Review of the SpyVerter Upconverter
The SpyVerter is a new upconverter that has recently gone on sale. It is created by Youssef (he programmed SDR# and worked on the development of the Airspy SDR) and Bob W9RAN (of rantechnology.com and youtube.com/user/ranickel). In this post we'll review the SpyVerter and compare it against some other up converters that we have used in the past.
Radio transmissions between 0 - 30 MHz can travel all the way around the world. At these frequencies many interesting signals such as international shortwave radio, ham radio communications and several military transmissions exist.
The RTL-SDR's lowest tunable frequency is 24 MHz, and so it can only receive a small portion of the interesting transmissions that occur between 0 - 30 MHz. In order to listen to frequencies below 24 MHz an upconverter is required (either that or perform the direct sampling mod). An upconverter works simply by shifting these lower frequencies up to a higher frequency that the RTL-SDR can receive. For example, a 5 MHz signal might be upconverted to 105 MHz.
To date, most decent upconverters (such as the popular ham-it-up upconverter) have been based on the double balanced mixer architecture implemented by the ADE-1 mixer chip from Minicircuits. The SpyVerter on the other hand is based on a different type of architecture which is inspired by the H-mode mixer design that was used in the unreleased HF7070 communications receiver. The expected major advantage that this design has over a ADE-1 based design is better IIP3 performance. This essentially means that strong signals will not cause overloading issues in the SpyVerter, meaning less noise and spurious images.
Another advantage of the SpyVerter is its use of a 120 MHz low phase noise/low jitter clock, meaning less reciprocal mixing and thus greater SNR and a lower noise floor. A low phase noise clock is essential for getting good performance when receiving the very narrowband signals that are typically found between 0 - 30 MHz. The other upconverters do not specify their phase noise performance as far as we can tell.
The SpyVerter comes in a metal box, with three SMA adapters. A metal box is great because it helps keep strong interfering signals from entering the signal path, as well as stabilizing the internal temperature, keeping frequency drift to a minimum. Most upconverters only come with a metal box as a paid add on, but the SpyVerter comes in one by default.
Although the SpyVerter is designed to be used with the Airspy, it is fully compatible with the RTL-SDR as well. The SpyVerter can be powered via a USB cable, or via 5V bias tee (and this is compatible with the bias tee used on the RTL-SDR Blog units sold by us).
|Upconverter||Clock||Bias Tee?||LNA?||Architecture||Enclosure?||Extras?||Price (USD)|
|Spyverter||120 MHz, low phase noise||Yes||No||H-mode mixer inspired||Yes, in metal box by default.||Three SMA adapters.|
|Ham-It-Up||125 MHz||No||No||ADE-1 (v1.0) / ADE-1 Reversed (v1.3+)||Available as extra add on for $21.95.||Optional noise source circuit (but not populated by default)||$44.95 + enclosure $21.95|
|Nobu||100 MHz||No||No||ADE-1 Reversed||Available as extra add on.||None.||$56 + enclosure $25|
|SV1AFN||200 MHz||No||Yes||ADE-1 Reversed||None available.||None.||$57|
The creators of the SpyVerter, Youssef and Bob (W9RAN), sent us a unit for review a few weeks ago and below we present some real world test results. We compare the SpyVerter against two ADE-1 based designs. A Ham-It-Up V1.0 and an upconverter designed by Nobu of Japan. The ham-it-up V1.0 uses the ADE-1 in a standard configuration, whereas Nobu's upconverter uses the ADE-1 in reverse. Using the ADE-1 in reverse allows better operation near DC, but can create impedance mismatch problems. If you want more info about this, see this post where Mile Kokotov reviews the SV1AFN upconverter (also uses the ADE-1 in reverse) and the ham-it-up and discusses a bit about the ways the ADE-1 can be connected. We note that newer versions of the ham-it-up (v1.3) use the ADE-1 in reverse as well, so the performance of a modern ham-it-up may be closer to that of Nobu and SV1AFN's design.
In each of these tests we used an RTL-SDR Blog dongle, SDR# and the special decimation driver for the R820T released by Vasilli. Vasilli's decimation drivers make viewing HF signals in SDR# much more pleasant as it allows for a high resolution decimated zoom and thus improved visual SNR. This makes spotting narrowband weak signals much easier. His plugin also has a "Use SpyVerter for HF" option which simply automatically sets the correct frequency offset value of -120 MHz for the SpyVerter. Each upconverter was powered by an external battery pack.
