Some newbie startup observations

[rcha...@att.net]

After reading Bob Nickels' QST article, I bought a Newsky dongle and downloaded SDR#.
I read through this forum; got a lot of useful pointers; too numerous to thank individually,
but my appreciation to all that have posted.

THe attachment contains some observations and a few measurements; perhaps the data will
be useful to someone.

Anyone with questions or comments, I am all ears (or eyes, or something)

Now, if I can change my WEFAX preamp from a 12 dB attenuator to a 12 dB amplifier, I will be in
Fat City.

73 and again, thanks

Bob Chadwick W1MTX
Plam Bay, FL

[stefan05]

Hello Bob, I have also recently started to experiment with SDR receiving.
My present set-up is:
Hardware:
Terratec TStick RC HD,  DELL 370 desktop and X31 Thinkpad.
Software:
SDR# Version 1.0.0.1089
Antenna: 65 cm vertical wire (about 1/4 lambda at 120 Mhz) close to a window.
Results:
1) WFM Broadcast (with voices distorted).  When I try to modify the frequency,
the program stalls (why?).
2) Air band: A few aircraft around 124 MHz (see attachment), the squelch works OK, but when a
pilot talks, there is much noise superimposed on his (her) voice.
As it is now, this set set-up works, but the performance is quite poor.
Maybe there is some overload due to local stations (Broadcast FM ?  TV?).
Clearly some improvement is needed.
Can you (or anyone in the group) make some suggestions?

[jdow]

Sample rate too high?

{^_^}

On 2013/01/08 14:15, stefan05 wrote:
> Hello Bob, I have also recently started to experiment with SDR receiving.
> My present set-up is:
> Hardware:
> Terratec TStick RC HD, DELL 370 desktop and X31 Thinkpad.
> Software:
> SDR# Version 1.0.0.1089
> Antenna: 65 cm vertical wire (about 1/4 lambda at 120 Mhz) close to a window.
> Results:
> 1) WFM Broadcast (with voices distorted). When I try to modify the frequency,
> the program stalls (why?).
> 2) Air band: A few aircraft around 124 MHz (see attachment), the squelch works
> OK, but when a
> pilot talks, there is much noise superimposed on his (her) voice.
> As it is now, this set set-up works, but the performance is quite poor.
> Maybe there is some overload due to local stations (Broadcast FM ? TV?).
> Clearly some improvement is needed.
> Can you (or anyone in the group) make some suggestions?
>
>

> Le lundi 31 d�cembre 2012 21:35:57 UTC+1, Bob Chadwick a �crit :

>
> After reading Bob Nickels' QST article, I bought a Newsky dongle and
> downloaded SDR#.
> I read through this forum; got a lot of useful pointers; too numerous to
> thank individually,
> but my appreciation to all that have posted.
>
> THe attachment contains some observations and a few measurements; perhaps
> the data will
> be useful to someone.
>
> Anyone with questions or comments, I am all ears (or eyes, or something)
>
> Now, if I can change my WEFAX preamp from a 12 dB attenuator to a 12 dB
> amplifier, I will be in
> Fat City.
>
> 73 and again, thanks
>
> Bob Chadwick W1MTX
> Plam Bay, FL
>

> --
> You received this message because you are subscribed to the Google Groups "Ultra
> Cheap SDR" group.
> To view this discussion on the web visit
> https://groups.google.com/d/msg/ultra-cheap-sdr/-/DPoHwSZD1OQJ.
> To post to this group, send email to ultra-c...@googlegroups.com.
> To unsubscribe from this group, send email to
> ultra-cheap-s...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/ultra-cheap-sdr?hl=en.
>
> Aircraft-traffic_2013-01-04.jpg
>
>

[stefan05]

thanks to jdow for the suggestion (too high sampling rate).
Starting from 2048 MPS (default sample rate), I tried 1024 and 0.25 MPS.
No change. Still very much noise.
---------------------------------------------------------------

[Douglas H Reed]

The fact that you are receiving anything at all means that you have
the software installed properly and the dongle is working. All else is
probably operator error because a general problem with FM audio would
have been reported the day after v1089 came out.

And you did say "distorted" for FM and "noisy" for AM, right? Not the
Snap-Crackle-Pop-Clicks, and silence of dropped packets?

Still, v1089 is almost 2 weeks out of date so you might want to update
the software again from nightly builds. I'd go as far as saying the
software is changing so fast that if I gave detailed step-by-step
instructions, they might not work for you because the program
interface was changed sometime before v1110.

I've started doing what some of the other old-timers (8 months is old
times, right?) suggested. Every time I get a new version installed and
working, I make a new folder with the SDR# version number and copy all
the files into it. So if the next version crashes, I can just run from
the backup directory and all is well. The software is "portable" so it
can be run from any drive once the drivers are installed. I don't keep
it on C-drive any more, I expect it could run direct from a USB stick
if I want to.

Anyway, now the useful stuff....

