Evaluation of SDR boards and toolchains [pdf] | Hacker News Reader

Evaluation of SDR boards and toolchains [pdf]

gitlab.com

92 pointswglb24 comments

A bit surprised that HackRF wasn't even mentioned, never mind included, given that it is fairly popular. My understanding is that HackRF's noise figure isn't that great so maybe that is why, but that wouldn't jive with the inclusion of the RTL-SDR.

The document at hand evaluates the use of SDRs for satellite communications (including Low Earth Orbit signals). HackRF is an awesome product and project, no doubt about that. Personally I would suggest anyone interested in a benchtop SDR get a HackRF. It's a great open-hardware project.

This is an evaluation of performance on bands used for satellite communications, i.e. VHF and above.

If anyone is looking for an SDR primarily for HF and below, I think the Airspy HF+ Discovery is considered the best affordable option right now. If you want a good general-coverage SDR, I'd recommend one of the SDRplay products like the one they reviewed or the cheaper RSP1A.

Or Flex Radio for the high end;

https://www.flexradio.com/

I'm interested in HF, but I'm interested in transmitting too. It's my understanding that Airspy is receive only?

I have the RTL-SDR and use it mostly for monitoring VHF/UHF, and it works just fine. There's nothing interesting in my area above 900 MHz so the frequency limitation above 1 GHz isn't an issue for me. It's a great way to get started with SDR if you're limited with your budget.

Holy crap, just skimming this it’s obvious this is going to be an amazing resource for folks making the agonizing decision of what kit to go with. Awesome!!!

I’d go for the widest real-time BW and best phase noise (clock jitter), as there are things you can do with offset tuning, low IF, and decimation to get around some of the spurs and dynamic range limitations that are inherent in these direct conversion transceivers.

A few external components (LNA, filter) can make a large increase in performance, so specs like noise figure are minimal for many applications.

That's good advice.

But yes, the best way to cure the ailments of zero-IF is to use low-IF, and the best way to cure the ailments of low-IF is to use high-IF and filters -- and we're back to standard RF architecture.

Nothing will make you respect traditional superheterodyne architecture quite like fighting with software image suppression / signal path correction.

Or direct sampling. Although they're expensive, there are 10 Gsps ADC's these days.

https://www.analog.com/en/products/ad9213.html

Holy moly you weren't kidding, they're nearly $6K USD apiece. Does anyone's SDR kit drive these?

On SDRs that transmit, I'd like to see the how much harmonic suppression there is. For example, if you transmit at 144 MHz, what's going on at 432 MHz?

If they use switching mixers I would expect a lot of power in the harmonics. If they use analog multipliers, less. There has been work to make the front end mixer more multiplier-like using "harmonic rejection mixers". Apple bought one such company, I forgot which. The advantage is that you can reduce the analog filtering needed.

I think it should be the same for receivers: I assume they pick up a lot of noise from the harmonics.

They're horrible. The 3rd harmonic is just a few dB down.

Yep. The killer for the AD9364 is the 5th harmonic of the 2m amateur band falling on this nation-wide (USA) LTE signal. That damn LTE seems to be everywhere, and it’s strong.

Hi people. I'm the vice chairman of Libre Space Foundation (https://libre.space) together with the European Space Agency (https://esa.int) we coordinate this report.

Just wanted to make sure that this document is still in review and it's results shouldn't be considered final yet.

Cheerio.

Why was the HackRF not included at all?

I don't have a good definitive answer to your question, other than we had to limit the scope of our work and we chose a representative set of devices amongst the ones that people showed interest in. Despite its popularity in wireless security research, I don't think the HackRF is widespread in the satcom or cubesat communities.

As far as I know (@csete or others might correct me here if I'm wrong), HackRF is primary a benchtop Software Defined Radio, and it's excellent for testing, prototyping, etc.

The paper @csete works on is focused on the use of such devices to enable satellite communications one should take into accounts the limitations and challenges of such signals (especially when working with Low Earth Orbit satellites).

HackRF is a great product with great value; if you are seeking to experiment with a versatile Software Defined Radio. In my personal opinion, it is the "swiss army knife" of SDRs.

Thank you for the explanation, I appreciate it :)

Does anyone have a good reccomendation for an HF transceiver? It seems like many popular SDRs are receive only or crazy expensive. I've been looking at the RS-HFIQ (https://www.hobbypcb.com/index.php/products/hf-radio/rs-hfiq) but it doesn't seem to be that popular and I can't tell if it's because it's got too much "hacker" in it for the average ham, or if something else is going on. Maybe I should just spend more money?

There is a duality between affordable and performance. For example, the Hack RF is pretty great and has a large frequency range and wide bandwidth for a relatively small cost. But for HF transceiving you need probably need better filtering, which has to be specific to the band and mode, amplifiers both ways, although, with transmission you probably need several stages of amplifiers and filtering in each stage as well. So all of those things would be costly and depending on what you want to do it may be better to buy a regular radio.

I agree with the duality that you describe. Nevertheless, wouldn't the software part of an SDR reduce the need for expensive physical electronic components, while increasing ability? An amplifier seems to be the only additional expense (as you pointed out). It seems like the amplifier negates any cost benefit, yet utility remains superior if one doens't consider multiple components and building blocks a liability. It seems like there are expensive SDR transceivers (e.g. Flexradio) which utilize SDR technology to beat out the competition in specs and performance. To get on the air with a budget SDR, I haven't seen lots of demand for the idea. is the utility of a budget SDR just not better than your basic radio? It seems like it to me but where's the demand?

Yeah I think most big name radio producers see amateur sales as a secondary market. There are some hobby projects like the uBitx and the mcHF, and both are low power. The amplification and filtering are in opposition to dynamic range and total power output. For instance without filtering, a very strong signal will blow out the dynamic range and quieter signals will get lost. If you say were to build a ZetaSDR and plug it into a 192khz 24bit sound card, you would have excellent dynamic range for weak signals, but you'd still want to filter out local FM and AM stations.

I think as more SDR stuff gets reduced to single chips for commerical or military reasons, then perhaps the amateur equivalent will be viable... But filtering and amplification is still something that is difficult to miniaturize due to heat and wavelength.

I've never tried any of them, but [Crowd Supply](https://www.crowdsupply.com) has a lot of SDR projects that look reasonable (like LimeSDR, etc). Maybe others can comment if they've tried any of them.

Reply on news.ycombinator.com