the dsp in Kenwood TS950
Doug Datwyler
Kenwood TS950, and remembering not what the ICOM adds contain, I
suspect that Kenwood is not using a digital filter at audio
frequencies, but is really using direct digital synthesis to get
l.o. signals. This is the claim of a spectrum analyzer picture in
the add. From my small understanding of the stuff inside, you can
filter the audio and i.f. as much as you want, but if the l.o. is
not clean, the output from the rig will still be junk. That is
why I suspect that Kenwood pulled a marketing ploy, calling DDS
(direct digital synthesis) DSP, so as not to be confused with
ICOM.
Thank you for your kind attention.
--
Douglas L. Datwyler WR7O
Evans & Sutherland Computer Corporation (801)582-5847
600 Arapeen Drive
Salt Lake City, UT 84108
Rick Karlquist
> Without looking at the QST/Ham Radio magazine adds for the
> Kenwood TS950, and remembering not what the ICOM adds contain, I
> suspect that Kenwood is not using a digital filter at audio
> frequencies, but is really using direct digital synthesis to get
> l.o. signals. This is the claim of a spectrum analyzer picture in
> the add. From my small understanding of the stuff inside, you can
> filter the audio and i.f. as much as you want, but if the l.o. is
> not clean, the output from the rig will still be junk. That is
> why I suspect that Kenwood pulled a marketing ploy, calling DDS
> (direct digital synthesis) DSP, so as not to be confused with
> ICOM.
> Douglas L. Datwyler WR7O
Sorry, the above is completely wrong. I asked Kenwood about this
and they said their DSP is *not* an audio filter and it is *not*
direct digital synthesis. It actually works at IF frequencies
to generate and detect SSB signals. Notice, when I said IF frequencies,
I didn't say 8 MHz. or even 455 kHz. It is a very low 4th IF frequency
at which DSP is practical. But it is not baseband (audio). Otherwise,
it wouldn't affect opposite sideband rejection, which Kenwood says
it does. The IF bandwidth of the analog filters in the 950 is 6 kHz.
and the DSP then sharpens this up as desired.
I specifically asked them about whether the 950 uses DDS, and they
said no, but their synthesizer scheme was claimed to be much better
than older designs like the 940, and "just as good" as DDS. Now
*that* may be a marketing ploy. But I haven't seen the numbers to
compare with Icom's DDS.
BTW, you can make the LO in your receiver as clean as you want, but
you will still be limited by the LO in the interfering transmitter.
Unless that happens to be DDS controlled, you are wasting your
money putting a DDS synthesizer in your receiver.
Rick N6RK
Doug Datwyler
just memory) statement:
> > Without looking at the QST/Ham Radio magazine adds for the
....
> > --
> > Douglas L. Datwyler WR7O
I appreciate being straightened out, wringed out, and being made the
fool ocasionally. Thank you for calling Kenwood and getting the low down
for the net. The only thing I know for sure is that I can not afford to
purchase any rice box or American made rig, so any argument for or
against "FEATURES" for me is just fun and helps me understand radio a
bit better. Thank you, Rick.
--
Douglas L. Datwyler WR7O
Bob Cutler
a low frequency. Almost everyone knows that a signal that is
bandlimited to some frequency, say F, can be sampled without aliasing
if the sample frequency is at least 2*F. What many people fail to
realize is that the sampling process is similar to mixing, and that
aliasing can be avoided provided the signal has a bandwidth of F.
It is not restricted to the span 0-F.
Consider an IF with a center frequency of 455kHz and a bandwidth of
7 kHz. If we sample at a 15 kHz rate, then anything in the span
450 - 457.5 kHz (450kHz = 30*15kHz ) will mix (alias if you will)
into the span 0 - 7.5kHz. Since the original signal didn't contain
energy at any other frequencies (say 0-7.5kHz), then aliasing really
hasn't occurred since there wasn't any overlapping of spectra. At
the input to the digital filter it appears as though the IF was operating
with a center frequency of 5kHZ, not 455kHZ.
The frequenies chosen for this example aren't very good, but you
get the idea. Although the ADC in this example would be running
at 15kHz (easy), the ADC's sample-and-hold must accurately sample a
460 kHz signal (harder).
