Texas Instruments TSW1200EVM User manual

User's GuideSLAU212A – April 2007 – Revised August 2008
TSW1200EVM: High-Speed LVDS Deserializer andAnalysis System
Contents1 Introduction ................................................................................................................... 32 Functionality .................................................................................................................. 33 Hardware Configuration ..................................................................................................... 54 Software Installation ....................................................................................................... 125 User Interface ............................................................................................................... 166 MATLAB™ Interface ....................................................................................................... 227 Schematics and Bill of Materials ......................................................................................... 238 Circuit Board Layout and Layer Stackup ................................................................................ 30
List of Figures
1 Position of Power Connections ............................................................................................ 52 Position of Switches and Jumpers ......................................................................................... 73 EEPROM Configuration Options ........................................................................................... 84 Position of LEDs ............................................................................................................. 95 Position of Input, Output, and USB Connections ...................................................................... 116 Pinout of Header Posts for Parallel Output Data ....................................................................... 127 TSW1200 User Interface Installation .................................................................................... 138 Hardware Device Manager ................................................................................................ 149 Found New Hardware Windows .......................................................................................... 1510 User Interface Initial Setup Screen ...................................................................................... 1611 User Interface Single FFT format ........................................................................................ 2012 User Interface Time Domain Format..................................................................................... 2213 Schematic Diagram Page 1 ............................................................................................... 2314 Schematic Diagram Page 2 ............................................................................................... 2415 Schematic Diagram Page 3 ............................................................................................... 2516 Schematic Diagram Page 4 ............................................................................................... 2617 Schematic Diagram Page 5 ............................................................................................... 2718 TSW1200C Layout Top Layer ............................................................................................ 3019 TSW1200C Layout Layer Two ........................................................................................... 3120 TSW1200C Layout Power Plane ......................................................................................... 3221 TSW1200C Layout Ground Plane ....................................................................................... 3322 TSW1200C Layout Layer 5 ............................................................................................... 3423 TSW1200C Layer 6 ........................................................................................................ 3524 TSW1200C Bottom Layer ................................................................................................. 3625 Circuit Board Stackup ...................................................................................................... 37
List of Tables
1 Bill of Materials ............................................................................................................. 28
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 1Submit Documentation Feedback

www.ti.com
Windows is a trademark of Microsoft Corporation.LabVIEW is a trademark of National Instruments Corporation.MATLAB is a trademark of The MathWorks, Inc..Xilinx is a trademark of Xilinx, Inc..
TSW1200EVM: High-Speed LVDS Deserializer and Analysis System2 SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

1 Introduction
2 Functionality
www.ti.com Introduction
The Texas Instruments TSW1200 High-Speed LVDS Deserializer and Analysis System provides acomprehensive set of hardware and user interface software to effectively evaluate the performance of awide range of Texas Instruments (TI) high-speed, analog- to-digital converters (ADC), particularly thosehigh-speed ADCs that employ an low-voltage differential signal (LVDS) format for the data converter’soutput data. With a high-quality, low-jitter clock and a high-quality input frequency provided to theevaluation module, the TSW1200 system can be used to demonstrate data sheet performancespecifications.
The TSW1200 hardware has a high-speed connector that plugs into an evaluation module (EVM) for theADC. Firmware for an FPGA on the TSW1200EVM has an interface to various LVDS data formats andFIFO memory sufficient to capture as much as 64K samples of data. A USB connection transfers thecaptured data to a personal computer for post-processing. The user Interface software controls theTSW1200 hardware and displays the FFT and important statistics related to the performance of the ADC.
Many TI high-speed ADCs have LVDS outputs for the digitized data. These ADCs are generally availableon an EVM that connects directly to the TSW1200EVM. The common connector between the ADC EVMand the TSW1200EVM is a Samtec high-speed connector suitable for differential pairs of pins separatedby ground. A common pinout for the connector across a family of EVMs has been established. At present,the interface between the ADC EVM and the TSW1200EVM has defined connections for 14 pairs of LVDSplus two clock lines. The connector pinout has an additional 14 LVDS pairs for future product support. Inmost cases, if a TI ADC supports LVDS outputs, then the TSW1200 system can be used to capture datafrom the ADC. If the ADC supports a CMOS single-ended format, then the TSW1100 system is intendedto interface to it.
The data format for the LVDS data bus can be in one of many formats, all supported by the TSW1200.For single-channel, high-speed ADCs, the data format is commonly parallel dual-data rate with one clockline. Dual-data rate means that both the rising and falling edges of the clock strobe data into theTSW1200. For multichannel ADCs, the data is commonly presented in a serialized format, whereindividual bits of the output data are presented on an LVDS line one bit at a time, at a higher data ratethan the sample rate of the ADC.
The firmware in the FPGA on the TSW1200 is designed to accommodate both parallel DDR formats andserial LVDS formats, although not at the same time. The EEPROM on the TSW1200EVM is large enoughto hold two distinct program files for the FPGA. One program bit file supports the parallel DDR format andthe other bit file supports serial LVDS formats. The TSW1200 can be set to support the desired dataformat by simply setting two jumpers and pressing the PROGRAM button.
The parallel DDR FPGA program supports several types of data formats. One common format presentsodd-numbered data bits on the bus on one clock edge and even-numbered data bits on the bus on theother clock edge. This format is commonly used for ADCs with sampling rates up to 250 MHz. For thisbit-wise DDR format, the parallel data bus uses half as many LVDS pairs as there are bits in the sample.For example, a 16-bit ADC uses eight LVDS pairs for data plus an LVDS clock pair for bit-wise DDR. Forhigher sample rates up to 500 MHz, a sample-wise DDR format is often used. For sample-wise DDR, thedata bus width has as many LVDS pairs as the bit resolution of the ADC. On one clock edge, a datasample from the ADC is registered; on the next clock edge, the next data sample from the ADC isregistered.
The serial FPGA program also supports several data formats. For one-wire serial formats, the data isserialized onto a single LVDS pair at a rate that is 12 times the sample rate for an ADC with a 12-bitresolution. A one-wire serialization format also is used for 14-bit and 16-bit data at data rates 14 or 16times the sample rate, respectively. For serial data formats, a DDR LVDS bit clock is used to strobe theserial data bits and to deserialize the data. An additional clock pair provided at the sample rate of the ADCidentifies the sample-word boundaries in the serial data. For multichannel ADCs, a single-bit clock and asingle sample-rate clock (frame clock) is used for all of the LVDS data channels. The other common serialdata format is two-wire serialization. Two-wire serialization is similar to one-wire serialization except that adata channel uses two LVDS pairs to carry the serialized data at a rate that is half of what it is for one-wireserialization. Two-wire serialization commonly is used for sample rates up to 125 MHz, whereas one-wireserialization generally is used for sample rates up to 65 MHz.
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 3Submit Documentation Feedback

