RHINO


The University of Cape Town in South Africa is developing an open source SDR project known as RHINO, which stands for Reconfigurable Hardware Interface for computiNg and radiO.

Reconfigurable Computing is an attractive computational technology, allowing an almost limitless set of possible solutions to any particular problem. It does this by providing an effective compromise between the performance of hardware and the flexibility of software.

Software Defined Radio is a discipline that has benefited greatly from this technology, allowing for complex, high speed digital signal processing algorithms to be implemented in cost effective field programmable gate array technology.

The Rhino Project is an Open Source effort at the University of Cape Town in South Africa to provide a hardware platform and software toolchain for SDR which is both easy to use, easy to learn and affordable to a broad audience. It is hoped that this effort will consolidate and enhance the teaching and research resources available for SDR.

Their hardware design is based on a Xilinx Spartan-6 FPGA teamed up with an ARM Processor running BORPH Linux. You can view the hardware schematic and project detail PDF on the RHINO project info page.

Via AllAboutOpenHardware.

Join the Conversation

13 Comments

  1. Oh my… I hope RHINO doesn’t turn into another over-priced white-elephant SDR project when it comes to hardware. We already have two that I know of off-hand:

    The first is TAPR’s High Performance Software Defined Radio (HPSDR), which has been dragging on forever as costs climb. HPSDR is here: http://www.tapr.org.

    The other is the Universal Software Radio Peripheral (USRP), which is available as hardware but it costs an arm and a leg and must be purchased from a for-profit firm called Ettus Research (www.ettus.com). It seems the USRP hardware is associated in some way with the GNURadio software project. But it should be noted however that you do not need to have a USRP to use GNURadio software with other hardware. So there’s some question in my mind about how strong the link is between Ettus Research and the GNURadio project.

    You can make a frequency agile direct conversion SDR with low power transmit (couple of Watts) for frequencies through around 50MHz for approximately $70 (e.g., Softrock). Connect that to a cheap PC with a half-way decent sound card (<$400), and for less than $500 you have an SDR that will perform as good as the over-priced dinosaurs noted above.

    Rgds, David

    1. True, but the Rhino IS a PC… Funcubes are also £100 compared to the Rhino’s projected cost of ~$1500.

  2. @Drone : For “around 50Mhz” you wouldn’t need a usrp anyway, so, what’s your problem ? The schematics are out there, the parts are known, if you think you can build it cheaper why don’t you do it ?

    1. I think Drone’s point is that you cannot build it cheaply. Alan Langman initially promised the Rhino hardware at US$250. Currently the hardware costs around US$2500 (not US$1500). Some of the parts are becoming obsolete and the CX4 standard, as far as network infrastructure goes, is dying out. There has been very little (read: virtually no) progressed on the promised software infrastructure. Bottom line is, it is cheaper (by US$1000) to buy a Xilinx development board than a Rhino and you can do more with the Xilinx board.

      See their presentation on Rhino here: http://www.ohwr.org/attachments/download/328/RHINO_CASPER_2010_Presentation.pdf

      1. “Lee Man”, I never promised anyone that the Rhino hardware would cost $250! And who ever is charging you $2500 for a Rhino is ripping you off! The University of Cape Town recently (March) obtained its last batch of 6 boards from Digicom for $1740 per board; which is not too far from the original guess of $1500 which excluded any manufacturer profits.

        Its is difficult to keep a open platform current and free; technology moves quickly and Rhino is rather old. But people like to give you old stuff, so it made sense to go CX4 – its tough funding a open hardware project!

        There has actually been great progress on the software. Rhino is supported by Borph/Migen/Mibuild and we have used it to build arbitrary waveform radars and a few other things.

        Although your argument is rather poor, your bottom line is correct; you are much better off buying something like the Zedboard.

      2. I never made any promise that Rhino would cost $250! Also, if you priced the board at $2500 you are being ripped off. The University of Cape Town recently (March) purchased 6 Rhino’s for $1749 each. Given that the manufacturer handles the complete manufacturing process and needs to make a profit, we were not too far off when we made our guess based on our initial prototyping cost.

        I am sorry that you believe that the Borph/Migen/Mibuild environment is “virtually no” progress. We built some pretty cool radar systems with the framework. Instead of critising the lack of progress, why don’t you just contribute?

        Although your argument is based on much hyperbole, your final conclusion is correct. Absolutely! you much better off buying a Zedboard than Rhino.

  3. Hi Dave,

    Unfortunately Dave, you are correct; the Rhino platform is not cheap. We found one manufacture in the US keen to provide them to people at a cost of around $2200; so its more expensive than USRP; however it does provide the flexibility of dual FMC-LPC; which allows you to tackle problems beyond simply classical SDR i.e. passive radar, “cognitive radar”, sideband receiver (1GSPS) for radio astronomy. You can also purchase FMC with complete configurable wireless front ends (check out LimeMicro) which allows you to explore SISO or even 2×2 MIMO systems, more than 2×2 if you couple multiple Rhino’s in a heterogenous network; the board includes IEEE1588 to allow synchronization across ethernet networks of multiple systems.. We will hope the Borph Linux (currently only on a very few systems) which provides a nice abstracted interface to work with code for the FPGA i.e. simply ./flash_led & to run your code. You can also interface to you registers on the FPGA using the /proc file system. The hardware is really the tip of the iceberg for the Rhino project. We hope in time to augment with some cool and different software tools.

    In addition its all FREE. You are free to make your own, modify the schematics or PCB, sell, learn or even not use it. The choice is yours.

    Best

    Alan

  4. An elephant is a mouse – designed by a committee. Judging by the pictures and schematics, this “radio” has the same problem – it is over-designed. The “affordable to a broad audience” standard has been ignored, forgotten, or someoene has perverse ideas of what “broadly affordable” really means. I would make the following suggestions:
    1. Eliminate the DDR3 Ram, and cut back on the DDR2 RAM.
    2. Toss away 1 ethernet, drop the other to 1G (or even 100M)
    3. Eliminate 1x FMC connector
    All this will help with the next step:
    4. Use a smaller (cheaper! less power!) FPGA chip,
    5. Drive it all with a Touch LCD panel. Self – contained cheapness !

    1. I agree with everything you suggest except the smaller FPGA. I think the FPGA is fine and the cost saving won’t be that much compared to other improvements like:

      1. Make the PCB smaller – there is a lot of open space. I understand this system is based on the ATX PCB layout spec, but it is a waste of money to have all the open PCB space. Dropping the other things suggested will get your size down even more and you can go for a MiniITX standard (for example) and get cheap enclosures etc.

      2. Get rid of the PSU management system. Who ever came up with this idea? You cannot hard power the board – you need to power it externally to program the PSU management system in order to switch on the board. Use a cheap sub-US$5 microcontroller for PSU monitoring etc but don’t make the power-up reliant on it.

      3. Try and don’t use FMC. The connectors are too expensive and the commercial FMC products are very expensive as well. Look for cheaper alternatives.

      1. Talking tech without really understanding the requirements for an application is rather pointless. But in answer to some of your questions;

        1) Yeh, a smaller board would have been nice; but challenging sometimes when you trying to route high speed signals without the luxury of expensive 3D em simulation packages to verify your routes.
        2) Board powers off a 12V supply without the need to program any management system. Do you have a board? For the applications we were aiming for, power supply control and monitoring is essential.
        3) You can’t discount FMC on the cost of a connector! Have you read the Vita 57.1 specification?

Leave a comment

Your email address will not be published. Required fields are marked *

Notify me of followup comments via e-mail. You can also subscribe without commenting.