National SW-3 Question
Bill M
I'm cobbling together an SW-3 with some "found" parts. The chassis is
the later model ACSW-3 with 58/58/27 tube line up and the change of the
RF gain to the cathode of the RF Amp stage.
It shows no evidence of ever having had the internal shielding
partitions and the various photos of this model on the web that I have
found seem to be the same.
However, I just noted one on ebay that looks the be "almost" the same
model ACSW-3 and it DOES have the partitions. I can't see how they are
mounted so I don't know if I'm looking for the right missing holes in my
set. Chassis and cabinet came from different sources and neither are
100% mint.
Can anyone clear this up for me? Did they do away with the partitions
at some stage of production? If not, can someone provide me with info
of how they were attached to the cabinet/chassis and I'll make repros?
Thanks,
Bill WX4A
exray at coqui. net
straydog
A ham friend told me that he heard that DSPs really work at least a little
bit. He doesn't have one and neither do I (although many of the newer and
more expensive rigs have them, and there are one or two outboard DSPs that
are available that can be added to older rigs in some ways). I don't know
all that they can do (auto null heterodynes? auto noise reduction [what
kind of noise reduction?]? audio enhancement [or audio bandpass shaping]
of some kind?) but if static crashes can be really removed or attenuated,
that would be of use to me. Auto null of heterodynes would be nice on those
occassions when tuner-uppers decide to tune up on my frequency in which I
am already in two way contact with others (on ssb phone). And, maybe there
are some other useful characteristics I don't know about. But, are these
really effective? For example, a 20 db heterodyne null is not much. Nulls
have to be about 50-60 db to be useful.
I'll take any kind of comments, suggestions, refernces to testing, or
personal experience, pro & con recomendations, ...anything...
Art, W4PON
Tom Holden
http://www.radio.bhinstrumentation.co.uk/html/demonstration.html . I have a
cheap Radio Shack DSP40 that performs at the lower end of the scale as far
as these go. It does a remarkable job of multiple heterodyne reduction and
selectable audio passband but its noise blanker is next to useless and it
does not do noise reduction other than through selectivity.
Tom
"straydog" <advo...@sdf.lonestar.org> wrote in message
news:Pine.NEB.4.61.05...@sdf.lonestar.org...
Michael A. Terrell
>
> There are some good demos at
> http://www.radio.bhinstrumentation.co.uk/html/demonstration.html . I have a
> cheap Radio Shack DSP40 that performs at the lower end of the scale as far
> as these go. It does a remarkable job of multiple heterodyne reduction and
> selectable audio passband but its noise blanker is next to useless and it
> does not do noise reduction other than through selectivity.
>
> Tom
The performance of a DSP is very dependant on the software it runs
under.
--
Former professional electron wrangler.
Michael A. Terrell
Central Florida
Leon Heller
"straydog" <advo...@sdf.lonestar.org> wrote in message
news:Pine.NEB.4.61.05...@sdf.lonestar.org...
>
If they didn't work, none of the current military radio systems would be any
use - they all use DSP! I've designed an audio processing DSP system using
one of the new dsPIC devices and a codec - just two small chips. The codec
includes a mic pre-amp and headphone amp.
73, Leon
K4YZ
straydog wrote:
> A ham friend told me that he heard that DSPs really work at least a
little
> bit.
Snip to...
> I'll take any kind of comments, suggestions, refernces to testing, or
> personal experience, pro & con recomendations, ...anything...
I've not had a lot of opportunity to use DSP for Amateur
applications, however I have used it in the Icom IC-718 for Civil Air
Patrol applications.
Several of CAP's 4mHz frequencies suffer from foreign hetrodynes
and other ectopics. The DSP has been over 90% effective in making the
channel "usable".
73
Steve, K4YZ
(TennesseeCAP 250)
xpyttl
news:Pine.NEB.4.61.05...@sdf.lonestar.org...
> A ham friend told me that he heard that DSPs really work at least a little
Of course they work. How well, as Michael points out, depends on the
software.
> are available that can be added to older rigs in some ways). I don't know
> all that they can do (auto null heterodynes? auto noise reduction [what
DSPs can be programmed to do almost anything, but in amateur radios, they
generally are used to reduce noise and null hetrodynes.
> of some kind?) but if static crashes can be really removed or attenuated,
In general, the point of DSPs tends not to be to remove static crashes.
Some of the newer noise blankers are, in fact, processors that might be hard
to distinguish from DSPs. They tend to do a very good job on pulse noise,
but I haven't tripped across any that are even decent on static crashes.
> really effective? For example, a 20 db heterodyne null is not much. Nulls
> have to be about 50-60 db to be useful.
Remember, these things are software rather than filters. The stronger the
heterodyne, the easier it is for the software to recognize it.
Unfortunately, it's not magic. A really strong heterodyne is going to be
driving your AGC and perhaps overloading some amps in your chain, so it
isn't like it magically disappears with no other effect. Still, very often
the DSP can make life more pleasant in the presence of lids.
..
Peter Barbella
its dynamic range, it's number of bits and it's sampling speed?