Functionality www.ti.com
The FPGA firmware for the TSW1200 generally consists of two major functions: the LVDS ADC interfaceand the FIFO capture. The LVDS ADC interface supports either the various parallel DDR interfaces or thevarious serial interfaces as previously described. The LVDS ADC interface code in the FPGA reformatsthe data into a standard single-ended parallel data word with sample clock. This parallel sample word plusclock is output continuously to header posts on the TSW1200 for capture by a logic analyzer. The paralleldata word with sample clock also is presented to the FIFO capture logic. Note that for parallel DDR dataformats, the header output posts are not enabled by default, as the data rate (up to a 500-MHz samplerate) is often too fast to be feasible for a single-ended CMOS output.
The TSW1200 FPGA has enough FIFO buffer to capture as much as a 65536-sample record length fromthe continuous sample data stream coming from the LVDS ADC interface. The TSW1200 FPGA designincludes a UART function that can transfer data to and from a USB interface device on the TSW1200board. The USB interface device on the TSW1200EVM connects to a personal computer (PC) runningWindows™ over a standard USB cable. The operation of the FIFO capture logic is controlled by writesfrom the PC USB port to a register map defined within the FPGA. The user interface software on the PCselects by register operations such things as record length of data to capture, which channel of an ADC tocapture from, and then the user interface software downloads the captured data from the TSW1200 forprocessing in the form of an FFT or time-domain display.
TSW1200EVM: High-Speed LVDS Deserializer and Analysis System4 SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

3 Hardware Configuration
3.1 Power Connections
TSW1200
Ground
J7
J14 J15
Selects 5 V
Output for J15
(Default)
Selects J7 for
6 V Input
(Default)
Selects J5
for 6 V Input
J22
JP8
6 V I/O
+6 V
Input
www.ti.com Hardware Configuration
In this section, the various portions of the TSW1200EVM hardware are described.
The TSW1200EVM hardware is designed to operate from a single-supply voltage of greater than 6 Vdc.For convenience, two options can supply power to the TSW1200EVM. A bench power supply can supplypower to banana jack connections on the TSW1200EVM, or a laptop-style power module that is includedwith the TSW1200 hardware can supply power. Figure 1 shows the relative position of the powerconnections on the TSW1200EVM (revision B or C of the hardware).
CAUTION
TI recommends that the black banana jack J14 be connected to a benchground even if the 6-V external power brick is connected to J7. Intermittent lossof the USB connection can sometimes be observed without a good ground fromthe TSW1200EVM to the bench ground reference.
Figure 1. Position of Power Connections
Care must be taken in selection of the input power supply. One jumper selects whether the 6-V inputpower comes from the banana jack, or whether it comes from the external power module. Another jumperselects whether to connect the onboard regulated 5 V to the red banana jack for output. If the red bananajack is used to input 6 V from a bench power supply, the onboard regulated 5 V must not be connected tothe red banana jack for output at the same time. Doing so causes the onboard 5-V regulator to fail overtime.
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 5Submit Documentation Feedback

