Hi George,
(and anyone else who is interested in an easy way to evaluate your satellite rx performance)
These remarks are with respective to linear birds, not FM
Some of the satellite ops I work with run various SDRs (software defined radios like the FunCube Pro+). The software we use with them is called SDR Console v3 (Google SDR Radio or look at this group:
main@sdr-radio.groups.io)
What we do so we can compare apples to apples is the following:
The software is designed to be able to measure relative signal levels at high accuracy (-dBM)
1. Connect a 50 ohm resistor to the antenna or preamp input. 2. Set the rx for maximum sensitivity (RF Gain) 3. Adjust IF Gain so that the 50 Ohm resistor produces a signal level of -95 dBM 4. This creates a common noise floor level calibration point so that we can compare local noise (environmental) as well as satellite signal levels.
5. Now disconnect 50 ohm resistor and change to "real" satellite antenna.
Observe the increase in noise floor. If there isn't one, you have poor rx sensitivity. If there is an increase and it is 6 dB or more, you have local noise problems (environmental). If there is a small increase in noise, you likely have relatively low noise, and at least it exceeds the internally generated noise of your radio/preamp combo.
Now, when we observe signals from the satellites, we can directly compare peak signal strength in dBm between our satellite setups. This includes, antenna polarity, antenna gain, feedline losses, differences in uplink power, differences in receiver noise figure, and locally generated noise like power line, WiFi hash, etc. All...easily and accurately done by having software that works correctly and a simple SDR Dongle using Simon's (G3ELI) software (which is free).
For example:
If KB7IJ has a noise floor of -76 and I have a noise floor of -90 (dBm), and he has a peak BEACON signal strength of -53 and I have a peak strength of --60, we can see two things; (these are real and typical measurements, btw)
1. His ambient noise is 14 dB higher than mine. He lives south of Dallas, Tx, and has to shoot thru the city to see sats. Lots of environmental noise! I live in rural Iowa. Much quieter.
2. His rx system has 7 dB greater gain than mine (antennas). That makes sense as he has 2x7 EL yagis and I have a single 5 EL Yagi. So all else being equal (and it almost never is), he should always see the birds 7 dB louder than I do (actual signal level, not SNR)
Now look at SNR (Signal to Noise Ratio):...a much different story. How well do we hear the birds?
His SNR (best): -76 dBm noise floor - (-53) signal = 23 dB SNR (signal is 23 above noise, loud!)
My SNR (best ): -90 dBm noise floor - (-60) signal = 30 dB SNR
So when we are both getting a satellite at our best signal level, I am hearing the bird 7 dB better than he is, even though his rx antennas are much better.
Look what happens when the ssb signal is much weaker:
If I am seeing the SSB signal at an SNR of 10 dB (the lowest signal strength that an SSB signal can be understood at):
He is seeing it at only 3 dB SNR which is unreadable.
In general, I nearly always hear both him and myself way better than he does, because my noise floor is so much better than his.
Keep in mind, these are all RELATIVE measurements not ABSOLUTE. They are not meant to be lab grade, they are meant to be able to understand and measure reasonably well, how a satellite station is performing in the real world. Common calibration between stations allows us to compare how our systems are working. It's not perfect, but it is useful. (There are other factors, like the practical limit of SNR from any bird based on passband noise). If you can consistently see/hear passband noise, you are good. If you only hear it now and then, your rx system needs work)
p.s. The software from Simon, also allows us to make MP4 recording (video and audio) of our satellite qsos, which can demonstrate some of the very dramatic ways to improve or ruin your ability to use satellites efficiently. That is to say, receive properly and not overdrive the satellite with excessive uplink power, because you can't hear worth a darn and don't know it. (Just continue to increase transmit power so you can hear yourself, taking more than your share of what is available, because it is a shared resource).
I have hours or recordings of assorted stations doing exactly this (running excessive uplink power and thus producing downlink signals 6 dB or more louder than the CW beacon) and therefore stealing the available downlink power of others, driving them down into the noise floor. It's embarrassing.
*The rule is simple: you should NEVER be louder than the CW beacon, period.*
73, N0AN Hasan
On Wed, Mar 4, 2020 at 7:48 PM George Sakai grsakai5120@yahoo.com wrote:
Thanks for the info Hasan. What do you use to measure signal levels in dBm?
George N3GS
On Wednesday, March 4, 2020, 03:43:43 PM CST, Hasan al-Basri via AMSAT-BB < amsat-bb@amsat.org> wrote:
K7VNE reported to me this morning that he had been on XW-2D (long since non-operational) and sigs were very good, but no one else was on.
