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

[cdev]

Showing how out of the loop I am I have only just noticed that there seem to be a bunch of cheap "MSI SDR" dongles out now selling for around $40. They are supposed to be compatible with software that works with the SDRPlay SDRs - basically a 12 bit SDR with range from LF to around 2 GHz.  Which is fairly respectable, definitely a huge step up from the RTLSDRs. The price makes it quite attractive from my perspective.

Does anybody own one? 

Unlike the original SDRplays they seem to have a 0.5 ppm TCXO - Software support now seems quite good, all the usual suspects make hooks that speak to it.  I am worried abour image rejection and performance in a noisy RF environment without the need for band stop filters.

 I am pretty sure the original SDRPlay has an issue with FM overload, which isnt surprising given how many strong signals are around.

[radiolistener]

reception quality of MSi receiver is much worse than RTLSDR, but MSi receiver has 9 MHz bandwidth mode.

Regarding to Chinese MSi.SDR, it has no filters on the input. Also it has a lot of spurs and annoying crackle sounds on some frequencies. It may be due to fake or defective MSi chips or bad PCB layout. For 145 MHz it has much worse reception sensitivity than RTLSDRv3. Also MSi.SDR has a lot of frequency gaps.

In short, if you want better reception quality, less spurs and less frequency gaps and less bugs, then RTLSDRv3 is much better choice.

The only MSi.SDR interesting feature is 9 MHz bandwidth mode. For all other things RTLSDRv3 is absolute winner in comparison with MSi.SDR.

[cdev]

The SDR chip from MSI has multiple internal (digital) filters but the receiver attempts to cover such a wide bandwidth that even if they were traditional analog filters they would have a hard time preventing spurii..

But your criticism seems likely to be true. So, compared to the SDRplay, its crap? Or are these problems inherent to SDRPlay also?

I have milked the most performace out of my 8 bit RTLSDR that it is possible to get. Its impossible to do much better without more bits.

145 MHz (lower part of two meters) is particularly problematic on the RTLSDR for me too. 144 MHz is even worse.

[radiolistener]

So, compared to the SDRplay, its crap? Or are these problems inherent to SDRPlay also?

I don't have SDRPlay, so I cannot say if these problems are present in SDRPlay or not. According to other users SDRPlay don't have some spur issues, which is present in MSi.SDR. So some difference between MSi.SDR and SDRPlay is definitely present.

I have milked the most performace out of my 8 bit RTLSDR that it is possible to get. Its impossible to do much better without more bits.

RTLSDR ADC works as expected, but MSi.SDR works a little different and strange, it can handle a little higher signal level on the input with no overload, but with higher noise floor. MSi.SDR has the same SNR or even worse. I suspect that MSi.SDR may use 8 bit ADC with dynamic attenuator which is enabled when input level is too high.

145 MHz (lower part of two meters) is particularly problematic on the RTLSDR for me too. 144 MHz is even worse.

No. I'm talking about problems with 145 MHz on MSi.SDR.
I cannot hear weak stations on 145 MHz with MSi.SDR.
RTLSDRv3 can hear it with no problem...

On short wave both have the same sensitivity and works almost the same. MSi.SDR don't have some mirror channels across 14 MHz which is present on RTLSDR. But MSi.SDR has a lot of huge spurs near 30 MHz.

I also forgot to say that MSi.SDR has a bit problem with DC offset between I/Q channels. It is so bad, that you cannot remove it even by using wideband DC filter in HDSDR. This issue is masked in SDRuno, because it has always enabled DC filter which cannot be disabled. But in HDSDR you can disable DC filter and see how it really bad.

RTLSDRv3 has smaller DC offset and it can be easily removed with IIR filter with just 1 Hz bandwidth. But for MSi.SDR you cannot remove it even with 25-100 Hz IIR filter.

Another issue present in my MSi.SDR is sound crackle pulses across entire bandwidth. It's frequency depends on LO frequency. I'm not sure if this is just defective MSi chips or this problem exists on all MSi.SDR devices. But it is very annoying and prevents comfortable radio listening.

And another issue with MSi.SDR (at least my one), it works unstable at 10 MHz bandwidth.
When I switch to 10 MHz BW, noise floor jumps like crazy. So, 10 MHz mode is unusable on MSi.SDR.

[cdev]

They were probably made with chips that failed quality control which were probably supposed to be trashed but werent. 

