Airspy HF+: An upcoming low cost yet high performance HF SDR
Over on the Airspy Yahoo forums and Twitter we’ve seen news of an upcoming new product from the developers of the Airspy SDR. The new product is called the Airspy HF+ and will be a low cost, yet extremely high performance HF specialty radio.
Preliminary specs:
- HF coverage between DC .. 31 MHz
- VHF coverage between 60 .. 260 MHz
- -138 dBm MDS
- -142 dBm MDS at 500Hz bandwidth in VHF
- +26 dBm IIP3 on HF at maximum gain
- +13 dBm IIP3 on VHF at maximum gain
- 110 dB dynamic range in HF
- 95 dB dynamic range in VHF
- 120 dB Image Rejection
- Very low phase noise PLL (-110 dBc/Hz @ 1kHz separation @ 100 MHz)
- +10 dBm Maximum RF input
- Wide Band RF filter bank
- Tracking RF filters
- Sharp IF filters
- Smart AGC with real time optimization of the gain distribution
- All RF inputs are matched to 50 ohms
- 2 x High Dynamic Range Sigma Delta ADCs @ 36 MSPS
- 600 kHz alias and image free output
- 18 bit DDC
- 0.5 ppm high precision, low phase noise clock
- 4 x Programmable GPIO’s
- No drivers required! 100% Plug-and-play on Windows Vista, Seven, 8, 8.1 and 10
- Industrial Operating Temperature: -45°C to 85°C
Basically, this addresses the lack of affordable and good performing receivers for HF and VHF.
Target price < $200
As with all Airspy products the SDR focuses on achieving extremely high dynamic range. From the specs is seems that the dynamic range and image rejection will be high enough so that even extremely strong broadcast AM or FM stations will not require any filtering or attenuation. They are also confident enough to say that no gain sliders will need to ever be adjusted to avoid overload.
For SWLers and MW DXers this seems like the ideal SDR as it should perform as well as high end SDRs like the Perseus, RFSpace and Elad SDRs, but at a fraction of the price.
Smoke testing Airspy HF+ The new HF/VHF High End receiver. Yes, no gain sliders needed! #AirspyHFPlus pic.twitter.com/T1T9OvpbcG
— prog' (@lambdaprog) February 18, 2017
The product is still in development and no release date has been offered yet, but judging from the Twitter feed the prototype is already working.
You need to take seriously the need NOT to leave gaps in coverage. Why accept a gap from 31 Mhz to 60 Mhz? Is the 6 meter band irrelevant? No it is not. I won’t buy this for that reason.
So that means you would never buy a Perseus SDR (10 KHz – 40 MHz) nor a RF Space SDR-IQ (0.0001-30 MHz) because they both fail to cover the 6 meter band ?
Sorry, but short of a finished product, final pricing and real life testing it is still just that… HYPE! I take the SDRPlay RSP1 or 2 any day and they are already “affordable” …LOL
It’s nice to see an attempt to fill this gap in the affordable SDR product space. The BW is OK, in-fact more and I would really start to question if the specs are real. Some things I’d like to see included:
1. Pads for a 10MHz (or similar) external reference input connector (e.g, MCX, SMA). This is really needed for small signal work like WSPR, beacon monitoring, moon-bounce, etc. Because of the low phase noise and broad frequency range, this is where the Silicon Labs Si571 (not Si570) is a game changer. To keep the Si571 cost down, use divide by four quadrature L.O. generation for the HF bands, an the same with subharmonic mixing for VHF.
2. A buffered input (e.g., LNA) to separate the receive performance dependence on antenna matching (as often seen with DDC SDR’s). This is a tough request to fill while still maintaining dynamic range, but worth it. The buffered front end would also fix the input characteristic impedance at 50+j0 Ohms which helps if you are going to add an external band-pass/reject filtering to further improve band-specific performance (something many will undoubtedly do, especially contesters). Finally (last but not least) you may find that buffering may be the easiest way to significantly reduce L.O. leakage out of the antenna port!
3. A switched attenuator on the input is needed. The amount of noise on the lower frequency HF bands is a lot more compared to the higher frequency bands. From 1MHz to 10MHz on a noisy night, having an RF attenuator to switch-in can be a game changer. One way to get this affect with multiple steps is to switch an attenuator in/out and allow the input LNA to be in/out switchable.
Will it offer an ExtIO interface ? Some users want to have freedom of choice for the software to use.