You may have too much antenna, or the RF gain may be set too high, or
maybe the FM stations are very strong. I expect that you are trying to
tune in on the image of a real FM station. On my dongle(s), I get best
results when I'm in the middle 70% of the tuning bar rather than near
the ends. You can also look at the waterfall or the spectrum display
and see if the signal has the normal bandwidth and shape. I see the
image most often when the station is just beyond the right side of the
frequency display but I'm clicking on the left end of the display.

These images are natural and expected, and will be worse if the RF
gain is set too high because it overloads the 8-bit range of the
dongle. If you have too much antenna, the images may be worse for the
same reason. After all, the dongle was designed to work with a 6"
antenna, right?

Oh, another reason the audio might sound distorted would be if the
filter bandwidth is too narrow. When you select WFM, it should
automatically set the filter bandwidth to 180,000 (180KHz). This is
what is required to pass all the essential sidebands to the decoder
section if you want good audio. If I manually set it to 20,000 then I
get badly distorted audio.

As for the Aircraft band sounding noisy, that would be the other side
of RF gain. If set too low the station is weak and noisy. It is an AM
transmission so noisy is what you get when the station is weak. It is
also what you get when the filter bandwidth is too wide. The default
for AM is set to 10,000 (10KHz). This will give you good sounding
decoded audio out to around 5KHz, which is probably correct for AM
broadcast stations. But for a communications receiver, you probably
want audio bandwidth of around 3KHz max, or around 5000 or 6000 filter
bandwidth. The narrower bandwidth will remove a lot of hiss and noise
from the audio but it also makes tuning more critical. Don't use
"Snap-to-Grid" with narrow-band modes unless you've accurately
calibrated the dongle and are sure the station is exactly where it
should be.

I've just started playing around with RF Gain and RTL AGC in the
Configure panel. I'm currently running with RTL AGC enabled and RF
Gain set to the max. This is probably not optimum. :-) When you go
back to the main screen, the waterfall will be solid red. Pull the
Contrast slider all the way to the bottom and it should look normal
again. There is no "Auto" function for Contrast so you will have to
play with the controls to balance things the way you like it.

That last is really the bottom line for SDR. Play with it. You will be
forced to learn more about SDR and the dongle if you want to get the
most out of the hardware and software.

Good luck with the dongle!

73, Doug Reed, N0NAS.

[jdow]

That picture simply suggests "not much signal". Discontinuities
in the audio suggest too high a sample rate. If you are trying to
receive weak signals a preamp MIGHT be helpful. (But you may want
to put in some attenuation after the preamp to avoid overload. It's
a tight balance because of the weak dynamic range of the 8 bits per
sample dongles.

{^_^}

On 2013/01/09 05:35, stefan05 wrote:
>
> thanks to jdow for the suggestion (too high sampling rate).
> Starting from 2048 MPS (default sample rate), I tried 1024 and 0.25 MPS.
> No change. Still very much noise.
> ---------------------------------------------------------------

> Le lundi 31 d�cembre 2012 21:35:57 UTC+1, Bob Chadwick a �crit :

>
> After reading Bob Nickels' QST article, I bought a Newsky dongle and
> downloaded SDR#.
> I read through this forum; got a lot of useful pointers; too numerous to
> thank individually,
> but my appreciation to all that have posted.
>
> THe attachment contains some observations and a few measurements; perhaps
> the data will
> be useful to someone.
>
> Anyone with questions or comments, I am all ears (or eyes, or something)
>
> Now, if I can change my WEFAX preamp from a 12 dB attenuator to a 12 dB
> amplifier, I will be in
> Fat City.
>
> 73 and again, thanks
>
> Bob Chadwick W1MTX
> Plam Bay, FL
>

> --
> You received this message because you are subscribed to the Google Groups "Ultra
> Cheap SDR" group.
> To view this discussion on the web visit

> https://groups.google.com/d/msg/ultra-cheap-sdr/-/o5_RIgChbQIJ.

> To post to this group, send email to ultra-c...@googlegroups.com.
> To unsubscribe from this group, send email to
> ultra-cheap-s...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/ultra-cheap-sdr?hl=en.
>

> Sampling-rates_1.jpg
>
>

[Opendous]

>weak dynamic range of the 8 bits per sample dongles

  Worse than that unfortunately, the DVB demodulators are 7-bit.  The Intel CE6230 datasheet provides insights into the specifications of similar devices.
http://download.intel.com/design/celect/datashts/D75934.pdf

[stefan05]

Many thanks to Jdow and Douglas who provided useful information.
At that time I'm not sure I understand all of it. (I'm a newbie, after all...).
I am waiting for a filter (ordered today) to alleviate overloading from WFM stations.
In the meantime, I still have to find and write down a clean method (call it a procedure) to download exactly the files that are needed for SDR# (any version) to work (and recognize my dongle right from the start, which is not always the case at this time).
-------------------------------

[kc koellein]

So there is no where to find an "End to end" "How-to" right now?

I was originally expecting to set this thing up in ubuntu linux (Raphael 820T-based from Nooelec)...

but a month later and a bazillion brain-cells fried... still no workie!

