2018 International Conference on radar, Antenna, Microwave, Electronics, and Telecommunications Software Defined Radio Applications for Mini GSM BTS and Spectrum Analyzer with BladeRF Ihan Martoyo Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia [email protected] Herman Y. Kanalebe Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia [email protected] Andi Coandi Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia Dave Pratignyo Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia Henri P. Uranus Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia [email protected] Marincan Pardede Department of Electrical Engineering Universitas Pelita Harapan Tangerang, Indonesia [email protected] available bandwidth, but to tune electronically to a certain bandwidth of interest for a particular system [5]. Abstract—Software Defined Radio (SDR) technology enables the flexibility of a programmable hardware platform for radio applications. A wideband SDR can be programmed to function with various radio systems: FM radio, GSM (2G), 3G or WiFi systems. In this paper, the BladeRF SDR with the frequency range of 300 MHz – 3.8 GHz and full-duplex transmission capability will be set up as a GSM BTS (base transceiver station). The BladeRF is working with the YateBTS software for the BTS operation. Raspberry Pi is utilized as the processor to provide further portability. A GSM repeater that is connected to the BladeRF was used to amplify the signal up to 41 dB, and can increase the coverage range to about 70 m. Although the reliability of voice and SMS communication is only about 50% and 85% respectively, the BladeRF can still provide an ad-hoc alternative communication system in time of emergency or in remote areas. Several kinds of SDR are already available: USRP, HackRF One and BladeRF. Table I displays the basic parameters of these SDR systems. USRP, the most expensive of the three, can have a bandwidth up to 61.44 MHz (depending on the specific model), has the highest sample rate of up to 128 Msps, and is operating in full-duplex. The HackRF One, which provides a cheaper alternative, is operating only in half-duplex, and has a bandwidth of 20 MHz (frequency up to 6 GHz). For the work in this paper, the BladeRF is chosen because it can already function in fullduplex, thus can be operated as a GSM BTS, but not as expensive as the USRP. This paper reports the setting up and characterization of a GSM BTS using the BladeRF SDR. The next section will describe the system, followed by measurement results and discussions in Section III. Section IV closes with some conclusions. Keywords— SDR, BladeRF, Radio Characterization, GSM BTS I. INTRODUCTION The term software radio was coined by Joseph Mitola, at first for military radio communication systems [1]-[2]. Since software radio is based on software, the flexibility for a multimode, multisystem, multiband radio communication should come as easy as downloading and installing the appropriate software – if needed – by over-the-air download [3]. Even if a military troop is deployed in an unknown area with unknown availability of communication systems, the Software Radio should be able to be reprogrammed for the need of that situation. II. SDR WITH BLADERF SYSTEM A. BladeRF Transmit/Receive with MATLAB BladeRF can be connected to MATLAB installed in a Laptop for transmit/receive testing as shown in Fig. 1. MATLAB works with the BladeRF software for windows from Nuand [6]. It is important to note the version of the suitable firmware for the BladeRF. For this work, the updated firmware version 2.0.0 is used [7]. Some basic commands available in the m-script bladeRF.m include: b.rx.frequency and b.tx.frequency to set the frequency parameter, b.rx.start and b.tx.start to start receiving and transmitting, and b.rx.stop and b.tx.stop to stop receiving and transmitting [8]. That was the initial dream. Soon, however, many realized the technological limit of the concept. An analog-to-digital converter (ADC) must be put right after the antenna so that all incoming signals can be converted to digital, fitting for further digital processing with software. However, such an ADC that can convert a signal with a very high bandwidth to cover various radio systems is prohibitively very expensive. TABLE I. SDR TYPES AND PARAMETERS SDR Then Software Defined Radio (SDR) came as a compromised version for the pure Software Radio. An SDR is hardware, thus not very flexible; but the FPGA-based hardware can be controlled or reprogrammed by software, therefore providing a certain amount of flexibility [4]. If the SDR were to cover the frequency band from 800 MHz to 5.5 GHz, an ADC of 12 bit and 11GS/s would be needed [5]. That would be impossible today and for the foreseeable future. The key for the front-end is not to receive the whole 978-1-5386-6519-0/18/$31.00 ©2018 IEEE Frequency Bandwidth Duplex Sample Rate 108 HackRF One 30 MHz – 6 GHz BladeRF 300 MHz – 3.8 GHz 20 MHz 28 MHz Half