SDRPlay - follow up.

FM receiver discussion
Post Reply
Posts: 1246
Joined: 07 Oct 2014 16:49

SDRPlay - follow up.

Post by daveB » 31 Jan 2015 23:19

I've been trying to condense down my experience with the SDRPlay over the past fortnight - and finding it difficult. So here it is for what it's worth and apologies if it is a bit long.

SDRPLay for FM DXing
A side by side comparison with my main analogue tuners – a Yamaha TX-930 and a Sony ST-SB920 – shows that in most cases the SDR+HDSDR+RDSSpy combination is more sensitive for decoding RDS pi codes than either analogue tuner feeding RDSSpy. [Conditions: SDR BW 140kHz; TX-930 2 x150kHz + 2 x 110kHz; SB920 2 x 180kHz + 2 x 150 kHz; avoiding adjacent channel stations. The Yamaha feeds directly from the audio pre-MPX into RDSSpy fro direct decoding at 192kHz].

The SDRPlay is my receiver of choice now – with the added bonus of MW and short-wave reception.

To get the best out of the system I have wrapped the SDRPlay in aluminium foil – grounded to the outer of the aerial f-connector. This has done a good job of eliminating leakage. As the box is tucked behind the PC and analogue tuners it is out of sight and a much easier solution than trying to remount it in a metal enclosure. I have also fitted clip on ferrite chokes on the aerial lead and the USB lead and in actual fact virtually every lead - RF and audio - in the room has a ferrite choke on it.

Gain settings used: 59dB LNA threshold (default=50); ADC setpoint = default; AGC = on.

I generally run 5 MHz as I find 8 MHz makes for a crowded spectrum. On marginal signals it helps to reduce bandwidth down to 600kHz or even 300 kHz then strong signals fall outside the IF passband, the AGC backs off, and improved S/N ratio is obtained. To give an example: Centre frequency 103.8 giving a span on 101.3-106.3. A very strong signal at 102.8 results in 61dB of gain reduction across the 5MHz. Reducing bandwidth to 300kHz and the gain reduction drops to 23dB. Fr. Inter on 103.7 goes from barely getting the RDS to getting a much more solid lock.

I prefer to run with AGC on rather than choosing a fixed gain reduction for the best compromise as for most of the band the gain reduction is around 40-45dB. As the LNA for 60MHz and up is either ON or OFF, by setting a threshold of 59dB (max) the LNA will stay on until 59dB or more gain reduction is set by the AGC thus preserving the best signal to noise ratio.

Under these conditions there is no sign of overload and no detrimental effect on weak signal reception. You can make it overload by turning off the AGC and reducing the gain reduction too far – but you won't find any benefit in S/N ratio before you reach that point that you couldn't achieve by using the AGC and reducing the bandwidth.

Zero IF Images
An inherent problem with this SDR receiver – in common with others that use Quadrature Switching Detectors at zero IF - is the presence of images appearing when there are strong enough signals within the IF passband. Phase and amplitude differences between the I and Q channels has to be very well balanced to get a good enough IF rejection (>65dB) and of the three software packages I've tried only HDSDR has manual slider controls that allow such precise adjustment. Without this precise adjustment a strong signal one one side of the centre frequency appears on the other side around 50dB lower. I set the phase/amplitude balance by setting a strong station to one side of centre frequency and listening the same difference the other side for the same station - just work out the frequency difference so the image will fall on a clear frequency. Then I adjust phase to minimise the signal and then amplitude. In practical terms amplitude balance is a once-off. Phase balance can need tweaking at the start of a new session and can be slightly different at each end of the band. But don't let this put you off! It is very easy to do and only needs to be really fine tuned if you are for example after the weakest signal and don't want the possibility of it being masked by a weak image. Note: This would most likely be the cause of the weak images MikeB reported as SDRSharp automatic I/Q balance doesn't appear to work on FM broadcasts.

HDSDR with SDRPlay
The EXTIO.dll for HDSDR is, I'm told, a work in progress, and needs some refinement. The most obvious missing bit is that, unlike the SDRSharp version, it doesn't yet remember custom settings, but goes back to the defaults.
Apart from HDSDR being the only one of the three software packages I've used that allows for precise setting of I/Q phase and balance it is also the only one that allows raw unfiltered audio to be output which is necessary for RDSSpy to work. It also has a built-in DRM decoder - but I've yet to find a DRM broadcast on SW to decode!
HSDR also has two independent VFO's which keeps the mode - so you could have one set to the FM broadcast Band and one set to AM Medium wave, or one set OIRT and one to FM Band.
HDSDR has a recording scheduler that works very nicely I've not yet tried it to see if it automatically starts the recording - but it certainly stops it and if the option is selected, to power down the computer.
Audio quality is 'brighter' than the Analogue tuners. As HDSDR is mono only the comparison was made in mono. Personally I find the HDSDR audio to be slightly clearer on a weak noisy signal.

Other SDR software
I've tried SDR Sharp and SDR console and queried image rejection on the forums and both have built-in software routines for automatically eliminating the I/Q imbalance. Currently they don't work on my set up on the FM broadcast band and I get one significant image from Heart 102.8 which is a 70dB signal and gives a 15dB image. Cross-checking on the HF Bands I can't see any images of shortwave broadcasters on SDR Sharp or SDR console from strong broadcast signals. I can force them to appear on HDSDR by unbalancing phase and or amplitude drastically so know where they would appear.

Medium/Short Wave listening
The only suitable aerial I have is the half-wave dipole cut for 70MHz – but it makes a nice diversion. I can't make any real meaningful comments about it on the LF/HF bands except to note that a low pass filter with a cut off above 1.6MHz would be needed to eliminate break-through from strong Short Wave broadcasters. But I gather this is not a new problem to MW-DX-ers. I came across this as I was getting Radio Taiwan on 3596 kHz at the low end of the MW band – below 500 Khz initially but when I changed frequency by 9kHZ it moved 36KHz. But to repeat the above - images of strong broadcast stations due to poor rejection of the unwanted side-band were not evident.

Finally - have a look at the Terms and Conditions for purchasing which include the option of returning it within 14 days if for any reason you don't like it. I got confirmation this meant I could trial it and if not satisfied with it performance I could return it - and all they would ask for is some feedback as to why. So if you are seriously considering buying an affordable SDR that is a big leap above an RTL-2832U dongle you have some peace of mind.

Apologies for any residual typos but it is late and I'm getting tired.


SDRPlay, AirSpyMini, RTL dongles, SB920s, Yamaha TX930. 5 ele compact yagi on rotator at 18ft agl. FM5 facing east, FM5 facing south, FM3 facing NE, OIRT dipole, 3-ele Moxon vertical facing east.

Post Reply

Return to “Receivers”