3.2 Switches and Jumpers
3.2.1 Pushbuttons
3.2.2 Jumpers
Hardware Configuration www.ti.com
Four pushbutton switches are mounted on the TSW1200EVM. Two pushbutton switches currently havedefined functions; two of the switches are reserved for future use.
The PROGRAM pushbutton (SW3) causes the FPGA to reload its bit file from the FPGA EEPROM on theTSW1200EVM. The EEPROM currently is large enough to hold two FPGA bit files. When the FPGA loadsa bit file, it loads the bit file indicated by the position of the jumpers J11 and J10. When power is firstapplied to the TSW1200EVM, the FPGA also loads its bit file from the EEPROM as determined by thejumpers J11 and J10. Therefore, pressing the PROGRAM pushbutton has the same effect aspower-cycling the FPGA. All register settings in the FPGA and any data held in the FPGA are lost onloading the bit file, and the FPGA then is in its initial default configuration.
The RESET (SW4) pushbutton causes the FPGA to clear the FIFO storage, but does not clear any of theregister settings in the FPGA. Any configuration of the FPGA done through register operations such assetting the UART baud rate will persist after pushing the RESET pushbutton. FPGA register settings onlyare cleared by pressing the PROGRAM pushbutton. For this reason, the RESET pushbutton has limitedutility in this revision of the TSW1200EVM; this pushbutton may have more utility defined for it in futurerevisions of the FPGA firmware.
The PROM RESET (SW2) is reserved for future use, as is pushbutton SW5.
Jumpers J10 and J11 select the bit file to be programmed into the FPGA as indicated in Figure 2 . TheEEPROM is currently large enough to contain two complete programming files for the FPGA, withprovision to later install a larger EEPROM with the capability to store four complete programming bit filesfor the FPGA.
Bit file CFG1 (jumper J10 set to LO and jumper J11 set to HI) is defined for use with ADC EVMs thatemploy a parallel LVDS DDR (dual-data rate) format. Bit file CFG2 is defined for use with ADC EVMs thatemploy a serialized data format, whether a one-wire serial format or a two-wire serial format. See Figure3.
Simply moving the position of the jumpers J10 and J11 has no immediate effect. The FPGA does not loadits programming bit file except on initial power up or until the PROGRAM pushbutton is pressed. Tochange the operation of the TSW1200EVM for use with a parallel DDR format or with a serial LVDSformat, the jumpers J10 and J11 must be set to the desired position, and then the board must bepower-cycled or have the PROGRAM button pushed.
Jumper JP8 selects whether the power supply to the TSW1200EVM is to be supplied by the external 6-Vpower module through power jack J7 or by an external 6-V bench power supply through the red bananajack J15.
A low drop-out linear power regulator on the TSW1200EVM generates a clean, low-noise 5-V supply fromthe 6-V power that is input to the TSW1200EVM. The jumper J22 can be used to connect this onboardregulated 5 V to the red banana jack J15 as an output. This option is a convenience for use with ADCEVMs that require a single 5-V supply input. In this way, the TSW1200EVM and ADC EVM combinationcan be powered completely from the laptop-style 6-V external power supply.
CAUTION
It is possible to select the red banana jack J15 as an input to be connected to a6-V bench supply and at the same time install jumper J22 to connect theregulated 5 V to the red banana jack as an output. This, however, over timecauses the 5-V regulator on the TSW1200 to fail.
Jumper J16 is used at the factory to program the small EEPROM for the USB port. The default position forthis jumper is to be installed, and in normal operation, J16 is left installed.
6TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

TSW1200
+6 V
Input
Ground
PROGRAM
RESET
PROM RESET
SW2 SW3
SW5
J17
J10
J11
J16
JP8
J22
open
USB EEPROM
(Default Shorted)
JTAG TDI =>TDO
Selects FPGA bit file
from EEPROM
(Default CFG1 for
Parallel DDR format)
(Default Shorted)
6 V I/O
J1
J12
SW4
www.ti.com Hardware Configuration
Jumper J17 can be used to disable the 1.2-V power regulator for the FPGA core logic. The defaultposition for this jumper is to be left uninstalled or open, and in normal operation, this jumper is leftuninstalled.
Jumpers J1 and J12 form part of a JTAG chain through the FPGA and the FPGA EEPROM. A chain ofJTAG devices form a loop, with the TDO from one JTAG device connected to the TDI of the next JTAGdevice. The normal setting of the JTAG jumpers is to connect the TDI of the JTAG connector to the TDIpin of the FPGA EEPROM through jumper J12 pins 2-4. Then, the TDO of the FPGA EEPROM connectsto the TDI of the FPGA through jumper J12 pins 1-3 and jumper J1 pins 2-4. The TDO of the FPGAconnects to the JTAG connector pin TDO through jumper J1 pins 1-2. If it desired to remove either thePFGA or the FPGA EEPROM from the JTAG chain, the jumpers J1 or J12 can be turned 90 degrees toconnect pins 1-2 and 3-4. See the TSW1200EVM schematics and layout sections for illustration of thisoption.
Figure 2. Position of Switches and Jumpers
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 7Submit Documentation Feedback

PRGM LOAD
CFG3
CFG4
CFG2
CFG1
J10 J11
LO LO
LO
LO
HI
HI
HI
HI
3.3 LEDs
Hardware Configuration www.ti.com
Figure 3. EEPROM Configuration Options
Six LEDs are on the TSW1200EVM to indicate the presence of power and the state of the FPGA. SeeFigure 4. LED D16 illuminates to indicate the presence of a 6-V power supply to the board. On power up,the FPGA loads its programming bit file from the FPGA EEPROM. Once the programming bit file hascompleted loading, the LED D7 illuminates to indicate that the FPGA is operational.
Once the FPGA has loaded its bit file, it internally resets the clock circuitry that locks to a 200-MHz clocksignal from a 200-MHz oscillator on the TSW1200 board. Once the internal reset is complete, LED D1(labeled ADC on the TSW1200 silkscreen) illuminates to indicate that the TSW1200EVM is ready to usewith an ADC EVM.
LED D4 flashes when the onboard 200-MHz oscillator is running. The 200-MHz oscillator clock ismultiplied up to 250 MHz within the FPGA. The pattern of the flashing is of the formflash-flash-pause-pause. That is, the LED illuminates for a time, then turns off for the same time, thenilluminates for the same time, then turns off 5 times. The internal time is derived from a division of the250-MHz reference clock derived from the onboard oscillator.
LED D2 (labeled DCM on the TSW1200 silkscreen, which is an abbreviation for the digital clock managerblock of logic in the FPGA) flashes when an LVDS clock from the ADC is present. The DCM clock flashesin the same flash-flash-pause-pause pattern, using the same clock division parameters as were used forthe 250-MHz reference clock LED. In this manner, the DCM LED can be used to see that the clock fromthe ADC is present and approximately the right frequency by comparing it to the 250-MHz LED. That is, ifthe sample rate clock to the ADC is 250 MHz, then the DCM LED and the 250-MHz LED flash at the samerate. If the sample clock to the ADC is 125 MHz, then the DCM LED flashes at half the rate of LED D4. Ifthe DCM LED is not flashing at all, then no clock is detected from the ADC. Common reasons for an ADCclock not to be detected are that the EVM is not powered on, or that the clock from the ADC is a CMOSclock instead of an LVDS clock due to an improper jumper setting.
LED D3 (labeled USB on the TSW1200EVM silkscreen) illuminates whenever the TSW1200 user interfacesoftware is accessing the FPGA by way of the USB connection. For longer FFT record lengths and lowerUART baud rates, the USB LED illuminates longer than for shorter FFT record lengths or higher UARTbaud rates.
8TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

