Subversion Repositories svnkaklik

Rev

Rev 1135 | Rev 1137 | Go to most recent revision | Details | Compare with Previous | Last modification | View Log

Rev Author Line No. Line
1133 hlavac 1
\chap Trial version of the receiver, design and implementation
1073 kaklik 2
 
1133 hlavac 3
The whole design of radioastronomy receiver digitalization unit is constructed to be used in a wide range of applications and tasks related to digitalization of signal from radioastronomy receivers. A good illustrating problem for its use is a signal digitalisation from multiple antenna arrays.
1083 kaklik 4
 
1132 kaklik 5
 
6
\midinsert
7
\clabel[expected-block-schematic]{Expected system block schematic}
8
\picw=\pdfpagewidth \setbox0=\hbox{\inspic ./img/Coherent_UHF_SDR_receiver.png }
9
\par\nobreak \vskip\wd0 \vskip-\ht0
10
\centerline {\kern\ht0 \pdfsave\pdfrotate{90}\rlap{\box0}\pdfrestore}
11
\caption/f Expected realisation of signal digitalisation unit.
12
\endinsert
13
 
1073 kaklik 14
\sec Required parameters
15
 
1134 fluktuacia 16
We require the following technical parameters, to supersede existing digitalization units solutions.
1135 kaklik 17
Primarily, we need wide a dynamical range and high IP3. \glos{IP3}{Third-order intercept point} The receiver must accept wide dynamic signals because a typical radioastronomical signal has a form of a weak signal covered by a strong man-made noise or other undesired noises as lighting, Sun emissions etc.
1073 kaklik 18
 
1133 hlavac 19
Summary of other additional required parameters follows
1098 kaklik 20
 
1083 kaklik 21
\begitems
1134 fluktuacia 22
  * Dynamical range better than 80 dB, see section \ref[dynamic-range-theory] for explanation
1133 hlavac 23
  * Phase stability between channels
1125 kaklik 24
  * Low noise (all types)
1083 kaklik 25
  * Sampling jitter better than 100 metres
1134 fluktuacia 26
  * Support for any number of receivers in the range of 1 to 8
1083 kaklik 27
\enditems
28
 
1133 hlavac 29
Now we analyze several of the parameters in detail.
1125 kaklik 30
 
1098 kaklik 31
\sec Sampling frequency
1133 hlavac 32
 
1135 kaklik 33
Sampling frequency is not limited by the technical constrains in the trial version. This parameter is especially limited by the sampling frequencies of analog-to-digital conversion chips available on the market and interface bandwidth. Combination of the required parameters -- dynamic range requiring at least 16bit and a minimum sampling frequency of 1$\ $MSPS \glos{MSPS}{Mega-Samples Per Second} leads to the need of high end ADC chips which does not support such low sampling frequencies at all. Their minimum sampling frequency is 5$\ $MSPS.
1083 kaklik 34
 
1135 kaklik 35
We calculated a minimum data bandwidth data rate for eight receivers, 2 bytes per sample and 5$\ $MSPS as $8 \cdot 2 \cdot 5\cdot 10^6 = 80\ $MB/s. Such data rate is at the limit of the actual writing speed of classical HDD \glos{HDD}{Hard disk drive} and it is almost double the real bandwidth of USB 2.0 \glos{USB 2.0}{Universal Serial Bus version 2.0}  interface. As a result of these facts we must use faster interface. Faster interface is especially needed in cases where we require faster sampling rates than ADC's minimal 5$\ $MSPS sample rate.
36
The most perspective interface for use in our type of application is USB 3.0 or PCI Express interface. However, USB 3.0 is a relatively new technology without good development tools currently available. We have used PCI Express \glos{PCI Express}{Peripheral Component Interconnect Express}  interface as the simplest and the most reliable solution.
1133 hlavac 37
 
1075 kaklik 38
\sec System scalability
1073 kaklik 39
 
1134 fluktuacia 40
For analogue channels' scalability, special parameters of ADC modules are required. Ideally, there should be a separate output for each analogue channel in ADC module. ADC module must also have separate outputs for frames and data output clocks. These parameters allow for conduction at relatively low digital data rates. As a result, the digital signal can be conducted even through long wires. Modular concept allows a separation from central logical unit which supports optimization of number analogue channels.
1073 kaklik 41
 
1135 kaklik 42
Clock and data signals will be then handled distinctively in our modular scalable design. Selected ADC chips are guaranteed to have defined clock skew between the sampling and data output clocks. This allows taking data and frame clocks from the first ADC module only. The rest of the data and frame clocks from other ADC modules can be measured for diagnostic purposes (failure detection, jitter measurement etc.), but these redundant signals are not used for data sampling. If more robustness is required in the final application, DCO \glos{DCO}{Data Clock Output} and FR signals may be collected from other modules and routed through an voting logic which will correct possible signal defects.
1075 kaklik 43
 
1133 hlavac 44
This system concept allows for scalability, that is technically limited by a number of differential signals on host side and its computational power.  There is another advantage of scalable data acquisition system -- an economic one. Observatories or end users can make a choice of how much money are they willing to spent on radioastronomy receiver system. This freedom of choice is especially useful for science sites without previous experience in radioastronomy observations.
1075 kaklik 45
 
1134 fluktuacia 46
\secc Differential signalling
1075 kaklik 47
 
1135 kaklik 48
The above mentioned concept of scalable design requires relatively long circuit traces between ADC and digital unit which captures the data and performs the computations. The long distance between the digital processing unit and the analog-to-digital conversion unit has an advantage in noise retention typically produced by digital circuits. Those digital circuits, such as FPGA \glos{FPGA}{Field-programmable gate array}, Ethernet or other flip-flops blocks and circuit traces, usually work at high frequencies and emit wide-band noise with relatively low power. In such cases any increase in a distance between the noise source and analog signal source increase S/N significantly. However, at the same time, a long distance brings problems with the digital signal transmission between ADC and computational unit. But this obstacle should be resolved more easily in free-space than on board routing. The high-quality differential signalling shielded cables should be used, such as massively produced and cheap SATA \glos{SATA}{Serial ATA} \glos{ATA}{AT Attachment} cables. This technology has two advantages over PCB \glos{PCB}{printed circuit board} signal routing. First, it can use twisted pair of wires for leak inductance suppression in signal path and second, the twisted pair may additionally be shielded by uninterrupted metal foil.
1075 kaklik 49
 