Pete
KB1LZH
"straydog" <advo...@sdf.lonestar.org> wrote in message
news:Pine.NEB.4.61.05...@sdf.lonestar.org...
>
straydog
Re: Peter's comment below; I don't know enough about DSPs to answer. I got
a few ideas and notions from all that responded. Me? I was most interested
to learn whether any DSPs could do noise reduction of any kind. Heterodyne
from lid tuner-uppers would be nice, too, and so would passband width and
position.
I did google "DSP & ham radio" and quite a bit came up and it will take me
some time to check out enough to figure out if it might be worth chucking
out $200+ on a decent outboard unit (I can't justify laying out a big
chunk of money on a really modern xcvr with built in DSP just for those
relatively rare moments when I have pulse noise, hash noise, or a
heterodyne, or someone with spillover from an adjacent frequency), so I
need something that works pretty good. One guy said there is no "magic"
meaning to me that maybe I should not expect rabbits to come out of hats
easily and quickly. My friend with whom I have weekly ragchew ssb QSOs
told me that just putting on headphones helped him read me when I was weak
and his noise level was high. Also, it helped him a lot when I bought a
linear and went from barefoot to four 811As. But, if a nearby QSO starts
and thier audio is broad, the spillover can be annoying if they are
stronger. Thus, if I could "DSP" a narrower passband and shift the
frequency (or at least shift the "knee"), that might be useful.
Regarding efficacy, I remember back some 3-4 decades ago they were coming
out with audio compandors that "compressed" 2.5+ khz of audio bandwidth
into a signal about 1.2 khz wide, thus saving bandwidth but the gadgets
never caught on bigtime and some gossip said that the extra "audio punch"
and thus power boost (3 db-ish) really didn't end up doing that much. So,
that is one example of a fancy-schmantzy gadget that didn't last long.
This new PSK-31 audio jazz is also out there (the box is about $500 ?) and
you need one at both ends of the voice QSO and I have not yet heard many
of them (the review said that if the signal drops too low, the voice info
just stops coming through the speaker, so it has its caveats, too).
Again, thanks to all those who made comments.
Art, W4PON
======= no change to below, included for reference and context ====
On Tue, 12 Apr 2005, Peter Barbella wrote:
> Date: Tue, 12 Apr 2005 18:55:08 GMT
> From: Peter Barbella <non...@verizon.net>
> Newsgroups: rec.radio.amateur.boatanchors, rec.antiques.radio+phono,
> rec.radio.amateur.equipment, rec.radio.amateur.homebrew,
> rec.radio.amateur.policy
> Subject: Re: Do DSPs (digital signal processors) really work?
Edward A. Feustel
"Peter Barbella" <non...@verizon.net> wrote in message
news:gQU6e.14294$nH4.4122@trndny05...
> Isn't the effectiveness of DSP critically dependent upon the A/D
> converter; its dynamic range, it's number of bits and it's sampling speed?
>
> Pete
> KB1LZH
See for example the organization of the Ten Tec Orion.
Ed,
N5EI
----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----
Paul Keinanen
<efeu...@direcway.com> wrote:
>
>"Peter Barbella" <non...@verizon.net> wrote in message
>news:gQU6e.14294$nH4.4122@trndny05...
>> Isn't the effectiveness of DSP critically dependent upon the A/D
>> converter; its dynamic range, it's number of bits and it's sampling speed?
>Yes, and the AGC system that makes maximum use of the A/D and D/A range.
>See for example the organization of the Ten Tec Orion.
If the DSP system really needs some kind of AGC, then it is a useless
toy :-).
If you have a strong and a weak signal in the ADC input bandwidth and
the system gain depends on the strong signal, then the weak signal is
going to be drowned into the (quantisation)noise whenever the
amplitude of the strong signal is increased.
Any properly working DSP system should be designed in a way that it is
capable of properly converting the thermal noise as well as the
combination of all strong signals in the input.
The only justification for using AGC in front of a DSP system is that
if antennas with dramatically different efficiencies are used. For
instance, if the receiver input is switched between a full sized
dipole on 160 m and a ferrite rod (with perhaps -40 dB gain), then it
is quite acceptable to modify the gain with some kind of AGC, but this
could be handled as well with some antenna specific gain presets.
If the system is working an hour on a specific band with a specific
antenna, you should not need any kind of AGC depending on the actual
signal level. If you need AGC for this, then you are going to have
problems, if you are trying to receive any weak signal close to a
strong signal.
Paul OH3LWR
robert casey
>
> If the system is working an hour on a specific band with a specific
> antenna, you should not need any kind of AGC depending on the actual
> signal level. If you need AGC for this, then you are going to have
> problems, if you are trying to receive any weak signal close to a
> strong signal.
>
What one really needs are ADCs with many bits, say 24 bits,
sampled at least twice the highest signal frequency to be digitized.
One has to avoid saturation (overflow) else the rest of the
signal gets blanked during the overflows.
Joel Kolstad
"Lloyd" <ll...@no.spam> wrote in message
news:ggg8701783...@tarrnews.net...
