Cheap "MSI SDR" (SDRPlay RSP1) clones, anybody have one?

[HB9EVI]

indeed the RTL-SDR v3 are much better than their reputation; for that price it's tricky or nearly impossible to find something performing on the same level

I also have an RSP1 and a RSPduo, which perform well; under certain circumstances you can profit from the higher adc resolution/higher dynamic range, but not always. I think they are too high priced and in the end closed source, what I consider the biggest minus of it.

[cdev]

Have you tried the rx-888?  I don't know much about it.

Seems like it could both receive and heat at the same time.
 
Useful in the winter. 

Of course in Italy, probably not needed?

The way I see it, If I am going to spend that much money, I should buy an HF ham transceiver kit with a general coverage receive capability. There are several options starting under $200.

As far as bitness, it is hugely important. I have a softrock which uses my desktop's sound card, so it has the ADC resolution of the soundcard its used with. There is a really really huge difference between it and any RTLSDR on HF, with either an upconverter or direct sampling. I only hear the strongest signals on the RTLSDRs. Most are just invisible/down in the noise.

When the softrock is properly calibrated (to cancel out the center spike) its much much more sensitive. All decent HF receivers are much more sensitive than any RTLSDR.  The RTLSDR is however quite comparable to consumer electronics shortwave receivers.

I hadn't really considered airspy recently, I forget why some reason I was not interested in them as a receiver. I forget what it is. Its been a long time since I compared all the cheap SDRs.

[vinlove]

I had been interested in the original AirSpy in metal enclosure, but the latest version AirSpy Discovery+ was in plastic casing. It looked and felt cheap, and prone to suffer from RFI easily. And talking about high priced SDR, AirSpy was highest priced among all on sale.
I looked elsewhere, and bagged a used RSP1A for the price of 2x new RTL dongles. I could't be happier than that. Although I am still not sure if AirSpy would be 4x times better performing than this used RSP1A.  I bet they are more or less the same or similar performance,  and really depend on what antenna they are fed with.

[HB9EVI]

a QSD is still a way to go for SDR reception; both Elecraft KX2 and KX3 work like that; it's also a possible solution for IF-tapped ham transceivers.

I'm running an OpenWebRX with 3 RTL-SDR v3 which work really well; digimode reception shows signals from all around the world; antenna is a simple active loop; additionally there are about 20m RG174 and 1:4 rf splitter in between antenna and every of the sdr sticks.

a classis heterodyne ham transceiver is not all the time better than a SDR; also you don't have the feature of e spectrum scope without IF-tapping.

[cdev]

They are both hugely important, and you're right in that its basically impossible to make the most of either one (antenna or receiver) if the other element is woefully inadequate. Are all of the upgrade from rtlsdr dongles  fairly comparable?

Only somebody like Leif Asbrink could tell us that. (expert on SDR receive parameters) They are probably within an order of magnitude of one another-

One big issue with USB receivers is USB noise. I have a lot of ferrite CM chokes on my USBs and antennas and could probably benefit from even more. They should be fed in a way that decouples them from the indoor house ground. A unun at the antenna end thats grounded there will do it.

I bet they are more or less the same or similar performance,  and really depend on what antenna they are fed with.

Is there anybody who has used the MSI.sdr receiver with the FOSS libraries who has had good luck with that?

In other words, without sdrplay anything. Under linux.

HB9EVI, or AZ, I am totally with you HB9EVI in wanting an open source SDR.   Curious, have either of you ever compared the open source library (Libmirisdr or libmirisdr4 from github) with the closed source sdrplay libs running under Linux receiving HF??

I should probably just keep my eyes peeled for a used SDRplay. If I could find one for the cost of the new MSI.sdr that would be much less of a gamble.


Not using Windows, my options are significantly different on receivers.

[cdev]

An active loop or magloop will pick up digimodes and I can receive signals from all around the world with my RTLSDRs and an upconverter or direct sampling. Just a lot fewer of them.

