Why are these SDRUno instances performing so differently?

Discussions about everything to do with SDRuno
Post Reply
franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Why are these SDRUno instances performing so differently?

Post by franko » Tue Oct 16, 2018 9:47 pm

I'm scratching my head over results I got when testing a couple of RSPduo units today, and would be interested in any ideas.

I have two RSPduo's (call them units A1 and A2) both connected to the same Windows 7 PC. Antenna port 2 of each RSP is connected via SMA cable to one of the two outputs of a Mini-Circuits ZFRSC-42 splitter. The antenna side of the splitter is attached to an RTL-SDR.com 2.6 MHz high-pass filter (due to strong MW broadcast signals in our urban area). The input of the high-pass filter is attached to a Wellbrook magnetic loop antenna.

On the PC, I start SDRUno v1.23, tell it to connect to RSPduo A1, open a receiver window and set it to the 20-meter FT8 band. The audio output goes via a Virtual Audio Cable channel to an instance of WSJT-X v1.9.1. I then start another instance of SDRUno, and it automatically connects to the second RSPduo unit (which I confirm by checking serial number in Main > Opt > About); I open a receiver window and also set it to the 20-meter FT8 band. The audio output goes via a second Virtual Audio Cable channel to a second instance of WSJT-X.

When I observe the waterfalls in SDRUno's Aux SP windows, the signals for the first RSPduo are visible brighter. The first instance of WSJT-X shows more than twice as many decoded FT8 signals, and any given signal is typically 2 to 4 dB better.

Now here's the odd part. I exit entirely from SDRUno, then launch a fresh instance. This time, I select RSPduo A2 for it to connect to first. I then open a second instance of SDRUno, and it automatically connects to RSPduo A1. Contrary to my expectation, RSPduo A1 did not continue to outperform RSPduo A2. Instead, whichever RSP was connected to the first application instance of SDRUno (designated as "0" in the upper righthand corner of the SDRUno Main window) always performed better.

I then started looking through all of the settings in SDRUno, Virtual Audio Cable and WSJT-X to see if I could find anything that explains this. In SDRUno I examined the settings in the Main, RX and EX Control windows, and could find no differences. However, I did notice that the RX Control window for the first instance of SDRUno ("0-00") consistently showed a dBM of about -85, while the RX Control window for the second instance of SDRUno ("1-00") showed a dBm of about -130. (Would this number be the signal-to-noise ratio?)

So I'm baffled. Are there any other settings I should checkin SDRUno? Or any other ideas for what would cause the performance difference?

Reason: No reason

sdrplay
Posts: 978
Joined: Wed Jan 07, 2015 7:58 am

Re: Why are these SDRUno instances performing so differently?

Post by sdrplay » Wed Oct 17, 2018 4:42 am

Can you post screenshots of the two instances of SDRuno? Things like the gain control settings are specific to the SDRuno instance and not the RSP. To be 100% certain of the same setup for each instance, you could perform a registry reset from the Main Window OPT menu.

Best regards,

SDRplay Support

Reason: No reason

franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Re: Why are these SDRUno instances performing so differently?

Post by franko » Wed Oct 17, 2018 2:39 pm

Thanks for the suggestions. As an unrelated issue, one of the two hard drives in the PC's RAID array appears to be gradually failing, so I'm leaving the computer off for a couple of days until a replacement drive arrives. By the weekend I'll probably be ready to resume working with SDRUno, and will try the registry reset. If that doesn't equalize the SDRUno instances, I'll post screenshots.

Reason: No reason

franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Re: Why are these SDRUno instances performing so differently?

Post by franko » Mon Oct 22, 2018 12:07 am

Okay, I have my hard disk issue resolved, and am ready to resume working on the question I posted about.

I gather that proceeding with Main > Opt > Reset to Default Settings will return SDRUno to its factory defaults. I'm reluctant to do this at this point, because there may be settings in the currently defined workspaces that I'll decide I want to retain.

I was hoping that I could load the identical workspace into both SDRUno instances 0 and 1, but it seems as though each SDRUno instance only has access to its own defined workspaces. Is there any way to share workspaces between them?

