Five-Watt SDR Transceiver For Hams

The availability of cheap SDR hardware created a flourishing ecosystem for SDR software, but a lot of the hardware driving the revolution was still “cheap”. In the last few years, we’ve seen quality gear replacing the TV dongles in many setups, and down-converters designed for them to allow them to work on the ham bands.

But something that’s purpose-built might be a better option if ham radio, particularly the shortwave portion thereof, is your goal. First off, you might want to transmit, which none of the TV dongles allow. Then, you might want a bit of power. Finally, if you’re serious about short-wave, you care more about the audio quality than you do immense bandwidth, so you’re going to want some good filters on the receiving end to help you pull the signal out of all the noise.

rs-hfiq_block_diagram_featuredThe RS-HFIQ 5 W SDR transceiver might be for you. It’s up on Kickstarter right now, and it’s worth looking at if you want a fully open source (schematics, firmware, and software) shortwave SDR rig. It’s also compatible with various open frontends.

The single-board radio isn’t really a full SDR in our mind — it demodulates the radio signal and sends a 96 kHz IQ signal across to your computer’s soundcard where it gets sampled and fully decoded. The advantage of this is that purpose-built audio rate DACs have comparatively high resolution for the money, but the disadvantage is that you’re limited to 96 kHz of spectrum into the computer. That’s great for voice and code transmissions, but won’t cut it for high-bandwidth data or frequency hopping applications. But that’s a reasonable design tradeoff for a shortwave.

Still, an SDR like this is a far cry from how simple a shortwave radio can be. But if you’re looking to build up your own SDR-based shortwave setup, and you’d like to hack on the controls more than on the radio itself, this looks like a good start.

50 thoughts on “Five-Watt SDR Transceiver For Hams

    1. The more common, the cheaper. So something intended for tv will be really cheap, and designed for a world where the desired signals will be the strongest. Something that transmits is going to be at the other end, since relatively little transmits. Well, cellphones sell in massive quantities, but I assume the ICs are designed for those.

      We saw the same thing with analog ICs. Lots of interesting parts in the seventies, but not so much demand for them since fancy communication equipment was a much smaller market than consumer electronics. The average am/fm radio used an IC intended for the purpose, less adaptable for other things.

      The TV SDR gizmos have wide range because tv is over a wide range. Something for cellphones, and I remember getting info about those 20 years ago, were really intriguing, and likely what was needed for non-broadcast use, but their tuning range was limited.

      If the digital part can be done with a computer, the RF side can be built up with analog components, just as any radio equipment was built in the past. So you’ll deal with traditional problems like a wide range stable local oscillator and front end selectivity, but tge former is what’s lacking on those TV SDRs.

      One reason better gets more expensive is that it is better. Higher resolution, which makes for better reception.

      Michael

    1. Unfortunately the most we have so far been able to get from RTL sticks is -75dBm for stock sticks and -45dBm for modified.
      That said, we got +100meter non-line-of-sight reception of 1270MHz RTTY between two RTL sticks.
      WA5VJB 850-6500MHz LDPA on TX and WA5VJB 3el 1296MHz Yagi on RX and we still had link budget left and got good decodes.
      There is plenty to be done optimizing the TX side with better “tap” for the carrier, better filtering and adding amplifiers.

      I really need to make a better video about a QSO between two RTL sticks.

    1. The second machine screw above the TO220 package is holding the heatsink to the enclosure. Likely is meant to be sold in levels with/without case.

      And $20 for the bare board and you can populate the RX section only and its probably one of the cheapest high quality SDR’s out there.

    2. Mechanical design is one of the hardest parts of putting out any product. The RS-HFIQ is designed to be run without the enclosure and therefore needs a heat sink onboard. There is a heat sink on both sides of the board, and when the board is slid into our enclosure, the bottom heat sink contacts the enclosure and transfers heat to the enclosure. Either way, the unit runs very cool with or without the enclosure.

      1. One of our beta testers did some extensive thermal testing for us:

        “I have just finished a one hour key down test at 4.75 watts. The case temperature of the final went to 159-160 degrees within the first five minutes and then stabilized. At one hour key down the temperature was still 160 degrees (F) and power was still 4.75 watts.”