If it supports Linux (which I honestly can’t imagine any new SDR device not doing in some way) then enough source code should be available that someone can write ExtIO dll for it, just like for the Airspy and mini.
https://github.com/amontefusco/extio-iw0hdv/releases
I would never expect hardware developers to support every single piece of SDR software in existence, but I would expect enough information to be released so that anyone who wanted to badly enough could run with the details and add their own support. But I’m only guessing, I’m sure we will all find out more when the hardware is available.
As always, Great exaggerations from Youssef!
You may find these words come back and bite you on your tail once the HF+ is released.
If first thought ‘wow’ and then I realized the max BW was “only” 600 kHz … (Unless I got it wrong)
It’s nice to see a big wide band of data as you’re scouting the airwaves and indeed a large bandwidth has its uses – running a webSDR, processing multiple signals simultaneously, using your SDR for spectrum analysis, processing wideband signals etc
But when it comes to actual reception of a given signal, things such as dynamic range, image rejection and so forth (as little as I am admittedly yet understand of the technical aspects) are the only things that matter.
Take an 8bit/12bit SDR – I was quite shocked at how much my SNR improved in various bands when I started applying filters to my signal chain – we’re all familiar with FM notch filters and the such but what’s available in terms of manufactured filters are to suit highest common denominators across a whole world of frequencies.
Think of all the powerful signals, interference etc at various frequencies that we can’t just buy a filter for – think of powerful signals that are so close to your desired signal that even the sharpest saw filter couldn’t cut them off without degrading what you really want to receive.
An SDR with the specs that all but eliminate signals bleeding into other frequencies? that alone surely improve matters across the whole spectrum – let alone other aspects I don’t yet understand that also contribute to a good SNR.
I’m sure when this device is available to buy we’ll see informative real-world tests and comparisons so that those of us unable to interpret the stats can see what this device could do for us.
I mean most of us just bought an RTL dongle because it’s really, really cool right?
But when it comes to actual reception of a given signal, things such as dynamic range, image rejection and so forth (as little as I am admittedly yet understand of the technical aspects) are the only things that matter.
Holy words
It’s nice to see a big wide band of data as you’re scouting the airwaves and indeed a large bandwidth has its uses – running a webSDR, processing multiple signals simultaneously, using your SDR for spectrum analysis, processing wideband signals etc
It could be a good exercise to imagine how a RTL dongle and the AirSpyHF could be coupled to offer a wide spectrum view and the best performance on the tuned signal. This should be handled by the software to switch between the two with a single click and reduce the system resources needed by the software. But probably optimizing such system we wouldn’t even need a 600Khz bandwidth.
Then you should try a bpf in front of a 8bit hf dongle ( ex. RTL blog V3) and be surprised how well it stand against its 12bit cousins (drplay / Funcube)
I’ve seen the SDR# code mention HFSpy, also, i noticed HFSpy.com redirects to sdrsharp.com
Very impressive specs – however $150-200 (<$200 is hardly going to be closer to 100$ is it?) is very steep given you can pick up a HF solution for a fraction of the price – the real-world performance would have to be exceptional to take the money of all but the most ardent of HAMs.
Existing R2+Spyverter customers? feels like a bit of a kick in the teeth – had I known this was in development I'd of happily waited and paid more than the existing cost of the R2+Spyverter instead now I feel a bit like I've been fleeced for old stock. They could of announced this was in development months ago and secured more customers. Such is life.
Existing customers will likely be looking to TX capable devices such as the LimeSDR for their next purchase. We could of been salivating over a new TX capable Airspy instead we're looking at a *very* good but not quite justifiable 'Spyverter Pro'
For the price you can get an RSP2 with good support and an open software interface…
The Airspy has, and the HF+ will have an open interface. Give Youssef a chance – he’s working every hour there is on this project.
I’ve examined what code is available for both the Airspy and SDRPlay fronts and looked at the problems people have had with both product families – I can’t attest to the quality of the hardware designs but as a developer I personally believe Airspy has the better quality code and cleanest API.
Competition is good! it has and will continue to refine both product families – I may be in two minds about purchasing a HF+ given what I’ve already spent (I eagerly await reviews and comparisons) but I’m very happy with the Airspy and it’s been an absolute breeze implementing all of its functionality in my own libraries.
I’d personally like to see more modular designs however – I’d rather have a separate bank of filters and separate TX/RX devices etc allowing for greater customisation and lower cost upgrade paths that I can tailor to my interests/goals.
it’s a big decision to buy a £200 device but I’ll splash out £15 / £30 on a new filter, LNA, Diplexer, Balun etc much more readily – I think a more modular family of products could be really popular and would represent a less risky and more future proof investment for hardware developers than one all singing all dancing device that could be made redundant tomorrow.