One of my upconverters has a built in LNA and it does very well compared to the one I have that does not have receive amplification. (but requires a variable attenuator)

To get better results I really need to get my antenna up a lot higher. I have two tall trees that since last year and a really big storm have been empty of antennas. Now they are bare and I should grab the opportunity to get a wire much higher up into them.

The Softrock is a very simple Tayloe detector design, and according to Dan Tayloe his receiving detector can be "operated in reverse" which is kind of intriguing..  This has made me wonder if there might be an inexpensive mod to turn it into a transceiver that uses my existing hardware more efficiently.

Need to get a ham license first, then start thinking about a rig.   

a QSD is still a way to go for SDR reception; both Elecraft KX2 and KX3 work like that; it's also a possible solution for IF-tapped ham transceivers.

I'm running an OpenWebRX with 3 RTL-SDR v3 which work really well; digimode reception shows signals from all around the world; antenna is a simple active loop; additionally there are about 20m RG174 and 1:4 rf splitter in between antenna and every of the sdr sticks.

a classis heterodyne ham transceiver is not all the time better than a SDR; also you don't have the feature of e spectrum scope without IF-tapping.

Is your OpenWebRX online?

[A.Z.]

ANY antenna will pick up "digimodes"

[fourfathom]

As far as bitness, it is hugely important. I have a softrock which uses my desktop's sound card, so it has the ADC resolution of the soundcard its used with. There is a really really huge difference between it and any RTLSDR on HF, with either an upconverter or direct sampling. I only hear the strongest signals on the RTLSDRs. Most are just invisible/down in the noise.

For what it's worth I've measured the direct-sampling HF sensitivity of the RTL-SDR Blog V3, with it powering a 20dB preamp, and I see a SSB noise floor around 0.005 uV (10 MHz single-tone SNR is 6dB with an input of 0.01 uV, using a 5 kHz audio bandwidth).  This is possibly better than my Icom IC-7200 transceiver, which has a spec'd sensitivity of "less than 0.13 uV". 

For some tests I've used a resistive splitter to connect an antenna to both my IC-7300 and the RTL-SDR-V3, with the IC-7300 connected to a PC running WSJTX (WSPR) and the RTL connected to a RPi also running WSJTX.  Running both rigs for 24 hours there was no significant difference in the number of signals decoded.  I'm in a fairly quiet radio environment, so the results might have been different had there been strong interfering signals.  I did put a narrow 10MHz bandpass filter in front of the preamp/SDR.

Of course sensitivity is only one parameter, and the RTL certainly has its shortcomings.  But it's good enough to be quite useful in some situations.  I've linked this before, but this shows more about my system: http://wb6cxc.com/wp-content/uploads/2019/04/Presentation-1-0.pdf  (The Finland station has since shut down due to RPi problems, and I haven't been back to fix things.  I have also added 40-meter receive gateways at my two USA locations)

[HB9EVI]

I tried running the RSP1 with libmirisdr; it's 'somehow' working, but far away from being satisfying; but that was a while ago already; maybe the lib got better in the meantime; maybe I'm going to try it once again

[radiolistener]

For what it's worth I've measured the direct-sampling HF sensitivity of the RTL-SDR Blog V3, with it powering a 20dB preamp, and I see a SSB noise floor around 0.005 uV (10 MHz single-tone SNR is 6dB with an input of 0.01 uV, using a 5 kHz audio bandwidth).  This is possibly better than my Icom IC-7200 transceiver, which has a spec'd sensitivity of "less than 0.13 uV". 

Here is my measurement. RTLSDRv3 with -50 dBm 10 MHz sine on the input.
S level is calibrated, so you can see actual dBm level on the grid.

Noise floor for CW bandwidth 500 Hz = -120 dBm.

It means MDS sensitivity for 500 Hz is -120 dBm = 0.224 uV.
10dB S/N sensitivity is -110 dBm = 0.707 uV.


Icom IC-7200 according to the specification at 10 MHz band and CW:
10dB S/N sensitivity is -122 dBm = 0.16 µV

As you can see, Icom IC-7200 has 12 dB better sensitivity at 10 MHz band than RTLSDRv3 in direct sampling mode.