...at this point I'll resurrect an old Athlon that has a legal copy of Winblows on it if I can just get it to WORK!

K9SPY

--
You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

To view this discussion on the web visit https://groups.google.com/d/msg/ultra-cheap-sdr/-/mqHH9bhrhvoJ.

To post to this group, send email to ultra-c...@googlegroups.com.
To unsubscribe from this group, send email to ultra-cheap-s...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/ultra-cheap-sdr?hl=en.

--

kc
K9SPY

- When dealing with the dishonest, BE dishonest. Honestly, this is the best policy.

[Adam Nielsen]

> >weak dynamic range of the 8 bits per sample dongles
>
> Worse than that unfortunately, the DVB demodulators are 7-bit.

If you mean the hardware DVB demodulator then that's ok, it's not used in SDR
mode.

I'm a bit confused about the suggestion that 8 bits per sample would lead to
overload though, maybe someone can explain? If you look at audio, you can
store the same signal with the same amplitude as either 8 or 16 bits per
sample with no effect on the overall amplitude. The bit depth only affects
the fidelity.

True, 16-bit will give you more dynamic range, but if the signal is scaled to
fit within 8-bits the only effect I can see is that very weak signals will be
lost in the presence of very strong ones. It shouldn't cause an overload if
the scaling (gain) is set correctly.

Cheers,
Adam.

[jdow]

What did the install.bat file miss?

{^_^}

[jdow]

Typical signal ranges in RF electronics range from no lower than -174 dBm
in a 1Hz bandwidth. Signal levels can, however, go to as high as 0 dBm in
bad cases and even higher in really bad cases. That is as much as 180 dB
or so dynamic range. Each bit of digital precision is 6 dB. So to get to
the 180 dB at 6 dB per shot means, in theory, you'd need about a 30 bit
ADC in really pathological situations. 16 bit ADCs give on the order of
96 dB of dynamic range. 8 bit ADCs give 48 dB of dynamic range.

Those are approximate numbers. Various digital effect modify that range
figure, a little. Also consider that I am describing the dynamic range
from smallest signal that can be seen directly (without averaging tricks)
to the highest signal that does not limit because it is too big. Since
limiting quickly introduces spurious signals into the system that's a
reasonably fair number for describing the dynamic range as a first
approximation.

{^_^} Joanne, W6MKU, who is too lazy to get more specific without a
very sincere monitory incentive.

[Arnie]

Monitory? No problem. Monetary? That's where there is severe limiting taking place.

--
You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

To post to this group, send email to ultra-cheap-sdr@googlegroups.com.
To unsubscribe from this group, send email to ultra-cheap-sdr+unsubscribe@googlegroups.com.

[jdow]

Blame the T'bird spell checker for that one.

In the ADC itself, of course. If it reads -1 v to 1 v as digital values
-128 to +127 and you feed the ADC 2 volts it's (best case) still going
to read +127. That's limiting, isn't it? It's like trying to get your
clock to read 2410 when it's a 0000-2359 type object.

{^_^}

>> To post to this group, send email to ultra-cheap-sdr@googlegroups.**com<ultra-c...@googlegroups.com>
>> .

>> To unsubscribe from this group, send email to ultra-cheap-sdr+unsubscribe@

>> **googlegroups.com <ultra-cheap-sdr%2Bunsu...@googlegroups.com>.
>> For more options, visit this group at http://groups.google.com/**
>> group/ultra-cheap-sdr?hl=en<http://groups.google.com/group/ultra-cheap-sdr?hl=en>
>> .
>>
>>
>

[Leif Asbrink]

Hi Adam, Joanne and All,

The singnal range actually goes well below -174 dBm/Hz on
microwave bands. People using the dongle with a low noise
amplifier may have a noise floor at the antenna input of
-180 dBm/Hz or even less while the strongest signal
could still be 0 dBm or more.

The dynamic range of maybe 190 dB is however not reached by
use of many bits only. There is also the bandwidth factor.

The rtlsdr can sample at 2 MHz. Recent versions work fine
at 3.2 MHz on my systems. That means that the noise floor
seen by the A/D converter is 63 to 65 dB above -180 dBm
which means something like 116 dB.

The 8 bit ADC should provide a dynamic range of 50 dB so
we need something like 66 dB from filters.

The designers of the E4000 chip were of course aware of this
so they have provided good filters. The problem is that
those filters are not well used because there is a compromise:

1) The dongle is used directly on an antenna to receive
a wideband signal.
In this case 8 bit is far more than required in the baseband
and it is perfectly OK to loose one or two bits by increasing
the noise floor by having a high IF gain which is applied
behind the filters. The high IF gain makes it possible to
use a low LNA gain and still get adequate sensitivity.

2) The dongle is used together with a selective preamplifier.
This is for example someone using it in the amateur bands
with a low noise amplifier and a good filter that will
only let about 2 MHz of bandwidth through. In this case
the filtering inside the dongle is not needed much so
one should set IF gain low to the extent that the noise
from IF amplifiers does not contribute at all. One can also
set a modest LNA gain because one wants the noise from the
external LNA to dominate.