The testing was performed sequentially, but all three units were tested within 5 minutes to ensure that HF conditions did not change too much. All tests used a roof mounted Miniwhip antenna (except for one test which used a Magnetic Loop).
AM Broadcast Band
Here we tested all three units on the AM broadcast band between DC and 2 MHz. Gain settings were kept the same for all tests. Click on the thumbnails to see screenshots from the other upconverters.
From these comparison images it is clear that the SpyVerter has the lowest noise floor and highest SNR on the signals.
With significantly less noise and weaker signals received by the Magnetic Loop the SpyVerter and Nobu's upconverter were similar. However, here the ham-it-up shows some significant overloading issues near DC, has an image of the DC signal near 2 MHz and also some FM band interference.
9 MHz Shortwave
In this test we used a Miniwhip and an RTL-SDR dongle to tune to the 9 MHz shortwave band. We used a sample rate of 2.56 MSPS and zoomed in using a decimation value of 4.
The screenshots show that the SpyVerter clearly has the lowest noise floor, best SNR and the least amount of noise artefacts. We think the artefacts and higher noise floors in ham-it-up and Nobu's upconverter may be caused by overloading in the mixer, and from reciprocal mixing from the use of clocks with higher phase noise. The artefacts in question are the yellow smudge like areas in the images.
11 MHz Shortwave
Again, this test at the 11 MHz international shortwave band shows similar results with the SpyVerter overloading much less and thus having a lower noise floor and less noise artefacts.
Non-Directional Beacons (NDBs)
Here we tested the upconverter on non-directional beacon (NDB) reception at VLF frequencies of around 250 - 300 kHz. We used a decimation value of 16 to get a nice zoom.
Non directional beacons are used to aid with aircraft navigation. All three performed similarly, but the SpyVerter had a slightly lower noise floor and less noise. These signals may not have been strong enough to notice any significant difference.
20 kHz VLF
In this test we tried to receive some military signals at a very low frequency of 20 kHz. We're not sure what these signals are, but they appear to be some sort of military direction finding signal. Here, the SpyVerter and Nobu's upconverter performed similarly, but the ham-it-up V1.0 could not receive the signal well due to it's standard use of the ADE-1.
Signal Generator Test
The final test we did was artificial in nature. We used a HackRF to inject a tone at 10 MHz and took screenshots. The more the noise floor rises up around the signal, the poorer the upconverter is performing.
Again the results show that the Spyverter had the least overload, followed by Nobu's upconverter and then finally the ham-it-up.
From the specs and results I think that it is quite clear that at the moment the SpyVerter is probably the best upconverter you can buy for the RTL-SDR. It is priced similarly when compared to the other upconverters, but comes with a low phase noise clock and metal enclosure by default. From our tests we think that its architecture appears to be superior compared to ADE-1 based designs when it comes to issues caused by strong signals. If we were able to receive even stronger signals in our tests then the difference would probably be even more significant.
So, if you already have an upconverter should you run out and upgrade it to the SpyVerter? That depends. You'll probably only see a significant advantage with the SpyVerter if you are able to receive strong signals. People with properly set up DX or ham antennas will probably see a significant improvement, but if you have just strung up a short wire antenna and your signal levels are low then you may not see much of a difference.
In a later review we will compare the performance of the SpyVerter on its designated SDR the Airspy.
The SpyVerter costs
$59 $49 USD, ships from China and can be bought from iTead Studio.
Update July 2016: The price of the SpyVerter has been reduced by $10, and now only costs $49 USD.
From what I’ve been reading it seems that they now make a Spyverter R2, but I can not seem to find a retailer that sells either of them anywhere. Has production of the Spyverter ceased?
My northern neighbours (France) love putting out in the air 500kW of power which vastly overloads my direct sampling RTL SDR V3 dongle. Would the spyverter solve this issue and get a cleaner spectrum? Here’s a picture of the spectrum. It gets worser though. All these signals are just the strong AM one which overloads and spreads all over.
The Spyverter has very high dynamic range, but it’s possible that your AM signals would still be far too strong. You might need a high pass filter.
