I have the SDRPlay since a few weeks now and I must say it has a good performance for the money it costs. Especially shortwave works fine.
However, beside the driver install problem with Win10 resp. tablet computers I'm not very satisfied with the band filter performance in the VHF range.
Have a look at this, it is mostly useless in this range: The antenna is an active wide band whip (Boni Whip). It does not produce these signals itself, I checked with the FDM-DUO, there are no signals from the antenna. I think I'll try to build some better VHF filter and if it works replace the internal filter.
Any other idea?
Have you been able to identify what would be the normal carrier frequency for these spurious signals? if so, please let us know as it will help us identify any possible work around strategy.
Reason: No reason
I did a quick test with the new API & DLL, maybe that helps.
LO @ 50MHz, 8MHz sample width
Switching from auto to 144MHz band plan eliminates most of the ghost signals. (waterfall: upper part 144MHz, lower part auto band plan) Switching from 144MHz to 168MHz band plan produces a lot more ghost signals. (waterfall: upper part 168MHz, lower part 144MHz band plan) So, it depends on the selected band plan. It seems that SDRConsole uses something like 168MHz band plan, right?
Could you please explain what happens there?
The RSP uses a block up-converter to cover the frequency range from 100 KHz to 60 MHz. This uses a fixed frequency first LO which can be selected to be 120 MHz, 144 MHz or 168 MHz in the new API/ExtIO. With the old API the first LO was fixed at 120 MHz and this is what is currently used by SDR Console.
So for example, a signal at 50 MHz will be mixed up to 170 MHz if the first LO is set to 120 MHz. The synthesizer is then set to 170 MHz, to mix the first IF to I/Q.
All block up converters are susceptible to spurious responses where unwanted signals may mix with harmonics of the first LO to produce either a spurious signal at the same frequency as the first IF or one at exact harmonics of the first IF. It tends to be odd order harmonics that create these responses.
So for example, if there was a interferer at 290 MHz that leaked to the first mixer, then it could mix down to 170 MHz (RF - LO1) which is the same as the first IF. Similarly a spurious signals at 190 MHz or 530 MHz could mix with the third harmonic of the first LO (360 MHz) to produce a signal at the same frequency of the first IF. Another example might be where a spurious signal at 150 MHz mixes with the 3rd harmonic of the 1st LO to produce a signal at 510 MHz at the 1st IF. This at the 3rd harmonic of the second LO and so will mix on top of the wanted signal at the I/Q mixer outputs.
There are an infinite number of possibilities, but the responses become weaker as the order gets higher. Most of the time, these responses will not be detectable because the rejection of the RF filter and antenna will effectively push their levels down into the noise floor. If the spurious signals are very strong or the antenna provides more gain at their frequencies, then these responses can be observable. By changing the frequency plan of the receiver, it may possible to move these spurious responses away from the band of interest, which is what you are observing. Another way of thinking about it is you have moved the spurious responses to a quiet area of spectrum.
As you have also correctly pointed out, it is also possible to help this situation by improving the RF filtering. If you know the carrier frequency of the interferers, a trap at that frequency will help matters considerably.
You will also observe similar phenomena in the range of 250 MHz to 430 MHz. This is because the block up-converter is now re-used as a block down converter to cover this band. In all other bands, the architecture used is a single conversion receiver which us much less susceptible to these responses.
We plan to put together a 'white paper' that describes in detail the architecture of the RSP. This will help people understand where they may be susceptible to spurious responses and devise strategies for effectively mitigating them. We expect the get this out and published on the website in the next few weeks.
Reason: No reason
thank you very much for your explanations, that clears up what happens.
So, the best solution will be to have another band pass filter with better stop band attenuation below 30MHz. Maybe just an additional 30MHz high pass filter will be enough. Let's heat up the soldering iron ...
I did it: I inserted a new bandpass filter into the SDRPlay and bypassed the existing one. It is a band pass filter 3rd order (47pF+220nH in series, 220pF||68nH to GND, and again 47pF+220nH in series). You can calculate it by yourself: http://www.electronicdeveloper.de/Filte ... ed_3O.aspx
Looks like this: ... and the result is that: Dear SDRPlay team: For the next revision, please spend a few cent more and give it better filters. Thanks!