Signals were of good quality, no FM'ing of my SSB signal.
I just checked it and had a pass of Max EL 2 deg to my East. It is there, it is loud. *The CW beacon is raspy*.
The passband was normal level and my signal was typical:
Max Sig: -64 dBm Noise Floor: -81 dBM
I made an mp4 recording of the pass, with me calling CQ and announcing Elevation and signal levels.
https://drive.google.com/open?id=1lKYvARp4SyHUraAX9u6HQRbmnCiamJAM
Download this mp4 file and play it back on any device (even a phone) and you will see and hear what I did.
I have seen no announcement about not using the satellite, but will watch for same. 73, N0AN Hasan _______________________________________________ Sent via AMSAT-BB@amsat.org. AMSAT-NA makes this open forum available to all interested persons worldwide without requiring membership. Opinions expressed are solely those of the author, and do not reflect the official views of AMSAT-NA. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: https://www.amsat.org/mailman/listinfo/amsat-bb
On 2020-03-05 04:06, Hasan al-Basri via AMSAT-BB wrote:
Keep in mind, these are all RELATIVE measurements not ABSOLUTE.
Indeed. I suggest not using the unit dBm, as that is decibels relative to 1 milliwatt. This is an absolute unit.
It is common to refer to some other standard when comparing relative power, such as dbFS (decibels relative to full scale, usually a negative number) for an ADC.
Perhaps dBANF might work, which would be dB relative to Agreed Noise Floor. I don't think I'd use "dBNF", as that might be confused with Noise Figure in dB.
--- Zach N0ZGO
On 2020-03-05 04:06, Hasan al-Basri via AMSAT-BB wrote:
- Connect a 50 ohm resistor to the antenna or preamp input.
- Set the rx for maximum sensitivity (RF Gain)
- Adjust IF Gain so that the 50 Ohm resistor produces a signal level of
-95 dBM 4. This creates a common noise floor level calibration point so that we can compare local noise (environmental) as well as satellite signal levels.
Hello Hasan,
Ah, noticed one more thing. This is only true if the measured thermal noise power bandwidth is the same on both setups.
Thermal noise power is k*T*B, where k is Boltzmann's constant (1.38E-23 J/K), K is "temperature" of your source in Kelvins (290 is about room temp), and B is the bandwidth of the noise in Hertz (1/seconds).
We assume that there is no current flowing through the resistor for this example.
(1.38E-23 J/K)(290 K)(1 Hz) = 4.002E-21 Watts
[Kelvins cancel out, leaving Joules/seconds, which is Watts -- math is cool!]
However, this is an annoying figure to remember, so we convert it to dBm:
(4.002E-21 W)(1000 mW/W) = 4.002E-18 mW
10 * log10(4.002E-18 mW / 1 mW) ~= -174 dBm
(~= is "approximately equal", rounded)
Now that's an easy number to remember, and all radio amateurs should!
Notice that we initially computed this noise floor with Bandwidth of 1 Hz. Not even the slowest CW would fit in that, so let's make it something more reasonable like 3kHz for voice.
Instead of doing all that math again, we can use logarithms to fix this up. Multiplication becomes addition when we're in a log scale (mmm, slide rules..), so:
-174 dBm + log(3000 Hz / 1 Hz) ~= -170 dBm
This is the thermal noise floor in a 3 kHz bandwidth from a noise source at 290 Kelvin (62 deg F).
Anyway, all of that was to show that bandwidth is very important when comparing noise power of two receivers.
If anyone is interested in learning more about noise floor, noise figure, or required SNR, here's a great article:
--- Zach N0ZGO
Hi Zach, Since we were speaking of SSB SNR, the bandwidth would be similar. In our cases we use 2.2 kHz. I forgot to even mention it.
The ...and the point of the whole process was to keep it simple. The more precise, the more complex and the less likely anyone is going to bother.
The perfect is the enemy of the good.
Suffice it to say, with hours and hours of measurements using the approach indicated, it is clear that if the antenna gain is increased by x dB or the uplink power is changed by x dB, the reference calibration between two setups holds nicely. There are a ton of other factors at any given moment that one can get lost in, but wash out with repeated measurement.
Using the approach I described, can a minimally equipped (test) station determine whether they have adequate sensitivity for their sat ops, or if they have improved their setup? Of course they can and they won't be stuck doing math they don't understand and has little bearing on the real world performance of their satellite station.