1076 kaklik 50
\secc Phase matching
51
 
1134 fluktuacia 52
For multiple antenna radioastronomy projects, system phase stability is a mandatory condition. It allows a precise, high resolution imaging of objects, increases signal to noise ratios in several observation methods and allows the use of advanced algorithms for signal processing.
1076 kaklik 53
 
1135 kaklik 54
High phase stability in our scalable design is achieved through centralized frequency generation  and distribution with multi-output LVPECL \glos{LVPECL}{Low Voltage Emitter-coupled logic} hubs (CLKHUB02A), that have equiphased outputs for multiple devices. LVPECL logic is used on every system critical clock signal distribution hub. LVPECL logic has an advantage over LVDS \glos{LVDS}{Low-voltage differential signaling} in signal integrity robustness. LVPECL uses higher logical levels and higher signalling currents. Power consumption of LVPECL logic is nearly constant over the operating frequency range due to the use of bipolar transistors. This minimizes voltage glitches which are typical for CMOS \glos{CMOS}{Complementary metal–oxide–semiconductor } logic. One drawback of its parameters is a high power consumption of LVPECL logic which easily reaches tens of milliamperes per device.
1076 kaklik 55
 
1133 hlavac 56
This design ensures that all system devices have access to the defined phase and known frequency.
1076 kaklik 57
 
1073 kaklik 58
\sec System description
59
 
1134 fluktuacia 60
This section deals with the description of the trial version based on Xilinx ML605 development board \ref[ML605-development-board]. The board had been used in a previous project and has not been used since then, but the FPGA parameters are more than sufficient of what we need for fast data acquisition system.
1073 kaklik 61
 
1133 hlavac 62
\secc Frequency synthesis
1075 kaklik 63
 
1135 kaklik 64
We have used a centralized topology as a basis for frequency synthesis. One precise high-frequency and low-jitter digital oscillator has been used \cite[MLAB-GPSDO], while other working frequencies have been derived from it by the division of its signal. This central oscillator has a software defined GPS \glos{GPS}{Global Positioning System}  disciplined control loop for frequency stabilization.\fnote{SDGPSDO design has been developed in parallel to this diploma thesis as a related project, but it is not explicitly required by the thesis itself and thus it is described in a separate document}
1134 fluktuacia 65
We have used new methods of software frequency monitoring and compensation in order to meet modern requirements on radioastronomy equipment which need precise frequency and phase stability over a wide baseline scales for effective radioastronomy imaging.
1133 hlavac 66
 
1135 kaklik 67
The GPSDO device consists of Si570 chip with LVPECL output. Phase jitter of GPSDO \glos{GPSDO}{GPS disciplined oscillator} is determined mainly by Si570 phase noise. Parameters of the Si570 are summarized in the following table \ref[LO-noise] (source \cite[si570-chip] ).
1112 kaklik 68
 
1134 fluktuacia 69
The GPSDO design, that is included in data acquisition system, has special feature -- it generates time marks for a precise time-stamping of the received signal. Timestamps are created by disabling the local oscillator's outputs, connected to SDRX01B receivers, for 100 us.  As result, a rectangular click in the ADC input signal is created which appears as a horizontal line in spectrogram.
70
Timestamps should be seen in image \ref[meteor-reflection] (above and below the meteor reflection).
1112 kaklik 71
 
1134 fluktuacia 72
Time-marking should be improved in future by digitalization of GPS signal received by antenna on observational station. Following that, the GPS signal can be directly sampled by a dedicated receiver and one separate ADC module. Datafile then consists of samples from channels of radio-astronomy receivers along with the GPS signal containing precise time information.
1126 kaklik 73
 
74
 
1134 fluktuacia 75
\midinsert \clabel[LO-noise]{Phase noise of the local oscillator}
1113 kaklik 76
\ctable{lcc}{
1114 kaklik 77
	&	 \multispan2 \hfil Phase Noise [dBc/Hz] \hfil 		\cr
78
Offset Frequency	&	$F_{out}$ 156.25 MHz	& $F_{out}$ 622.08 MHz \cr
1113 kaklik 79
100 [Hz]	&	–105	&	–97 \cr
80
1 [kHz]	&	–122	&	–107 \cr
81
10 [kHz]	&	–128	&	–116 \cr
82
100 [kHz]	&	–135	&	–121 \cr
83
1 [MHz]	&	–144	&	–134 \cr
84
10 [MHz]	&	–147	&	–146 \cr
85
100 [MHz]	&	n/a	&	–148 \cr
86
}
1134 fluktuacia 87
\caption/t Phase noise of the used Silicon Laboratories Si570 chip. Offset frequency is measured from carrier frequency. Values shown in the table are given for two different carrier frequencies.
1113 kaklik 88
\endinsert
1112 kaklik 89
 
1134 fluktuacia 90
Every ADC module will be directly connected to CLKHUB02A module which takes sampling clock signal delivered by FPGA from the main local oscillator.  This signal should use high quality differential signalling cable -- we should use SATA cable for this purpose. FPGA may slightly affect the clock signal quality by adding a noise, but it has a negligible effect on the application where developed system will be used.
1112 kaklik 91
 
1098 kaklik 92
 
1133 hlavac 93
\secc Signal cable connectors
1073 kaklik 94
 
1126 kaklik 95
\label[signal-cables]
96
 
1133 hlavac 97
Several widely used and commercially easily accessible differential connectors were considered to be use in our design.
1073 kaklik 98
 
99
\begitems
1077 kaklik 100
* HDMI % [[http://en.wikipedia.org/wiki/Hdmi|HDMI]]</del>
101
* SATA  		%{http://en.wikipedia.org/wiki/Serial_attached_SCSI#Connectors|SAS]]/[[http://en.wikipedia.org/wiki/Serial_ATA|SATA]]
102
* DisplayPort 		%[[http://en.wikipedia.org/wiki/Display_port|DisplayPort]]</del>
103
* SAS/miniSAS
1073 kaklik 104
\enditems
105
 
1134 fluktuacia 106
At the end, MiniSAS connector was chosen as the best option to be used in connecting together multiple ADC modules. A transition between SATA and miniSAS is achieved by SAS to SATA adapter cable which is commonly used in servers to connect SAS controller to multiple SATA hard disc in RAID systems and thus is commercially easily available. It is compatible with existing SATA cabling systems and aggregates multiple SATA cables to a single connector. It also has SPI configuration lines which can be seen in the following picture \ref[img-miniSAS-cable] as standard pinheader connector.
107
The main drawback of miniSAS PCB connectors lies in the fact, that they are manufactured in SMT versions only. SMT design may eventually decrease the durability of the connector even if outer metal housing of connector is designed to be mounted using a standard through-hole mounting method.
1073 kaklik 108
 