TSW1200
6 V InputGround
D 7
ADC D1
D16
D4
D3
D2
DCM
USB
Lit when 6V
present
Lit after FPGA
reset completed
Flashes when clock
from ADC present
Lit during
USB access
Flashes while
200 MHz osc present
Lit when FPGA
finishes loading bit file
6 V I/O
3.4 Input Connections
3.4.1 Samtec LVDS Connector
www.ti.com Hardware Configuration
Figure 4. Position of LEDs
Figure 5 illustrates the position of the various input and output connections on the TSW1200EVM.
The connection between the TSW1200EVM and the ADC EVM to be tested is through a 120-pin Samtecconnector. Fourteen LVDS data pairs plus two LVDS clock pairs have a defined position in the connectorpinout that is common between the TSW1200EVM and many TI ADC EVMs. For the parallel LVDS DDRdata format, the bit clock runs at the same rate as the sample clock to the ADC. For the serial LVDS dataformat, the bit clock runs at a higher multiple of the ADC sample clock and is used to strobe the serialdata into the TSW1200EVM and then deserialize the data. For the serial LVDS data format, a secondclock is provided, called the frame clock or FCLK, that runs at the sample rate and is used to delineate the
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 9Submit Documentation Feedback

3.4.2 JTAG Connector
Hardware Configuration www.ti.com
sample boundaries in the serial data stream. The frame clock line is unused in the parallel LVDS DDRformat. In addition, 14 extra LVDS pairs are defined in the connector and routed to the TSW1200EVMFPGA for future expansion for larger parallel bus widths needed for faster data converters. The datadirection for the LVDS data pairs is always defined as the ADC EVM driving the signal through theconnector to the TSW1200EVM FPGA, with integrated 100- Ωtermination in the FPGA.
For one-channel parallel DDR bit-wise data formats, eight of the LVDS data pairs are used to support upto 16-bit-resolution ADCs at up to 250-MHz sampling rates. For one-channel parallel DDR sample-wisedata formats, 14 of the LVDS data pairs are used to support up to 14-bit-resolution ADCs at up to500-MHz sampling rates. For two-channel parallel DDR bit-wise data formats, 14 of the LVDS data pairsare used to support two channels of 14-bit resolution at up to 250-MHz sampling rate.
For serial data formats, eight of the LVDS data pairs support up to eight channels of one-wire serial ADCsat up to 65-MHz sampling rate or four channels of two-wire serial ADCs at up to 125-MHz sampling rates.All other LVDS pairs are ignored for this release.
Five extra CMOS single-ended signals are defined in the Samtec connector that are sourced from theFPGA through the connector to the ADC EVM. These signals are optionally defined to allow the FPGA(under control of the TSW1200 user interface software) control the SPI serial programming of the ADC forthose ADC EVMs that support this feature. For those ADC EVMs that support this feature, the SPI signalsSEN (SPI Enable), SCLK (SPI Clk), and SDATA (SPI Data) are sourced by the TSW1200EVM FPGA toallow the TSW1200 user Interface to configure the operational mode of the ADC under evaluation. ThereSPI signals are by default not connected on the ADC EVM until a 0- Ωresistor is installed on the EVM toenable control of the SPI port from the TSW1200 user Interface software. Two additional signals, SPIReset and SPI Power Down, are defined for possible future use.
The Samtec connectors snap together with no screws or other mechanism to hold the TSW1200EVM andthe ADC EVM together. The TSW1200EVM comes with standoff posts for setting the TSW1200EVM flaton a bench or table. The ADC EVM has shorter standoff posts so that the TSW1200EVM and ADC EVMwill lay flat on a bench or table and stay snapped together during use.
The TSW1200EVM includes an industry-standard JTAG connector that loops the JTAG ports of the FPGAand the FPGA EEPROM. Jumpers on the TSW1200EVM allow for either the FPGA or the FPGAEEPROM to be removed from the JTAG chain. The most frequent use for the JTAG connector is toprogram the TSW1200EVM FPGA. An FPGA programming pod can be purchased inexpensively fromXilinx™ to program the FPGA or the FPGA EEPROM.
The FPGA programming pod can be used to load a programming bit file directly into the FPGA for debugand development. However, once the FPGA is power-cycled or programmed by the PROGRAMpushbutton, this loaded FPGA bit file will be lost and the FPGA will revert to the bit file that is stored in theFPGA EEPROM. The FPGA programming pod also can be used to store a new FPGA programming bitfile in the FPGA EEPROM so that the TSW1200 can be upgraded as new revisions of FPGA firmwarebecome available.
The part number of the Xilinx Platform Cable USB programming pod that can be used to program orupgrade the TSW1200EVM is DLC9G. The programming pod operates from a USB port of a PC andconnects directly with the TSW1200 JTAG connector through a ribbon cable supplied with theprogramming pod.
10 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