[radiolistener]

They were probably made with chips that failed quality control which were probably supposed to be trashed but werent.

I have the same thoughts. It also may be bad quality clones of MSi chipset.

[A.Z.]

Hello, if you are referring to those "blue dongles" with a dolphin, that is the ones like this

https://www.rtl-sdr.com/photos-of-the-msi-sdr-dongle-a-new-sdrplay-rsp1-clone/



then I own one, and two friends of mine another pair (we ordered them together); the unit is just like the one depicted above, although it hasn't that electrolytic capacitor soldered on the board



as for performances and defects, I must admit that I didn't perform any measurement, all I can say is what came out from using the unit for some months now

First of all, the chipset is really the mirics one, that is the same used in SDRplay units, but according to some informations floating on the 'net, the board design detaches from the RSP1 one, even if plugging the dongle it's identified as an RSP1 (not "a"); performance has been quite good on all bands where I tried the unit, and never had audio issues or the like as reported by others here, same goes for the other two units, so my guess is that there may be some defective or badly assembled units around; as for issues, yes, even if it should have frontend filters, the unit suffers from overloading, I had a bunch of out-of-band interferences generating ghost signals which I was able to pinpoint to a nearby (about 30Km) FM tower, at that point I went on and bought a NooElec "FlaminGo" FM bandstop filter and that solved the issue, other than that, even if it has a metal box, the unit needs good grounding to avoid noise and also choking at both the coax and the USB cables, as for the latter, the one which comes with the unit is quite crappy and it would be a good idea picking a better USB cable (possibly with chokes); in my case I'm using the unit with a random wire antenna fed using a NooElec 9:1 unun/balun V1 (the V1 can be configured either way by cutting a trace on the PCB - see here) and around 25m of RG-58, it has been serving me pretty well for these months and I believe it's definitely worth its price, just give it a decent antenna (the whip shipped with some units is worth nothing) and I believe it will serve you well

[radiolistener]

A.Z., could you please perform some tests with your one?

1) Leave input open

2) Start HDSDR (with SDRPlay ExtIO plugin). I'm using HDSDR, because it allows you to configure more parameters manually

3) Open ExtIO setup and select the following settings (Zero IF, 8 MHz bandwidth, Sample Rate 8MHz, IF gain=45):


4) Set LO = 500'000 Hz and take the picture of spectrum, please share it here. Do you hear some low frequency pulse sound hum (about 50 Hz)? I'm interesting if your device has parasitic spurs the same as my one. See my result in attachment

5) Set LO = 6'000'001 Hz and Tune = 6'400'000 Hz, do you hear crackle sound pulses with frequency about 3-5 Hz? This is the same issue with pulses.

[A.Z.]

Will try, but I've to admit that I'm using the dongle using the SDRuno software and not through the extIO, so I'll probably have to fiddle a bit with my system and install some more stuff... so, well, I'll try but don't hold your breath !

Also, those "spurs" (vertical lines evenly spaced) WERE present in my unit, until I didn't connect the unit to a good ground and improved the antenna counterpoise and choking system, after that the "lines" (almost) disappeared; also, and since we're at it, here's the unit I bought (although back then the cost was lower)

https://www.amazon.com/KKmoon-Broadband-10KHz-2GHz-Panadapter-Compatible/dp/B07YYFJ94N

not sure if it does any difference, but the above is working ok for me; I reused the pigtail cable which comes with the unit to connect it to the NooElec FM filter (added chokes to the pigtail !) which in turns goes to the antenna coax feedline (chokes there as well, both on the SDR and the antenna side)

[A.Z.]

Didn't yet find the time to run the requested tests (to tell the truth, hadn't time to play with the SDR at all), but given the interest in those MSi.SDR dongles, I thought that people here may be interested in some more details about those chips and their implementations, so here are some links which will offer some more details

https://www.rtl-sdr.com/forum/viewtopic.php?t=31

http://blog.palosaari.fi/2013/08/mirics-msi3101-sdr-linux-driver.html

http://blog.palosaari.fi/2013/10/naked-hardware-13-logitec-ldt-1s310uj.html

I hope that the above infos could be helpful and clear some confusion about the mirics chipsets architecture and design choices

[radiolistener]

HDSDR: http://www.hdsdr.de/

It needs ExtIO driver in order to communicate with MSi.SDR: https://www.sdrplay.com/software/SDRplay_ExtIO_Installer_1.2.exe

Here is also both together (I didn't tested this version):
https://www.sdrplay.com/software/SDRplay_HDSDR_Installer_2.76a.exe

Also, those "spurs" (vertical lines evenly spaced) WERE present in my unit, until I didn't connect the unit to a good ground and improved the antenna counterpoise and choking system, after that the "lines" (almost) disappeared; also, and since we're at it, here's the unit I bought (although back then the cost was lower)

it is impossible to remove these spurs with any kind of antenna, because the source is inside MSi.SDR device and it always present. You can put strong signal on the input to hide these spurs. But if you want to receive weak radio, these spurs and noises will be an obstacle.

