Application Note 1725 Authors: Gary Hendrickson and Michael Steffes Ultra Low Power Broadband 8 to 14-Bit Data Acquisition Platform ISLA112P50/55210EV1Z High Speed ADC/AMP Evaluation Board Evaluation Platform Overview • ISLA112P50 High Speed, Low Power ADC [12Bits,500MSPS] • ISL55210 High Performance, Low Power, Differential Amplifier • Compatible with Existing Intersil High-Speed ADC Evaluation Platform • Optional Measurement Port to ADC Inputs • Pin Compatible Family of 8- to 14-bit ADCs can be used. Performance • Clock Rate Range: 200MSPS to 500MSPS • 200mVP-P Input (-10dBm) for -1dBFS ADC Input • ±0.5dB flat response: 100kHz to 100MHz The ISLA112P50/55210EV1Z is an evaluation platform featuring Intersil’s high-speed FDA (Fully Differential Amplifier) (ISL55210) and High Speed, Low Power 12-bit, 500Msps ADC (ISLA112P50). The PCB is compatible with Intersil’s existing high speed ADC evaluation platform allowing for easy performance measurement and analysis. (See Intersil’s Application Notes AN1433, AN1434 for more information). The ADC evaluation platform consists of custom designed hardware and software. The function of the hardware is to provide power to the ADC and to excite and/or measure the appropriate analog and digital inputs and outputs. The software is required to configure the device for initial operation, to modify the device functionality or parameters, and to process and display the output data. Konverter software version 1.22c (or later) supports the ISLA11XP50 family and the ISLA112P50/55210EV1Z PCB. CONTACT THE FACTORY FOR ASSISTANCE IN USING THE KONVERTER SOFTWARE TO MODIFY THIS BOARD TO A DIFFERENT ADC. • Typical SNR: 65dBFS • Typical SFDR: 82dBC @ -1dBFS Input [10MHz to 100MHz Analog Input] Operating Precautions System Requirements IT IS STRONGLY RECOMMENDED TO INSERT THE +5V PLUG AT THE MOTHERBOARD PRIOR TO PLUGGING IN THE AC ADAPTER TO REDUCE THE POSSIBILITY OF POWER SURGES WHICH CAN DAMAGE THE PCB. PROBING ON THE PCB SHOULD BE DONE WITH CARE USING PROPER ESD TECHNIQUES WHILE HANDLING. • ISLA112P50/55210EV1Z Evaluation Board • KMB-001LEVALZ Intersil Motherboard (5V Supply Provided with Motherboard) • Intersil Konverter Software http://www.intersil.com/converters/adc_eval_platform/ • Low Jitter Clock Source • Bandpass Filter(s) USB To Host PC Mezzanine Connector REF IN Analog Input Low-Jitter RF Generator Test Bandpass Filter Daughter Card SRAM Optional Attenuator AMP 10MHz Reference Motherboard KMB001 +5V USB ADC FPGA Analog Input Low-Jitter RF Generator REF OUT Clock (200-500MHz) SRAM Optional Measurement Port Clock Inputs FIGURE 1. TYPICAL CHARACTERIZATION SETUP March 8, 2012 AN1725.0 1 CAUTION: These devices are sensitive to electrostatic discharge; follow proper IC Handling Procedures. 1-888-INTERSIL or 1-888-468-3774 | Copyright Intersil Americas Inc. 2012. All Rights Reserved Intersil (and design) is a trademark owned by Intersil Corporation or one of its subsidiaries. All other trademarks mentioned are the property of their respective owners. Application Note 1725 Hardware Motherboard There are two components in the hardware portion of the evaluation platform: the daughter card and the motherboard (Figure 1). The FDA and ADC are contained on the daughter card, which routes power from the motherboard and contains the analog input circuitry, clock drive and decoupling. The daughter card interfaces to the motherboard through a mezzanine connector. The motherboard contains a USB interface, an FPGA and SRAM. The motherboard serves as the interface between the host PC and the ADC daughter card. Most of the ADC functionality is controlled through the motherboard by the Konverter software. The FPGA accepts output data from the ADC and buffers it in the SRAMs before passing it to the PC at a lower speed for post-processing. The maximum buffer depth is approximately one million words. The only connections required for the motherboard are +5V power and a USB connection to the PC running Konverter Analyzer. No additional configuration of the motherboard is required. IT IS STRONGLY RECOMMENDED TO INSERT THE +5V PLUG AT THE MOTHERBOARD PRIOR TO PLUGGING IN THE AC ADAPTER TO REDUCE THE POSSIBILITY OF POWER SURGES, WHICH CAN DAMAGE THE PCB. PROBING ON THE PCB SHOULD BE DONE WITH CARE USING PROPER ESD TECHNIQUES WHILE HANDLING. The user must supply a low-jitter RF generator for the clock input to achieve the SNR shown here. Recommendations of suitable generators can be found in “Appendix A: RF Generators” on page 13. Note, a low cost, low jitter 500MHz clock source also is available from Crystek - the RFPRO33-500. Many low-jitter RF generators exhibit high harmonic spectral content relative to the ADC performance. A band-pass filter is recommended to attenuate the harmonics when testing the analog path. Software The