TSW1200
6V InputGround
HEADER POSTS
HEADER POSTS
HEADER POSTS
HEADER POSTS
CLK
GND
GND
CLK
J6 ch 4
J21 ch 8
JTAG
LVDS
LVDS
USB
HEADER POSTS
HEADER POSTS
HEADER POSTS
HEADER POSTS
CLK
GND
CLK
GND
J18 ch 5
6 V I/O
J3 ch 1
3.5 Output Connections
www.ti.com Hardware Configuration
Figure 5. Position of Input, Output, and USB Connections
Two ways are available to output the parallel clock and sample data from the TSW1200EVM. The ADCsample data can be presented as a continuous stream of CMOS single-ended data on output headerposts, or a set record length of ADC parallel data samples can be captured in the TSW1200EVM FIFOsand output to a PC through the USB serial port. The data capture by the FIFOs and TSW1200 userinterface is the most convenient way to capture data from an ADC, but sometimes the continuous streamof data is desirable. For example, an application may require a larger capture depth for an FFT on amillion continuous data samples or more. For this, the output header posts are available so that a logicanalyzer can be used to capture ADC data in real time.
The pinout of the output data headers is shown in Figure 6 . In all cases, the output header is a standardtwo-row header of square 0.025-inch posts on 0.1-inch centers. One of the two rows of posts areconnected to ground down the whole row of posts, whereas the other row of posts are signal. Thesample-rate clock is presented on the first post, and after skipping one no-connect post (or three posts forChannel 1) the parallel data bus is presented from the least-significant bit (bit D0) through themost-significant bit. Two of the channels allow for as much as 16-bit data resolution whereas the other sixchannels provide for up to 14-bit data.
By default, the output headers are not enabled for parallel DDR data formats due to the potential highsample rates of up to 500 MHz.
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 11Submit Documentation Feedback

GND
GND
GND
GND
CLK
NC
D1
D2 D4
D3 D5
D6
D7
D8
D9
D10
D11
D12
D13
D14
D15
D0
J18 Channel 5
J19 Channel 6
J20 Channel 7
J21 Channel 8
J5 Channel 2
J4 Channel 3
J6 Channel 4
J3 Channel 1
CLK
NC
D1
D2 D4
D3 D5
D6
D7
D8
D9
D10
D11
D12
D13
D0
NC
NC
CLK
NC
D1
D2 D4
D3 D5
D6
D7
D8
D9
D10
D11
D12
D13
D0
NC
NC
NC
NC
CLK
NC
D1
D2 D4
D3 D5
D6
D7
D8
D9
D10
D11
D12
D13
D0
NC
NC NC
NC
3.6 USB I/O Connection
4 Software Installation
Software Installation www.ti.com
For the current release, four of the header posts are enabled for serial data formats. For up tofour-channel ADCs, the parallel deserialized sample data is presented on the corresponding outputheader. For eight-channel serial data formats, the TSW1200 user interface software selects whether tooutput the lower four channels of data on the output headers for Channels 1 through 4 or whether tooutput the data for Channels 5 through 8 on the output headers for Channels 1 through 4. Output headersfor Channels 5 through 8 are unused for this release.
Figure 6. Pinout of Header Posts for Parallel Output Data
Control of the TSW1200EVM is through a USB connection to a PC running Windows operating system.
For the computer, the drivers needed to access the USB port are included on the TSW1200 InstallationCD and are installed during the installation process. The USB is accessed as a virtual communication port(VCP) and shows up in the Hardware Device Manager as TSW1200 under COM ports and as TSW1200EVM under Multi-Port Serial Adapters. See Figure 8. On the TSW1200EVM, the USB port acts as a bridgeto UART control of the FPGA. Control of the FPGA is managed by reads and writes to a register map ofcontrol registers defined in the design of the FPGA. Normally, register writes from the TSW1200 userinterface software sets up the mode of operation of the FPGA. These register writes define such things asthe depth of FIFO to use for data capture or from which channel of an ADC to capture data. Then, a singleregister access triggers the filling of the capture FIFOs. Immediately after the capture FIFOs havecaptured the desired amount of data, the FIFO data is streamed back up the USB connection to theTSW1200 user interface software. The UART data rate between the FPGA and the USB port can be setto 115K, 230K, or 460K baud. A UART baud rate of 920K is not recommended for reliable data transfer.On first connection of the USB port to a computer, the Microsoft Found New Hardware Wizard appears.Follow the dialog box prompts as covered in the Software Installation section of this User’s Guide.
TSW1200EVM: High-Speed LVDS Deserializer and Analysis System12 SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

4.1 Installation CD
www.ti.com Software Installation
A software installation CD is included with the TSW1200 kit. This Installation CD includes all that isnecessary to install the TSW1200 user interface software on a computer that is running the Microsoft Vistaor Microsoft XP operating system. Included on the installation CD is this user’s guide, a setup program toinstall the user interface software, and a suite of MATLAB™ routines that can be used to interface to theTSW1200EVM hardware. All drivers necessary for the USB connection and the user interface softwareare included.
The user interface software is installed in two parts, which is done automatically once the setup program isrun. The actual user interface software is installed first along with the run-time engine that underlies theuser interface. Then, Virtual Com Port (VCP) is installed for the USB connection to the hardware.
Figure 7. TSW1200 User Interface Installation
If the TSW1200 user interface software is being installed on a machine that has an older version, TIrecommends that you first remove the old TSW1200 installation, using the Microsoft Add/RemovePrograms function in the Control Panel. The TUSB3410 VCP (Virtual Com Port) can be left installed, butduring the installation of the new TSW1200 user interface software, a dialog box appears to warn that theprevious TUSB3410 is about to be uninstalled. Click cancel for this uninstall to leave the old installation ofthe TUSB3410 VCP in place.
If the installation program encounters an older revision of the TSW1200 user interface software that wasnot removed with Add/Remove Programs, it removes the older version, and the Setup.exe must beexecuted again to install the newer version. If an existing TUSB3410 Virtual Com Port is on the machine,the Installation program may remove this also unless the removal is canceled when the dialog boxappears.
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 13Submit Documentation Feedback

