Sun Noise

The sun emits RF energy. It's the strongest radio noise signal in the sky and is often used to measure the performance of a receiver using a measurement knows as "Sun Noise". The measurement process consists of pointing the antenna at the sun and measuring the strength of RF signal received at the wavelength of interest. There are a number of ways to measure the strength of the signal. The simplest method is to measure the receiver audio noise power (with AGC OFF over the audio frequency bandwidth of the reciver (typically about 2.5KHz in SSB mode). It can also be measured over a wider bandwidth using, for example, and SDR and suitable software. Whatever method you use, you get some sort of power reading. Then you point the antenna at the cold sky and take a second reading. The difference between these two reading is "sun noise", sometimes knows as the "Y-factor". Sun noise is normally measured in dB. A 3m dish on 1296 might give a power reading that's 10dB higher with the antenna pointed at the sun then with the antenna pointed at the coldest part of the sky. You would them say the system shows 10dB of sun noise.

The higher the gain of your antenna and the lower the noise figure of your receiving system, the greater the noise difference between cold sky and the sun will be. So if you tweak your system and see higher sun noise, then either your antenna gain has gone up or your noise figure has gone down, both of which are "a good thing".

Sound easy (and it is), but what's the catch when comparing your reading with someone else's reading, or readings taken in different days of your system. Well, there are several:

Fluctuations in Solar Flux

  • Note: One solar flux unit = 10E-22 watt per square meter-hertz. 1 sfu = 10,000 Jansky.
  • The total RF power emitted from the sun at radio wavelengths isn't constant and it is generated by a number of different mechanisms.
    • There's a fairly constant background level of noise that is always present. This is noise from the "Quiet Sun". At wavelengths shorter than abut 1cm this radiation is that expected from a black body at a temperature of about 6000K. Visible sunlight, for example, follows the black body cured. At wavelength longer then 1cm the RF emitted by the sun is higher than that of a 6000K black body and doesn't follow the wavelength dependence of a black body curve. At 1296 MHz for example, the solar flux is around 10x that you would expect from a 6000K black body. Noise from the quite sun has a constant wavelength dependence, so if you know the flux at one wavelength, you can estimate it pretty accurately at another. Quite sun radiation is randomly polarized.
    • There is also a slowly variable component of RF emission which tracks pretty closely with sunspot number. The frequency distribution of the slowly varying component is different from that of the quiet sun and peaks in the 10cm wavelength region. The origin of this component is sunspots and the area around them and extends up from the photosphere (visible "surface" of the sun") into the chromospher and corona. Polarization is a mixture of random and circular (due to intense magnetic fields). This radiation typically varies over days and weeks as sun-spots are carried across the visible face of the sun due to the 28 day solar rotation.
    • There are also rapidly varying components if solar RF emission. These change over time scales of seconds to hours. They are caused to a number of phenomena such as solar flars. Their magnitude can be significantly higher than the background quiet sun and and the RF emission can be quite strongly wavelength dependent. A large flare may release as much energy as 1 million 100 megaton hydrogen bombs. Polarization is random and circular.
  • The upshot of all this is that the sun is not a constant RF noise emitter. The RF power varies form year to year, month to month, day to day, hour to hour and minute to minute. In addition the emission spectrum can change, so that the ratio between, say, emission at 10cm and emission at 23cm is not always the same. This makes using sun noise as an absolute reference quite tricky. It's possible to measure 10dB one day and 11dB the next day, and maybe 13dB 5 years later. It's even possible to measure 10dB at 12:00 and 11dB at 12:30.

So unless you know how much RF power the sun is emitting at the frequency of interest and at the time you make your measurement, you can't tell if changes you measure in sun noise are due to changes in your antenna and receiving system or to changes in the amount of RF the sun is emitting. RF emission from the sun are constantly monitored by a number of observatories at a wavelength at 10.7cm (the value of 10.7 is a standard for historical reason - it's where the first measurement were made) and those numbers are published online. This is known as the "10.7cm Solar Flux". While this is a useful number for solar research, it's not all that useful for estimating the solar flux at other wavelengths for use at amateur radio operators. Fortunately a number of observatories make measurements at multiple frequencies - 245, 410, 610, 1450, 2695 (10.7cm), 2800, 4995, 8800 and 15400 MHz. From these numbers, values for the amateur bands can be interpolated with reasonable accuracy.

https://spaceweather.gc.ca/solarflux/sx-4-en.php will give you a recent 10.7cm solar flux measurement, and the time it was taken. The measurements are reported for 1700UTC, 2000UTC and 2300 UTC (10am, 1pm and 4pm at the Dominion Observatory in British Columbia). In those 3 hours between readings the value might not have changed at all, or it might have increased or decreased by 10% or 20%. If there's a solar flare it could change by 100% or even more. You can see some historical solar flux data at https://spaceweather.gc.ca/solarflux/sx-5-en.php so you can get an idea of how much and how quickly the numbers can change. See also About the Solar Flux Data for an explanation of the three different values for 10.7cm solar flux which are given.

EMECalc provides Solar Flux data at multiple frequencies from a number of sources including NOAA and Learmont observatory in Australia. It does not seem to be able to access the Canadian measurements since they are always displayed as "-1". The time of the last NOAA update is given. Updates seem to be about every 6 hours and historical data is available (though only as single data points index by how many readings before the current reading). The solar is given in as a number rounded(?) to the nearest integer. Not all data is received from all stations due to format and url changes, but there is still usually some good data.

A few links for solar data are given below. Not all of them may work, and not all of them may contain data at all frequences, or any data at all. However if you go through them you should find somewhere that seems to be reporting good data, at least at 10.7cm.