Bob Cutler
Hewlett-Packard
Lake Stevens Instrument Division
hplsla!bobc
Rick Karlquist
> Sampling the IF output does not imply the the IF is operating at
> a low frequency. Almost everyone knows that a signal that is
> bandlimited to some frequency, say F, can be sampled without aliasing
> if the sample frequency is at least 2*F. What many people fail to
> realize is that the sampling process is similar to mixing, and that
I, for one, do realize this.
> aliasing can be avoided provided the signal has a bandwidth of F.
> It is not restricted to the span 0-F.
>
> Consider an IF with a center frequency of 455kHz and a bandwidth of
> 7 kHz. If we sample at a 15 kHz rate, then anything in the span
This is the 6 kHz. (not 7) bandwidth analog filter I mentioned.
> 450 - 457.5 kHz (450kHz = 30*15kHz ) will mix (alias if you will)
> into the span 0 - 7.5kHz. Since the original signal didn't contain
> energy at any other frequencies (say 0-7.5kHz), then aliasing really
> hasn't occurred since there wasn't any overlapping of spectra. At
> the input to the digital filter it appears as though the IF was operating
> with a center frequency of 5kHZ, not 455kHZ.
This 5 kHz. is effectively the low fourth IF frequency I was referring to.
>
> The frequenies chosen for this example aren't very good, but you
> get the idea. Although the ADC in this example would be running
> at 15kHz (easy), the ADC's sample-and-hold must accurately sample a
> 460 kHz signal (harder).
The difficulty of building an acceptable sample-and-hold is why Kenwood
probably (just a guess) didn't implement it this way. You would need
an aperture time under 7 picoseconds to get 12 bits of accuracy. In all
likelyhood, they use a simple mixer and a traditional fourth IF feeding
a digital-audio-type A/D. This is based on the fact they shyed away
from a DDS synthesizer, and past Japanese design philosophy.
Anyone know how they implemented this for sure?
>
>
> Bob Cutler
> Hewlett-Packard
> Lake Stevens Instrument Division
> hplsla!bobc
> ----------
Rick N6RK
HP Santa Clara
rka...@hpscd.hp.com
Henk Peek
CQ DL 12/89 page 750-760 is the first part of a test of the TS-950SD
by DL1BU. He writes an good technical article about any new piece of
HF equipment. He has done a lot of measurments on the TS-950.
CQ DL is the QST of the DARC (german) The TS-950 has only a small
Englisch summary:
The new kenwood tranceiver TS-950 SD is the first to make use
of digital signal processing in amateur radio on the audio
section both for receiving and transmitting. The improvement
in signaL quality is remarkable.
This is an in-depth review of a prototype model with many
sophisticated measurments of receiver and tranmittor specifications
screen photos of morse signal, modulation evaluation,
selectivity curves both at the front-end and in the if and af
stages, sensitivity and large signal capabilities. It wil be the
first evaluation with such data worldwide, one of a later model
might follow. (DL1BU)
Some technical info TS-950SD
Tranceiver:
Pa 150Watt, 50V. The intermodualtion is more the 40db down (SSB).
The PA is capable to run one hour at full power.
Receiver:
first if 73.05Mc, second if 8.83Mc, third if 455kc
15 different bandpas filters before the first mixer.
Intercept point between 20 and 24dbm. Dynamic range 105db.
The phase noise is very low. This the best ham equipment synthesizer
DL1BU has seen on his table. It isn't a DDS.
The dsp has his own low if. The if freq was not in the first article.
Receiver modes SSB, CW, AM, FM, and FSK.
The second receiver can be used 500Kc around the main receiver.
First if 40.055Mc, second if 10.695Mc
There is more information but it is imposible to include the pictures.
I haven't seen the second article. I get the CQ DL in an circulation map.
When you can read german and want your copy of the CQ DL's
DARC e.V, PB 1155, 3507 Baunatal, Germay.
Phone +49 561 492004 Mailbox +49 561 498669, Fax +31 561 4911850
Henk PA0HZP he...@nikhef.nl