4.2 USB Port and Drivers
Software Installation www.ti.com
To install the TSW1200 user interface software, run the setup.exe file that is contained on the installationCD. Figure 7 shows the more important dialog boxes that appear during the installation process. Foldersare created for the TSW1200 user interface and for the USB Virtual Com Port software. The default foldernames can be accepted in the dialog box as shown in Figure 7 , or a folder name and directory path canbe entered into the dialog box. A license agreement is required for the National Instruments LabVIEW™runtime engine that is used by the TSW1200 user interface software. It is unnecessary for the targetcomputer to have LabVIEW installed; the TSW1200 user interface software comes with the runtime engineand is completely self-contained.
When the Virtual Com Port is installed and the USB cable has connected the TSW1200EVM to the PC,the TSW1200 can be located in the Hardware Device Manager as shown in Figure 8 . The TSW1200appears as TSW1200 as a COM port. Each time the USB cable is connected, it is possible that a differentcommunication port is assigned. The TSW1200 user interface software locates the TSW1200 com portand opens communications with the TSW1200EVM. If the TSW1200 software is unable to find the port forthe TSW1200EVM, it displays an error message that the TSW1200 is not found. Usually, this is anindication that the user interface software was opened before the USB cable was plugged in or that poweris not supplied to the TSW1200EVM. In this case, ensure that power is applied to the TSW1200EVM andthat the USB cable is plugged in. Then close and re-open the TSW1200 user interface software.
Figure 8. Hardware Device Manager
When the USB cable connects the TSW1200 to the PC for the first time, the Microsoft Found New
14 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

4.3 Device ini Files
www.ti.com Software Installation
Hardware Wizard appears along with a series of pop-up windows. Select the default settings as shown inFigure 9 . If the warning dialog box appears and indicates that the software has not passed Microsoft LogoTesting, select Continue Anyway. The Found New Hardware Wizard usually appears twice, once to installthe com port and once to install the TSW1200 under Multi-Port Serial Adapters. The response to thedialog box prompts is the same in both cases.
When the TSW1200EVM USB cable is unplugged and plugged in again to a different USB port on thecomputer, the Found New Hardware Wizard may appear again, depending on whether that USB port iscontrolled by a different USB controller in the computer and depending on the registry configuration for thePC. If the Found New Hardware Wizard does appear again when the USB cable is reconnected, follow thesame responses to the dialog boxes as when the TSW1200 was first installed.
Figure 9. Found New Hardware Windows
Included in the installation for the TSW1200 user interface software is a subdirectory of ini files for eachcategory of ADC that is supported by the TSW1200EVM. TI strongly recommends that these files are notedited except at the factory. These files contain necessary information for the user interface software toproperly configure the TSW1200EVM FPGA registers for proper operation with the desired ADC EVM.Some of the entries within the ini file are obvious, such as defining the bit resolution for a device to be 11,12, 14, or 16 bits. Other entries in the ini file define for the FPGA which LVDS pairs within the Samtec
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 15Submit Documentation Feedback

5 User Interface
5.1 Toolbar
5.1.1 File Options
User Interface www.ti.com
connector define the data bus, and correct operation may not be possible if these entries are edited. Theuse of ini files allows for new device types to be supported by the TSW1200EVM as they becomeavailable without having to modify, re-release, or re-install the TSW1200 user interface software. Newdevice types may be supported at a later date simply by adding a new ini file to the proper subdirectory,using a script that TI will make available at that time.
When the TSW1200 user interface software is started, the initial setup screen of Figure 10 appears. Thestatus pane in the lower left reports on the firmware selection of the TSW1200EVM, the revision of thefirmware, and the revision of the user interface software. Many of the TSW1200 software controls areavailable from this setup screen, such as ADC device selection. However, a test must be selected beforethe TSW1200 software is ready for data capture from the TSW1200EVM. The tests available for thisrevision of the TSW1200 software are a logic-analyzer-style Time Domain test and a single-tone FFT test.At any time, this setup screen can be chosen from the test pulldown menu.
Figure 10. User Interface Initial Setup Screen
The toolbar contains options and settings that are independent of the device selected for test or the test tobe performed, such as configuration options and save/recall operations. The operations available underthe toolbar are grouped in categories of File, Instrument Options, Data Capture Options, and Test Options
Instrument Options contains all of the options for saving or importing test results and saving or recallingthe setup of the TSW1200 User Interface software.
The Save Capture operation can save the results that are displayed in the test window for export orarchival purposes. If the Single Tone FFT test is active, then the FFT plot is be saved, along with theperformance statistics and setup information. If the Time Domain test is active, then the Time Domain plotis saved along with the time domain statistics. The saved data plot can be saved in jpeg, png, or bmpformat.
The captured data from the ADC under test also can be saved to a file, or previously saved data can beimported back into the TSW1200 software and re-displayed in the Single-tone FFT or Time Domainwindows. Data can be saved in the format of binary, comma separated format, or both.
TSW1200EVM: High-Speed LVDS Deserializer and Analysis System16 SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