software component is a Konverter Analyzer, a graphical user interface (GUI) created with MATLAB™. A MATLAB Component Runtime engine is supplied, which executes a compiled version of the m-files. Therefore, a stand-alone version of MATLAB is not required to run the Konverter Analyzer. The GUI controls the ADC configuration through the SPI port, reads data from the motherboard and performs post-processing and display of the output data. Data can be viewed in the time or frequency domain, and can be saved for later processing. Critical performance parameters such as SNR, SFDR, harmonic distortion, etc. are calculated and displayed on-screen when viewing in the frequency domain. Different regions of the FFT output may be easily viewed by zooming in using the settings available in the Konverter software. Initial Start-Up Referring to Figure 1, connect the daughter card to the motherboard by aligning the two matching mezzanine connectors. Four screws on the motherboard (not shown) align with mounting holes in the daughter card. Next, connect the clock source (~10dBm) which is required for communication to the Konverter software, then connect a test source or signal of interest coming from your signal channel at a maximum VP-P < 200mV (-10dBm). With the RF generators on, apply +5V power (minimum 5W supply) to the motherboard. The daughter card is powered by linear regulators on the motherboard. A USB cable should then be connected after +5V power has been applied. (Use the same USB port that was used when Konverter was initially installed so the hardware is recognized). 2 Software Start-Up The FPGA clock is derived from the ADC output clock, and the FPGA clock must be active for the software to run properly. Therefore, it’s important that the evaluation platform is powered and receiving a conversion clock prior to initializing the Konverter Analyzer Software. The compressed MATLAB files are unpacked the first time the GUI is invoked after installation. Subsequently, the graphical window will open more quickly. Complete information can be found in the KMB-001 Installer manual. http://www.intersil.com/converters/adc_eval_platform/ The main Konverter Analyzer window is shown in Figure 4. The application opens in FFT mode by default, but other modes can be selected using the radio buttons in the lower left corner. In each mode, relevant parameters are displayed in the data box on the left side of the window. The following parameters are displayed in the data box in all operating modes: • Fsamp: Sample clock frequency, automatically detected • Ffund: Input frequency, automatically detected (assumes sine wave single tone input) • Fund: Input amplitude (in dBFS) • Samples: Record length • Power: Total ADC power dissipation (FDA Power is in addition to this), as well as the voltage and current of each ADC supply Data Acquisition Press the Run button in the lower left corner of the screen to produce a single FFT plot. The Continuous check box can be selected to capture and display successive FFT plots. Multiple acquisitions can be averaged by selecting the Average check box. The number of averaged acquisitions defaults to 10, but can be changed in the Setup Conditions dialog. The FFT window also allows for tuning the number of frequency bins used in the FFT calculations. Menu items and the toolbar buttons may not function properly if data is being captured in Continuous mode. Stop the acquisition before selecting a menu item or using a toolbar button. AN1725.0 March 8, 2012 Application Note 1725 Menus AN1433 should be consulted for more information on the File, Setup, and Help software menus. AN1434 has information on software installation if needed. FIGURE 5. KONVERTER MENUS FIGURE 2. ISLA112P50/55210 DAUGHTERBOARD Analog Signal Path Description . The amplifier implementation on this daughterboard is intended to: 1. Terminate a single-ended input source with an AC-coupled broadband 50Ω impedance. 2. Convert the single-ended source to differential inputs for the ADC. 3. Provide extremely low noise and distortion amplification from a nominal -10dBm (200mVP-P single tone) input to a differential VP-P at the ADC inputs of approximately 1.3VP-P. This 6.45V/V gain (16.2dB) gain is a combination of input transformer step up, amplifier gain, and interstage filter insertion loss. FIGURE 3. DAUGHTERBOARD WITH KMB MOTHERBOARD 4. The interstage filter between the amplifier’s differential outputs and the ADC, also level shifts the Vcm output voltage from the amplifier to a lower Vcm voltage required by the ADC. ISLA112P50 (SN−000011) 30−Jan−2012 15:14:38, 25C 0 −20 The signal path circuit can be broken into an input side operation and then an output interstage filter. −40 dBfs Fsamp: 500 MHz Ffund: 70.12 MHz Fund: −1.12 dBFS SNRFS: 64.9 dBFS SNR: 63.8 dBc SFDR: 