In the meantime, below is a screen shot of two SDRUno instances, 0 and 1. Each is connected to a separate RSPduo, which share a single antenna. The only differences I can see in the way I have them set up is:

-- Instance 0 has one VRX defined. Instance 1 has three VRX's defined, but only one is active.

-- The RF ATTEN slider on both instances is set to three notches up from the bottom. Instance 0's Main window shows Atten: 46 dB, while instance 1's Main window shows Atten: 44 dB. I haven't been able to get the two sliders set in a way to get these numbers to match.

-- The FM mode settings are different, but that shouldn't matter since the active mode is USB.

As noted, instance 0's RX Control window shows -88.8 dBm next to the frequency display. Instance 1 shows -123.6 dBm. Instance 0's waterfall shows visibly stronger signals, and the instance of WSJT-X it's connected to produces more than twice as many FT8 decodes.

Short of doing an SDRUno factory default reset, can you see any settings that would account for the difference in output? As noted, the difference is associated with the SDRUno instances, not with the RSPduos.
instances60.jpg
instances60.jpg (298.98 KiB) Viewed 23434 times

Reason: No reason

sdrplay
Posts: 978
Joined: Wed Jan 07, 2015 7:58 am

Re: Why are these SDRUno instances performing so differently?

Post by sdrplay » Mon Oct 22, 2018 5:46 am

Can you also show the SP1 (Main SP) windows and change the sample rate to 2 MHz in both instances. It will make it easier to see what's going on in the images. Thanks.

SDRplay Support

Reason: No reason

franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Re: Why are these SDRUno instances performing so differently?

Post by franko » Mon Oct 22, 2018 10:13 pm

Hmm, well, I seem to have my issue basically resolved now, though I'm not sure exactly what the cause was, or what fixed it.

When I changed the sample rate to 2 MHz and clicked to display the SP1 windows, I noticed that the LO was somewhat different in one instance, so I sync'd them up. Now the two instances appeared to be producing very similar results, with nearly identical dBm figures in the frequency window (see screen shot below). So, basically, I was no longer able to replicate the problem. Changing the LO around (even making it the same as the tuned frequency) doesn't replicate the performance degradation I'd seen previously, so I'm baffled as to what made it better).

However, after doing this I was still getting very different results within WSJT-X, so I started doing more debugging on that end. Although the problem initially seemed to follow the SDRUno instances and not WSJT-X, now it seemed as though the problem followed WSJT-X and not SDRUno instances. I could not, however, find any differences in the way the WSJT-X instances were configured. One of the members of the WSJT-X development team suggested that I copy the .ini file for the "good" WSJT-X instance into the directory for the "bad" WSJT-X instance, which he said would make them truly identical. After doing this, I found that the two instances of WSJT-X produced identical decode results when taking their input from a shared SDRUno instance.

Finally, I separated them out so that one SDRUno instance passed audio to one WSJT-X instance, and the other SDRUno instance passed audio to the other WSJT-X instance. The result was what I'd call "nearly identical" decodes -- very minor variations occurring once every few dozen decodes, but almost the same overall.

So for my purposes this is "close enough." I'm still in touch with the WSJT-X team to find out exactly what was buried in the .ini files that caused the different behavior of the WSJT-X instances.

Meanwhile, I did want to ask you the following:

1) I can't find anything in the SDRUno manual that states exactly what the dBm number in the frequency display in the RX Control window signifies. Is this signal-to-noise ratio? What settings affect it, and what would be optimal?

2) Even though the RF ATTEN sliders in the Main windows of my two SDRUno instances are both set to the main notch level (three notches up from the bottom), one instance reports Atten: 45 dB and the other reports Atten: 44 dB. This amount of variation probably isn't worth worrying about, but is there a way besides the sliders to set it more precisely?

Thanks again for the help.
instances3.jpg
instances3.jpg (318.78 KiB) Viewed 23370 times

Reason: No reason

Tech_Support
Posts: 499
Joined: Mon Jun 01, 2015 7:00 pm

Re: Why are these SDRUno instances performing so differently?

Post by Tech_Support » Tue Oct 23, 2018 4:32 am

Hello Franko,
The dBm meter measures the total RF power (all signals plus noise) within the SP2 bandwidth. dBm is a standard definition for absolute power and is relative to 1mW, so 0 dBm = 1mW, +30 dBm = 1 watt, -30 dBm = 1uW and so on.