1085 kaklik 109
\midinsert
1086 kaklik 110
\clabel[img-miniSAS-cable]{Used miniSAS cable}
1118 kaklik 111
\picw=5cm \cinspic ./img/miniSAS_SATA_cable.jpg
112
\caption/f An example of miniSAS cable similar to used.
1085 kaklik 113
\endinsert
114
 
1086 kaklik 115
\secc Signal integrity requirements
1126 kaklik 116
 
1105 kaklik 117
\label[diff-signaling]
1085 kaklik 118
 
1134 fluktuacia 119
We use ADC devices that have DATA clock frequency eight times higher than sampling frequency in a single line output mode, implying a 40 MHz output bit rate. This implies a $ 1/4 \cdot 10^7 = 25\ $ns time length of data bit, which is equivalent to 7.5m light path in free space. If we use copper PCB with FR4 substrate layer or coaxial/twinax cable, we could obtain velocity factor of 0.66 at worst condition. Then the light path for the same bit rate $t_s$ will be 4.95 m. Although we do not have any cables in the system with comparable lengths, the worst data bit skew described by data sheets of the used components is $0.3 \cdot t_s$, which is 1.485 m. Therefore the length matching is not critical in our current design operating on lowest sampling speed. Length matching may become critical in future versions with higher sampling rates, where the cable length must be matched. However SATA cabling technology is already prepared for that case and matched SATA cables are a standard merchandise.
1086 kaklik 120
 
1093 fluktuacia 121
\secc ADC modules design
1073 kaklik 122
 
1124 kaklik 123
\midinsert
1126 kaklik 124
\clabel[adcdual-preview]{Preview of designed ADCdual PCB}
1132 kaklik 125
\picw=10cm \cinspic ./img/ADCdual01A_Top_Big.JPG
126
\picw=10cm \cinspic ./img/ADCdual01A_Bottom_Big.JPG
1134 fluktuacia 127
\caption/f Realised PCB of ADCdual01A modules. Differential pairs routings are clearly visible.
1124 kaklik 128
\endinsert
1083 kaklik 129
 
130
\secc ADC selection
131
 
1134 fluktuacia 132
There exist several standard ADC signalling formats currently used in communication with FPGA.
1083 kaklik 133
 
134
\begitems
135
  * DDR LVDS
136
  * JEDEC 204B
137
  * JESD204A
138
  * Paralel LVDS
139
  * Serdes
140
  * serial LVDS
141
\enditems
142
 
1134 fluktuacia 143
As a result of our need to use the smallest number of cables possible, the choice fell on the serial LVDS format. Small number of differential pairs is an important parameter determining the construction complexity and reliability\cite[serial-lvds]. No many currently existing ADC devices have this kind of digital interface. An ultrasound AFE device chips seem to be ideal for this purpose -- the chip has integrated both front-end amplifiers and filters. It has a drawback though - it is incapable of handling differential input signal and has a relatively low dynamic range (as it consists only of 12bit ADC) and has many single ended ADC channels. Consequently, the scaling is possible only by a factor of 4 receivers (making 8 analogue single ended channels).
1083 kaklik 144
 
1134 fluktuacia 145
If we add a requirement of separate output for every analogue channel and a 16bit depth we find that there are only a few 2-Channel simultaneous sampling ADCs currently existing which meet these criteria. We have summarized those ADCs in the following table \ref[ADC-types]
1083 kaklik 146
 
1133 hlavac 147
\midinsert
1126 kaklik 148
\typosize[9/11] \def\tabiteml{ }\let\tabitemr=\tabiteml
149
\clabel[ADC-types]{Available ADC types}
1116 kaklik 150
\ctable{lccccccc}{
1086 kaklik 151
\hfil ADC Type & LTC2271 & LTC2190 & LTC2191 & LTC2192 & LTC2193 & LTC2194 & LTC2195 \cr
1126 kaklik 152
SNR [dB] & 84.1 & 77 & 77 & 77 & 76.8 & 76.8 & 76.8  \cr
153
SFDR [dB] & 99 & 90 & 90 & 90 & 90 & 90 & 90  \cr
154
S/H Bandwidth [MHz] & 200 & \multispan6 550 \strut \cr
155
Sampling rate [MSPS] & 20 & 25 & 40 & 65 & 80 &  105 & 125  \cr
156
Configuration & \multispan7 SPI \strut \cr
157
Package & \multispan7 \hfil 52-Lead (7mm $\times$ 8mm) QFN \hfil \strut \cr
1086 kaklik 158
}
1134 fluktuacia 159
\caption/t The summary of the currently available ADC types and theirs characteristics.
1086 kaklik 160
\endinsert
1083 kaklik 161
 
1134 fluktuacia 162
All parts in this category are compatible with one board layout. The main differences lay in the sampling frequency and signal to noise ratio, with the slowest having a maximum sampling frequency of 20 MHz. However, all of them have a minimal sampling frequency of 5 MSPS and all are configurable over a serial interface (SPI). SPI seems to be a standard interface used in high-end ADC chips made by the largest manufacturers (Analog Devices, Linear technology, Texas instruments, Maxim integrated..).  We have selected two slowest types for our evaluation design. Following that, a PCB for this part have been designed.
163
We have decided that ADCdual01A modules will have a standard MLAB construction layout with four mounting holes in corners aligned in defined raster of 400 mils.
1083 kaklik 164
 
1134 fluktuacia 165
Data serial data outputs of ADC modules should be connected directly by LVDS signalling levels conducted by SATA cables to FPGAs for the basic primary signal processing. The ADC chips used in the modules have a selectable bit width of data output bus and thus the output SATA connectors have signals arranged to contain a single bit from every ADC channel.  This creates a signal concept enabling a selection of proper bus bit-width according to the sampling rate (higher bus bit-width downgrades signalling speed and vice versa.)
1105 kaklik 166
 
1126 kaklik 167
In order to connect the above mentioned signalling layout, miniSAS to multiple SATA cable should be used as described in section \ref[signal-cables].
1105 kaklik 168
 
1134 fluktuacia 169
A KiCAD design suite had been chosen for PCB layout. However, the version, despite having integrated CERN Push \& Shove routing capability, is slightly unstable as it sometimes crushes due to an exception during routing. On the basis of these stability issues, the design had to be saved quite often. On the other hand, compared to commercially available solutions, such as MentorGraphics PADS or Cadence Orcad,  the Open-source KiCAD provides an acceptable option and it easily surpasses a widely used Eagle software.
1105 kaklik 170
 