        The RD16HHF1 MOSFET is pretty tough when it’s running at 5 watts.

    3. Never limit your options in the design phase unless you’re 100% certain of the final design or if you are comfortable locking down your supplychain.

      You can cut holes in any case to provide airflow past a heatsink. You can’t heat sink every case.

    4. Yes, there are all sorts of ways to do design. But, have a look at the Globemaster SDR. Tiny enclosure with a big grunt. Up to 38 watts PEP on the 1.8 to 72 MHz bands and to boot it has 144 to 1296 MHz at 2 Watts PEP included.
      Oh, I almost forgot, it has its own sound card in the same box. A totally integrated design. You can check it out on my qrz.com and at globemaster@groups.io

      Regards Ed GM3SBC

  1. We hams are extraordinarily fortunate in that while a license is required, type acceptance of the equipment is not.

    In most other services (notably apart from Part 15 devices), type acceptance is the norm, where the manufacturers of the equipment have to have their products undergo testing to insure that they comply with emissions standards when operated. Type acceptance would basically be the kiss of death for amateur radio, as it would require any homemade equipment to be individually type accepted. It would reduce the entire hobby down to just a slightly fancier version of Citizens’ Band.

    Type acceptance did manage to enter into the hobby in one particular way – HF amplifiers must be type accepted to insure that they don’t work on the 11 meter band, which usually also makes them useless on 10 meters. But the rules have accommodations for individual hams to modify their type accepted HF amps so they can be used on 10 meters, and allows for very limited resale of such amplifiers within the Amateur community (to try and keep them away from CBers).

    1. Yes it is a Softrock! with nine significant differences:

      1. 80/60/40/30/20/17/15/12/10M bands (it would take 3 Softrocks to cover the same range)
      2. 5W power output (nothing you can do to a Softrock to make this happen except add an amp)
      3. Lowpass filters so it’s legal on all bands (the Softrock you have to plug in a LPF on one of it’s bands)
      4. RF, AUDIO and Digital grounds (the Softrock ties the RF and Audio together so the antenna ground isn’t ground)
      5. MACOM wideband transformers for excellent phase and amplitude balance (the Softrock uses hand wound BALUNs)
      6. SI5351 LO generator with 3 outputs (the Softrock has a SI570 with a single output)
      7. Open source FW running on an Arduino (the Softrock uses a difficult to update ATTiny85)
      8. LNA in the RX path for improves RF sensitivity (The Softrock hs no gain in the RX path before the mixer)
      9. The board is fully assembled and tested (lots of SMT assembly and coils to wind with the Softrock)

      The Softrock is a great little radio but some of it’s limitations limit it’s appeal. Not every Softrock user will be interested in our RS-HFIQ but the RS-HFIQ will appeal to some hams who wouldn’t consider building a Softrock.

      1. I guess everything is relative, but I wouldn’t call the ATTiny series hard to program. You can even get serial bootloaders (but you’d be unlikely to use them given the small flash space on most of the tinies).

        1. Agreed it’s all relative. But try this experiment. Walk into a classroom of 7th graders give 1/2 of them an Arduino Nano, give the other half a ATTIny based development board and in 1/2 hour ask both groups which platform is hard to work with and which one is easy.

          Give someone an Arduino Nano, Uno, or Mega, and even if they have never used a microprocessor before with a USB cable and a computer most people can get ‘Blink’ running in under 5 minutes and start modifying the code in less than 10.

          Now let’s expand this to the RS-HFIQ. One of the beta testers wanted to run the rig with a PowerSDR variant that requires a 10 KHz offset in the LO frequency. We were able to build it into the Arduino Sketch in 2 minutes (5 minutes if you include downloading/installing the Arduino IDE), and in about 1/2 hour add it to the control structure so it can be turned on/off, adjusted for different offsets and stored in EEPROM. How long would to make a similar change in the Softrock’s processor?

          I’m a big Arduino fan.

          Jim.

    1. Err… this is the most HAM thing I keep hearing:

      I want SOFTWARE defined radio. But please keep me away from computers. And Software.