With the osmocom library you use a compromise. If you
would use your dongle with Linrad you would get a 20 dB
improvement in the first case or a 6 dB improvement
in the second case. The library provided with Linrad
is fully compatible with the osmocom library. The difference
is that besides the two gain modes 0=agc, 1=manual there
is also 2=linearity and 3=sensitivity for the cases
1 and 2 above.

For details on dynamic range performance, have a look
here:
http://www.sm5bsz.com/linuxdsp/hware/rtlsdr/rtlsdr.htm

The latest version of the modified library is here:
http://www.sm5bsz.com/linuxdsp/hware/rtlsdr/rtl-sdr-linrad2.tbz

I am not a programmer, the code requires some polishing
to perhaps become accepted by osmocom. What the code does is
trivial however.

The other manufacturers face the same problem and it seems
to me that the other tuner chips are likely to allow different
gain distributions to make the tuner fit in different
environments. It might be useful to have more than one
mode for manual gain setting for them as well.

Linrad is a bit different from other SDR packages and
some find it too difficult. To use the modified rtl-sdr
library one has to use Linrad under Linux. (Under Windows
one would have to use an ExtIO dll)

I think it would be good for all users who have an
E4000 dongle if osmocom could include an appropriately
polished version with more gain modes. A couple of reports
from independent people using Linrad might make it happen.

In the Linrad linearity mode the dongle can tolerate
a signal level of -30 dBm at a frequency separation of
10 MHz with maximum gain (NF=8.3 dB)

The compromise code in the osmocom library causes saturation
at -53 dBm with maximum gain (NF=7.9)

At reduced gain with about 13 dB loss of sensitivity,
the linearity mode saturates at -18 dBm while the
standard implementation saturates at -38 dBm.

The 20 dB difference might make filters to block the
FM band needless.

73

Leif / SM5BSZ

> --
> You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

[jdow]

Brown vaguely stinky material such as emanates from the South facing
end of a North facing fertile male bovine.

Note I said dBm/Hz. within a small fraction of a dB -174 dBm is
the Boltzmann noise limit for noise figure. ANY technique that
digs down further is looking for very small noise temperature
changes. But absolute zero is the bottom of the well. All you can
do is look for very slight increases in noise which must be
averaged over time to find. That is often done by switching between
an antenna and a very cold dummy load and looking for very slight
increases in noise correlated with the switching waveform.

You can reach as low as you want if you average long enough, which
means the bandwidth is less than 1Hz. But that is not any longer
measured in a 1 Hz bandwidth. In a 1/10th Hz bandwidth the bottom
of the well becomes -184 dBm, for example.

So signals may go down below -174 dBm. They do NOT go down below
it in a 1/Hz bandwidth. That is why I specified -174 dBm/Hz aka
-204 dBw/Hz. That is a 0 dB noise figure input. You by definition
cannot get less than 0 dB noise figures.

(I do at least try for accuracy when putting out numbers like that.)

{^_^} Joanne, W6MKU

[kc koellein]

As if a female cow's poop duddnt stink...!?

As IF!

:-P

K9SPY

kc
K9SPY

http://www.10bitworks.com
"COME AND MAKE IT"

To post to this group, send email to ultra-cheap-sdr@googlegroups.com.
To unsubscribe from this group, send email to ultra-cheap-sdr+unsubscribe@googlegroups.com.

For more options, visit this group at http://groups.google.com/group/ultra-cheap-sdr?hl=en.

--
You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

To post to this group, send email to ultra-cheap-sdr@googlegroups.com.
To unsubscribe from this group, send email to ultra-cheap-sdr+unsubscribe@googlegroups.com.

[jdow]

Of course, but that's not the traditional expletive I circumlocuted.

{^_-} (And ANY noun can get verbed....)

On 2013/01/11 16:31, kc koellein wrote:
> As if a female cow's poop duddnt stink...!?
>
> As IF!
>
> :-P
>
> K9SPY
>
> kc
> K9SPY
>
> http://www.10bitworks.com
> "COME AND MAKE IT"
>

> On Jan 11, 2013 6:13 PM, "jdow" <jd...@earthlink.net <mailto:jd...@earthlink.net>>

> http://www.sm5bsz.com/__linuxdsp/hware/rtlsdr/rtlsdr.__htm

> <http://www.sm5bsz.com/linuxdsp/hware/rtlsdr/rtlsdr.htm>
>
> The latest version of the modified library is here:

> http://www.sm5bsz.com/__linuxdsp/hware/rtlsdr/rtl-sdr-__linrad2.tbz

> ultra-cheap-sdr@googlegroups.__com
> <mailto:ultra-c...@googlegroups.com>.

> To unsubscribe from this group, send email to

> ultra-cheap-sdr+unsubscribe@__googlegroups.com
> <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.

> For more options, visit this group at

> http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.

>
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Ultra Cheap SDR" group.

> To post to this group, send email to ultra-cheap-sdr@googlegroups.__com
> <mailto:ultra-c...@googlegroups.com>.