1133 hlavac 171
As a part of work on the thesis, new PCB footprints for FMC, SATA, ADCs a and miniSAS connectors have been designed and were committed to KiCAD github library repository. They are now publicly available on the official KiCAD repository at GitHub.
1105 kaklik 172
 
1126 kaklik 173
ADCdual01A module has several digital data output formats. Difference between these modes lays in the number of differential pairs used.
1105 kaklik 174
 
175
\begitems
176
    * 1-lane mode
1114 kaklik 177
    * 2-lane mode
178
    * 4-lane mode
1105 kaklik 179
\enditems
180
 
1133 hlavac 181
All of the above-mentioned modes are supported by the module design. For the discussed data acquisition system, the 1-lane mode was selected. 1-lane mode allows a minimal number of differential pairs between ADCdual01A and FPGA. Digital signalling scheme used in 1-lane mode is shown in the following image \ref[1-line-out].
1105 kaklik 182
 
183
\midinsert
184
\clabel[1-line-out]{Single line ADC output signals}
185
\picw=15cm \cinspic ./img/ADC_single_line_output.png
1122 fluktuacia 186
\caption/f Digital signalling schema for 1-line ADC digital output mode.
1105 kaklik 187
\endinsert
188
 
1134 fluktuacia 189
ADCdual01A parameters can be set either by jumper setup (referred to as a parallel programming  in the device's data sheet) or by SPI interface. SPI interface has been chosen for our system, because of the parallel programming's lack of options (test pattern output setup for example).
1105 kaklik 190
 
1133 hlavac 191
Complete schematic diagram of ADCdual01A module board is included in the appendix.
1105 kaklik 192
 
193
 
1073 kaklik 194
\secc ADC modules interface
195
 
1124 kaklik 196
\midinsert
1132 kaklik 197
\picw=10cm \cinspic ./img/FMC2DIFF_Top_Big.JPG
198
\picw=10cm \cinspic ./img/FMC2DIFF_Bottom_Big.JPG
199
\caption/f Realised PCB of FMC2DIFF01A module.
1124 kaklik 200
\endinsert
201
 
1126 kaklik 202
Both of the ADCdual01A modules were connected to FPGA ML605 board trough FMC2DIFF01A adapter board. The design of this adapter expects the presence of FMC LPC connector on host side and the board is, at the same time, not compatible with MLAB. It is, on the other hand, designed to meet the VITA 57 standard specifications for boards which support region 1 and region 3. VITA 57 regions are explained in the picture \ref[VITA57-regions].
1133 hlavac 203
This industry standard guarantees the compatibility with other FPGA boards that have FMC LPC connectors for Mezzanine Card. Schematic diagram of designed adapter board is included in the appendix.
1073 kaklik 204
 
1133 hlavac 205
The primary purpose of the PCB is to enable the connection of ADC modules located outside the PC case with ML605 development board. (In PC box analog circuits cannot be realized without the use of massive RFI mitigation techniques).
1134 fluktuacia 206
Differential signalling connectors should be used for conducting digital signal over relatively long cables. The signal integrity sensitive links (clocks) are equipped with output driver and translator to LVPECL logic for better signal transmission quality.
1083 kaklik 207
 
1134 fluktuacia 208
LVPECL level signal connectors on FMC2DIFF01A board are dedicated to transmit the clock signals. We have selected  the SY55855V and SY55857L dual translators. Dual configuration in useful due to fact, that SATA cable contains two differential pairs.
1119 kaklik 209
 
1134 fluktuacia 210
The SY55855V is a fully differential, CML/PECL/LVPECL-to-LVDS translator. It achieves LVDS signalling up to 1.5Gbps, depending on the distance and the characteristics of the media and noise coupling sources.
211
LVDS is intended to drive 50 $\Omega$ impedance transmission line media such as PCB traces, backplanes, or cables. SY55855V inputs can be terminated with a single resistor between the true and the complement pins of a given input \cite[SY55855V-chip].
1119 kaklik 212
 
1125 kaklik 213
The SY55857L is a fully differential, high-speed dual translator optimized to accept any logic standard from single-ended TTL/CMOS to differential LVDS, HSTL, or CML and translate it to LVPECL. Translation is guaranteed for speeds up to 2.5Gbps (2.5GHz toggle frequency). The SY55857L does not internally terminate its inputs, as different interfacing standards have different termination requirements\cite[SY55857L-chip].
1119 kaklik 214
 
1134 fluktuacia 215
Inputs of both used chips are terminated accordingly to used logic. The LVDS input is terminated differentially by 100 $\Omega$ resistor between positive and negative inputs. PECL input is terminated by Thevenin resistor network. Thevenin termination method was selected as optimal one, due to the absence of a proper power voltage (1,3 V) for direct termination by 50 $\Omega$ resistors. Termination on FPGA side is realized directly by settings the proper digital logic type on input pins.
1119 kaklik 216
 
1073 kaklik 217
\midinsert
1126 kaklik 218
\clabel[ML605-development-board]{ML605 development board}
1073 kaklik 219
\picw=10cm \cinspic ./img/ML605-board.jpg
1094 fluktuacia 220
\caption/f FPGA ML605 development board.
1073 kaklik 221
\endinsert
222
 
1096 kaklik 223
\midinsert
224
\clabel[VITA57-regions]{VITA57 board geometry}
225
\picw=10cm \cinspic ./img/VITA57_regions.png
226
\caption/f Definition of VITA57 regions.
227
\endinsert
228
 
1126 kaklik 229
% doplnit presny pocet konektoru
1116 kaklik 230
 
1133 hlavac 231
Several SATA connectors and two miniSAS connectors are populated on this board.  This set of connectors allows a connection of any number of ADC modules within the range of 1 to 8. ADC data outputs should be connected to the miniSAS connectors, while other supporting signals should be routed directly to SATA connectors on adapter.
1076 kaklik 232
 
1136 fluktuacia 233
Lengths of differential pairs routed on PCB of the module are not matched between the pairs. Length variation of differential pairs is not critical in our design according to facts discussed in paragraph \ref[diff-signaling]. Nevertheless, signals within differential pairs themselves are matched for length. Internal signal trace length matching of differential pairs is mandatory in order to minimize jitter and avoid a dynamic logic hazard conditions on digital signals, that represents the worst scenario. Thus the clocks' signals are routed in the most precise way on all designed boards.
1104 kaklik 234
 