84 dBc + SINAD: 63.7 dBc THD: −82.2 dBc HD2: −88 dBc HD3: −84 dBc HD4: −95 dBc HD5: −98 dBc HD6: −100 dBc HD7: −98 dBc FIS −112 dBc OS −87 dBc ENOB: 10.3 ENOBFS: 10.5 Samples: 200000 Window: BH4T Power: 553 mW Ovdd 1.80V @ 123 mA Vdd2 1.80V @ 183 mA Figure 6 shows the input circuit for this daughterboard. Elements in green are optional and not populated for the nominal design board as delivered. −60 −80 +3 2 7 −100 6 10 9 8 OS 4 5 FIS −120 0 50 100 150 Frequency (MHz) 200 250 FIGURE 4. FFT OUTPUT WITH 70MHZ ANALOG INPUT Figure 4 shows a typical FFT for the PCB with a 70MHz Analog Input with the drive level adjusted to ~ -9.4dBm at the card edge resulting in a -1dBFS signal for the ADC. The ISLA112P50 is a low power 500Msps 12-bit ADC, more detailed information on the ADC can be found in the datasheet (http://www.intersil.com/data/fn/fn7604.pdf) FIGURE 6. POWER SUPPLY DECOUPLING AND INPUT SIDE TRANSFORMER/FILTER BEFORE ADC 3 AN1725.0 March 8, 2012 Application Note 1725 The input is AC-coupled through a 4.7μF blocking cap into the primary side of a 1:2 turns ratio step-up transformer; a Minicircuits ADT4-6T device. The blocking cap protects the source from accidental DC shorts while adding a 1.4kHz highpass pole (if the source is 50Ω). The 2nd transformer (ADTL1-475) acts as a common mode choke improving the differential balance for the amplifier stage. Ignoring it for a moment, the two 100Ω resistors (R1006 and R1007) act as both the gain resistors for the ISL55210 Fully Differential Amplifier (FDA) and the input termination elements. They drive into the differential summing junction of the FDA, which will appear as a broadband virtual ground for the differential signal. Since the input transformer is a 1:4 ohms ratio step up, those two 100Ω elements will add together to look like a 200Ω termination. They input refer through the transformer to appear as a 50Ω termination to the source. Detailed specifications on the 4GHz, 0.85nV/√Hz input noise ISL55210 may be found at – http://www.intersil.com/data/fn/fn7811.pdf Frequency Response Flatness Issues Each of the elements in the signal path have fine scale rolloffs that need to be considered to achieve the final ±0.5dB flatness through the 100kHz to 100MHz range. The ADT4-6T input transformer was selected mainly for its low frequency performance. While specified as -1dB flat from 150kHz to 200MHz, it actually does better than this in lab tests. Figure 7 shows the measured S21 response for several of the transformer options (for 1:2 turns ratio) possible on this board. Here, the ADT4-6T is clearly very flat down to 100kHz while hitting about -1dB at its specified 200MHz on this 2dB/div plot. These measurements included a resistor output network that looks like 200Ω to the transformer but 50Ω to the network analyzer – giving an insertion loss to the measurement but more accurately measuring the S21 flatness. The purely differential signal at the output side of 2nd transformer then gets through to the amplifier differential output gained up by the two 412Ω feedback resistors. A complete description of the noise, loop gain, and signal path balance of this implementation may be found in this series of 4 articles. • Part 1. Advantages to transformer input in a single to differential AC-coupled application. http://www.eetimes.com/design/analogdesign/4215415/Deliver-the-lowest-distortion-and-noise-in-alow-power--wideband--ADC-interface--Part-1-of-4• Part 2. Calculating integrated noise at the ADC for different filters and input pin SNR with the ADC for a net result. http://www.eetimes.com/design/analogdesign/4215416/Deliver-the-lowest-distortion-and-noise-in-alow-power--wideband--ADC-interface--Part-2-of-4• Part3. Distortion issues and combining SFDR at input pins with ADC for a net result. http://www.eetimes.com/design/analogdesign/4215417/Deliver-the-lowest-distortion-and-noise-in-alow-power--wideband--ADC-interface--Part-3-of-4• Part 4. Summary amplifier + ADC data on the Rev. A daughterboard and transformer modeling. http://www.eetimes.com/design/analogdesign/4215418/Deliver-the-lowest-distortion-and-noise-in-alow-power--wideband--ADC-interface--Part-4-of-4?Ecosystem=analog-design This article series was not including the 2nd common mode choke transformer (T2). Significant HD2 improvement at higher frequencies was found in adding this to the design. Thus, going from the Rev. A board described in the article series to the Rev. C final implementation, added that component. One of the options on this board is to eliminate that transformer and short across its pins with 0Ω resistors. 