> To unsubscribe from this group, send email to

> ultra-cheap-sdr+unsubscribe@__googlegroups.com
> <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.

> For more options, visit this group at

> http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.

>
> --
> You received this message because you are subscribed to the Google Groups "Ultra
> Cheap SDR" group.

> To post to this group, send email to ultra-c...@googlegroups.com.

> To unsubscribe from this group, send email to

> ultra-cheap-s...@googlegroups.com.

[Leif Asbrink]

Hi Joanne,

It is a well known fact that the galactic background noise
temperature is about 5 K. Those who use the rtlsdr for
radio astronomy or for amateur radio EME (earth-moon-earth)
communication may have a system noise temperature around
30 K on 1296 MHz.

You assume room temp 290 K to arrive at -174 dBm/Hz.
Some people have a 10 times lower temperature at their
feed point.

If you read my posting carefully a couple of times and
perhaps take the trouble of actually doing measurements
yourself you may perhaps realize what mistakes you have made.

Hopefully there are readers on this list with a technical
interest and some basioc knowledge. The E4000 chip together
with the RTL2832 is a much better radio than its reputation says.
The reason is inappropriate software.

There is an alternative solution in the FunCube dongle which
allows an appropriate gain distribution between LNA and IF
but which has an additional LNA in front of the E4000 that
degrades the dynamic range by something like 15 dB (I do not
remember exactly) The reason for this extra LNA is a bad PCB
layout. The FunCube had a bad noise figure on 144 MHz without
it due to pick-up of noise from the USB. The proper solution
would have been in a better layout, but the addition of an
extra LNA was also solving the poor NF problem (easier) but
with a degraded dynamic range as a result.

My native language is not English (ASmerican) but I understand
(assisted by the posting from K9SPY. To others you just demonstrate
your ignorance.

73

Leif / SM5BSZ
73

Leif / SM5BSZ

[kc koellein]

Ooops!

Sorry all! I hadn't spotted the tension... Just an opportunity for a joke... I'll be backing out now...!

:-)

kc
K9SPY

http://www.10bitworks.com
"COME AND MAKE IT"

            <mailto:ultra-cheap-sdr@googlegroups.com>.

            To unsubscribe from this group, send email to
            ultra-cheap-sdr+unsubscribe@__googlegroups.com
            <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.
            For more options, visit this group at
            http://groups.google.com/__group/ultra-cheap-sdr?hl=en
            <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.



    --
    You received this message because you are subscribed to the Google Groups
    "Ultra Cheap SDR" group.
    To post to this group, send email to ultra-cheap-sdr@googlegroups.__com

    <mailto:ultra-cheap-sdr@googlegroups.com>.

    To unsubscribe from this group, send email to
    ultra-cheap-sdr+unsubscribe@__googlegroups.com
    <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.
    For more options, visit this group at
    http://groups.google.com/__group/ultra-cheap-sdr?hl=en
    <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.

--
You received this message because you are subscribed to the Google Groups "Ultra
Cheap SDR" group.

To post to this group, send email to ultra-cheap-sdr@googlegroups.com.

To unsubscribe from this group, send email to

ultra-cheap-sdr+unsubscribe@googlegroups.com.

For more options, visit this group at
http://groups.google.com/group/ultra-cheap-sdr?hl=en.

--
You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

To post to this group, send email to ultra-cheap-sdr@googlegroups.com.
To unsubscribe from this group, send email to ultra-cheap-sdr+unsubscribe@googlegroups.com.

[jdow]

If they cool their feed points, significantly, then, yes, they can
get lower noise figures. I should have considered that point. I am
not sure many hams invest in liquid nitrogen cooling systems to get
the additional 6 dB drop in noise. You are possibly the only one here
who might do that. (Merely aiming at a cold spot is not enough. You
also have to have a VERY low noise, usually cooled, LNA.)

Out of curiosity what devices are you using and what temperature are
you working at? That has always interested me; but, I admit a VERY
strong bias towards communications rather than astronomy. And virtually
all communications takes place at temperatures within human bearable
temperature ranges. That is the bias I took in approaching your note.

There is still the issue of dynamic range. I should be more accurate
and cite "spurious free dynamic range".

With these dongles without external bandpass filters this is basically
"0 dB" in some pathological cases. If I tune to WX broadcasts, which
are fairly strong here, and move the "LO" frequency such that the WX
broadcasts appear right at the right hand end of the spectrum display
and then move that little bit to move them off the end, they appear,
basically full strength, at the bottom end of the display as "images".
So discussions of dynamic range are fraught with problems, too.

{^_^} Joanne, W6MKU

[jdow]

No problem.

I am so biased by "communications" uses that Leif's radio astronomy use
simply did not occur to me. It does take extraordinary equipment, last
I knew, to get noise temperatures as low as Leif has just cited. It also
requires really good antennas with very little spill, side lobes, and
the like. To get figures he cites you also must take into consideration
the noise that may get radiated into the feed system from the physical
antenna itself such as the parabolic antenna's reflector. It's a whole
fascinating set of systems design issues that never much interested me.