1125 kaklik 235
Signal configuration used in our trial design is described in the following tables \ref[minisas-interface], \ref[SPI-system] and \ref[clock-interconnections].
1076 kaklik 236
 
1073 kaklik 237
 
1125 kaklik 238
\midinsert \clabel[minisas-interface]{miniSAS differential pairs connections}
1121 kaklik 239
\ctable {cccc}
240
{
241
miniSAS	&	SATA pair	&	FMC signal	&	Used as	\cr
242
P0	&	1	&	LA03	&	 not used 	\cr
243
P0	&	2	&	LA04	&	 not used 	\cr
244
P1	&	1	&	LA08	&	 not used 	\cr
245
P1	&	2	&	LA07	&	 not used 	\cr
246
P2	&	1	&	LA16	&	ADC1  CH1 (LTC2190)	\cr
247
P2	&	2	&	LA11	&	ADC1  CH2 (LTC2190) 	\cr
248
P3	&	1	&	LA17	&	ADC2 CH1 (LTC2271)	\cr
249
P3	&	2	&	LA15	&	ADC2 CH2 (LTC2271)	\cr
250
}
1133 hlavac 251
\caption/t miniSAS (FMC2DIFF01A J7) signal connections between modules.
1121 kaklik 252
\endinsert
253
 
1116 kaklik 254
 
1125 kaklik 255
\midinsert \clabel[SPI-system]{SPI configuration interface connections}
1121 kaklik 256
\ctable {ccc}
257
{
258
SPI connection J7	&	FMC signal	&	Connected to	\cr
259
SAS-AUX1	 &	LA14\_N	&	SPI DOUT	\cr
260
SAS-AUX2	 &	LA14\_P	&	SPI CLK	\cr
261
SAS-AUX3	 &	LA12\_N	&	CE ADC1	\cr
262
SAS-AUX4	 &	LA12\_P	&	CE ADC2	\cr
263
SAS-AUX5	 &	LA13\_N	&	soldered to GND	\cr
264
SAS-AUX6	 &	LA13\_P	&	not used	\cr
265
SAS-AUX7	 &	LA09\_N	&	not used	\cr
266
SAS-AUX8	 &	LA09\_P	&	soldered to GND	\cr
267
}
1133 hlavac 268
\caption/t SPI system interconnections
1121 kaklik 269
\endinsert
270
 
1133 hlavac 271
SPI interface is used in an unusual way in this design. SPI Data outputs from ADCs are not connected anywhere and read back is not possible, thus the configuration written to registers in ADC module cannot be validated. We have not observed any problems with this system, but it may be a possible source of failures.
1121 kaklik 272
 
1126 kaklik 273
 
1125 kaklik 274
\midinsert \clabel[clock-interconnections]{System clock interconnections}
1121 kaklik 275
\ctable {lccc}
276
{
277
Signal	&	FMC signal	&	FMC2DIFF01A	&	ADCdual01A	\cr
278
DCO	&	CLK1\_M2C	&	J5-1	&	J13-1	\cr
279
FR	&	LA18\_CC	&	J10-1	&	J12-1	\cr
280
ENC	&	LA01\_CC	&	J2-1(PECL OUT)	&	J3-1	\cr
281
SDGPSDO01A LO	&	CLK0\_M2C	&	J3-1 (PECL IN)	&	N/A	\cr
282
}
1133 hlavac 283
\caption/t Clock system interconnections
1121 kaklik 284
\endinsert
285
 
1133 hlavac 286
\secc FPGA function
1116 kaklik 287
 
1136 fluktuacia 288
Several tasks in separate FPGA blocks are performed by FPGA. In the first block the FPGA prepares a sampling clock for ADCdual01A modules by dividing the signal from the main local oscillator. This task represents a separate block in FPGA and runs asynchronously to other logical circuits. Second block is a SPI configuration module, which sends configuration words to ADC modules and it is activated by opening of Xillybus interface file. Third block represents the main module, which resolves ADC - PC communication itself and it communicates via PCIe, collect data from ADC hardware and creates data packet \ref[xillybus-interface]. Last block is activated after the ADC is configurated via SPI.
1116 kaklik 289
 
1136 fluktuacia 290
Communication over PCIe is managed by proprietary IP Core and Xillybus driver, which transfers data from FPGA registers to host PC. Data appear in a system device file named  "/dev/xillybus_data2_r" on the host computer. Binary data which appear in this file after its opening are shown in the table below \ref[xillybus-interface].
1116 kaklik 291
 
1133 hlavac 292
\midinsert
1127 kaklik 293
\def\tabiteml{ }\let\tabitemr=\tabiteml
1126 kaklik 294
\clabel[xillybus-interface]{Grabber binary output format}
295
\ctable {lccccccccc}{
1127 kaklik 296
\hfil & \multispan9 \hfil 160bit packet \hfil \strut \crl \tskip4pt
297
Data name &  FRAME  & \multispan2 \hfil ADC1 CH1 \hfil & \multispan2 \hfil ADC1 CH2 \hfil & \multispan2  \hfil ADC2 CH1 \hfil & \multispan2 \hfil ADC2 CH2 \hfil \strut  \cr
1073 kaklik 298
Data type & uint32 & int16 & int16 & int16 & int16 & int16 & int16 & int16 & int16 \cr
299
Content & saw signal & $t1$ &  $t_{1+1}$ &  $t1$ &  $t_{1+1}$ &  $t1$ &  $t_{1+1}$ &  $t1$ &  $t_{1+1}$ \cr
300
}
301
\caption/t System device "/dev/xillybus_data2_r" data format
302
\endinsert
303
 
1136 fluktuacia 304
Data packet block which is carried on PCI Express isa  described in the table \ref[xillybus-interface]. The data packet consist of several 32bit words. The first word contains FRAME number and it is filled with saw signal for now, with incremental step taking place every data packet transmission. The following data words contain samples from ADCs' first and second channel. Samples from every channel are transmitted in pairs of two samples. Number of ADC channels is expandable according to the number of physically connected channels. An CRC word may possibly be added in the future to the end of the transmission packet for data integrity validation.
1073 kaklik 305
 
1136 fluktuacia 306
FRAME word at the beginning of data packet, now filled with incrementing and overflowing saw signal, is used to ensure that no data samples ale lost during the data transfers from FPGA. FRAME signal may be used in the future for pairing the ADC samples' data packet with another data packet. This new additional data packet should carry meta-data information about the sample time jitter, current accuracy of the local oscillator frequency etc.
1116 kaklik 307
 