4 FIGURE 7. OPTIONAL INPUT 1:2 TURNS RATIO TRANSFORMERS MEASURED IN 50Ω SYSTEM The ADT4-6T is a good match to this 100kHz to 100MHz design. To modify this board to higher frequencies, the Macom MABA0096-CF48A0 seems to give a little better flatness range vs the ADT4-1WT from Mini-circuits. The Macom device measures very flat over a 1MHz to 200MHz range. The 2nd common mode choke transformer (ADTL-4-75, T2 in Figure 6) is extremely broadband, showing < 0.25dB insertion loss from DC to 100MHz. While not specified by MiniCircuits, since these elements are in fact a DC short, the loss below the minimum specified frequency of 500kHz actually decreases. For higher frequency designs, the pin compatible ADTL1-12 holds lower insertion loss through 200MHz. The amplifier will have its own frequency response from these source impedances and gain settings. One of the subtle advantages of this configuration is that the signal gain is higher than the amplifier noise gain. In the default circuit, the amplifier is operating at a signal gain from the output of the transformer that is simply the ratio of the feedback to gain setting resistors (4.12V/V here). However, looking at the noise gain for this Voltage Feedback Amplifier (VFA) implementation will also see the source element coming through the transformers as another 100Ω element on each side if the source signal is coming from a 50Ω impedance. Thus, the amplifier believes it is in a 1+ 412/200 = 3.06V/V noise gain while delivering a signal gain of AN1725.0 March 8, 2012 Application Note 1725 4.12V/V. These issues are described and explored in detail in the ISL55210 EVM board users guide available at http://www.intersil.com/data/an/an1649.pdf From that, a series of bandwidth vs Rf value curves were taken. These are using a different input transformer, ADT2-1T, that has a bit broader response on the high end. Stepping Rf is stepping the gain, with fixed 50Ω Rg resistors, which provide the match for the 1:2 ohms ratio input transformer used for this data. The Rf = 300Ω curve (between the 200Ω and 400Ω curves in Figure 8) is the same noise gain as that described on the ADC daughterboard where we see a very flat response through 100MHz. Amplifier to ADC Interstage Circuit and Vcm Issues Figure 9 shows the signal path from the amplifier outputs to the ADC inputs. 1 NORMALIZED GAIN (dB) 0 Rf = 200Ω -1 -2 Rf = 800Ω -3 -4 Rf = 400Ω Rf = 600Ω -5 -6 -7 -8 -9 1.E+06 1.E+07 1.E+08 1.E+09 FREQUENCY (Hz) FIGURE 8. MEASURED AMPLIFIER FREQUENCY RESPONSE vs GAIN SETTING (1VP-P OUTPUT) Increasing the amplifier gain does of course start to rolloff the frequency response as you move higher for the VFA design. The two input node capacitors in Figure 6 on page 3 (green, not populated 2.2pF), can be used as a fine tune on the output rolloff to the amplifier output pins. They were not used here as the response using the ADT4-6T and this relatively low gain setting on the ISL55210 did not require any equalization. This technique should be used with caution as it is also peaking up the output spot noise, which can quickly degrade the overall SNR for the solution. The ISL55210 Device information page includes a Spice Model that can be used to test these issues. http://www.intersil.com/products/deviceinfo.asp?pn=ISL55210 Power Supply Decoupling Issues The Daughterboard schematic of Figure 6 shows one example of good power supply decoupling for this single 3.3V supply device. Where possible, a large valued capacitor isolated towards this 4GHz amplifier with a high frequency ferrite and then another 1μF element provides a Pi filter on the board. Right at the device pins, the daughterboard uses two X2Y capacitors on each side of the package to get the best high frequency decoupling. Standard 0.01μF can also be used, which may reduce the HD2 performance at higher frequencies. FIGURE 9. INTERSTAGE CONNECTIONS FROM THE AMPLIFIER TO THE ADC The basic signal path is from the two amplifier output pins on the left through the differential RLC filter to the two ADC input pins on the right (Vinp and Vinm). The circuit here looks a little more involved than an actual end equipment implementation as it includes some optional features. The ISL55210 includes an internally regulated 1.2V common mode voltage generator. The circuit implemented here uses that and no connection to the external Vcm adjustment pin on the ISL55210 is required. These are the resistor and pot elements at the top of Figure 9. If a very high insertion loss filter needs to be tested on this board, where the required VP-P at the amplifier outputs exceeds 3VP-P, moving the default amplifier Vcm up will give more output swing range. This can be done by populating the optional amplifier Vcm adjustment elements (R1001 and R1005). On the lower part of Figure 9 is a logic gate to drive the disable line on the ISL55210. With the 50Ω termination tied to ground with no input signal, this 74AHC1G04 single inverter holds the amplifier enable line high (keeping it turned on) but it can certainly