5.1.2 Instrument Options
5.1.3 Data Capture Options
5.1.4 Test Options
www.ti.com User Interface
The Instrument Options menu tab contains two options: Reinitialize Instrument and Read ADC EVM SetupProcedure.
The Reinitialize Instrument option causes the TSW1200 software to reset itself to initial conditions and toread the firmware settings from the TSW1200EVM again. The firmware and software revisions then aredisplayed in the status window. This command is useful for cases where the jumper settings on theTSW1200EVM may have changed since the TSW1200 software was opened, or in cases where the USBconnection to the PC has been dropped and the COM port must be reopened.
The Read ADC EVM Setup Procedure command causes the TSW1200 software to read a comment stringfrom the ini file for the ADC that is currently selected and then display that comment string in the statuspane. This comment string generally contains necessary setup information pertaining to the ADC EVM,such as possibly requiring a non-default data format or required jumper setting for the EVM tocommunicate properly with the TSW1200EVM.
The Data Capture menu tab contains two options: Capture Options and UART Baud Rate.
Capture Options allows for continuous capture that can be checked or unchecked. Continuous capturecauses the capture button to continuously capture, process, and display test data. Once the results aredisplayed, the capture process automatically begins again at the desired capture interval The defaultcapture interval is zero, which means that the next capture is started once the current results aredisplayed. The capture interval can be set in seconds to make the continuous capture less frequent.
The UART baud rate can be set to one of four data rates. This controls the data rate of unloading theFIFO capture from the TSW1200EVM FPGA to the USB port. The UART baud rate is the limiting factor forthe time it takes to capture long record lengths of data. The UART data rate is commonly set to 460800,as 921600 is not recommended for reliable data transfer.
The Test Options menu tab allows for setting the parameter options for the Single Tone FFT test and theTime Domain test. Later revisions of the TSW1200 software will allow for setting test parameters for aDual Tone FFT test and the ACPR test.
For a Single Tone FFT test, the RMS line may be enabled or disabled. When enabled, a horizontal markeris displayed over the FFT plot to indicate the RMS average of the noise floor of the FFT plot. The RMSaverage is computed over all of the FFT bins except the bin containing the input frequency. Moreprecisely, the RMS line = SINAD + FFT Record Length Process Gain where FFT Record Process Gain =10log(number of points/2).
SNR, SFDR, and SINAD can be expressed in either dBc or dBFS as selected by the dBFS selectionunder the Single Tone FFT options.
By default, the noise calculations for SNR and SINAD do not include the five FFT bins around theexpected input frequency or the first five FFT bins at DC. The rest of the FFT bins out to the Nyquistfrequency are included in the calculation of the total noise. Vertical marker cursors are present in the FFTdisplay that indicate the beginning and the end of the bandwidth of interest for noise calculations. Becausethese vertical markers are located at the extreme left-most and right-most positions on the FFT display,these vertical markers often are not noticeable. It is possible to compute the total noise power over anarrower range than the default DC through Nyquist band of frequencies. An integration band for the noisecalculations can be set in two ways. First, the mouse can be used to click on one of the vertical cursorsand drag it to a new desired location. Second, the Cursor Band Location window under the Single ToneFFT Test Option in the tool bar can be used to set a new frequency band of integration for the calculationof the total noise power.
Also excluded from calculation of the SNR is the power in the first five harmonics of the input frequency.These first five harmonics are included in calculation of SINAD (signal to noise and distortion) and thusthis is the principal difference between SNR and SINAD. (SINAD is sometimes called SNDR, signal tonoise and distortion ratio.) The number of harmonics to exclude from SNR can be set to a value other thanthe default 5 in the Number of Harmonics window in the Single Tone FFT Test Option in the toolbar.
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 17Submit Documentation Feedback

5.2 Status Window
5.3 Device Specific Selections
5.4 Single Tone FET
5.4.1 Parameter Controls
User Interface www.ti.com
The lower left portion of the TSW1200 user interface software window under the TI logo is reserved forreporting status, warnings, errors, and informational output. When the TSW1200 software is first run, itqueries the TSW1200EVM and displays the revision of the FPGA firmware and the type of ADC interfacethe TSW1200EVM is expecting to see based on jumper settings J10 and J11. At any time, this initialinformation can be displayed again by selecting the Reinitialize Instrument option in the InstrumentOptions tab of the toolbar.
During operation of the TS1200 software, warnings may appear in the status window if selections madefrom the drop-down menus of the interface are incompatible with the hardware selections or settings. If anADC selection is made that is not supported by the TSW1200EVM jumper settings of J10 and J11, then amessage in the status window prompts the user for a different device selection or to set the jumpers to theproper position on the hardware. If a sample rate is entered that is faster than that supported by aparticular ADC data sheet, then a warning is displayed in the status window.
Drop-down menus that are specific to a particular ADC device selection are located along the top of thedisplay under the toolbar.
The first selection a user makes is to select a type of ADC device from the device selection drop-downmenu. Each ADC that has an ini file installed in the proper directory automatically has an entry in thedevice selection drop-down menu.
Once an ADC device part number is selected, the ADC Channel can be selected in the Channel selectiondrop-down menu. The proper number of channel selections are made available based on the ADC deviceselection.
The format for display of the captured data is chosen in the Test Selection drop-down menu. Single ToneFFT displays the power spectrum of the captured data with calculated AC performance statistics. TimeDomain displays the raw captured data in the format of a logic analyzer display and output level over time.
In the Window Display drop-down menu, the user chooses a windowing function to be applied to thecaptured data. Rectangular Window applies a unity gain to all data points of the captured data. A HanningWindow, Hamming Window, or Blackman-Harris Window function can be applied to the captured data forsituations where the sample rate and the input frequency are not or cannot be set precisely to capture aninteger number of cycles of the input frequency (sometimes called coherent frequency).
The Capture button initiates a data capture once all other selections are made. The data capture can be asingle capture and display, or a continuous repeating capture.
The Single Tone FFT test is shown in Figure 11 . The larger central pane displays the FFT powerspectrum, whereas the calculated statistics are grouped into categories on the right of the screen. Settingsand inputs relevant to the test are entered in drop-down menus or text input boxes on the left portion ofthe window.
The sampling rate is entered in the ADC Sampling Rate text box, also called the Sampling Frequency FS.The number is entered in Hertz (Hz), although the letter M may be appended to represent the samplingrate in MHz. For example, 125M = 125 MHz or 125,000,000 Hz.
The expected input frequency is entered in the ADC Input Frequency input box, also known as FC. If theAuto Calculation of Coherent Input Frequency mode is enabled, then this input frequency is adjusted up ordown slightly away from the input frequency automatically. If coherent input frequency is required, thesignal generator used to source the input frequency must be set to this exact calculated coherentfrequency. The coherent frequency calculation takes the sampling frequency, the input frequency asentered by the user, and the FFT record length and adjusts the input frequency so that the captured datastarts and ends on the same place of the sine wave of the input frequency. This avoids an artifact of the
18 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback

