To deliver the next generation of satellite services, spacecraft operators are increasingly using high-throughput payloads with larger bandwidths at higher frequencies. Characterising transponder performance such as signal-to-noise ratio (SNR), spurious free dynamic range (SFDR), and flatness over hundreds of MHz or several GHz can be very difficult for OEMs and equally challenging for suppliers of test and measurement equipment. Arbitrary waveform transceivers could offer a solution.
Satellite manufacturers are looking for a measurement solution that will allow them to generate wideband, microwave carriers to test their transponders, as well as a broadband, RF receiver to characterise the performance of the payload transmitter. The architecture of a high-throughput payload is shown in Figure 1.
Figure 1 This diagram shows the architecture of a digital high-throughput payload.
Wide bandwidths are often split into multiple channels and dynamic range problems occur when non-linearities within the signal chain, e.g. amplifiers, ADCs, and DACs, generate intermodulation products between the input frequencies. These appear within other channels, causing distortion.
Noise power ratio (NPR) is a wideband test that measures the ‘quietness’ of an unused channel accounting for intermodulation-distortion products generated by non-linearities within the signal chain.
Traditionally, NPR characterisation of a payload was an analog process: Gaussian noise from a white source would be bandlimited using a low-pass anti-aliasing filter and then a band-stop filter was used to create an empty notch (channel) at the desired centre frequency, as illustrated in Figure 2. As the amplitude of the analog input noise is increased, there is a linear relationship between its power and the measured NPR. At some value, determined by the ADC’s hard-limiting behaviour, this begins to clip, creating intermodulation products that raise the quantisation noise floor rapidly reducing NPR.
Figure 2 Traditionally, NPR characterisation of a payload was an analog process. Source: Analog Devices
The problem with analog broadband signal generation is that many hardware filters are required to accommodate different notch widths and positions, and complete in-band characterisation becomes a slow and manual process. A quicker, automated, repeatable, and more controllable test procedure is needed, independent of mission frequencies, information bandwidths, and individual notch requirements.
With the advent of high-resolution, fast DACs, any arbitrary signal that can be described mathematically can be generated flexibly and accurately. A block diagram of an arbitrary waveform generator (AWG), also affectionately known as an ARB, is shown in Figure 3, where digital samples stored in a waveform memory are read and converted to analog at the desired frequency. The sampling rate determines the bandwidth and the maximum frequency that can be output, while the resolution of the DAC determines the dynamic range.
Figure 3 This is the block diagram of an AWG.
An AWG is a universal signal source and is first used to generate a series of single tones at different frequencies to measure the payload’s receiver in-band SNR, harmonic, and spurious performance. Continuous-wave (CW) characterisation allows OEMs to simultaneously differentiate between device-level artefacts and system issues, e.g. an ADC interleaving spur vs. noise coupling from the routing of the sampling clock, power supply, or poor grounding. Once the single-tone performance of the payload has been understood, its linearity and wideband operation can be characterised using stimuli such as multi-tone or NPR carriers to provide a measure of intermodulation distortion.
Following CW and wideband measurements, the complete payload is then tested using representative stimuli such as modulated carriers to verify operational performance. Unlike a conventional signal generator, an AWG can be programmed to output all of these waveform types.
AWGs have sophisticated sequencers for storing and playing waveform data as well as markers and triggers that can interface with the external environment. Memory segmentation allows seamless sequencing without gaps between the last and first samples of contiguous segments, and complex signal scenarios can be created comprising multiple sequences.
The output from a transponder is typically a modulated carrier and a receiver is required to acquire this, followed by analyses to generate a constellation diagram and calculate bit-error rate (BER). While testing one of Spacechips’ high-throughput payloads, I realised we were running out of lab space and needed a compact, single-box, multi-channel transceiver to generate carriers to test our transponder’s input, as well as a receiver to characterise the fidelity of its transmitter.
We didn’t have anything and after a search, I discovered the Proteus arbitrary waveform transceivers (AWT) from Tabor Electronics. I had never heard of an AWT and as an owner of an SME, I know how expensive test equipment can be and became very intrigued by the potential financial saving, specifications, and capabilities of the AWTs being offered. After a few phone calls, I managed to get my hands on a Model P9082D-AWT, which contains an AWG offering two 16-bit, 9 GSPS, 9 GHz (2nd Nyquist) DACs as well as a digital receiver containing two 12-bit, 2.7 GSPS 9 GHz ADCs.
The Proteus AWG can offer four modes of operation: direct, NCO, IQ, and streaming. The first replays the signal stored within the internal DDR4 memory while NCO mode outputs a sinewave. Each channel has an independent NCO to select a different frequency if required. IQ mode mixes the signal stored within the waveform memory with the NCO to generate an AM modulated carrier, while streaming mode bypasses internal memory outputting signal data directly from the controlling host PC.