be connected to control signal input to test this feature. Interstage Filter Design and Vcm Bias Figure 10 shows an iSim PE simulation circuit for the filter and Vcm bias circuit implemented on the board as shown in Figure 9. This is a free Spice and Power simulator available in the Intersil web site at http://web.transim.com/intersil/iSimPE.aspx 5 AN1725.0 March 8, 2012 Application Note 1725 ISL55210 + ISLA112P50 daughterboard Interstage filter design and Vcm bias 47n 5% 1% R7 20k C4 L1 AC 1 V1 29.4 1% 1% R8 20k 4.7u 10% 47n 5% R4 C2 L2 1% R6 216 1% R5 216 -5 0.0013 I2 1% C1 10p 20% C8 1.9p -10 20% R2 500 0.0013 I1 578.655m dBV(I2-pos - I1-pos) / dB 4.7u 10% 29.4 1% R1 578.655m ISLA112P50 Input characteristic at 500MSPS -15 -20 ISLA112P50 Vcm model -25 422 1% 3.3 V2 107 1% R3 R11 859.455m 20 1% 1% R10 130 R9 0.535 V3 -30 10k 20k 50k 100k 200k 500k 1M 2M 5M 10M 20M 50M 100M 200M 500M 1G Frequency / Hertz FIGURE 10. SIMULATION MODEL FOR THE INTERSTAGE FILTER AND VCM SETUP The numbering does not follow Figure 9, but this is the same circuit with the ADC input and Vcm terms added. The amplifier output shows up as AC1 and the R7 and R8 elements are simulation artifacts to get DC operation points set. Aside from simply implementing a 2nd order differential RLC filter, this circuit also provides the bias voltage for the ADC input Vcm pins. The ISLA112P50 has a clock rate dependent common input current (shown in Figure 10 as two 1.3mA current sources). The network on the lower part of Figure 10 generates a DC bias voltage as a Thevenin equivalent from the converter’s Vcm output pin (modeled above as a 0.535V and 20Ω internal element) and the +3.3V supply. This voltage ends up being greater than the required voltage at the ADC inputs and it is pulled down through the two 216Ω resistors that are part of the filter design to be in range for the ADC. It is important to recognize that the filter design must be including the differential input impedance of the ADC – shown in Figure 10 as 500Ω in parallel with 1.9pF. Running the simulation with the DC probes included, shows this circuit provides a Vcm at the ADC of approximately 0.578V for this 500MSPS Icm. Remember the actual amplifier outputs are sitting at approx. 1.2Vcm and that is being level shifted to this ADC voltage through the two 4.7μF blocking caps. Figure 11 shows the simulated response for this design. FIGURE 11. SIMULATED DIFFERENTIAL FILTER RESPONSE The key information in this response is the midband insertion loss of only 2dB and the moderate peaking in the nominal response. It is very desirable to limit the insertion loss as this will reduce the maximum amplifier output VP-P to reach -1dBFS at the ADC. For instance, here we need 1.3VP-P at the ADC to be delivering a -1dBFS signal, which moves back through this filter to be a (10(2/20))*1.3V = 1.63VP-P at the amplifier outputs. This simulated response is peaking approximately 1.5dB but in the actual board, there are parasitic C’s in the interstage path that rolls this off to be perfectly flat. Measurement Port Option The Vcm bias setup resistors in the simulation circuit (216Ω) are actually implemented on the board as a passive sense path for the signal response right up to the ADC input pins. Going back to Figure 9 on page 5, those resistors are split into 4 elements that show the correct differential impedance for the filter, but divide down the differential signal at the ADC inputs while presenting two 25Ω source impedance elements into a very broadband 1:1 transformer (ADT1-1WT), where it is converted back to single-ended and delivered to a measurement port on the board. Using this technique, a direct network analyzer measurement of the response to the ADC inputs can be made. End equipment implementations would not use this feature and the sense path transformer can be eliminated from the design. However, this provides a relatively quick and easy way to measure the frequency response to the ADC inputs. Figure 12 shows a typical small signal measurement with the ADC clocking at 500MSPS, while Figure 13 shows the same thing running the ADC at 200MSPS. Note that the board as shipped does not have the measurement path populated; to install it add 0.1µF input capacitors C1017, C1018, the transformer and the transformer output DC blocking capacitor C1019, as shown in the schematic at Figure 39. Nearly identical results occur at the minimum clock rate of 200MSPS. 6 AN1725.0 March 8, 2012 Application Note 1725 2. -0.3dB midband insertion loss for the ADT4-6T 3. -0.25dB Insertion Loss for the ADT1-4-75 4. +12.3dB amplifier gain from the transformer outputs to the amplifier outputs 5. -2dB interstage filter insertion loss, which comes to 15.75dB nominal gain – very close to the measured 16.0dB result. Figures 14 and 15 show typical AC performance for the ADC alone as a function of analog input frequency. These figures are from the ISLA112P50 datasheet. 