Has anyone attempted to put the spyverter and SDR dongle in a small temp stabilized cabinet and check then for stabilty. It would be easy to do for a room temp application.
I’ve been looking for an upconverter and found this site. I was in the process of building my own but for the money the spyverter is hard to beat.
With regards to the oscillator and spikes, did anyone check the decoupling capacitors around the supplier oscillator? Before swapping the oscillator you should try changing out any power decoupling capacitors around the oscillator with type x7r. Those spikes as well as sidebands are a known problem with certain oscillators when not properly decoupled with low ESR caps. The caps would be much less than a new oscillator. Also, I’ve found some very high quality Valpey Fisher oscillators on ebay at 140Mhz that are really inexpensive, have very low phase noise and great PPM stability. I think they listed at $120USD per when new. I assume the spyverter has a filter after the MO and if so that would have to be changed to support the higher frequency. You would also have to change sdr# as well.
Got mine a couple weeks ago. I’m really happy, with one qualification. The linearity seems pretty good — I’m able to sweep through the very strong AM band without any spurious signals showing up, unlike the *multiple* spurious signals I see with — ahem — another upconverter.
The bad is that the oscillator used to generate 120MHz (a Fox XpressO) is not very temperature-stable and generates spikes spaced at 5 MHz throughout the spectrum, easily visible as lines at 5 MHz, 10 MHz, 15 MHz, etc (or in upconverted frequency, 125, 130, etc). The temperature stability is poor enough that without thermally isolating the unit from drafts, it was useless for the digital HAM bands, swinging by as much as 100 Hz.
I was able to solder in a new TCXO from Abracom despite it being larger than the original package, and the performance is much improved now. The spikes every 5 MHz are gone, and the temperature stability is rock-solid.
Overall, I think the product is great — it only suffers from a poor LO, which is easily fixed if you have the patience and soldering skills.
What’s the part number you used for the TCXO?
Abracon LLC ASGTX-C-120.000MHZ-1 from digikey, not really a cheap part but it’s hard to find a TCXO at 120 MHz without ordering in quantity.
Not what I would call a cheap mod. That oscillator is more expensive than the SpyVerter itself!
Specifically, I removed the original oscillator and the two zero ohm resistors at the outputs, covered pads 4,5, and 6 and the “near” 0 ohm pads with Kapton, soldered the new oscillator pins 1,2, and 3 directly to the existing pads 1,2, and 3, added a C0G 1000 pF across pins 2 and 3 as recommended by Abracon. I then ran a tiny wire from Pin 6 to VDD at the cap near the oscillator, and ran tiny wires from the output at pin 4 to the “far” pads of the 0 ohm resistors.
It would be interesting to see a performance comparison of Airspy/Spyverter vs. SDRPlay in the long-, medium- and shortwave bands…
Many thanks for the test. I like the way you showed the screenshots in the same situation with the different converters. The Spyverter really seems to be outstanding from the other two tested devices.
From what I have been reading previously my personal favourite so far is the SV1AFN converter, which was not part of the test, though it is in a similar price range – any chance you can compare that with the Spyverter?
What I am missing the most is the converters oscillator stability and temperature sensitivity, which can render the device useless and it has not been mentioned at all. I once build a older kit (v3.1) of the CT1FFU HF converter and apart from its poor design it would barely be able to be stable enough to receive an AM station!!! if the room temperature doesn’t change! Now think about SSB or any of the narrow band teletype modes or whisper … no chance to receive any of that if the oscillator is unstable.
The dongles stability even without a TCXO is quite acceptable for all of this when used in indoor temperature ranges, once the device has warmed up after a few minutes and the calibration factor has been adjusted.
So I wonder if you can add the oscillator stability to this comparison?
Many thanks and best regards
I don’t have a SV1AFN upconverter so couldn’t compare it, but I think out of the ADE-1 based upconverters SV1AFN is the best as it is similarly priced to the others and comes with a switchable LNA and has the most filtering out of them all. The SpyVerter is probably still the best though if there are strong signals around. The comparison was mainly to show the difference between the SpyVerter design and the ADE-1 designs in both connection orientations. I think all ADE-1 designs would show similar performance.