Full 20 Msps 40 Msps USRP 50 MHz – 2.2 GHz / 6 GHz 16 MHz / 61.44 MHz Full / 2x2 MIMO 64 Msps / 128 Msps III. MEASUREMENT RESULTS AND SDR APPLICATIONS A. BladeRF Transmit/Receive with MATLAB BladeRF can function as a spectrum analyzer or a signal generator with Matlab. Fig. 4 and Fig. 5 show a measurement comparison between BladeRF and a spectrum analyzer at 375 MHz. We see in the measurement with BladeRF some visible spurious frequencies, for example, around 384 MHz (Fig. 4) that cannot be seen with the spectrum analyzer (Fig. 5). These spurious frequencies could be an artifact of the down-mixing strategy of the SDR, e.g., artifact in the digitizing of the IF signals due to its periodic nature [11]. The SDR also always displays the center frequency of the measurement (DC Spike). Thus, using the BladeRF as an affordable spectrum analyzer, we should be aware of the possible appearance of the spurious frequencies. It is perhaps good to shift the center frequency and adjust the span while measuring frequencies with the BladeRF to shake-off any spurious frequencies that might be visible at certain settings. Fig. 6 shows another view on the frequencies that include the signal at 375 MHz. Here, even more spurious frequencies are visible. Fig. 1. BladeRF connected to MATLAB for Tx/Rx testing. B. BladeRF with Raspberry Pi, GSM Repeater and YateBTS To function as a mini GSM BTS, the BladeRF can be connected to Raspberry Pi 3 as a single board processor, which runs on Linux Jessie Debian as the operating system [9]. YateBTS can then be installed on the Raspberry Pi. The NiPC (Network in a PC) performs the function of a regular GSM BTS for YateBTS [10]. Fig. 2 and Fig. 3 show the BladeRF connected to Raspberry Pi and a GSM repeater. Apache, PHP, and the necessary libraries for USB interface and GSM libraries should also be installed in the Raspberry Pi, so that the GSM BTS system can be accessed and controlled via a web-interface. Fig. 7 and Fig. 8 compare two measurements of a signal at 400 MHz with the BladeRF. In Fig. 7, the reading from the BladeRF does not show any significant spurious frequencies; however, Fig. 8 shows the spurious frequency at 384 MHz again (compare with the clean spectrum of Fig. 5 and Fig. 9 with a spectrum analyzer). Fig. 2. BladeRF (top) connected to Raspberry Pi (middle) and GSM repeater (bottom). Fig. 4. BladeRF measurement of a signal at 375 MHz. Fig. 3. GSM mini BTS block diagram with BladeRF. Fig. 5. Spectrum analyzer measurement of a signal at 375 MHz. 109 Fig. 9. Spectrum Analyzer measurement on the signal at 400 MHz. Fig. 6. Another angle of BladeRF on the signal at 375 MHz. calls were heavily delayed, some calls got connected after the caller hung up, some calls got connected without any sound going through, and some calls did not get listed in the online subscriber list in YateBTS. However, 50% of the calls went through successfully. It was however difficult to make a second call while one voice connection was still running through the BTS. The limitation can come from the rather small computing capability of the Raspberry Pi, or some constrains with the YateBTS system. The SMS (Short Message Service) was tested by sending groups of 100 messages, each saying “Spinx of black quartz, judge my vow,” which was chosen because it consists all the alphabets. If the 100 messages were sent at once, the number of perfectly arriving messages was around 20% and could be as low as 7%. If the messages were sent in 2 groups of 50 messages, the success rate could increase to around 40%. Sending the messages in groups of 20 messages or lower, increased the success rate to more than 85%. Fig. 7. BladeRF measurement on a signal at 400 MHz. C. Alternative Applications with BladeRF GSM BTS The transmit signal of the BladeRF was also tested and measured. Fig. 10 compares the measurement with cable connector (top) and through antenna (bottom). The signals through cable connector were about 20 dB higher than through the antenna. The frequency response seems to show that the BladeRF is optimized for the GSM systems in 800900 MHz and around 1800 MHz. Thus, the BladeRF system can be used as a cheap RF signal generator (although an RF wideband amplifier is needed to operate as a proper and adjustable signal generator) and spectrum analyzer for education and research. Fig. 8. Another angle of BladeRF on the signal at 400 MHz. B. BladeRF with Raspberry Pi, GSM Repeater and YateBTS BladeRF is also setup as a GSM BTS on Raspberry Pi 3 as a single board processor running Linux Jessie Debian for the OS and YateBTS for GSM BTS functionalities. The addition of a GSM repeater can increase the signal of bladeRF up to 41 dB (uplink 890-915 MHz; downlink 935960 MHz) and increase the coverage of the BTS from about several meters to about 70m. The voice communication with GSM phones through the BladeRF BTS was tested with groups of 10 attempts of call each time. The success of attempts was about 50%. Some