{^_^} Joanne, W6MKU

On 2013/01/11 19:08, kc koellein wrote:
> Ooops!
>
> Sorry all! I hadn't spotted the tension... Just an opportunity for a joke...
> I'll be backing out now...!
>
> :-)
>
> kc
> K9SPY
>
> http://www.10bitworks.com
> "COME AND MAKE IT"
>

> On Jan 11, 2013 6:39 PM, "jdow" <jd...@earthlink.net <mailto:jd...@earthlink.net>>

> wrote:
>
> Of course, but that's not the traditional expletive I circumlocuted.
>
> {^_-} (And ANY noun can get verbed....)
>
> On 2013/01/11 16:31, kc koellein wrote:
>
> As if a female cow's poop duddnt stink...!?
>
> As IF!
>
> :-P
>
> K9SPY
>
> kc
> K9SPY
>
> http://www.10bitworks.com
> "COME AND MAKE IT"
>
> On Jan 11, 2013 6:13 PM, "jdow" <jd...@earthlink.net

> <mailto:jd...@earthlink.net> <mailto:jd...@earthlink.net

> http://www.sm5bsz.com/____linuxdsp/hware/rtlsdr/rtlsdr.____htm
> <http://www.sm5bsz.com/__linuxdsp/hware/rtlsdr/rtlsdr.__htm>

> <http://www.sm5bsz.com/__linuxdsp/hware/rtlsdr/rtlsdr.__htm
> <http://www.sm5bsz.com/linuxdsp/hware/rtlsdr/rtlsdr.htm>>
>
> The latest version of the modified library is here:

> http://www.sm5bsz.com/____linuxdsp/hware/rtlsdr/rtl-sdr-____linrad2.tbz
> <http://www.sm5bsz.com/__linuxdsp/hware/rtlsdr/rtl-sdr-__linrad2.tbz>

> ultra-cheap-sdr@googlegroups.____com
> <mailto:ultra-cheap-sdr@__googlegroups.com
> <mailto:ultra-c...@googlegroups.com>>.

> To unsubscribe from this group, send email to

> ultra-cheap-sdr+unsubscribe@____googlegroups.com
> <http://googlegroups.com>
> <mailto:ultra-cheap-sdr%__2Buns...@googlegroups.com
> <mailto:ultra-cheap-sdr%252Buns...@googlegroups.com>__>.

> For more options, visit this group at

> http://groups.google.com/____group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/__group/ultra-cheap-sdr?hl=en>

> <http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>>.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups
> "Ultra Cheap SDR" group.
> To post to this group, send email to

> ultra-cheap-sdr@googlegroups.____com
> <mailto:ultra-cheap-sdr@__googlegroups.com
> <mailto:ultra-c...@googlegroups.com>>.

> To unsubscribe from this group, send email to

> ultra-cheap-sdr+unsubscribe@____googlegroups.com
> <http://googlegroups.com>
> <mailto:ultra-cheap-sdr%__2Buns...@googlegroups.com
> <mailto:ultra-cheap-sdr%252Buns...@googlegroups.com>__>.

> For more options, visit this group at

> http://groups.google.com/____group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/__group/ultra-cheap-sdr?hl=en>

> <http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>>.
>
> --
> You received this message because you are subscribed to the Google
> Groups "Ultra
> Cheap SDR" group.
> To post to this group, send email to ultra-cheap-sdr@googlegroups.__com

> <mailto:ultra-c...@googlegroups.com>.

> To unsubscribe from this group, send email to
> ultra-cheap-sdr+unsubscribe@__googlegroups.com
> <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "Ultra Cheap SDR" group.
> To post to this group, send email to ultra-cheap-sdr@googlegroups.__com

> <mailto:ultra-c...@googlegroups.com>.

> To unsubscribe from this group, send email to
> ultra-cheap-sdr+unsubscribe@__googlegroups.com
> <mailto:ultra-cheap-sdr%2Bunsu...@googlegroups.com>.
> For more options, visit this group at
> http://groups.google.com/__group/ultra-cheap-sdr?hl=en
> <http://groups.google.com/group/ultra-cheap-sdr?hl=en>.
>
> --
> You received this message because you are subscribed to the Google Groups "Ultra
> Cheap SDR" group.

> To post to this group, send email to ultra-c...@googlegroups.com.

> To unsubscribe from this group, send email to

> ultra-cheap-s...@googlegroups.com.

[Leif Asbrink]

Hi Joanne,

> I am so biased by "communications" uses that Leif's radio astronomy use
> simply did not occur to me.

Well, it is also communicatons. There is something like a thousand
EME operators on 432 MHz and above. Not all have the kind of
equipment required to get noise temperatures around 30 K, but have a
look here:
http://www.vhfdx.ru/apparatura/accurate_noise_figure_measurements_1296_mhz
You will find a long article by RW3BP who demonstrates how he can use
a horn antenna to make good measurements of low noise temperatures.
Note that here he uses an ordinary amplifier to establish the noise
temperature in his horn antenna 14 K when the antenna points to cold sky.