1136 fluktuacia 308
Detailed description of the currently implemented FPGA functions can be found in a separate paper \cite[fpga-middleware]. HDL source codes for FPGA at a state in which it was used are included on the enclosed CD. More recent development versions are publicly available from MLAB sources repository.
1125 kaklik 309
 
1126 kaklik 310
% doplnit odkaz na mlab repozitar
311
 
1133 hlavac 312
\secc Data reading and recording
1073 kaklik 313
 
1136 fluktuacia 314
In order to read the data stream from the ADC drive, we use Gnuradio software. Gnuradio suite consists of gnuradio-companion package which is a graphical tool for creating signal-flow graphs and generating Python flow-graph source code. This tool has been used to create a basic RAW data grabber to record and interactively view waterfall plots using the data streams output from ADC modules.
1073 kaklik 315
 
316
\midinsert
1127 kaklik 317
\clabel[grabber-flow-graph]{Gnuradio flow graph for signal grabbing}
318
\picw=\pdfpagewidth \setbox0=\hbox{\inspic ./img/screenshots/Grabber.grc.png }
319
\par\nobreak \vskip\wd0 \vskip-\ht0
320
\centerline {\kern\ht0 \pdfsave\pdfrotate{90}\rlap{\box0}\pdfrestore}
321
\caption/f The ADC recorder flow graph created in gnuradio-companion.
1073 kaklik 322
\endinsert
323
 
324
\midinsert
325
\picw=15cm \cinspic ./img/screenshots/Grabber_running.png
1094 fluktuacia 326
\caption/f User interface window of a running ADC grabber.
1073 kaklik 327
\endinsert
328
 
1136 fluktuacia 329
The interactive grabber-viewer user interface shows live oscilloscope-like time-value display for all data channels and live time-frequency scrolling display (a waterfall view) for displaying the frequency components of the grabbed signal. Signal is grabbed to file with exactly the same format, as described in the table \ref[xillybus-interface].
1073 kaklik 330
 
1116 kaklik 331
 
1132 kaklik 332
\chap Achieved parameters
1116 kaklik 333
 
1136 fluktuacia 334
The trial version construction was tested for proper handling of sampling rates in the range of 5 MSPS to 15 MSPS, but it should work above this limit. System works on i7 8 cores computer with Ubuntu 12.04 LTS operating system.  Data recording of input signal is impossible above the sampling rates around 7 MSPS due to bottleneck at HDD speed limits, but it should be resolved by the use of SSD disk drive. However, such design has not been tested in our setup.
1127 kaklik 335
 
1132 kaklik 336
\sec Measured parameters
1073 kaklik 337
 
1136 fluktuacia 338
Two prototypes of ADC modules were assembled and tested. The first prototype, labeled ADC1, has LTC2190 ADC chip populated with LT6600-5 front-end operational amplifier. It also has a 1kOhm resistors populated on inputs which give it an ability of an internal attenuation of the input signal. The value of this attenuation $A$ is calculated by the following formula \ref[ADC1-gain]
1078 kaklik 339
 
1105 kaklik 340
\label[ADC1-gain]
1103 kaklik 341
$$
1126 kaklik 342
A = {806 \cdot R_1 \over R_1 + R_2} \eqmark
1103 kaklik 343
$$
344
 
1136 fluktuacia 345
Where the letters stand for the following:
1103 kaklik 346
\begitems
1122 fluktuacia 347
  * $A$ -  Gain of an input amplifier.
1103 kaklik 348
  * $R_1$ - Output impedance of signal source (usually 50 Ohm).
1105 kaklik 349
  * $R_2$ - Value of serial resistors at operational amplifier inputs.
1103 kaklik 350
\enditems
351
 
1136 fluktuacia 352
We have $R_2 = 1000 \Omega$ and $R_1 = 50 \Omega$ which imply that $A = 0.815$. That value of A was further confirmed by the measurement.
353
In our measurement setup we have H1012 Ethernet transformer connected to inputs of ADC. We have used this transformer for signal symetrization from BNC connector at Agilent 33220A signal generator. Circuit diagram of the used transformer circuit is shown in picture \ref[balun-circuit]  and circuit realization in photograph \ref[SMA2SATA-nest].
1103 kaklik 354
 
1132 kaklik 355
\midinsert
356
\clabel[balun-circuit]{Balun transformer circuit}
357
\picw=10cm \cinspic ./img/SMA2SATA.pdf
1133 hlavac 358
\caption/f Simplified balun transformer circuit diagram.
1132 kaklik 359
\endinsert
1105 kaklik 360
 
1136 fluktuacia 361
The signal generator Agilent 33220A which we used, does not have optimal parameters for this type of dynamic range measurement. Signal distortion and spurious levels are only -70 dBc according to Agilent datasheet \cite[33220A-generator]. We have managed to measure an ADC saturation voltage of 705.7 mV (generator output) with this setup, mostly due to an impedance mismatch and uncalibrated measurement setup, with 1V ADC range selected by sense pin. This is a relatively large error, but the main result of our measurement, seen as a FFT plot shown in image \ref[ADC1-FFT], confirms $>$80 dB dynamic range at ADC module input.
1127 kaklik 362
 
1080 kaklik 363
\midinsert
1105 kaklik 364
\clabel[ADC1-FFT]{ADC1 sine test FFT}
365
\picw=15cm \cinspic ./img/screenshots/ADC1_CH2_FFT.png
366
\caption/f Sine signal sampled by ADC1 module with LTC2190 and LT6600-5 devices.
1080 kaklik 367
\endinsert
368
 
1103 kaklik 369
 
1136 fluktuacia 370
Similar test was performed at ADC2 module. For ADC2 we have to use formula with a different constant \ref[ADC1-gain]. The ADC2 module has LT6600-2.5 amplifiers populated on it with a gain equal to $A = 2.457$ and uses the same $R_2$ resistors. We measured saturation voltage of 380 mV (generator output) at channel 1 on this ADC. It is well within the parameter tolerances of the used setup.
1078 kaklik 371
 
1105 kaklik 372
\label[ADC2-gain]
1098 kaklik 373
$$
1126 kaklik 374
A = {1580 \cdot R_1 \over R_1 + R_2} \eqmark
1098 kaklik 375
$$
1073 kaklik 376
 