Note that different observatories may well report slightly different values of solar flux, so if you are trying to standardize your own readings, always use the solar flux values from the same observatory. The differences reflect calibration differences of the various systems. That all use different antennas and different receiving systems and it's not easy establishing an absolute standard. They should all report the same changes in the numbers though. As I write his I'm looking at 10.7cm solar flux numbers of 78 from stations in Australia and Hawaii, 70 from Massachusetts and 73.1 from British Columbia. As I said, always chose the 10.7cm SFU number from the same source if you want to meaningfully compare your measurements.

There is a long term variation in solar flux which follows the 11/22 year solar activity (sunspot) cycle. The long tern trend is shown in the plot below. The flux at other wavelengths also follows this general trend

There are also short term variations in solar flux over a matter of a hours, even during times of a "quiet" sun. Here's a plot illustrating this point (base date taken from spaceweather.gc.ca). Each day 10.7cm measurements were made at 10am, 1pm and 4pm local time (shown joined by the white lines). No data is available for the period between the measurements. Solar flux at other wavelengths e.g. 23cm might be expected to show similar variation. A 2.5% variation would correspond to about a 0.1dB difference in a sun noise measurement.

This plot shows how the variation of solar flux over a 7 days (144 hours) period would affect a particular sun noise measurement. It shows how the measured sun noise from the same system might vary from a high of about 11.14dB at around the 48 hour point to a low of 10.43 at around the 146 hour point. This is due to a variation in solar flux from a high of 78 to a low of 73. You can see that the reading can change significantly over just a few hours. If you're looking at a "daily average" value for solar flux, measurements can be in error unless you are also averaging your sun noise measurements over a day! The bottom line is that any sun noise measurement has some amount of inherent uncertainty since there's always some uncertainty about the exact value of the solar flux at the time of the measurement. I can measure the value of sun noise with a resolution of 0.01dB, but the uncertainty in the accuracy of the measurement is probable at least +/- 0.05dB even if my measurement is taken close to the time of a published 23cm SF measurement.

Differences in Measurement Techniques

See also - Sun Noise Measurement

  • Many of the sun noise measurement techniques presume the power measurement is well calibrated. The techniques measuring audio power assume linearity of the system (which is why AGC MUST BE OFF). If you put 2x the RF noise power into the antenna they assume that you get 2x the audio noise out of the receiver (3dB noise increase). This may or may not be true. Similarly when measuring RF power may any other means, you're assuming the system is consistent calibrated. If you always use the same equipment and use the same technique to measure sun noise, then your own readings will be self consistent, but comparing your numbers with those of another station.
  • It's possible to eliminate any system non-linearities by the use of a calibrated (step) attenuator. If an attenuator is inserted in the receive line after the preamp - and the preamp has enough gain - then the following procedure can be used. First a measurement of cold sky is taken. Finding the antenna direction which generates least noise is a function of both the temperature of the sky seen by the main lobe of the antennas, and noise picked up by spillover (in the case of a dish) and side and back lobes. That noise may be comprised of both ground noise (due to objects at ambient temperature) and any man made noise from electronic devices. Second, the antenna is pointed at the sun, which will give a higher noise reading. The in line attenuator is then adjusted to bring the noise level back down to that which was measured for the cold sky. The value of the attenuator is then equal to the measured value of sun noise. Of course this depends on accurately knowing the value of attenuation at the measurement frequency.
  • How the professions do it https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/swe.20064

For more about measuring sun noise the the page here on Measurement of Sun Noise

The bottom line

The "take away" from all this is that you need to be careful when comparing sun noise between stations, and even comparing sun noise measurements made by the same station with the same equipment on different days or at different times. Any sun noise measurement has to include the value of the solar flux at the time of measurement - and that's not easy to find. There's a list sun noise measurements (in Microsoft Exel format) at http://www.ok2kkw.com/next/nl_k2uyh/sun_table.xls which is maintained by OK1TEH. It includes solar flux numbers, but the accuracy of the measurements and how the solar flux at the time measurement was obtained in unknown. Also they are presumably 10.7cm solar flux values, and as noted above, the solar flux at over wavelengths is not exactly proportional to the 10.7cm flux. So while the difference between a 10dB listing and 11dB listing is probably real, the difference between, say, 10.4dB and 10.6dB may not be. Without knowledge of the actual solar flux at the wavelength of measurement and at the time of measurement, comparing sun noise measurements is difficult.

Notes

The usual way of measuring sun noise gives does not give a signal to noise ratio. It's a (signal+noise/noise) ratio. The two are not the same, and the difference gets larger as the signal gets smaller.

WSJT10 has a very basic noise measurement mode built in. It's not very flexible, but will give you a readout of receiver noise (0.1 degree resolution) once a second using whatever averaging routine is built into the program. It's very easy to use.

Maximizing sun noise maximizes the RECIEVE performance of your system. It says nothing about transmit performance. The feed pattern for illuminating a dish which produces the lowest sky temperature is different from that which produces maximum gain. For lowest sky noise you want minimum spillover (without sacrificing too much gain), giving the best G/T (Gain/Temperature) ratio. This corresponds to dish illumination which may be around 13dB lower at the edge of the dish than the center. However for maximum gain the feed should illuminate the edge of the dish only about 10dB down on the center. So modifying the feed for best sun noise (best G/T. best receive performance) will probably not give you the maximum possible dish gain (strongest transmit signal).

An excellent article of the nitty gritty details of sun noise measurement was written by VK3UM (SK). It can be found here http://www.vk3um.com/SunNoise_Measurements.pdf

See also: