the 340 a few weeks and am not ready to tear into
> it yet, but I probably will eventually.
> I tried a couple of weeks ago to get a DSP-599 from the HRO store up in
> NH but they did have any and had several back orders in place with no
> idea when they would be filled. So that was that.
> I will be very interested in all the information you learn from your new
> combination. For what its worth I don't intend to get rid of my Omni
> VI+ not the RX340!
> It is always good to read your posts.
> Cheers Chip Brown KR1P
> Carl Moreschi wrote:
> > Jim,
> > Fast AGC should not be used for SSB or CW. These modes
> > work best on the RX340 with slow AGC. The slow AGC setting
> > seems perfect for these modes. The fast AGC setting is so fast,
> > it increases the gain noticeably between words or CW characters.
> > I rarely find the noise reduction setting on the 599zx helps.
> > When it does help, it's not for random noise, but more for
> > man made noises such as power pole buzzes.
> > Carl Moreschi N4PY
> > Franklinton, NC
> > ----- Original Message -----
> > From: "Jim Reid" <firstname.lastname@example.org>
> > To: <DennisKT5D@aol.com>; <email@example.com>
> > Cc: <firstname.lastname@example.org>
> > Sent: Thursday, February 14, 2002 7:13 PM
> > Subject: [TenTec] Re: RX-340 and Noise Reduction
> > > Hi again,
> > >
> > > Well, later today should know what the Timewave
> > > 599zx does for the 340. UPS site advises it is already
> > > on a delivery truck here on Kauai, from HRO in two days.
> > >
> > > Anyway, have learned a bit more about the audio noise
> > > level from the 340 -- the amplitude is AGC dependent!
> > > With the 599zx I think I will be able to measure the actual
> > > audio amplitudes, I hope also of the noise. In FAST agc,
> > > the noise output is "considerably" louder than in SLOW,
> > > and a bit louder in MEDIUM agc. Carl, N4PY has an
> > > explanation as follows:
> > >
> > > " At slow agc, the gain will set itself to the peaks of the random
> > > noise. At fast agc, the gain is constantly varying at a very fast
> > > rate giving a higher average gain value and hence more noise."
> > >
> > > Well, I never noticed this with my FT-1000D which had fast,
> > > med, and slow agc. But the 340 AGC is DSP generated, not
> > > just feedback from the product detector to the IF stages.
> > >
> > > The companion Pegasus in this system does have DSP noise
> > > reduction, which is spec'd to be about 15 dB; it works, but
> > > causes distortion to SSB signals. The FAST agc setting of the
> > > 340 makes SSB copy very unpleasant; huge noise between
> > > every word spoken on low level signals. I have checked this
> > > on both SSB and CW modes and across the IF bandwidth
> > > selection options available in the 340. Once the BW is below
> > > about 300Hz or so, the impact of the FAST agc to noticeably
> > > increase the noise amplitude is down.
> > >
> > > What I hope to yield from the zx with the 340 is:
> > >
> > > 1. Lower noise,
> > > 2. Auto notch to eliminate heterodynes in crowded SSB bands,
> > > 3. More constant audio levels via the rumored super agc of the zx.
> > >
> > > Am told the NR function in both Omni VI+ and the Pegasus
> > > follows the product detector, so is totally an audio "base" band
> > > function; in the 340 I wonder if NR can be included in an IF
> > > DSP firmware revision -- don't know, but ought to be able
> > > to subtract non-coherent from coherent content even at
> > > the relatively low final 340 IF of 16 2/3rds kHz at which
> > > A/D function into the DSP occurs and returns via D/A.
> > >
> > > Be nice if possible to add to the 340.
> > >
> > > Have also been told about a product: the "Clear Speech"
> > > loudspeaker. Anyone know how this works, believe it
> > > is not a DSP device, but some other technique to reduce
> > > noise levels.
> > >
> > > Will advise results later about my 340/zx experiments.
> > >
> > > 73, Jim KH7M
> > >
> > >
> > >
> > > _______________________________________________
> > > TenTec mailing list
> > > TenTec@contesting.com
> > > http://lists.contesting.com/mailman/listinfo/tentec
> > _______________________________________________
> > TenTec mailing list
> > TenTec@contesting.com
> > http://lists.contesting.com/mailman/listinfo/tentec