This is why I'm prefer to use RTLSDRv3 which is twice cheaper and provide much better reception quality (more clean, less spurs, less noise and less distortions)

[A.Z.]

it is impossible to remove these spurs with any kind of antenna, because the source is inside MSi.SDR device and it always present. You can put strong signal on the input to hide these spurs. But if you want to receive weak radio, these spurs and noises will be an obstacle.

All I can say (at least util I'll find time to run the test - thanks for the links), is that by connecting the MSi.SDR dongle SMA connector gnd to a good ground those spurs totally disappear from the spectrum, then ok, in general you will also want to choke the USB and possibly run over batteries to remove switching PSU noise, but that's a different issue and has nothing to do with the SDR, then if you prefer the RTL unit, that's fine with me, it's just an SDR not a religion or whatever else  all I can say is that I owned an RTL unit and I liked it, but I like the MSi one more, but again that's just me

[cdev]

All SDRs have certain issues. They do need a good ground. Just about any unun also will help a lot with a long wire antenna.

For HF you should try a different receiver than an RTLSDR if you can. Adding additional bits gives you substantial improvements in dynamic range.

Make a unun if you have a lot of spurs. You can use any small toroid or ferrite core. Even a ferrite split bead for noise reduction will do great. If you have two identical split beads that makes a really good unun. Use it like you would any binocular core. . Twist the wires you use together to make a bifilar or trifilar winding.  This will take five minutes and give you much less noise.


Certain kinds of sound card SDRs based on the "Soft Rock" design are easy to build.

[radiolistener]

by connecting the MSi.SDR dongle SMA connector gnd to a good ground those spurs totally disappear from the spectrum

No, it don't disappears. It always present.  Their source is inside MSi.SDR device, so it doesn't depends on grounding or antenna type. It is present with any type of antenna, with any type of ground and with any type of power supply. This is an issue of device, not environment.

then ok, in general you will also want to choke the USB and possibly run over batteries to remove switching PSU noise

the source of these noises is not USB and not PSU. The source is MSi.SDR itself.

I tried to power MSi.SDR from external very clean ultra-low noise linear PSU. Also I tried to add ferrite filters on USB cable. The result - I don't see any visible difference. These spurs are still present independently. It doesn't matters how good power supply and how low is interference from PC. Because the noise source is receiver istelf.

At the same time RTLSDRv3 works good on the same USB port and don't have such spur and  noise issues.

Also I tried to connect MSi.SDR to RTLSDRv3 and can see a lot of parasitic emission from MSi.SDR on it's input connector.

Here is example - RTLSDRv3 SMA connected to MSi.SDR SMA connector. So, on the left side you can see parasite emission of MSi.SDR (it is received by RTLSDRv3). And on the right side you can see parasite emission of RTLSDRv3 (it is received by MSi.SDR).

These spurs are not processing issue. These are real parasite RF emission from MSi.SDR on it's SMA connector.

Also you can see crackle pulses around entire bandwidth of MSi.SDR (green spikes on the waterfall).


Also, MSi.SDR has a noise issue from USB transmission, you can catch it in some modes. See the second screenshot for example.

[radiolistener]

For HF you should try a different receiver than an RTLSDR if you can. Adding additional bits gives you substantial improvements in dynamic range.

The bits don't tell you all info. Many bits may works much worse than less bits receiver. You're needs to take into account ADC speed, SNR, IMD and other specification for ADC and RF frontend. This is something like compare 5 megapixel professional camera with high quality optics and large matrix and cheap 10 megapixel camera from smartphone with a small and noisy sensor. It has much more megapixels, but works much worse

RTLSDR and MSi.SDR are example.

RTLSDR has 8 bit ADC working at 28.8 MHz but it has better performance than MSi.SDR which has 12 bit ADC working at 24 MHz.