      Sure, I can put the software to your SDR inside a box and hide it from you. give you one big knob to play around with. I’ll build a device that can be much better with the same analog filters than a pure analog receiver. And lets you do exactly the same thing that you could have done with a more expensive analog receiver.

      But in the end, why people on Hackaday, and Hams should be excited about SDR, is that you can extend/modify the signal processing done in software *exactly* because that can run on a normal PC, where I have actual access to it. That means that your transceiver isn’t actually only AM/SSB/FM only anymore, with maybe three different configurations for the bandwidth. It can be anything. It can be any digital mode. It can be an FM of an arbitrary frequency deviation and preemphasis. It can actually carry pictures alongside with digital speech, or a ticker.

      If you want your radio to interact with you only by means of a single button, please leave me alone with that desire. I like to be excited about technology. I’m not excited about the load of things I can’t chose when my radio is just a boring device that can only do the things it was wired up to do.

  2. Can anyone tell me the sample rate of this device? Really, this is the *core* specification of an SDR transceiver. It defines the amount of bandwidth you can get at once, defines how much you can apply digital filters (instead of relying on, worse by orders of magnitudes, analog filtering), and thus pretty much says how much better than a good ol’ analog receiver you can get.

    Also, effective sample bitdepth (==dynamic range), and Noise Figure would be appropriate. This feels like someone is telling me

    Kickstarter Campaign: “Buy my great device!”

    Me: “We’ll see. Can you give me the core specs?”

    Kickstarter Campaign: “Sure! Core Spec: it’s great!”

    Me: “Thank you very much.” [goes somewhere else]

    1. I have characterized the RX performance of the RS-HFIQ and the HDSDR software using four different sound devices including the SoundBlaster X-FI, ASUS XONAR, the low-cost Startech that we sell and the lowest price 96 kHz device I could find. You can see how they fared here:

      https://sites.google.com/site/rshfiqtransceiver/home/technical-data/rx-performance

      The biggest difference if the large signal handeling for the most part all of the units could pull in the weak (<-120 dBm) signals without issue.

      Jim.

  3. I’m delighted to see this Kickstarter project and wish it every success. But the requirement to work with a computer will turn off those who want a transceiver that’s less complex and more portable, meaning something built and priced more like this:

    http://qrznow.com/bitx40-a-fully-assembled-45-ssb-qrp-transceiver/

    It is a hacker’s delight at $45. That’s a 40-meter SSB transceiver assembled as a board and absent only features such as a case, the power supply, and an antenna matcher. It’s so appealing, I’ve begun to scout around for those parts myself.

    Anyone with more time on their hands than I have might want to research ways to turn it into a full-flegged transceiver including a weather-proof case. It’d be great for hikers, bikers and travelers to have a transceiver that’d fit in less space than a cigar box and costs under $100. I would also be great for young hams with tight budgets. And being SSB should open up it up to digital modes.

    –Mike Perry, KE7NV/4

    1. The bitx40 only works in one frequency band, so it’s not really comparable to an SDR. You’re metaphorically saying an LCD screen should be as cheap as a monochrome box light.

    1. Beat me to it.

      That’s like the clods who say “I’m married to a good women”……
      or “I could care less”.

      I’ve thrown applicant resume’s in the trash for grammatical
      errors or typos. It demonstrates a lack of care/due dilligence –
      or plain ignorance (ie. stayed home from class that day).

      However, in this instance one typo is excused.

      Carry on

  4. Since this is not a piece of laboratory test equipment (USRP//HackRF/BladeRF/LimeSDR) which generally have a maximum output power of in and around ~10dBm (~10 mW). Does that mean that this hardware needs to go through FCC testing, CE compliance for the EU and the C-Tick mark for Australia New Zealand ?

    1. The FCC rules about the type certification of amplifiers are fairly clear. Less so with transceivers and even less with a product that is sold as a PCB. I tried to find examples in the FCC database where a SDR had been through certification and couldn’t find any. Since the technical applications are public record some manufacturers create ‘cover’ entities to file with the FCC to protect their proprietary data.