Then look here:
http://www.vhfdx.ru/apparatura/rw3bp_1296mhz_lna_optimization
Here RW3BP demonstrates what modifications he is doing to the
standard amplifier to improve its noise figure. He reports
he found NF=0.132 dB which means that his noise floor is
22.9 K.

After the addition of a relay he might reach 30 K. Admittedly
there would be some spillover when the horn is mounted in the
focal point of a parabola, so the real system temperature
might be more loke 60 K than the 30 K I wrote.

Modern GAAS FETs give extremely low noise figures even at room
temperature. A good terrestrial microwave station has a noise
temperature well below 290K because the antenna looks at the
average of the sky (15K) and the ground (290K.) To that comes
the losses and the LNA noise. Amateurs with a LNA at the
feedpoint may reach well below 200 K in terrestrial communication.
Today a standard LNA is rated NF=0.5 dB (=36 K.) Optimized
amplifiers are below half that up to at least 1296.

Data sheets tell us interesting things. For example MGF4941AL
a low noise GaAs HEMT gives a noise figure of 0.35 dB (Typ.)
at 12 GHz with an associated gain of 13.5 dB. That is
official from Mitsubishi.

73

Leif

[Leif Asbrink]

Hi Joanne,

> There is still the issue of dynamic range. I should be more accurate
> and cite "spurious free dynamic range".

That is not quite appropriate. Most of the spurs move relative to
the desired signal when the center frequency is changed so if the number
of spurs is reasonable we can avoid them. At least in amateur radio
where the bandwidth of the desired signal is small.

> With these dongles without external bandpass filters this is basically
> "0 dB" in some pathological cases. If I tune to WX broadcasts, which
> are fairly strong here, and move the "LO" frequency such that the WX
> broadcasts appear right at the right hand end of the spectrum display
> and then move that little bit to move them off the end, they appear,
> basically full strength, at the bottom end of the display as "images".
> So discussions of dynamic range are fraught with problems, too.

Well, if you set your dongle to sample at 3.2 MHz you find that by
the time this alias appears at the opposite side it is attenuated
by 20 dB, and when it enters the center region +/- 0.5 MHz it
is at -40 dB and it moves in the opposite direction relative to the
desired signal so you can easily avopid it.

And, when one listens for a desired signal within the center
1 MHz region, any threat from aliases are generated by signals that
are 2.2 MHz or more separated from the desired signal. In
amateur radio where the interest is in narrow band segments
we can easily attenuate such signals by another 40 dB (or more
if required) to avoid interference from them.

But, of course there are many more spurs of other types.
Operating with one of these dongles puts some extra requirements
on the skils of the operator.-)

Surely the dongle has a limited performance, but not quite to the
extent currently reported. (Nearly everyone seems to need a filter
to reduce the levels from the 88-108 MHz band) The software could
be improved.

73

Leif / SM5BSZ

[Simon G4ELI/HB9DRV]

Leif,

I see in the library that there are user-definable stage gains for the
e4000, is this what you are referring to below?

What are you applying / setting for the e4000 to enhance linearity /
sensitivity.

Simon G4ELI/HB9DRV
http://dit-dit-dit.com

[jdow]

I'll reply off list.

{^_^}

[Simon G4ELI/HB9DRV]

What?

I've already bought the popcorn and an extra-big bottle of Chianti, and now
you're only doing a private screening?

Simon G4ELI/HB9DRV
http://dit-dit-dit.com


-----Original Message-----
From: ultra-c...@googlegroups.com

[jdow]

On 2013/01/12 03:46, Leif Asbrink wrote:
> Hi Joanne,
>

>> There is still the issue of dynamic range. I should be more accurate
>> and cite "spurious free dynamic range".
> That is not quite appropriate. Most of the spurs move relative to
> the desired signal when the center frequency is changed so if the number
> of spurs is reasonable we can avoid them. At least in amateur radio
> where the bandwidth of the desired signal is small.

For Moon bounce I can see moving. For channelized (military) comms
you're stuck AND the signal spacing is "worst case" almost by
definition.

>> With these dongles without external bandpass filters this is basically
>> "0 dB" in some pathological cases. If I tune to WX broadcasts, which
>> are fairly strong here, and move the "LO" frequency such that the WX
>> broadcasts appear right at the right hand end of the spectrum display
>> and then move that little bit to move them off the end, they appear,
>> basically full strength, at the bottom end of the display as "images".
>> So discussions of dynamic range are fraught with problems, too.
> Well, if you set your dongle to sample at 3.2 MHz you find that by
> the time this alias appears at the opposite side it is attenuated
> by 20 dB, and when it enters the center region +/- 0.5 MHz it
> is at -40 dB and it moves in the opposite direction relative to the
> desired signal so you can easily avopid it.

2.4 MHz and it's only five or ten dB down. By the center of the
display it is noticeably down. I never bothered to measure it
since I can't fix it easily. I'm more playing with this thing as
a scanner than anything serious. It's a learning tool.