> Not really. All you need is any modern floating point ADC, and the
> truncation problem is overcome.
I wasn't aware such devices were common place. Do you have any favorite
part numbers that are of this type? (E.g., from Analog Devices, TI, etc.)
Paul Keinanen
>On Wed, 13 Apr 2005 19:31:11 GMT, robert casey <wa2...@ix.netcom.com> wrote:
>>
>> >
>> > If the system is working an hour on a specific band with a specific
>> > antenna, you should not need any kind of AGC depending on the actual
>> > signal level. If you need AGC for this, then you are going to have
>> > problems, if you are trying to receive any weak signal close to a
>> > strong signal.
>> >
>>
>> What one really needs are ADCs with many bits, say 24 bits,
>> sampled at least twice the highest signal frequency to be digitized.
The LSB should be less than the band or thermal noise in the _input_
bandwidth, so that it would act as dithering noise. For terrestrial
VHF/UHF/SHF systems, the antenna temperature is close to 300 K and
thus a noise density of -174 dBm/Hz. Assuming 10 kHz ADC input
bandwidth, the equivalent noise power would be -134 dBm. In order to
be able to correctly convert a strong signal _and_ the broad band
noise riding on it correctly, the maximum signal level at the antenna
for 16, 20 and 24 bits would be -38 dBm, -14 dBm resp. +10 dBm or
assuming 50 ohm antenna impedance 2.7 mV, 44 mV resp. 710 mV. Not
very common values on VHF/UHF.
At HF, the band noise is typically at least 20 dB stronger, so the
gain in front of the ADC can be reduced and thus the maximum antenna
voltage would be at least 10 times larger, so even the 16 bit
converter would tolerate a single S9+55 dB signal.
However, in order to be able recover weak narrow band that would be
buried in the 10 kHz broadband noise, quite good linearity is required
for the dithering process. The dithering noise properties are also
critical, e.g. if it contains mostly impulse noise instead of white
thermal noise. You may have to move the ADC LSB deeper into the noise
floor, which would reduce the maximum input voltage allowed or require
more bits.
>> One has to avoid saturation (overflow) else the rest of the
>> signal gets blanked during the overflows.
>
>Not really. All you need is any modern floating point ADC, and the
>truncation problem is overcome. Back in the old days of integer ADCs,
>you had the weak / strong signal problem as well as the overflow /
>underflow problem. Modern ADCs typically generate IEEE format
>floating point numbers "on the fly" by using triple or even quadruple
>banks of digitizers, all contained in a small amount of chip real
>estate, and all controlled by a small microprocessor.
The exponent is just the gain in front of the ADC and the mantissa is
the real converted data bits. If the exponent bits varies constantly,
this just reflects the "AGC" control voltage of the preamplifier. The
number of mantissa bits is the critical thing for a radio receiver.
The mantissa bits must still be able to represent the broadband noise
riding on strong signals. In practical terms, the exponent could only
be used to reflect the broadband noise and antenna gain on a
particular band and remain constant as long as the receiver is on that
band.
Of course, if we could get a true IEEE float data type ADC with 24
significant bits in the mantissa and nearly 1500 dB dynamic range,
there would not be much problems in data conversions :-).
While the use of floating point formats greatly simplifies range
problems associated with multiplications and divisions, it still does
not do any good for additions and subtractions, so you still have to
watch for repeated additions of very small values to a large value and
possibly regroup the repeated small number additions.
Paul OH3LWR
Scott Dorsey
Most of the folks doing FP (for radar work, for instance) use a pair
of converters and successive-approximation logic to do scaling.
I'm also going to point out here that there are two different kinds of
DSP systems being discussed here.
First of all there are audio processors, which take the demodulated audio
from the receiver, digitize it, and process it. This is what the original
poster is talking about, and these gadgets work pretty well to deal with
heterodyne whistles. Some of them also work pretty well on impulse noise,
at least better than the noise limiter on an R-390. There is a wide
variety of them on the market, each of which has an audio-speed D/A and
A/D pair. The dynamic range required really isn't very high since the
AGC of the receiver has done all the hard part for you. 16-bit converters
are cheap and plentiful and more than enough. For the most part, though,
these devices are noise reduction units and can't do much about improving
receiver selectivity or reducing passband group delay.
SECONDLY, and a totally different animal, there are DSP systems that
directly digitize the receiver IF, and do IF filtration and demodulation
in the digital domain. They may _also_ do some audio processing, but
the big deal with these is that they now have very sharp IF filters
without substantial ringing, and they have extremely good demodulation
without a lot of group delay problems in general. These systems require
much wider input dynamic range, which is usually accomplished with
prescaling. These can be found built into the latest generation of
communications receiver, although folks like Watkins-Johnson also make
retrofit units that you could hang off your old Hallicrafters if you
don't mind paying a few hundred times what the radio costs.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
John Smith
your computer sound card and try some processing for yourself?
Regards,
John
"straydog" <advo...@sdf.lonestar.org> wrote in message
news:Pine.NEB.4.61.05...@sdf.lonestar.org...
>