It’s hard to measure the upconverters stability because we can’t be sure if we’re measuring the dongles stability or the upconverters. The SpyVerter should be very stable however as it is basically ovenized inside it’s metal case.
Looking at the IP3, the Spyverter looks quite impressive and regarding the oscillator, it may even be that the SV1AFN uses the same … look at this data sheet
page 5 shows you the phase noise and is the same as specified looking at each oscillator frequencies.
SV1AFN does a lot of filtering behind the oscillator to keep the signal very clean.
The PPM factor describes the possible oscillator offset due to slight manufacturing differences in the crystal, it does not say anything about the frequency drift versus temperature. The fixed offset factor doesn’t matter, because that something you calibrate in the SDR software, but the drift can make things impossible when the circuit warms up, it stays in a certain temperature range and what matters is how high the drift is in this small operational range after warming up.
You say the Spyverters oscillator is “ovenized”, I doubt it. If it were so, I am sure Youssef would advertize that it uses a TCXO, because that would be a significant feature for the user and even more for the product marketing, because so far I believe there is no converter out that can be ordered optional with a TCXO. It is also almost impossible to squeeze it into the very good price, considering it comes with a solid metal case, some adapter cables and maybe even a bias-T, because it lacks any external power connectors?!
The only thing I don’t like about all SV1AFN designs is the fact that his RF connectors of all of his devices are always directly mounted flat onto the pcb, not sideways standing out, because that implies if you want to install it into a metal case you need some extra coax and 4 more RF connectors to make it work. Or maybe I am wrong and the parts on the PCB are so low that you can mount it directly into the case, but I doubt it.
What might do the trick is to unsolder the connectors and resolder them on the other side of the PCB, because that side is flat!
It is questionable if the IP3 parameter is still a major factor for a receiver today, where the big power broadcast stations on long- medium and shortwave are all being take off air and only a very small percentage still remain.
My major concern would be the choice of the oscillator frequency of 120MHz with the Spyverter. Even if you have effective cascaded low pass filters for VHF rejection in the front end, the very strong local FM- and aircraft radio signals might still get into the circuit, so I wonder if you experienced any of that yet?
E.g. 88 until 108MHz > 12-32MHz IF … so in that range you might experience some remains from very strongest local FM stations interfering or random aircraft communication signals (118-137MHz > 0-17MHz IF) from planes over your house sending with up to 100W AM with line of sight signals, I am not sure if that is realistic keeping these completely out! This is a big plus for the 200MHz oscillator of the SV1AFN design, since there are not so many VHF2/DAB signals around and they are even further away from the actual HF frequency.
It is tricky to find a good choice, because some 50-60MHz oscillators gives you problems with either 88MHz or 30MHz and so maybe something around 300MHz would be the best solution, because there is no broadcast or HF signal that might mix, only a very little bit of military aircraft activity and that is it. Satcom is to weak to interfere, and the next ground based service starts somewhere above 410MHz here in Germany I believe. But at 300MHz you have to create your frequency from a TCXO, else you are lost, so there will be a higher price tag, because you need a multiplier and plenty of filters to get a stable and clean 300MHz signal.
Bottom line, there is still plenty of potential for even better HF converters, so lets see when they will be released and at what price point. These will also be interesting to people who use more sophisticated SDRs, because some of these start at 70MHz, some even above 200MHz, and they can digitize the entire HF band up until 6m in one stream at 12Bit, which will be a great base for a Web SDR, but it requires raw CPU power to process and distribute this … so there will be more than just the low cost SDR market that requires quality converters.
Anyway many thanks for your thoughts.
Hi Joe, thanks for the comments. When I commented about it being ovenized I just meant that it may act that way because the entire circuit is mounted inside an RF tight metal case. Its a similar effect to what happens when you put the RTL-SDR in a metal case, the oscillator temp drift stabilizes significantly because it becomes like an oven inside the case. I didn’t notice any FM interference and I have some pretty strong stations nearby, but it could be that stronger stations might still get through.
> It is questionable if the IP3 parameter is still a major factor for a receiver today
Do you have any clue about RF dude?
> so lets see when they will be released and at what price point.
Yep. Appliance operator thinking.
I own a Spyverter, a Ham-it-Up, and a homebrew upconverter based loosely on a QST design. The Spyverter is my upconverter of choice. 🙂