In its desktop or benchtop form factor, the Proteus AWG can be configured to offer up to 12 phase-coherent outputs synchronised to the same sampling clock. In PXI format, the number of channels is unlimited with each having its own output stage to independently set amplitude and offset. There are two options for the output: differential DC coupled with an amplitude up to 1.3V peak-to-peak and bandwidth limited to the first Nyquist zone, and direct output, offering an AC coupled output up to 600 mV peak-to-peak optimised for dynamic range and linearity. Its usable bandwidth extends up to the second Nyquist zone, 9 GHz.
The internal DDR4 SDRAM is where the waveforms to be generated are stored, with the duration of the desired signal determined by the size of the memory. You can use this entirely to store a single waveform or split into smaller segments, with each containing different stimuli. Arbitrary mode generates the signals stored within the waveform memory one at a time while task mode allows you to output sequences of segments in a pre-defined order to define complex scenarios.
The Proteus AWG offers internal and external triggers to control when waveforms start and stop, and digital outputs synchronised to the analog channels, known as markers, are also available to clock or control external peripherals.
By definition, DACs are non-linear, generate harmonic distortion, and offer a dynamic range based on their resolution. For signal generation, this behaviour needs to be transparent or understood before testing high-throughput payloads. In terms of output purity, the datasheet of the P9082D-AWT specifies harmonic distortion from < -70 to < -50 dBc and SFDR from -85 to -70 dBc, dependant on the measurement bandwidth and output frequency. Phase noise at a 10 kHz offset ranges from -134 to -104 dBc/Hz as the output varies from 140 MHz to 4.5 GHz respectively.
For multi-tone outputs, Tabor offers a script that flattens and equalises individual amplitudes to compensate for the frequency response of the AWG, i.e. the DAC’s sinc roll-off and its analogue bandwidth, as well as external cables, amplifiers, etc. AWGs can control the phase of each tone to preset the carrier peak-to-average power ratio (PAPR) and crest factor to the desired value.
A range of AWTs are available offering 2, 4, 8 and 12 AWG outputs with internal DAC sampling rates ranging from 1.25 to 9 GSPS and memory sizes up to 16 G samples. The receiver comprises two 12-bit, 9 GHz, 2.7 GSPS ADCs, which can be interleaved to create one 5.4 GSPS digitiser to double the effective input bandwidth.
The Proteus AWTs are available in three different form factors as shown in Figure 4, with the benchtop version having a touch-screen display to directly control the instrument:
Figure 4 The Proteus AWTs come in benchtop, desktop, and PXI form factors. Source: Tabor Electronics
Depending on the form factor, the AWTs must be used with its embedded PC or an external host computer. The instrument can be controlled using its proprietary Wave Design Studio (WDS) GUI software, SCPI commands, or application-specific IVI drivers. WDS allows different types of signals to be created easily and intuitively, as well as the creation of complex waveform scenarios, as illustrated by the screen captures in Figure 5.
Figure 5 These WDS screen captures show sine, digital noise, chirp, and 16-QAM GUI-based signal generation. Source: Tabor Electronics
You can also command the arbitrary waveform transceivers using MATLAB, LabView, and Python. From the perspective of a manufacturer of satellite payloads, the Proteus AWT is a complete, single-box test solution replacing traditional RF signal generators and a spectrum analyser. The AWT offers the following functional measurements:
- AWG single-tone, wideband NPR, and modulated carrier generation for multi-channel payload receiver testing. The WDS software allows phase coherent stimuli to be created quickly and intuitively.
- AWT’s wideband ADCs to measure the performance of the payload transmitters. The data captured by the ADCs can be streamed to the controlling host PC for analyses or stored within the internal waveform memory.
- Complete payload back-to-back testing: AWG to generate input carriers followed by AWT’s RF ADCs to characterise the performance of the processed output from the transponder transmitters.
To share my experiences with you, I have made a short video of me using the Proteus AWT to test space-grade ADCs and DACs. Further information can be obtained from the manufacturer’s website.
Spacechips has just introduced a new test as a service offering where we test your EM-grade space electronics for you or allow you to remotely and securely access our lab equipment from anywhere in the world, so you can measure and characterize the performance of your satellite avionics.
We also advise companies on how to test their space electronics (from prototyping and qualification measurements to full in-orbit verification), can architect a test and measurement solution for your avionics, and teach a one-day training course on Testing Satellite Payloads. Our lab facilities include full RF, space-grade mixed-signal and FPGA functional, test and measurement capabilities. Contact me if you would like to learn more.
Until next month, how would you use arbitrary waveform transceivers for payload testing? The best answer will win a Courses for Rocket Scientists World Tour t-shirt. Congratulations to Julie from Hawthorne, USA, the first to answer the riddle from my previous post.
Dr. Rajan Bedi is the CEO and founder of Spacechips, which designs and builds a range of advanced, L to Ku-band, ultra-high-throughput on-board processors and transponders for telecommunication, Earth-observation, navigation, Internet, and M2M/IoT satellites. Spacechips’ Design Consultancy Services develop bespoke satellite and spacecraft sub-systems, as well as advising customers how to use and select the right components, and how to design, test, assemble, and manufacture space electronics. The company will be offering a three-day course on space electronics from May 17-19.