Johnny,
That’s not bad suggestion at all. Technically that’s a very nice solution. But what about entry price for newcomers? I expect interoperability testing of various hardware and software modules would increase the cost, hence the price. I have no figures, but RTL-SDR being popular has definitely something to do with the price the average SDR user want to pay. I think < $200 for an integrated SDR HF receiver with high quality HF reception is what SWL's are willing to pay for a(nother) HF radio.
D. B. Gain has to admit to a slight bit of drool on the keyboard.
please do a “anti” spurii design with a lot of EMI Filter and decoupling from the USB bus…
Yes, this is the biggest problem of the original Airspy, ridiciolous amounts of spurii and USB noise.
Not a new R2 version but a complementary higher level SDR for HF and VHF only.
I will not consider it untill the old memory display is reintroduced in SDR#
Hi! This is new version R2?
From what I have read this is not a R2/mini (12bit ADC), it is something very different (18bit ADC). If I reach far and make make a leap of faith that a post I read elsewhere is related to this device, then I’ll quote someone who I suspect has a early prototype, “Well this radio will really annoy the big players”
A Perseus has a 14 bit 80 MS/s ADC, if you decimate that by 256 (4^4) to increase the dynamic range to18 bits that would be 18bit @ 312.5 kHz and the Airspy HF+ internally is 18bit @ 600kHz.
Upside
price/performance – wow
Downside
600kHz bandwidth
DC-~30MHz and ~54MHz to ~270MHz (if I link this post to the other one I read).
The funny thing is if it did not cover ~54MHz to ~270MHz I would not consider 600kHz bandwidth a downside. In fact the only thing I can think of that would be more than 600kHz would be DAB/DAB+ ( https://en.wikipedia.org/wiki/Band_III#Radio ) which is not that much of a loss, at least to me. Oh and terrestrial television which would require 6-8MHz bandwidth DVB-T/DVB-T2/ATSC which again are not really of interest to me. Is there anything else ?
Did I mention that the Perseus , considered by many as a gold standard SDR receiver, costs about a grand.
At the risk of being a dummy, why is there a gap in low VHF between 30 MHz and 54 MHz?
Slight bummer that the amateur 6 meter band is not covered; especially since there is a significant amount of weak-signal ops there. Not a deal-killer though.
The post I read elsewhere (I’m not linking to it because the person said they should not have posted it in the post!) was about a device that had a dynamic range of 120dB+, covered HF and ~54MHz to ~ 250MHz. I’m only guessing the two devices are the same, I could totally be wrong, I’m taking the low end from that post and the high end from this post and merging them into ~54MHz to ~270MHz and replacing HF with DC-30MHz. It could be two totally different devices.
Beware, the bit number in ADC (Analog Digital Converter) is unknown, it can be 12, 14 or 16 bit. According to the diagram link only the DDC (Digital Down Converter) with 18 bits is shown. However this translates into about 110 dB of Dynamic Range.
https://www.rtl-sdr.com/wp-content/uploads/2017/04/ashfp.png
Yes it does not say the bits, but it does say “2 x High Dynamic Range Sigma Delta ADCs @ up to 36 MSPS” ( ref: http://airspy.com/airspy-hf-plus/ ). If you assume that one ADC is for I and one is for Q, then you have two pieces of information that at 600kHz of bandwidth after DDC you have 18-bits, but before that you had unknown-number-of-bits at up to 36 MSPS. Also if you look at their spyserver configuration file ( spyserver.config ) they have a comment that the Airspy HF+ has a sample rate of 744192 Hz (and 0.8 times that would be almost 600kHz). 744192 will divide into 35721216 Hz (~36MSPS) exactly 48 times. Which is more than 2 bits (2^2=16) and less than 3 bits (2^3=64). So it could be a 15-bit 36 MSPS Sigma Delta ADC, and they are gaining that missing fraction of a bit by DDC’ing to 600kHz. I am reading between the lines, but that is my thoughts anyhow.
No
As always, Great work from Youssef!
This might be the SW receiver I have been waiting for. 600 khz at 16 bits most will work fine with my current PC’s. More bandwidth would need an investment in a higher powered CPU. Well, that would reduce the potential buyers, don’t you think? Anyway, with Spectrumspy it only takes 50 samples to scan from 0 – 30 Mhz; and that will be high quality samples!
Cheers, Jaap
PS, the Icom R8600 shifted down on my wishlist 🙂
I expect you could buy the HF+ and a good PC for ~ the price of the R8600, I haven’t seen the R8600 price though.
Maybe even 1 pc and 2 HF+. ?Great to experiment with antenna diversity.
Cheers, Jaap