5.4.2 FFT Power Spectrum
5.4.3 Single FFT Statistics
S
N
P
SNR = 10Log 10 P
(1)
S
N
P
SINAD = 10Log 10 P+PD
(2)
www.ti.com User Interface
FFT calculation from presenting a smeared power spectrum due to the fact that the FFT presumes thesample of the input is part of a continuous input signal. If the input and sampling frequency is notcoherent, and the sampled data is appended end to end to form a continuous input signal, then there is anapparent phase discontinuity at the beginning and the end of the sampled data. Making the sampling andinput frequencies coherent avoids this apparent discontinuity. If the input frequency cannot be madecoherent, then the windowing functions other than Rectangular can be used to process out this effect tosome degree.
The FFT record length can be set in the FFT Record Length (NS) input text box. The TSW1200EVMsupports FFT record lengths of as much as 65536 samples, or as little as 4096 samples.
The FFT power spectrum of the captured data is displayed in the major center portion of the window. TheTSW1200 software automatically scales the horizontal axis from DC through the Nyquist frequency,although the scale of the horizontal axis can be changed simply by highlighting the text and typing in anew value. For example, the display in Figure 11 can be used to zoom in on the input frequency byhighlighting the 0MHz and typing 25M, and then highlighting the 62.5M and typing in 35M. This causes theportion of the power spectrum from 25 MHz through 35 MHz to fill the power spectrum display.
The vertical scale of the power spectrum is automatically scaled to display the noise floor of the FFT resultup through 0 dBFS. The vertical scale can also be manually adjusted by highlighting the limits of thevertical scale and typing in new limits.
By default, the first few harmonics of the input frequency are marked in the display, as well as anadditional marker that can be placed by dragging the marker to any place in the power spectrum, such asa noise spur that is not already marked as a harmonic. By default this additional marker initially goes tothe highest spur that is not identified as a harmonic.
Display properties can be edited by using the mouse to right-click in the power spectrum display. Visibleproperties such as the graph palette or plot legend can be edited, and auto-scale of the vertical andhorizontal axises can be enabled or not.
For the Single FFT test, a number of calculated statistics and AC performance measurements aredisplayed to the right of the power spectrum display, grouped into several categories.
AC
Signal-to-Noise Ratio – SNR is the ratio of the power of the fundamental (P
S
) or input frequency to thenoise floor power (P
N
), excluding the power at DC and the first five harmonics. SNR is either given in unitsof dBc (dB to carrier) when the absolute power of the fundamental is used as the reference, or dBFS (dBto full scale) when the power of the fundamental is extrapolated to the converter’s full-scale range.
Signal-to-Noise and Distortion (SINAD) – SINAD is the ratio of the power of the fundamental (P
S
) to thepower of all the other spectral components including noise (P
N
) and distortion (P
D
), but excluding DC.
Spurious-Free Dynamic Range (SFDR) – SFDR is ratio of the power of the fundamental to the highestother spectral component (either spur or harmonic). SFDR is typically given in units of dBc (dB to carrier).
SFDR w/o 2,3 – Spurious-Free Dynamic Range without the second or third harmonic. Commonly thelargest spectral components after the fundamental are the second and third harmonics of the inputfrequency, and commonly the input frequency can contain significant power in the second and thirdharmonics. SFDR w/o 2,3 reports the SFDR with these two harmonics ignored.
Total Harmonic Distortion (THD) – THD is the ratio of the power of the fundamental (P
S
) to the power inthe first five harmonics (P
D
). THD is typically given in data sheets in units of dBc (dB to carrier).
SLAU212A – April 2007 – Revised August 2008 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System 19Submit Documentation Feedback

S
P
THD = 10Log 10 PD
(3)
SINAD 1.76
ENOB =
6.02
−
(4)
5.5 Time Domain
User Interface www.ti.com
Effective Number of Bits (ENOB) – The ENOB is a measure of a converter’s performance as compared tothe theoretical limit based on quantization noise
Time Domain
Several of the statistics of the Time Domain test are repeated here, particularly the minimum andmaximum sample values in the FFT record length, as well as the mean and standard deviation of thesample values
Signal
The frequency of the expected input signal is reported, as well as the power level of the signal in eitherdBc or dBFS. The amplitude of the input frequency for typical data sheet measurements is commonly setexternally to be about 1 dB below Full Scale, or –1 dBFS.
Distortion
The power values for the second, third, fourth, and fifth harmonics of the input frequency and theuser-selectable marker are displayed in either dBFS or dBc.
Test setup
Input parameters relevant to the test are repeated, particularly FFT record length, sample rate, and theend points of the bandwidth of integration for noise calculations. The lower end point for the bandwidth ofintegration is normally not zero because the first few FFT bins are not included to remove any DC biasingcomponent of the signal.
Figure 11. User Interface Single FFT format
The Time Domain test is shown in Figure 12 . The larger central pane displays the raw sampled datawhereas the calculated statistics are grouped into categories on the right of the screen. Settings andinputs relevant to the test are entered in drop-down menus or text input boxes on the left portion of thewindow.
20 TSW1200EVM: High-Speed LVDS Deserializer and Analysis System SLAU212A – April 2007 – Revised August 2008Submit Documentation Feedback