90 SNR (dBFS) AND SFDR (dBc) FIGURE 12. AMPLIFIER FREQUENCY RESPONSE AT 500MSPS 85 80 75 SFDR 70 65 60 SNR 55 50 45 40 0M 200M 400M 600M 800M 1G INPUT FREQUENCY (Hz) FIGURE 14. SNR and SFDR vs ANALOG INPUT FREQUENCY AT 500MSPS (ADC ONLY) HARMONIC MAGNITUDE (dBc) -40 -50 HD3 -60 HD2 -70 -80 -90 -100 0M 200M 400M 600M 800M 1G INPUT FREQUENCY (Hz) FIGURE 13. AMPLIFIER FREQUENCY RESPONSE AT 200MSPS The markers show very small deviation from 1MHz gain through this 100MHz span. These measured responses are from the board input SMA to the sense port output SMA. While not shown in Figures 12 & 13, the response going down to 100kHz remains flat to within 0.5dB to the 1MHz gain. The total insertion loss in the 4 resistor network, including two 50Ω impedance reflected through from the output pin load, is approximately 23.5dB for the resistor divider and then another 0.5dB insertion loss for the ADT1-1WT transformer. Adding this 24dB insertion loss to the -8.0dB 1MHz number measured above, gives a nominal gain from the input of the board to the ADC inputs of 16.0dB. Going from the input SMA to the ADC inputs, we should be seeing: 1. +6dB for the input ADT4-6T 7 FIGURE 15. H2 and H3 vs ANALOG INPUT FREQUENCY AT 500MSPS (ADC ONLY) The SNR curves are always referenced to full scale. Thus, if an SNR is developed at -1dBFS, the resulting number is increased by 1dB to project the SNR if the exact full scale input swing was being delivered. Generally, up through 100MHz, the ISLA112P50 typical data shows: 1. SNR of about 65.5dBFS 2. HD2 of about -90dBc 3. HD3 of about -86dBc These are remarkable numbers for an ADC consuming < 500mW at this maximum 500MSPS clock rate. AN1725.0 March 8, 2012 Application Note 1725 Amplifier/ADC PCB AC Performance The frequency response achieved here has exceptional flatness over this 100kHz to 100MHz range and then rolls off with a bit more than a 2nd order rolloff above that. This filtering is very important to: Very little degradation in ADC SNR is observed here. This is the combined result of very low output spot noise from this low input noise (0.85nV/√Hz) wideband FDA and good noise power bandwidth control in the interstage filter. -75 1. Control the integrated noise power bandwidth for the noise out of the amplifier stage. -80 -85 (dBc) 2. Help attenuate the HD terms present at the amplifier outputs. For instance, the filter shows about 10dB attenuation at 200MHz. This will be knocking down the amplifier HD2 by 10dB for a 100MHz input signal while the HD3 will be getting >10dB attenuation for frequencies above 67MHz. -90 The board supports a 3rd order filter design as well with the C16 element shown in Figure 9. The measurement port path provides an easy way to build and test alternate interstage filter designs. Contact the factory for assistance in redesigning this filter. Tested SNR and SFDR at 500MSPS The goal for this analog signal path before the ADC is to greatly reduce the required input signal level while suffering minimal degradation from lab tested ADC performance and provide a very broadband solution from 100kHz to 100MHz inputs. The ISLA112P50 shows an SNR over frequency (and these are always -1dBFS single tone test curves) in Figure 14 and HD2 and HD3 curves in Figure 15. These are normally taken with a 2 transformer input circuit using the ADTL1-12. Thus, the signal source is driving a very high amplitude single-ended test signal but is heavily filtered by narrow bandpass filters for each data point in Figure 15. The performance numbers shown are implicitly very narrowband due to the bandpass filter used in ADC characterization. Figures 16 through 18 show tested SNRFS, HD2 and HD3 for 3 example amplifier + ADC daughterboards tested at 500MSPS. BOARD #3 BOARD #2 -95 -100 10 30 110 -75 BOARD #1 -80 -85 BOARD #3 -90 -95 BOARD #2 65.1 10 30 50 70 90 ANALOG FREQUENCY (MHz) 110 FIGURE 18. TESTED HD3 FOR 3 BOARDS AT -10dBm BOARD EDGE SINGLE TONE POWER 64.9 (dBFS) 90 Quite a bit of variability in the measured HD2 is seen in Fig. 17 and should be expected due to the cancellation nature of HD2. However in any case, all 3 boards over 10MHz to 100MHz hold lower than the -80dBc combined performance. This is exceptional performance for a 115mW signal path solution. -100 65.3 BOARD #2 64.5 BOARD #3 64.3 Figure 18 shows that the boards are much closer in HD3 performance and we can detect a clear pattern in the response shape. As the frequency moves above 60MHz, the filter rolloff above 180MHz is really starting to attenuate the HD3 at the ADC inputs. Above 85MHz, the HD3 out of the amplifier appears to be coming up faster than the filter is rolling it off. In any case, all 3 boards hold about -80dBc