I have DDC receiver on  FPGA with fair 14 bit AD6645 ADC. So, I can compare it with RTLSDR. RTLSDRv3 works very good for it's  price. Yes, it has some mirror channel issues and some spur, it's SNR is about 80 dB. But I'm very skeptic that you can get better performance for less than 100 USD. For comparison, my FPGA setup cost me a little more than 100 USD. Actually I spent more than 200 USD for experiments.

Now it can be possible to build setup with FPGA and fair 12 bit AD9226 ADC for about 55 USD + good power supply. It will be much better than RTLSDR or MSi.SDR. But since I already have AD6645, I'm experimenting with 14 bit.

Also there is a very interesting receiver on aliexpress:
https://www.aliexpress.com/item/4001186751605.html

I didn't tried it, but it looks very interesting. It uses noisy oscillator, but it can be replaced with ultra-low noise one from Abracon or other known brand.

Make a unun if you have a lot of spurs.

unun or antenna cannot fix self noise of receiver. RF choke, baluns  and good antenna allows  to reduce noise from your home  electronic equipment,  but it cannot remove noise which is already present in your receiver. It's already here, so you cannot block it.

You can put very strong signal on the input to hide receiver self noise, but it still present and when you try to receive weak station you will catch noise of your receiver.

[A.Z.]

ok folks, if it is a religion war, then I'm giving up, all I can say is that the dongle is serving well, the remainder is just chatter, all the best.

andrew

[radiolistener]

ok folks, if it is a religion war, then I'm giving up, all I can say is that the dongle is serving well, the remainder is just chatter, all the best.

Sorry, but this is not religion war, there is no even a small piece of religion. Just technical talks. I just tell you what issues I found on MSi.SDR.

Regarding to spurs and noises, it doesn't depends on receiver type. This is technical fact that self noise of receiver cannot be removed with some magic ground or magic antenna.

The goal of RF choke and balun is to prevent common mode currents on the outer braid of coax cable. The goal is to protect your receiver from noises collected by the cable from home equipment. There is no magic and no religion. This is just technical issue.

But if noise source is a receiver itself, then there is no way to protect your receiver from these noises, because they are already inside of receiver... So, RF choke, balun and antenna cannot help to remove it.

The reason why they are less noticeable with antenna is because your antenna has some strong signal. Since dynamic range is very limited it leads to higher noise floor. Also there are a lot of noise from air. This is why these spurs is less noticeable on the background of real noise from the air.

For example, I taken two screenshots. The first one is open input. And the second one is connected antenna. As you can see these spurs are still here, it just becomes less visible in the background of air noise. But these spurs still here with exactly the same level.

[A.Z.]

again, you aren't getting the point, but it's ok with me, if your custom SDR is doing well, then it's fine with me; that said, maybe I'm crazy, but I won't sell my MSI.SDR for an RTLSDR one, that said, have a nice day.

[radiolistener]

but I won't sell my MSI.SDR for an RTLSDR one, that said, have a nice day.

I don't propose to sell it Just talking about issues.
I have both of them. Also I have different RTLSDR versions.

But most of the time I'm using RTLSDRv3, because it has better sensitivity and less self noises, also it don't have frequency gaps. So, when I want to listen some radio, I'm using RTLSDRv3.

On the other hand MSi.SDR is useful when you want to see all 8 MHz bandwidth on the spectrum. Because RTLSDR provides max 3.2 MHz.

Also there is difference for 14.4-28.8 MHz band. Since RTLSDRv3 use direct sampling for ADC running at 28.8 MHz, it's bandwidth is just 14.4 MHz. All signals above 14.4 MHz will be folded into 0-14.4 MHz (through aliasing). On RTLSDRv3 it's better to use bandpass filter on the input if you want to listen 14.4-28.8 MHz (in order to cut-off broadcast stations from 0-14.4 MHz band). On MSi.SDR there is no such issue (it uses up-conversion inside). But since most of broadcast stations are below 14.4 MHz, this is not a big deal.