[radiolistener]

I've linked this before, but this shows more about my system: http://wb6cxc.com/wp-content/uploads/2019/04/Presentation-1-0.pdf

interesting, can you show your mods for bias-tee from RTLSDRv3 on that LNA (page 17)?

[fourfathom]

For what it's worth I've measured the direct-sampling HF sensitivity of the RTL-SDR Blog V3, with it powering a 20dB preamp, and I see a SSB noise floor around 0.005 uV (10 MHz single-tone SNR is 6dB with an input of 0.01 uV, using a 5 kHz audio bandwidth).  This is possibly better than my Icom IC-7200 transceiver, which has a spec'd sensitivity of "less than 0.13 uV". 

Here is my measurement. RTLSDRv3 with -50 dBm 10 MHz sine on the input.
S level is calibrated, so you can see actual dBm level on the grid.

Noise floor for CW bandwidth 500 Hz = -120 dBm.

It means MDS sensitivity for 500 Hz is -120 dBm = 0.224 uV.
10dB S/N sensitivity is -110 dBm = 0.707 uV.


Icom IC-7200 according to the specification at 10 MHz band and CW:
10dB S/N sensitivity is -122 dBm = 0.16 µV

As you can see, Icom IC-7200 has 12 dB better sensitivity at 10 MHz band than RTLSDRv3 in direct sampling mode.

This lines up fairly well with my measurements, where I have about 20dB preamplification (amp gain minus filter loss).  Of course we have to consider the atmospheric noise floor and antenna gain before we know if any of this makes a difference in real-world sensitivity.  My WSPR tests showed no significant difference.

Have you made or seen any IMD / overload measurements for the RTLSDRv3?  We can calculate using the A/D resolution, but I would like to see actual measurements.  I still haven't done IMD tests, but have looked at the aliasing in the direct mode, which can be a big issue in noisy environments (and even the atmospheric noise gets aliased and added to the effective noise floor).

[A.Z.]

regarding noise figures and IMD, the datasheet carries some infos

https://rtl-sdr.com/wp-content/uploads/2013/04/R820T_datasheet-Non_R-20111130_unlocked.pdf

then, for sure, measuring "the real thing" gives the effective values

[radiolistener]

Have you made or seen any IMD / overload measurements for the RTLSDRv3?

Yes, I tested it some years ago, it's about -49.6 dBc.
Here is two tone IMD test with 2250 Hz between carriers.

Also you can see dynamic range about 90 dB, because ADC input level is tuned to be about -1 dBFS.
This is extreme measurement

The signal source is si5351 with two channel outputs connected together through SMA-T adapter.
Each channel is configured for dedicated PLL: CH0=PLLA and CH1=PLLB.
Carriers: 10000 kHz + 10002.250 kHz.
Level is tuned with SMA attenuators, don't remember exactly, something like 6 dB at each channel then SMA-T and then some attenuator to fit full ADC range.

[radiolistener]

regarding noise figures and IMD, the datasheet carries some infos

https://rtl-sdr.com/wp-content/uploads/2013/04/R820T_datasheet-Non_R-20111130_unlocked.pdf

R820T is not used in direct sampling mode

[cdev]

Wonder how feasible it would be to take the data coming out of the tuner and sample it using another method, leveraging the tuner, but replacing the ADC ?

Since the tuner chip is used in other similar devices, and the RTL is there to control it, simply bypass that part of the chain ?

At the same time, replace the DC-DC converter also?

I am not saying that would necessarily be better, although it might. I guess the USB data pipeline incolving the RTL could not be used because too much data. So it would be a real Frankenstein device.

--- How does the HackRF One compare to the RTLSDR as far as SN and overload performance?

Lets not forget that at its core the RTL was made to be sold profitably for just a few dollars.. It does remarkably well considering that. The RTLSDR blog version improves the things that it can improve and sells it for a slightly higher price but still the BOM cost must be fairly low. 

HackRF One components room is a bit higher.

