[Moon] [Moon-Net] SDR IQ stability

[Peter Blair]

Thanks Bob, thats a useful input. I was really just looking to eliminate the SDR/ SV combo, which I think from your comments I can do. I’m aware of the other sources of drift. In the past I have insulated the LNA at the feed to reduce that variation. My receive system (on all bands) is all HB right down to the final 14MHz IF, all run from properly stabilised supplies and the most variation comes fom the LNA which is out in the open. 
Its a bit off topic but I wonder if youve done any zenith drift thrus over  a long period, say 12-15  hours? Ive done that on both 23 and 432 with interesting results which I will send you. After I did this, Doug, VK3UM put a feature into the “sky noise display” tab in the VK3UM planner which will plot the sky temperature against RA  for any given declination, click on the “chart noise” button at the lower right corner and select a declination and plot. My dish won’t look exactly at zenith in the stowed position, its about 10 degees off. So, interestingly, you can export the plot as a csv file and compare it to the csv file generated by the Spectravue. (I think youhave to time reverse it, I cant remember now) When I took out the gain change slope I got extremely good agreement on 432. On 1296 you have to scale the temperature values of course.
For me Taurus is the most useful source on 23,13 and 9cm , although its much weaker than Cygnus, I can see it over a much wider area of sky due to the construction of my dish mount.
73 Peter G3LTF

From: Bob Atkins 
Sent: Thursday, April 09, 2020 7:54 PM
To: freetalk moonnet ; Moon-Net ; Peter Blair 
Subject: Re: [Moon-Net] SDR IQ stability

I run a FunCube pro+ with MAP65 and SpectraVue all the time - looking at a 90kHz slice of the spectrum around 1296. A laptop PC (always on and plugged into an AC power supply) supplies the power. Sometimes I leave the antenna pointed at cold sky for hours (overnight) with SpectraVue set on the 0.2dB per division range. Unless someone turns on a TV or other broadband noise generating device (or a Galileo satellite passes through the bean...)  it's very stable over hours. Certainly better then 0.05dB. The FunCube pro plus sits inside an insulated case and is powered up 24/7/365 to minimize any temperature induced instability. The preamp is also powered up 24/7/365.

However, pointing at an area of sky that isn't the coldest spot, where presumably I'm seeing some terrestrial noise, stability can be significantly worse, and moving the antenna 10 degrees from my "cold spot" not only increases the noise floor but lowers the stability. My best "cold spots" are all at 55-65 degrees elevation.

My conclusion is that the SDR/SpectraVue combination is very stable, but there are significant variable terrestrial noise sources that the antenna sees when looking at the noise floor with 0.1dB resolution with a stationary antenna. I don't see any power supply or temperature effects. The main instability problem appears to be related to natural and man made noise sources being picked up by the antenna.

BTW I see about 8.4dB between my coldest spot in the sky and a 50 ohm termination.  I don't know if the shift from the feed to a 50 ohm termination results in an impedance change which affects the noise. The best I can see with the dish (which won't do down below 0.5 degrees elevation), is around 7.4dB, but I don't have any thick trees close enough to full the pattern, so I'm sure I'm still seeing some sky.

I have not yet positively detected moon noise (or other celestial broadband radio source except for H emission at 21cm) with my 3m dish. It should be possible, put whenever I have tried there has been enough noise floor variation that even with the antenna in none position and allowing the moon to drift through the beam, I still can't positively identify the moon noise contribution. Moon noise of 0.2dB can clearly be seen with me 85cm offset dish on 10Ghz. The noise floor on 10Ghz with that dish seems much more stable with respect to dish position, probably because there's not much man made noise at 10Ghz and spillover/sidelobes may be less of an issue.


73
Bob, KA1GT

--------------------------------------------------------------------------------

From: Moon-net <moon-net-bounces at mailman.pe1itr.com> on behalf of Peter Blair via Moon-net <moon-net at mailman.pe1itr.com>
Sent: Thursday, April 9, 2020 7:49 AM
To: freetalk moonnet <moon at moonbounce.info>; Moon-Net <moon-net at mailman.pe1itr.com>
Subject: [Moon-Net] SDR IQ stability 

To save me repeating what others may have already done....has anyone actually looked at the gain stability (at the 0.1 / 0.05 dB level)  of an SDR IQ with a Spectravue continuum display with time? Im thinking about supply voltage effects, for instance is it better supplied directly rather than via USB from PC, or temperature effects.

I’m looking at radio stars for calibration, like Taurus at 13cm, so typically 0.4dB or so with 190kHz BW. I dont think this is what SDR IQ was designed for! I know from other work that things like the 78 series voltage regulators have a dV/ dT change which can affect stage gain. 
I know everything else in the system is prone to gain drift but I want to know if the final “measuring instrument” is stable and over what period. 
73 Peter G3LTF