Many, many years ago Amsat Journal published an article on Link Budget Analysis that did all the math and painstaking analytics you describe. How many people made use of it....next to none. I know, because I wrote it.
What I was aiming at was a simple, quick and dirty "rough estimate" of : 1. Is my system sufficiently sensitive 2. Am I running too much power.
Boltzman is not needed for that. We aren't doing EME. Sat ops are a relatively strong signal mode. It says a lot that many receive setups are performing so poorly that "rough and dirty" (and simple) approaches like I outlined can make a big difference in overall efficiency. I just hope that it helps people hear better. (and subsequently reduce their uplink power)
73, N0AN Hasan
On Thu, Mar 5, 2020 at 11:16 AM Zach Metzinger via AMSAT-BB < amsat-bb@amsat.org> wrote:
On 2020-03-05 04:06, Hasan al-Basri via AMSAT-BB wrote:
- Connect a 50 ohm resistor to the antenna or preamp input.
- Set the rx for maximum sensitivity (RF Gain)
- Adjust IF Gain so that the 50 Ohm resistor produces a signal level of
-95 dBM 4. This creates a common noise floor level calibration point so that we
can
compare local noise (environmental) as well as satellite signal levels.
Hello Hasan,
Ah, noticed one more thing. This is only true if the measured thermal noise power bandwidth is the same on both setups.
Thermal noise power is k*T*B, where k is Boltzmann's constant (1.38E-23 J/K), K is "temperature" of your source in Kelvins (290 is about room temp), and B is the bandwidth of the noise in Hertz (1/seconds).
We assume that there is no current flowing through the resistor for this example.
(1.38E-23 J/K)(290 K)(1 Hz) = 4.002E-21 Watts
[Kelvins cancel out, leaving Joules/seconds, which is Watts -- math is cool!]
However, this is an annoying figure to remember, so we convert it to dBm:
(4.002E-21 W)(1000 mW/W) = 4.002E-18 mW
10 * log10(4.002E-18 mW / 1 mW) ~= -174 dBm
(~= is "approximately equal", rounded)
Now that's an easy number to remember, and all radio amateurs should!
Notice that we initially computed this noise floor with Bandwidth of 1 Hz. Not even the slowest CW would fit in that, so let's make it something more reasonable like 3kHz for voice.
Instead of doing all that math again, we can use logarithms to fix this up. Multiplication becomes addition when we're in a log scale (mmm, slide rules..), so:
-174 dBm + log(3000 Hz / 1 Hz) ~= -170 dBm
This is the thermal noise floor in a 3 kHz bandwidth from a noise source at 290 Kelvin (62 deg F).
Anyway, all of that was to show that bandwidth is very important when comparing noise power of two receivers.
If anyone is interested in learning more about noise floor, noise figure, or required SNR, here's a great article:
--- Zach N0ZGO _______________________________________________ Sent via AMSAT-BB@amsat.org. AMSAT-NA makes this open forum available to all interested persons worldwide without requiring membership. Opinions expressed are solely those of the author, and do not reflect the official views of AMSAT-NA. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: https://www.amsat.org/mailman/listinfo/amsat-bb
On 2020-03-05 13:00, Hasan al-Basri wrote:
Since we were speaking of SSB SNR, the bandwidth would be similar. In our cases we use 2.2 kHz. I forgot to even mention it.
The ...and the point of the whole process was to keep it simple. The more precise, the more complex and the less likely anyone is going to bother.
Hello Hasan,
I definitely applaud your efforts to measure relative system performance, but I did want to make it clear to others on the list that there was another factor which needed to be taken into account when doing an apples-to-apples. This additional factor isn't too hard, when you understand why it matters.
Boltzman is not needed for that. We aren't doing EME. Sat ops are a relatively strong signal mode. It says a lot that many receive setups are performing so poorly that "rough and dirty" (and simple) approaches like I outlined can make a big difference in overall efficiency. I just hope that it helps people hear better. (and subsequently reduce their uplink power)
I also completely agree with you on the need for fewer alligators on the satellites. With GOLF, where we're building bigger vehicles with larger orbits and footprints, excessive uplink power becomes an even bigger issue for more simultaneous users.
That said, GOLF satellites will also incur more path loss, so getting LEO operators more familiar with the "down in the noise" issues will make for better MEO/HEO operators, too.
Keep up the good work! I enjoyed your video.