[cdev]

That would be great, you likely know how it should be fairly well owning the sdrplay hardware and knowing how it performs using that.

The open version on github, libmirisdr4 incorporated several peoples work. I think it has not been developed past that version.

I tried running the RSP1 with libmirisdr; it's 'somehow' working, but far away from being satisfying; but that was a while ago already; maybe the lib got better in the meantime; maybe I'm going to try it once again

I'd like to build some kind of VLF MW receiver with more bits, maybe I will investigate whats needed to build a simple SDR using a Tayloe mixer, homemade filters  and my Feeltech siggen (its a DDS siggen)  and my computers sound card line input. That might be educational.

[cdev]

I was wondering if you had done this, Ive been thinking about getting a "malachite" which uses an msi front end, and was wondering what its weak signal performance looked like to somebody with a lot of HF experience. (with external antenna of course)


Thank you!

I tried running the RSP1 with libmirisdr; it's 'somehow' working, but far away from being satisfying; but that was a while ago already; maybe the lib got better in the meantime; maybe I'm going to try it once again

[hampolar]

Imagine that hundreds of people are using cheap MSI SDR USB radio scanner and no one had problems like you describe. Anyway from what I see in your images you have a perturbation source nearby your house or in your region. If you want to be sure just scan the band with other scanner or ask someone to do it, you could have some band jamming from a power supply or other industrial equipment in your area or you just had some bad luck and get a broken sdr radio.

[radiolistener]

Anyway from what I see in your images you have a perturbation source nearby your house or in your region. If you want to be sure just scan the band with other scanner or ask someone to do it

If you read it carefully, the source of these spurs is MSi.SDR itself and it was tested with open input (nothing connected to it). There is no such interference in the house, you can see that RTLSDRv3 receiver don't show it. The same there is no such spurs on other receivers. I have no idea where you see "perturbation source nearby your house", because these spurs and beats are present on MSi.SDR only and it's frequency depends on LO setting.

or you just had some bad luck and get a broken sdr radio.

this is possible, but I was seen screenshots and videos from other MSi.SDR owners and it shows the same issues. And their owners don't notice it, just because don't know how it should works or just don't have other radios to compare with... 

So, at least half of issues that I know are present on other MSi.SDR instances and some issues was confirmed on genuine RSP1

It's more likely that you're just had good lack and get instance that works well, not like others. But I suspect that you're not very demanding and are not used to high quality sdr radio, so you're simply do not notice all these issues

If you have MSi.SDR, we can test it, just let me know if you're ready to do test, I will prepare for you settings, you will record wav file with IF from your MSi.SDR instance and share it here, I will check if your device has these issues or not. So, we can check if you have the same "perturbation source"

As a start point you can do this test: https://www.eevblog.com/forum/rf-microwave/cheap-msi-sdr-(sdrplay-rsp1)-clones-anybody-have-one/msg3158922/#msg3158922

I was wondering if you had done this, Ive been thinking about getting a "malachite" which uses an msi front end, and was wondering what its weak signal performance looked like to somebody with a lot of HF experience. (with external antenna of course)

Malachite uses MSi tuner chip only and capture it's output with audio ADC, it don't uses MSi receiver chip. So, this is not correct to compare it with MSi.SDR or RSP1, they using different SDR stack. But it is well known that Malachite has frequency gaps and worse sensitivity than other portable sdr radio.

You can find sensitivity and dynamic range comparison of Belka with Malachite on youtube.

[p.larner]

are these the same talked of on this thread?
https://www.aliexpress.com/item/1005003954483936.html?spm=a2g0o.productlist.main.5.1e9c290aRwpve4&algo_pvid=a04813b5-2429-4483-845f-d565500210f0&algo_exp_id=a04813b5-2429-4483-845f-d565500210f0-2&pdp_npi=4%40dis%21GBP%2115.09%2115.09%21%21%2117.87%21%21%40210387a116990270547681667e90bc%2112000034890967339%21sea%21UK%212308464272%21&curPageLogUid=R6qAhBzFB77u