1136 fluktuacia 377
Where the letters stand for the following:
1098 kaklik 378
\begitems
1122 fluktuacia 379
  * $A$ -  Gain of an input amplifier.
1103 kaklik 380
  * $R_1$ - Output impedance of signal source (usually 50 Ohm).
1105 kaklik 381
  * $R_2$ - Value of serial resistors at operational amplifier inputs.
1098 kaklik 382
\enditems
383
 
1105 kaklik 384
\midinsert
385
\clabel[ADC2-FFT]{ADC2 sine test FFT}
386
\picw=15cm \cinspic ./img/screenshots/ADC2_CH1_FFT.png
387
\caption/f Sine signal sampled by ADC2 module with LTC2271 and LT6600-2.5 devices.
388
\endinsert
1098 kaklik 389
 
1136 fluktuacia 390
Computed FFT spectra for measured signal are shown in the images \ref[ADC2-FFT] and \ref[ADC1-FFT].  Both images confirm that ADCdual01A modules have input dynamical range of at least 80 dB.
1098 kaklik 391
 
1127 kaklik 392
\midinsert
393
\clabel[SMA2SATA-nest]{Used balun transformer}
394
\picw=15cm \cinspic ./img/SMA2SATA_nest1.JPG
1133 hlavac 395
\caption/f Balun transformer circuit used for ADC parameters measurement. It is constructed from H1012 transformer salvaged from an old Ethernet card.
1127 kaklik 396
\endinsert
397
 
398
 
399
 
1132 kaklik 400
\sec Example of usage
1098 kaklik 401
 
1133 hlavac 402
For additional validation of system characteristics a receiver setup has been constructed.
403
 
1132 kaklik 404
\secc Basic interferometric station
1073 kaklik 405
 
1136 fluktuacia 406
Interferometry station was chosen to serve as the most basic experimental setup. We connected the new data acquisition system to two SDRX01B receivers. Block schematics of the setup used is shown in the image \ref[block-schematic]. Two ground-plane antennae were used and mounted outside the balcony at CTU building at location 50$^\circ$ 4' 36.102'' N,  14$^\circ$ 25' 4.170'' E.
407
Antennae were equipped with LNA01A amplifiers. All coaxial cables had the same length of 5 meters. Antennae were isolated by common mode ferrite bead mounted on cable to minimise the signal coupling between antennas. Evaluation system consisted of SDGPSDO local oscillator subsystem used to tune the local oscillator frequency.
1098 kaklik 408
 
409
\midinsert
1105 kaklik 410
\clabel[block-schematic]{Receiver block schematic}
1132 kaklik 411
\picw=\pdfpagewidth \setbox0=\hbox{\inspic ./img/Basic_interferometer.png }
1127 kaklik 412
\par\nobreak \vskip\wd0 \vskip-\ht0
413
\centerline {\kern\ht0 \pdfsave\pdfrotate{90}\rlap{\box0}\pdfrestore}
1105 kaklik 414
\caption/f Complete receiver block schematic of dual antenna interferometric station.
415
\endinsert
416
 
1116 kaklik 417
% doplnit schema skutecne pouziteho systemu
1105 kaklik 418
 
1133 hlavac 419
Despite of the schematic diagram proposed at beginning of system description....
1136 fluktuacia 420
We have used two separate oscillators -- one oscillator drives encoded signal to ADCs still through FPGA based divider and the other one drives it to SDRX01B mixer.
421
The reason for this modification was an attempt to simplify the frequency tuning during the experiment. A single oscillator may be used only with a proper setting of FPGA divider and this divider may be modified only by recompilation of FPGA code and loading/flashing a new FPGA schema. Due to fact that the FPGA is connected to PCI express and kernel drivers with hardware must be reinitialized, reboot of PC is required every time a FPGA scheme is changed. Instead of this complicated procedure, we set the FPGA divider to a constant division factor of 30 and used another district oscillator for ADCdual01 sampling modules and for SDRX01B receiver.
422
We have used ACOUNT02A MLAB instrument for frequency checking of correct setup on both local oscillators.
1105 kaklik 423
 
424
\midinsert
1098 kaklik 425
\clabel[meteor-reflection]{Meteor reflection}
426
\picw=10cm \cinspic ./img/screenshots/observed_meteor.png
1136 fluktuacia 427
\caption/f Meteor reflection received by an evaluation setup.
1098 kaklik 428
\endinsert
429
 
430
\midinsert
1124 kaklik 431
\clabel[phase-difference]{Phase difference}
1098 kaklik 432
\picw=10cm \cinspic ./img/screenshots/phase_difference.png
1131 fluktuacia 433
\caption/f Demonstration of phase difference between antennae.
1098 kaklik 434
\endinsert
435
 
1136 fluktuacia 436
For the simplest demonstration of phase difference between antennae, we have analysed part of the signal by complex conjugate multiplication between channels. Results of this analysis can be seen in the following picture \ref[phase-difference]. Points of the selected part of the signal create a clear vector, which illustrates the presence of the phase difference.
1124 kaklik 437
 
1098 kaklik 438
 
1132 kaklik 439
\secc Simple passive Doppler radar
1073 kaklik 440
 
1132 kaklik 441
% doplnit popis
1105 kaklik 442
 
1132 kaklik 443
 
444
\secc Simple polarimeter station
445
 
446
% doplnit popis
447
 
1131 fluktuacia 448
\chap Proposition of the final system
1073 kaklik 449
 
1136 fluktuacia 450
The construction of the final system, that is supposed to be employed for real radioastronomy observations will be described in this chapter. It is mainly a theoretical analysis of the data handling systems. Realization of the described ideas might be possible as a part of our future development after we fully evaluate and test the current trial design.
1073 kaklik 451
 
1136 fluktuacia 452
The system requires proper handling of huge amounts of data and either huge and fast storage capacity is needed to store the captured signal data, or enormous computational power is required for online data processing and filtering. Several hardware approaches currently exist and are in use for data processing problem handling. Either powerful multi gigahertz CPUs, GPUs, FPGAs, or specially  constructed ASICs are used for this task.
1107 kaklik 453
 
1075 kaklik 454
\sec Custom design of FPGA board
1073 kaklik 455
 
1136 fluktuacia 456
In the beginning of the project, a custom design of FPGA interface board had been considered. This FPGA board should include PCI express interface and should sell at lower price than the trial design. It should be compatible with MLAB internal standards  which are further backward compatible with the existing or improved design of ADC modules. For a connection of FPGA board to another adapter board with PCIe we expect a use of a PCIe host interface.
457
Thunderbolt technology standard was expected to be used in this PC to PCIe module communication which further communicates with MLAB compatible FPGA module. Thunderbolt chips are currently available on the market for reasonable prices \cite[thunderbolt-chips]. However, a problem lies in the accessibility to their specifications, as they are only available for licensed users and Intel has a mass market oriented licensing policy, that makes this technology inaccessible for low quantity production. As a consequence, an external PCI Express cabling and expansion slots should be considered as a better solution, if we need to preserve standard PC as a main computational platform.
1085 kaklik 458
 