Fig. 10. Frequency response of BladeRF transmit signal 110 Although the BladeRF does not seem to support a very high reliability in voice and SMS communications, the system could be applied for alternative communication systems in remote areas or emergency settings. The availability of cheap GSM handphones and the affordability of BladeRF are important promising factors. BladeRF can be set up with Matlab to function as an affordable spectrum analyzer. Spectrum measurements with BladeRF might show some spurious frequencies as artifacts of the internal frequency mixing or sampling. Shifting the center frequency and span might give more clarity to the spectrum being measured. It is also possible to envision other alternative applications of the GSM BTS system that do not require high reliability. An example for such application would be to use the GSM BTS for free advertisement purposes. Fig. 11 shows some modified SMS welcome messages that can be used for free SMS advertisement. GSM handphones will be attracted automatically to a strong nearby BTS signals, so passing-by handsets might get connected automatically to the standalone BladeRF BTS. However, care should be taken to release the handsets again, so that they can regain communication with the GSM network. BladeRF has been set-up for a standalone Mini GSM BTS successfully. To increase portability, a Raspberry Pi was used as a single board processor to work with the BladeRF. A GSM repeater was used to increase the signal of BladeRF to about 41 dB and the coverage of the GSM BTS to about 70 m. The success rate of voice communication is around 50% and for SMS can be as high as 85%. The spectrum measurement capability of BladeRF seems to be limited by some artifacts of internal mixing and sampling effects. Despite the shortcomings, the BladeRF can be used as an affordable spectrum analyzer for education or even research, considering that a conventional spectrum analyzer could cost up to ten times the BladeRF. Frequency response of the transmit signal of BladeRF seems to be optimized for GSM systems at 800-900 MHz and around 1800 MHz. The strength of SDR is its reconfigurability and multimode operation. Thus, the BladeRF can be reconfigured rapidly to function and communicate with several communication systems: Handy Talky systems, GSM, even detecting car key transmissions. Such capability may open other potential applications besides the traditional GSM voice and SMS communications. Although the reliability of the GSM BTS with BladeRF is rather low for voice & SMS communication, the system might be enough for a standalone alternative communication system in remote areas or emergency settings. Other alternative application, such as free SMS advertisement system, can also be explored and applied. IV. CONCLUSIONS After the text edit has been completed, the paper is ready for the template. Duplicate the template file by using the Save As command, and use the naming convention prescribed by your conference for the name of your paper. In this newly created file, highlight all of the contents and import your prepared text file. You are now ready to style your paper; use the scroll down window on the left of the MS Word Formatting toolbar. ACKNOWLEDGMENT This work is supported by the Indonesian Ministry of Research and Higher Education no. 021/KM/PNT/2018, March 6, 2018; Kontrak Penelitian Dasar Unggulan Perguruan Tinggi no. 149/LPPM-UPH/IV/2018. REFERENCES [1] J. Mitola, “The Software Radio Architecture,” IEEE Communications Magazine, vol. 33, no. 5, pp. 26-38, 1995. [2] J. Mitola & G.Q. Maguire, “Cognitive Radio: Making Software Radios More Personal,” IEEE Personal Communications, vol. 6, no. 4, pp. 13-18, 1999. [3] E. Buracchini, “The Software Radio Concept,” IEEE Communications Magazine, vol. 38, no. 9, pp. 138-143, 2000. [4] T. Hentschel, M. Henker, G. Fettweis, “The Digital Front-end of Software Radio Terminals,” IEEE Personal Communications, vol. 6, no. 4, pp. 40-46, 1999. [5] A. A. Abidi, “The Path to the Software-Defined Radio Receiver,” IEEE Journal of Solid-State Circuits, vol. 42, no. 5, pp. 954-966, 2007. [6] “Windows Installer,” Nuand. Internet: http://nuand.com/support.php [Jul. 23, 2018]. [7] “BladeRF Windows Install Guide,” Nuand. Internet: http://www.nuand.com/bladeRFdoc/guides/bladeRF_windows_installer.pdf, Jun. 29, 2016. [Jul. 23, 2018]. [8] “BladeRF.m” Nuand. Internet: https://github.com/Nuand/bladeRF/blob/master/host/libraries/libblade RF_bindings/matlab/bladeRF.m Jun. 30, 2016. [Jul. 23, 2018]. [9] “Raspbian Download,” Raspberry Pi. Internet: https://www.raspberrypi.org/downloads/ [Jul. 23, 2018]. [10] “Network in a PC,” YateBTS Wiki. Internet: https://wiki.yatebts.com/index.php/Network_in_a_PC Oct. 10, 2017. [Jul. 23, 2018]. [11] M. D. McKinley et al., “Eliminating FFT artifacts in vector signal analyzer spectra,” Microwave Journal, vol. 49, no. 10, 2006, p. 156. Fig. 11. Modified SMS messages for free advertisement purposes 111