worst case over this 10MHz to 100MHz span. BOARD #1 64.1 63.9 63.7 63.5 10 70 FIGURE 17. TESTED HD2 FOR 3 BOARDS AT -10dBm BOARD EDGE SINGLE TONE POWER 65.5 64.7 50 ANALOG FREQUENCY (MHz) (dBc) It is of course also possible to use the ADC to measure the frequency response through the FFT. A splitter is used at the input to measure the change in input signal power required over frequency to hold the output FFT amplitude for the single tone input constant over frequency. BOARD #1 30 50 70 90 110 ANALOG FREQUENCY (MHz) FIGURE 16. TESTED SNRFS FOR 3 BOARDS AT -10dBm BOARD EDGE SINGLE TONE POWER 8 Thus, the amplifier interface has degraded the SNRFS by about 0.5dB and the HD2 a varying amount but no more than 9dB, while the HD3 has been degraded by no more than 6dB. AN1725.0 March 8, 2012 Application Note 1725 Tested Performance at Lower ADC Clock Rates As the clock rate is reduced on the ADC, many fine scale changes in performance will be observed. The very simple interface of Figure 9 on page 5 develops the Vcm voltage for the ADC as a fixed Thevenin source that is then adjusted down by the common mode input current of the ADC. Measured results on this board show a 2.8µA/MSPS common mode current into each input of the ISLA112P50 ADC. As the clock rate decreases, the resulting Vcm voltage at the ADC inputs will be shifting up. This also has a lot of fine scale interaction with the ADC performance. Figure 19 shows a typical ADC Vcm vs Clock rate curve for the design shown here. Figures 20 through 37 give the typical SNRFS, HD2 and HD3 as the clock is stepped down in 50MHz steps for 3 different boards. And remember, all of these require a very low phase noise (or jitter) clock to get these results. These plots are mainly showing the variation coming from the ADC performance as the clock rate and input Vcm voltage is changing. The signal path up to the ADC inputs is not changing at all as the clockrate is being swept. These seem to show the SNR degrading more at low clock rates while the HD performance actually improves significantly. 1.0 0.8 Vcm (V) The goal here was to keep the interface as simple as possible while holding performance. A more sophisticated approach would be to include a Vcm servo loop to target a fixed ADC Vcm vs clock rate. It is, however, this increase in Vcm that limits the lower end of operation to 200MSPS before the ADC common mode voltage goes out of range for this particular set of values. However, 200MSPS also matches up with this 100MHz maximum application bandwidth for a 1st Nyquist zone implementation. 0.6 0.4 0.2 0 200 250 300 350 400 450 500 SAMPLE RATE (MSPS) FIGURE 19. ADC Vcm vs CLOCK RATE 9 AN1725.0 March 8, 2012 Application Note 1725 450MSPS 400MSPS 66.0 65.5 65.3 65.5 65.1 BOARD #2 64.9 BOARD #3 BOARD #1 (dBFS) (dBFS) 64.7 64.5 BOARD #2 65.0 BOARD #3 BOARD #1 64.5 64.3 64.1 64.0 63.9 63.7 63.5 10 63.5 30 50 70 90 110 10 30 ANALOG FREQUENCY (MHz) 110 FIGURE 21. SNR at 400MSPS FIGURE 20. SNR at 450MSPS -75 -75 -80 -80 -85 50 70 90 ANALOG FREQUENCY (MHz) -85 BOARD #3 (dBc) (dBc) BOARD #3 -90 -90 BOARD #2 BOARD #2 -95 -95 BOARD #1 -100 10 30 50 70 90 -100 10 110 BOARD #1 30 ANALOG FREQUENCY (MHz) -75 -80 -80 -85 (dBc) -85 (dBc) 90 110 FIGURE 23. HD2 at 400MSPS -75 BOARD #3 BOARD #1 BOARD #2 BOARD #1 BOARD #3 BOARD #2 -90 -95 -100 10 70 ANALOG FREQUENCY (MHz) FIGURE 22. HD2 at 450MSPS -90 50 -95 30 50 70 ANALOG FREQUENCY (MHz) FIGURE 24. HD3 at 450MSPS) 10 90 110 -100 10 30 50 70 90 110 ANALOG FREQUENCY (MHz) FIGURE 25. HD3 at 400MSPS AN1725.0 March 8, 2012 Application Note 1725 350MSPS 300MSPS 66.0 66.0 65.5 65.5 BOARD #3 (dBFS) (dBFS) BOARD #2 65.0 BOARD #1 64.5 BOARD #2 BOARD #3 BOARD #1 64.5 64.0 63.5 10 65.0 64.0 30 50 70 90 63.5 10 110 30 50 70 90 110 ANALOG FREQUENCY (MHz) ANALOG FREQUENCY (MHz) FIGURE 27. SNR at 300MSPS FIGURE 26. SNR at 350 -75 -75 -80 -80 BOARD #2 (dBc) (dBc) BOARD #1 BOARD #3 -85 -85 -90 -90 BOARD #3 -95 -100 10 30 50 70 90 BOARD #2 -95 BOARD #1 -100 10 110 30 ANALOG FREQUENCY (MHz) -75 90 110 -75 -80 -80 BOARD #2 -85 (dBc) (dBc) -85 -90 -90 BOARD #3 -100 10 70 FIGURE 29. HD2 at 300MSPS FIGURE 28. HD2 at 350MSPS -95 50 ANALOG FREQUENCY (MHz) BOARD #2 BOARD #3 BOARD #1 BOARD #1 -95 30 50 70 ANALOG FREQUENCY (MHz) FIGURE 30. HD3 at 350MSPS 11 90 110 -100 10 30 50 70 90 ANALOG FREQUENCY (MHz) 110 FIGURE 31. HD3 at 300MSPS AN1725.0 March 8, 2012 Application Note 1725 250MSPS 200MSPS 66.0 66.0 65.5 65.5 BOARD #2 BOARD #1 65.0 65.0 (dBFS) (dBFS) BOARD #2 BOARD #3 64.5 BOARD #3 64.5 BOARD #1 64.0 64.0 63.5 10 60 ANALOG FREQUENCY (MHz) 63.5 10 110 FIGURE 32. SNR at 250MSPS 30 50 70 90 ANALOG FREQUENCY (MHz) 110 FIGURE 33. SNR at 200MSPS) -75 -75 -80 -80 BOARD #1 -85 (dBc) (dBc) -85 -90 -90 BOARD #3 BOARD #3 -95 -95 BOARD #2 -100 10 BOARD #1 BOARD #2 -100 30 50 70 90 10 110 30 50 70 90 110 ANALOG FREQUENCY (MHz) ANALOG FREQUENCY (MHz) FIGURE 34. HD2 at 250MSPS FIGURE 35. HD2 at 200MSPS -75 -75 -80 -80 BOARD #2 -85 (dBc) (dBc) -85 -90 BOARD #3 BOARD #1 BOARD #3 -90 BOARD #1 BOARD #2 -95 -95 -100 10 30 50 70 90 ANALOG FREQUENCY (MHz) FIGURE 36. HD3 at 250MSPS 12 110 -100 10 30 50 70 90 110 ANALOG FREQUENCY (MHz) FIGURE 37. HD3 at 200MSPS AN1725.0 March 8, 2012 Application Note 1725 Conclusions TABLE 1. PCB SUPPORTED ADCs (Continued) This combination amplifier + DAQ/ADC board is essentially delivering a much lower full scale input solution with minimal degradation from ADC only performance. A rough set of specifications might be in the following: 1. 