1136 fluktuacia 459
However, these PCI express external systems and cables are still very expensive. Take Opal Kelly XEM6110 \cite[fpga-pcie] as an example, with its price tag reaching 995 USD at time of writing the thesis. Therefore, a better solution probably needs to be found.
1085 kaklik 460
 
1133 hlavac 461
An interface problem will by probably resolved by other than Intel ix86 architecture. Many ARM computers have risen on market due to an increased demand of embedded technologies, which require high computation capacity, low power consumption and small size -- especially smartphones. Many of those ARM based systems have interesting parameters of signal processing. These facts make Intel's ix86 architecture unattractive for future projects.
1116 kaklik 462
 
1075 kaklik 463
\sec Parralella board computer
464
 
1136 fluktuacia 465
<<<<<<< .mine
466
Parallella is a new product created by Adapteva, Inc. \cite[parallella-board]. It represents a small supercomputer, that has been in development for almost two years with only testing series of boards produced until now (first single-board computers with 16-core Epiphany chip were shipped in December 2013) \cite[parallella-board]. The board has nearly ideal parameters for signal processing (as it provides around 50 GFLOPS of computational power). It is is equipped with Epiphany coprocessor which has 16 High Performance RISC CPU Cores,  Zynq-7020 FPGA with Dual ARM® Cortex™-A9 MPCore™ and operating frequency of 866 MHz, 1GB RAM, 85K Logic Cells, 10/100/1000 Ethernet and OpenCL support \cite[parallella16-board]. In addition to this, the board consumes only 3 Watts of power if both Zynq and Epiphany cores are running simultaneously.
467
=======
1135 kaklik 468
Parallella is a new product created by Adapteva, Inc. \cite[parallella-board]. It represents a small supercomputer, that has been in development for almost two years with only testing series of boards produced until now (first single-board computers with 16-core Epiphany chip were shipped in December 2013) \cite[parallella-board]. The board has nearly ideal parameters for signal processing (as it provides around 50 GFLOPS  of computational power). It is is equipped with Epiphany coprocessor which has 16 High Performance RISC CPU Cores,  Zynq-7020 FPGA with Dual ARM® Cortex™-A9 MPCore™ and operating frequency of 866 MHz, 1GB RAM, 85K Logic Cells, 10/100/1000 Ethernet and OpenCL support \cite[parallella16-board]. In addition to this, the board consumes only 3 Watts of power if both Zynq and Epiphany cores are running simultaniously.
1136 fluktuacia 469
>>>>>>> .r1135
1075 kaklik 470
 
1133 hlavac 471
The main disadvantage of Parralella board is its unknown lead time and an absence of SATA interface or other interface suitable for data storage connection. Fast data storage interface would be useful and would allow bulk processing of captured data. Following that, the results of data processing may be sent over the Ethernet interface to data storage server.
1117 kaklik 472
 
1116 kaklik 473
\midinsert
474
\clabel[img-parallella-board]{Parallella board overview}
475
\picw=15cm \cinspic ./img/ParallellaTopView31.png
476
\caption/f Top view on Parallella-16 board \cite[parallella16-board].
477
\endinsert
478
 
1136 fluktuacia 479
If Parallella board will be used as a radioastronomy data interface, there would be a demand for new ADC interface module. The interface module will use four PEC connectors mounted on the bottom of the Parallella board. This daughter module should have MLAB compatible design and should preferably be constructed in the form of separable modules for every Parallella's PEC connector.
1116 kaklik 480
 
1133 hlavac 481
\sec GPU based computational system
1075 kaklik 482
 
1133 hlavac 483
A new GPU development board NVIDIA K1, shown in the following picture \ref[img-NVIDIA-K1], has recently been released. These boards are intended to be used in fields including computer vision, robotics, medicine, security or automotive industry. They have good parameters for signal processing for a relatively low price of 192 USD.  Unfortunately, they are currently only in pre-order release stage (in April 2014).
1075 kaklik 484
 
1086 kaklik 485
\midinsert
486
\clabel[img-NVIDIA-K1]{NVIDIA Jetson TK1 Development Kit}
487
\picw=15cm \cinspic ./img/Jetson_TK1_575px.jpg
1132 kaklik 488
\caption/f The NVIDIA Jetson TK1 Development Kit \cite[nvidia-k1].
1086 kaklik 489
\endinsert
490
 
1136 fluktuacia 491
NVIDIA board differs from other boards in its category by a presence of PCI Experess connector. If we decide to use this development board in our radio astronomy digitalisation system, the PCI express  should be used for FPGA connection. A new FPGA board with PCI Express direct PCB connector
1127 kaklik 492
 
1133 hlavac 493
% doplnit popis pripojeni FPGA desky s HDMI Kabelem.
1127 kaklik 494
 
1133 hlavac 495
\sec Other ARM based computation systems
1127 kaklik 496
 
1131 fluktuacia 497
Other embedded ARM based computers, for example ODROID-XU, lack a suitable high speed interface. Their highest speed interface is USB 3.0 which has currently unsettled development support and needs commercial software tools for evaluation and testing.
1127 kaklik 498
 
499
 
1136 fluktuacia 500
From the summary analysis mentioned above, the Parrallella board seems to be a best candidate for computational board in radioastronomy data acquisition system, as it is optimised for high data flow processing. On one hand, Parrallella does not have much memory to cache the processing data but on the other hand it has wide bandwidth data channels instead. Other boards might provide much more computational power -- 300 GFLOPS in case of NVIDIA K1, but they are optimised for heavy computational tasks on limited amount of data which represents a typical problem in computer graphics. However, in our application we do not need such extreme computation power at data acquisition system level.
501
As a result we should presumably wait until Parallella becomes widely available. Following that, a new ADCdual interface board should be designed and prepared to be used in new scalable radio astronomy data acquisition system. In the meantime, before suitable computing hardware become accessible, the required applications and algorithms should be optimised using the proposed trial version with FPGA development board on standard PC host computer (having a PCI Express interface to development board).
1127 kaklik 502