      That being said, we extensively tested the RS-HFIQ and all spurious signals are more than 43 dB below the PEP output of the transceiver as required by the FCC. If the transceiver is overdriven, some spurious signal could be out of spec so the RS-HFIQ has a ‘CLIPPING’ LED that illuminates when an over-drive condition occurs this makes it easy to set the TX level, insuring spectral purity without the need for test equipment.

      I am not in a position to comment on the rules and regulations of countries other than US.

      Jim.

      1. The rules on type acceptance for amateur transceivers are quite clear – there is no requirement for type acceptance or certification whatsoever. That’s what makes the hobby so great. It is up to the individual operator to insure that the emissions meet requirements, and because of the relatively stringent licensing (certainly compared to other services, where there’s bugger all in terms of proficiency testing anymore), it makes sense to hold the operators to those standards.

    2. Hi,

      Would you be able to explain why a piece of laboratory test equipment (USRP/HackRF/BladeRF/LimeSDR) does not require a C-Tick? Does ACMA have any documents about the rules for using SDR?

      Thanks
      Long

  5. Question: Is there applicable transceiver software available that will run on Linux boxes? This is really not a Linux fanboy question but one rooted in a practical concern.

    When you invest in SDR hardware that’s is tied to Windows, your use of that hardware is effectively contingent upon Microsoft’s permission. If, three years from now, your hard drive dies and requires replacement, there is no guarantee that Microsoft will allow you to reinstall and reactivate your OS, which means you can’t be sure that you can continue to use your radio. Imagine purchasing and “owning” a GM automobile where some third party retains the power, at any time, to suddenly decline your use of the engine control software. Who would buy a car under those terms?

    Don’t get me wrong. I’m not bashing Windows. There is a place for it in business, and home, and elsewhere. I’m simply saying that the Windows licensing model is antithetical to the open-source and cooperative-development ethos that has characterized the ham radio community since day one.

    Windows XP support ended in 2014. Vista will end in 2017. Windows 7 will end in 2020. Windows 8 in 2023. Windows 10 in 2025 (data from Microsoft’s web site.) That’s obsolescence on a 3 year cycle. My BC-348 is 80 years old and works fine. I repaired/refreshed it a few years ago (didn’t need anyone’s permission to “activate” any of the parts) and it will probably still be running long after I am gone.

    In my opinion, the only SDR transceiver worth investing in will at least allow for the option of running application code on Linux.

    1. Absolutely! There are many excellent SDR software packages for Linux including Pylink-SR, GNU Radio and Quisk. Currently our only ‘supported’ configuration is Windows running HDSDR with OmniRig for radio control. We are actively working on a ‘supported’ Linux configuration as well It will be based on Quisk we’ll base on something like a Raspberry PI so it will be our portable platform.

      In this case, the definition of ‘supported’ means that we provide step-by-step instructions on how to get up and running and will help you with any issues the arise.

      Jim.

    1. So you can remove it entirely! One of the important applications of this piece of RF hardware is stand-alone, no computer SDRing. There are many mature IQ processors just waiting for a decent RF front-end; the STM-32 SDR, Cumbria Designs, I even built one from a Teensy 3.2 with a Teensy Audio.

      https://www.youtube.com/watch?v=rtdGiA5sZSY

      In these instances, a USB based control system is not particularly useful. Most stand-alone applications will have enough GPIO to control the RS-HFIQ without any intervening processor.

      Another good reason is cost. We can get Arduino Nano clones for less than we could duplicate the circuitry on the board.

      Jim.

      1. Oh, ok. It looked in the picture like it was soldered in.

        It’s quite surprising that a nano clone costs less than just embedding the equivalent, though, unless the nano suppliers are cutting a TON of corners.

          1. Another good reason to put it in a socket! The only time I’ve fried one is powering it from the 12V input. THey seem pretty robust if you run them off the USB power as intended.

            The beat boards did have them soldered to the PCB but the Rev D boards for the Kickstarted and subsequent builds will all have bottom entry header receptacles on the bottom of the PCB so removing the Nano will be quick and painless.

            Jim.

Leave a Reply to Michael BlackCancel reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.