73,
--- Zach N0ZGO
My pleasure, Zach. KB7IJ and I have logged probably 1000 hours of measurements to arrive at the "rough and dirty" approach listed and despite all the confounding variables like:
1. Rotating Polarization 2. Faraday 3. Satellite Body Shading 4. Broken sats (Like XW-2F with it's 20 dB tumble) etc., etc.
it became apparent that if we were running the same software, same bandwidth (as you noted), and calibrated to a simple dummy load, we could isolate variables quite quickly and see if what we "did" made a predictable difference. Having the ability to record both video and audio and play back with announcements of the "real time' change, is quite an advantage.
Operating sats without an SDR or a 9700 is operating blindfolded. Once a person can "see" the entire passband in full duplex in real time, a new world of what is going on opens up. Of special importance is the ability to see the beacon 100% of the time and adjust uplink power accordingly.
It is quite interesting to see the CAS series (4A/4B) with not only their CW beacon, but also the Batwing PSK beacon. Seeing that beacon fade 20 dB on one polarization and then return to full strength immediately upon changing to another polarization is quite awakening. There are also profound pre and post TCA (Zenith) signal strength patterns that can be strikingly correlated to the "nature of the pass". Very Shallow, Shallow, Medium, Overhead etc.
There are so many variables involved, even the 3 dB worst case loss from linear to circular does not hold ...in fact it is very rare that my 2m EggBeater is within 3 dB of my 5 EL vertically polarized yagi. Close in vegetation loss, ordinary obstruction loss (EB is only up 7', Yagi up 65') Of course at very high elevations the EB shines...but only for a very few minutes.
Bob Bruninga hit on a great find with his Fixed Elevation of 15 degrees for LEO birds. You can look at my pix on my qrz page and see what his approach looks like. The performance has been phenomenal and no elevation rotor is required. I routinely work every one of the Mode B LEO birds down to an elevation of -0.7 degrees with that setup, and only at EL > 60 deg do I see the EB outperforming Bob's simple suggestion.
Sats are fun for the curious. Anyone who loves satellites, get an SDR ...you will never go back. 73, N0AN Hasan
On Thu, Mar 5, 2020 at 1:33 PM Zach Metzinger via AMSAT-BB < amsat-bb@amsat.org> wrote:
On 2020-03-05 13:00, Hasan al-Basri wrote:
Since we were speaking of SSB SNR, the bandwidth would be similar. In our cases we use 2.2 kHz. I forgot to even mention it.
The ...and the point of the whole process was to keep it simple. The more precise, the more complex and the less likely anyone is going to bother.
Hello Hasan,
I definitely applaud your efforts to measure relative system performance, but I did want to make it clear to others on the list that there was another factor which needed to be taken into account when doing an apples-to-apples. This additional factor isn't too hard, when you understand why it matters.
Boltzman is not needed for that. We aren't doing EME. Sat ops are a relatively strong signal mode. It says a lot that many receive setups are performing so poorly that "rough and dirty" (and simple) approaches like I outlined can make a big difference in overall efficiency. I just hope that it helps people hear better. (and subsequently reduce their uplink power)
I also completely agree with you on the need for fewer alligators on the satellites. With GOLF, where we're building bigger vehicles with larger orbits and footprints, excessive uplink power becomes an even bigger issue for more simultaneous users.
That said, GOLF satellites will also incur more path loss, so getting LEO operators more familiar with the "down in the noise" issues will make for better MEO/HEO operators, too.
Keep up the good work! I enjoyed your video.
73,
--- Zach N0ZGO _______________________________________________ Sent via AMSAT-BB@amsat.org. AMSAT-NA makes this open forum available to all interested persons worldwide without requiring membership. Opinions expressed are solely those of the author, and do not reflect the official views of AMSAT-NA. Not an AMSAT-NA member? Join now to support the amateur satellite program! Subscription settings: https://www.amsat.org/mailman/listinfo/amsat-bb
On 2020-03-05 10:38, Zach Metzinger via AMSAT-BB wrote:
-174 dBm + log(3000 Hz / 1 Hz) ~= -170 dBm
This is the thermal noise floor in a 3 kHz bandwidth from a noise source at 290 Kelvin (62 deg F).
Thanks go to WB4APR for spotting my math error; I added Bels to deciBels. At the time, I thought "that seems a bit low", but never went back to check my work. :-)
The correct answer is:
-174 dBm + 10*log(3000 Hz / 1 Hz) ~= -139 dBm
--- Zach N0ZGO
participants (2)
-
Hasan al-Basri -
Zach Metzinger