[A.Z.]

all that is fine with me, I do mainly HF bands and MW/LW plus some 79 and 118 MHz (ok around that) listening, and once I properly set up the whole antenna system (including grounding and feedline) I had no issues, plus the better dynamic range of MSi.SDR allowed me to handle conditions which a bare RTLSDR unit wouldn't gracefully handle, again, I'm not pretending it's the best SDR around (I'd be crazy), but for sure it's serving me better than, say, the. nooelec RTL dongle then ... 70% of whatever station (tx too by the way) is the antenna system, but then that's a different topic

best

[radiolistener]

the better dynamic range of MSi.SDR allowed me to handle conditions which a bare RTLSDR

Here is comparison for dynamic range left is RTLSDRv3, right is MSi.SDR.
RTLSDRv3 is connected through 30 dB attenuator (because RTLSDRv3 has better sensitivity).
MSi.SDR is connected directly (with no attenuator).

[fourfathom]

The SDR-Play has a full set of switchable analog (L/C) filters ahead of the digital sampling stuff.  You can find the filter design on the SDR-Play website (https://www.sdrplay.com/wp-content/uploads/2018/01/RSP1A-Technical-Information-R1P1.pdf).  The RTLSDR and similar SDRs have much simpler and lower-performance filters so I suspect that the spur performance will be worse.

I used the "RTL-SDR Blog V3" in direct-sampling mode for 10 MHz and 7 MHz ham band receivers, and ended up putting an external low/bandpass filter and preamp on it.  It worked well enough for my needs, but wasn't nearly as bulletproof as a good ham receiver.

[cdev]

The SDR-Play has a full set of switchable analog (L/C) filters ahead of the digital sampling stuff.  You can find the filter design on the SDR-Play website (https://www.sdrplay.com/wp-content/uploads/2018/01/RSP1A-Technical-Information-R1P1.pdf).  The RTLSDR and similar SDRs have much simpler and lower-performance filters so I suspect that the spur performance will be worse.

I used the "RTL-SDR Blog V3" in direct-sampling mode for 10 MHz and 7 MHz ham band receivers, and ended up putting an external low/bandpass filter and preamp on it.  It worked well enough for my needs, but wasn't nearly as bulletproof as a good ham receiver.

The direct sampling mode on the RTLSDR works particularly well with a tuned magnetic loop antenna. (the kind that is just a loop of wire and a variable capacitor, with a pickup loop)

The selectivity of the loop makes up for the lack of selectivity of the RTLSDR. I found that that using that combination was when I got the best performance out of my RTLSDR.

Direct sampling it also makes a very good probe for tracking down RFI.

[radiolistener]

The RTLSDR and similar SDRs have much simpler and lower-performance filters so I suspect that the spur performance will be worse.

No, spur performance don't depends on input filters. It depends on internal circuit LO purity and processing. I think that worse spurious performance of MSi.SDR is due to worse purity of LO NCO. As I mention before, these spurs depends on LO frequency, so I think these problems is due to bad spurious performance of NCO. Probably it uses smaller phase accumulator or something like that.

And probably these sound pulses is also due to spurs of NCO. But it also may be caused by mixing of some frequencies used inside of MSi.SDR.

I'm interesting if these issues with pulses are present on genuine MSi2500 chip. Because it is possible that MSi.SDR use Chinese clone chips which can have some issues, for example with NCO, which may be missing in genuine chipset.

The SDR-Play has a full set of switchable analog (L/C) filters ahead of the digital sampling stuff.

I suspect that SdrPlay use a lot of filters on the input just because MSi chipset has a big issues with dynamic range or intermod distortions. Because good SDR don't needs a lot of filters. Just proper antialias filter before ADC is good enough.

Also I hear that msi001 tuner has just 85dB dynamic range, so this is the reason why MSi.SDR cannot have better dynamic range than RTLSDR. Just because dynamic range of analog frontend of MSi.SDR is limited to 85 dB. So, 12 bit cannot help here, and spurious LO makes things even worse...

For example:

RTLSDR:
8 bit ADC = 6.02 * 8 + 1.76 = 49.92  dB
Processing gain: 28.8 MHz => 3 kHz = 10*log(28800/(3*2)) = 36.8 dB
SNR = 49.92 + 36.8 = 86.72 dB

MSi.SDR
12 bit ADC = 6.02 * 12 + 1.76 = 74 dB
Processing gain: 24 MHz => 3 kHz = 10*log(24000/(3*2)) = 36 dB
SNR = 74+36 = 110 dB
But it is limited with msi001 tuner to 85 dB, so the total SNR = 85 dB
Almost the same as RTLSDR...

[fourfathom]

No, spur performance don't depends on input filters.

Input anti-aliasing and preselector filters do indeed affect IMD-related and aliasing spurious responses.  But of course so do the A/D and NCO and datapath.