Serveral Issues here - pls help

[Lentas]

Hi @all,
for the last months i was working with a nooelec r820t2 stick on frequencies greater than 100MHz. i´m receiving and decoding radiosondes between 400-406 MHz, Weather Sats on 137-138 MHz and all kind of stuff on 2m and 70cm.
Antenna is a Diamond X-50N Groundplane mounted on the roof, 12m Coax, Software is the latest SDR-Console.

3 Days ago i received the RSP and tried the same things i did with the r820t2 stick, and now i´m very disappointed.

DC Spike & Image Frequencies


Normally the band between 400-406 MHz is empty with weak signals. After "turning on" the receiver a have a DC spike in the middle, that doesnt vanish. And there are two huge intermodulation products/image frequency signals. this is the picture when i change the 1st LO from auto to 168MHz. as far as i understand, this should be the right LO. when the 1st LO is on "Auto" there are even more image frequency signals.

Receiving Radiosonde
This is the normal picture, when a sonde is in the air. the picture was taken 1.5 hours after the start, therefore the signal was quite weak.


With the RSP, i tried different Gain Reductions to establish a readable signal, but failed


Volmet
On 126.000 MHz the DC Spike distorted the AM-Volmet


So please could you give me some hints, what i should do? The results on 2m and 70cm are not better. From my experience now and the evaluation between the nooelec stick an the RSP i want to send the item back. There´s no apparent improvement in reception, maybe my expectation where to high. But maybe i´m doing something wrong ?

Best, Lentas

[13dka]

Hi,

I'm basically in the same "coming from the RTL dongle"-boat and I feel your pain but I'm also a few weeks ahead of you.

1. The only way to get rid of the spike on QRGs higher than 30 MHz is currently not using SDR Console but HDSDR or SDR# v1361 instead. They make use of the ExtIO.DLL that lets you switch to "Low IF" mode, which has a lower spike up there and the imbalance correction of those programs takes care of it. Sensitivity on the air band is generally better than the (my) RTL dongle, but there will be some images too.

2a. 137 MHz should be fine on the RSP and better than the dongle as well. The LNA should probably be on with your discone (or any other passive wideband antenna) to get all of the improved SNR and sensitivity on that frequency range, and the gain reduction should be set accordingly high (higher GR = less gain, less gain is usually better until you reduce it too much).

2b. The only way of getting rid of the bad FM images between 300 and 420 MHz (I get one in the 2m band too) would be using narrow band antennas and/or an FM band stop filter. You can try to switch the LO, but in my case that only replaces the FM band images with DVB-T images. However I can receive radiosondes fine because the FM images never fall on the exact sonde frequencies, they are just dangling around left and right which is optically annoying. Mind you, unlike conventional receivers that have a birdie here and there too (usually ignored by the owners for the most part), SDRs let you see them and annoy you. YMMV of course, you may have bad luck and get an image right on your QRG of interest and a high FM station density at your location may make things worse.

2c. I have not had any issues on 70cm yet. SNR is generally better than on the RTL dongle, but the difference between them might be the least pronounced on 70cm. The RSP gets better the higher you go from there, minus the images.

2d. Some images can be suppressed or shifted by changing the IF bandwidth to a lower or higher value, or changing from Zero IF mode to Low IF mode or vice versa, or a combination of both methods. I couldn't find a situation where turning off the LNA helps with that so far, but you gotta make your own experiences.

That being said, I think VHF/UHF is quite a mixed bag on the RSP1 but if you manage to ignore its shortcomings, you may find out that it excels the dongle all the time up there. Or in other words, due to the images it's probably not the greatest device to find new frequencies, but it's the better receiver once you found them (maybe with the dongle). On the other hand, I could find frequencies with the RSP1 that I missed on the dongle because they were drowning in the noise there, and showing up nicely on the RSP.

[jon]

We were just about to post this, but 13dka beat us to it with some very helpful comments. So this slightly duplicates some of what they said......

The image frequencies you are seeing occur as a result of the use of a block frequency converter to cover the frequency range of 250 - 420 MHz. The block frequency converter uses a fixed LO and the FM signals mix with harmonics of the LO to produce an unwanted IF image. These unwanted images can then mix with either the fundamental or harmonics of the synthesizer to put the images within the bandwidth of the IF.

There are two ways to deal with this. You can change the frequency of the LO you are using from “Auto”to one of the other options. This has the effect of moving around the particular interferes you are suffering with. As you are experiencing that does not always solve you problems and so the other way is simply to add extra external filtering to reduce the FM signal present at the input to the block converter. For broadcast FM signals, external trap filters are readily available at a low cost.

The central DC spike is the effect of DC offsets in the I/Q signal path. This is a function of the direct conversion architecture and will be present in all receivers using this architecture. This spike is largely removed by applying post processing DC offset compensation., but there is a limit to its effectiveness and so there is always some residual spike present. By switching to a program like HDSDR you can switch to using a low IF mode which will remove the problem of the DC spike you are seeing. Anecdotally we have heard of people showing significant improvement in the image signals when operating in this mode. At the moment these settings are currently not available in SDR-Console but we are working with the developer of SDR-console and hope to include the updates within the coming few months.

Best regards, SDRplay support

[g1hbe]

I notice that in all these cases of poor reception, the wanted signal is lined up with the DC spike. Surely this is a no-no? Why not just move the wanted frequency to one side? Apologies if I've mis-read the situation (I'm good at that...).

[13dka]

So this slightly duplicates some of what they said......

Hey, I'm only one. Are you saying I'm too fat?

[13dka]

I notice that in all these cases of poor reception, the wanted signal is lined up with the DC spike. Surely this is a no-no? Why not just move the wanted frequency to one side? Apologies if I've mis-read the situation (I'm good at that...).

I'd say the DC spike is a no-no and pretty annoying. All softwares let you offset the "dial pointer" temporarily to avoid it while you tune (SDR Console doesn't even let you scroll the spectrum) but with all softwares there are situations where you can't help ending up right on the spike, when you recall a saved frequency (which is particularly bad) or when you need to center the spectrum. Maybe there is some global spectrum offset function I haven't found yet in any of those, but that's what I'd love to see, an offset setting that helps avoiding these situations reliably.

[Lentas]

@13dka and @jon: thank you for your hints!
a few words about an additional LNA. i´ve ordered a kit from SV1AFN that has the possibility to implement a highpass filter before the pga-103 and even (if wanted) after the amplifier. so i soldered a HPF with to remove the FM-band (Fg around 115 MHz).
will be interesting how it plays out, i´ll run this LNA with a homebrew QFH antenna (calculated for 403 MHz).

for HF i ordered an upconverter by SV1AFN that gives back very low noise and some kind of prefiltering. need some housing and cables before 247 running, but first tests were very impressive.

so you can see, i´m not really dependant on the RSP, but i´ll take the time for further testings. i really appreciate your hints, will try now.
73 de OE3MAA

[13dka]

i´ve ordered a kit from SV1AFN that has the possibility to implement a highpass filter before the pga-103 and even (if wanted) after the amplifier. so i soldered a HPF with to remove the FM-band (Fg around 115 MHz).
will be interesting how it plays out, i´ll run this LNA with a homebrew QFH antenna (calculated for 403 MHz).

for HF i ordered an upconverter by SV1AFN that gives back very low noise and some kind of prefiltering. need some housing and cables before 247 running, but first tests were very impressive.

I'd be curious about your findings on the LNA and the upconverter, please keep us posted!