> And, when one listens for a desired signal within the center
> 1 MHz region, any threat from aliases are generated by signals that
> are 2.2 MHz or more separated from the desired signal. In
> amateur radio where the interest is in narrow band segments
> we can easily attenuate such signals by another 40 dB (or more
> if required) to avoid interference from them.
>
> But, of course there are many more spurs of other types.
> Operating with one of these dongles puts some extra requirements
> on the skils of the operator.-)

Indeed.

> Surely the dongle has a limited performance, but not quite to the
> extent currently reported. (Nearly everyone seems to need a filter
> to reduce the levels from the 88-108 MHz band) The software could
> be improved.

Welllll, the FM station interference is serious since it simply
results from limiting in the ADC in those cases. Very nearby
broadcast stations can put a lot of voltage on the antenna.

And, yes, the software can be improved. And I am not entirely
sure the algorithms in SDRSharp are the best in this regard. I do
not really see "enough" sample combining to give the extra bits of
dynamic range - particularly when I play with strong signals and
the dongle's spectrum display. It is really easy to see limiting
when playing with a signal generator - at rather low signal levels.

{^_^}

[jdow]

I'm pretty sure it would be. The dongle data I have available describes
how the gain steps can be implemented stressing sensitivity (high gain
up front) or intercept point (less gain up front.)

{^_^} Joanne, W6MKU

[Simon G4ELI/HB9DRV]

Care to share?

[jdow]

Heh - it was microwave talk not SDR talk. I made a couple quibbles about
measurement technique and what the measurement results really meant on
one paper. On the other I wondered if he was feeding the transistors he
substituted into his noise temperature testing circuit with the
optimum impedance. That can make a heck of a difference. And I'd not
noticed the author mentioning that step.

{^_^}

[Leif Asbrink]

Hello Simon,

> I see in the library that there are user-definable stage gains for the
> e4000, is this what you are referring to below?

Yes.

> What are you applying / setting for the e4000 to enhance linearity /
> sensitivity.

The chip has variable gain all along the processing chain.
The standard library from osmocom provides a compromise
which is a reasonable default.

In case one wants maximum tolerance for strong signals
in the 88 to 108 MHz range when listening at 118 MHz or above
one should increase the gain of the IF stages to make the
system NF low in all the IF stages so the mixer would
send its signal into a stage with a low system NF. That allows
a good sensitivity (=low NF) with lower mixer gain.
The high gain is behind the IF filters so signals
more than about 10 MHz away are well suppressed. The
improvement relative to the compromise of today is 20 dB,
a factor of 100 in power.

The disadvantage is that the noise floor within the
selected passband is higher which means that the
close range dynamic range is degraded.

If one wants sensitivity one would go the other
way. Low gain in the IF chain and high gain in LNA
and mixer. That provides a very low noise floor.
Actually below 1 bit so an external LNA can be added
without too much loss in dynamic range and give
very interesting performance for 144 MHz for example.
In this case one would need a very good filter for
the 144 MHz band because the dynamic range for signals
at large separations would be very poor due to the
high gain in front of the IF filters.

It seems to me that the designers of other tuner chips
than the e4k faced the same problem so I think the same
feature might be available in the other chips also
but I do not have access to any data sheet.

If you could provide this functionality in SDR Radio
I am sure users would appreciate it and then I think
we would find that the osmocom maintainers would
find a way they like to implement different gain
distributions. Some way that does not require
the application programmer to know what dongle he
is using. I just did it in a simple-minded way by
trial and error using a e4k dongle and two signal
generators searching for the best compromise when
the interferer was 400 kHz below respectively 10 MHz
above the desired weak signal.

I currently maintain a clone of librtlsdr for Linrad
users but I would very much prefer to have Linrad
using the standard library.

The possibillity to set different gain distributions
is available in FunCube, but since the A/D has 16 bit
the advantage is small.

73

Leif / SM5BSZ

>
> Simon G4ELI/HB9DRV
> http://dit-dit-dit.com
>
>
> -----Original Message-----
> From: ultra-c...@googlegroups.com
> [mailto:ultra-c...@googlegroups.com] On Behalf Of Leif Asbrink
>
> The difference is that besides the two gain modes 0=agc, 1=manual there is
> also 2=linearity and 3=sensitivity for the cases 1 and 2 above.
>

[Simon G4ELI/HB9DRV]

Thanks Leif.

[murrij]

I'm with Simon - taking the discussion private? In all seriousness though, I've got a bunch of notes down to research & understand what the heck some of what u were talking about actually meant. With my budding SDR knowledge growing rapidly I eat this stuff up. I got sidetracked with my Ubertooth (sniffing Frequency Hopping Bluetooth - there's some fun and opportunity for learning too), now I'm back onto the rtl sdr and Funcube Dongle Pro+ side of things again.
I didn't see the new Funcube Dongle Pro+ mentioned in this discussion, only the original dongle. How does that dongle change the discussion?

--Jeff KK4ETK

[Leif Asbrink]

This might be of interest:
https://www.youtube.com/watch?v=NYhVvtbqaD0

Regards

Leif