All values are referred to the antenna connector on the RSP and because the RSP is calibrated, they give you an accurate reading of signal strength. There is also an SNR meter, but this only gives a reading for values of 6dB or greater. The reason for this is that is becomes impossible to get accurate SNR estimates with analog signals for SNR values of less than 6 dB. So for weak signals, the SNR will probably just display --

Regarding your question on gain, the slight variation in values that you will have seen will be down to the fractionally different signal conditions that each instance of SDRuno will have been experiencing when the IF AGC converged. It is impossible to start two instances of SDRuno simultaneously and so the IF AGC for each instance will experience slightly different signal conditions for initial convergence.

The IF AGC is simply intended to help prevent strong signals from overloading the ADCs and causing distortion. The IF gain setting has very little effect on the receiver performance as long as the ADCs are not overloaded and certainly 1 dB difference in IF gain setting will have no measureable impact on the performance.

Sincerely

Tech_Support

Reason: No reason

franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Re: Why are these SDRUno instances performing so differently?

Post by franko » Tue Oct 23, 2018 4:49 am

Thanks much for the explanations -- they are very clear. My objective is, first, to test RSPduo's against each other to ensure that there is no significant difference in their performance, and then use them to compare different antennas on a given band in real-time. This information will help enable me to do that.

Reason: No reason

Roger
Posts: 217
Joined: Fri Dec 29, 2017 7:59 pm

Re: Why are these SDRUno instances performing so differently?

Post by Roger » Tue Oct 23, 2018 4:38 pm

franko wrote:Thanks much for the explanations -- they are very clear. My objective is, first, to test RSPduo's against each other to ensure that there is no significant difference in their performance, and then use them to compare different antennas on a given band in real-time. This information will help enable me to do that.
I have compared the two tuners on an RSPduo and they are very well matched. You can record the audio output from both tuners simultaneously using Audacity ( a free audio recording program). You will need to install VB cable so you can transfer the audio to Audacity. Send the output of Tuner 1 to the left channel and Tuner 2 to the right channel. Then record in stereo in Audacity. You can then use Audacity to analyze and play the recording. You will find the two tracks will be time shifted by several milliseconds due to lack of IQ streaming synchronization between the tuners. Audacity does have tools to shift one track in time so you can manually align them to be in sync. When I used the same antenna on both tuners and listened to WWV I could easily align the tracks and it was almost a perfect match between tracks if I used the same bandwidth and volume level.

With different antennas connected to the tuners you should see some interesting observations. Sounds like fun...

Roger

Reason: No reason

franko
Posts: 55
Joined: Mon Aug 28, 2017 9:33 pm

Re: Why are these SDRUno instances performing so differently?

Post by franko » Tue Oct 23, 2018 4:59 pm

Hi Roger, this sounds a little like the diversity mode offered in some transceivers, where the radio makes use of two separate antennas to increase its ability to pull weak signals out of noise. I wonder if there is (or will ever be) a way to do that in SDRUno or other software making use of more than one RSP/antenna?

I have a total of three RSPduo's, two of which I've used a fair amount while the third at this point is still in the box. One of my motivations is that both of the RSPduo's I've used have been briefly exposed to strong RF from short, inadvertent transmission nearby, and I wanted to see if I could detect any impact on the tuner in each one that was active at the time. So far I've tested the 50-ohm ports for tuner 1 and tuner 2 on both actively used RSPduo's, and their performance is all pretty much identical. My next step will be to test them against the as-yet-unused RSPduo, but given the results so far I think they're all fine. After that, I'll probably run tests comparing the 50-ohm port vs the Hi-Z port (with 1:9 balun) on one or more of the RSPduo's to see how they compare.

Then after that I'll test out various antennas. Last night I briefly benchmarked a Wellbrook mag loop and W6LVP mag loop against each other and against a dipole cut for 40 meters that's set at a height of 35 feet. The two mag loops were close to each other, but were blown away by the dipole, which did quite a few dB better. However, this may be affected by placement of each of the antennas, so before coming to any conclusions I'll do further tests using various locations and orientations.

Reason: No reason

Post Reply