200mVP-P full scale input 2. 100kHz to 100MHz flat response 4. <600mW combined power dissipation 5. 65dB SNRFS 6. -80dBc worst case HD2, -80dBc worst case HD3 Considerable flexibility exists on the board to modify the input transformer selections, amplifier gain and interstage filter design to target different frequency ranges for digitization. The ADC itself can also be swapped out for numerous pin-compatible alternatives. These include 8-, 10-, and 12-bit pin-compatible parts in the ISLA1XXP50 family as well as 10-, 12-, and 14-bit pin compatible parts in the KAD55XX family, See Table 1 for a list of ADCs that this PCB supports. TABLE 1. PCB SUPPORTED ADCs RESOLUTION RESOLUTION MAXIMUM SAMPLE RATE(Msps) KAD5514P 14 125/170/210/250 grades KAD5512P 12 125/170/210/250 grades KAD5512HP 12 125/170/210/250 grades KAD5510P 10 125/170/210/250 grades Swapping out the ISLA112P50 for these other options may require a re-design of the interstage filter and Vcm setup circuit values. 3. 200MSPS to 500MSPS clock rate PART NUMBER PART NUMBER MAXIMUM SAMPLE RATE(Msps) In addition, the board will need to be re-programmed to allow for operation with the Konverter ADC software, contact the factory if you need assistance. In summary, this board offers a good breadboarding platform to get performance measures over a large range of designs before committing to end equipment layout. Appendix A: RF Generators Intersil uses the following RF generators as clock and signal sources when characterizing high-speed ADCs: • Rohde & Schwarz: SMA100A ISLA112P50 12 500 • Agilent: 8644B (with Low-Noise option) KAD5512P-50 12 500 ISLA110P50 10 500 KAD5510P-50 10 500 ISLA118P50 8 500 These generators provide very low jitter to optimize the SNR performance of the ADC under test. Other generators with similar phase noise performance can also be used. Contact Intersil Technical Support for recommendations. Bill of Materials DESIGNATOR C1, C6, C7, C31 COMMENT 33µF DESCRIPTION SIZE NUMBER MANUFACTURER Capacitor (Polarized) C-SIZE 4 Kemet MAN. PART NUMBER T491C336K016AT C2, C3, C1005 0.1µF Capacitor 0603 3 Murata GRM188R71E104KA01D C1007 4.7µF Capacitor 603 1 TDK C1608X5R1A475K/0.80 C4, C5, C12, C13, C14, C15, 0.1µF C20, C22, C24, C25, C26, C27, C33, C34, C42, C43, C44, C45, C48, C50, C52, C1014 Capacitor 0402 22 TDK C1005X7R1C104K C28, C29, C49, C51 Capacitor 0402 4 Panasonic ECJ-0EB1H102K 1000pF C30 10000pF Capacitor 0402 1 TDK C1005X7R1C103K C32 1000pF Capacitor 0603 1 AVX 06035C102KAT2A C1010,C1012 4.7µF Capacitor 402 2 TDK C1005X5R0J475K C53 220pF Capacitor 0603 1 TDK C1608C0G1H221J C1001 4.7µF Capacitor (Polarized) 3528 1 Kemet T491B475K010AT C1002 1.0µF Capacitor 1206 1 TDK C3216X7R1C105K/1.15 C1004, C1009 0.047µF Capacitor 0603_X2Y 2 Johanson Dielectrics Inc 160X14W473MV4T Cdiff 10pF Capacitor 0603_X2Y 1 Johanson Dielectrics Inc 500X14N100MV4T 13 AN1725.0 March 8, 2012 Application Note 1725 Bill of Materials (Continued) DESIGNATOR COMMENT DESCRIPTION SIZE NUMBER MANUFACTURER MAN. PART NUMBER INPUT SMA END LAUNCH Edge Launch SMA 1 Emerson 142-0701-851 J1 2MM HDR 14P SMT JTAG 1 Molex 87832-1420 J4 SMA SMA 2 Amphenol 901-144-8RFX J6 1 Molex 53475-1879 L3, L4, L13, L14, L15, L16, Bead L17, L18 connector 0805 8 TDK MMZ2012R102A L1001 BEAD 1206 1 Laird signal HZ1206E601R-10 L1002, L1003 47nH Inductor 0603 2 Coilcraft 0603CS-47NXGLU R1016, R1017 28.7Ω Resistor 0402 2 Panasonic ERJ-2RKF28R7X R1, R3, R4, R5, R6, R7, R8, 1k R9, R10, R11, R12, R13, R30, R35, R36, R38, R40, R41, R42 Resistor 0402 19 Vishay CRCW04021K00FKED R2, R29, R31, R37, R39 4.7k Resistor 0603 5 Yageo RC0603FR-074K7L R14, R28 1k Resistor 0603 2 Yageo RC0603FR-071KL R27, 200 Resistor 0402 1 Panasonic ERJ-2RKF2000X R1008, R1011 412Ω Resistor 0402 2 Panasonic ERJ-2RKF4120X R32, R33, R1000 0Ω Resistor 0603 7 Stackpole RMCF0603ZT0R00 R1013, R1014 29.4Ω Resistor 0402 2 Panasonic ERJ-2RKF29R4X R1019 392Ω Resistor 0402 1 Panasonic ERJ-2RKF3920X R48, R1006, R1007, R1021 100Ω Resistor 0402 4 Yageo RC0402FR-07100RL R1020 118Ω Resistor 0402 1 Panasonic ERJ-2RKF1180X R34 10k Resistor 0603 1 Yageo RC0603FR-0710KL R46 10k Resistor 0402 1 Stackpole RMCF0402FT10K0 R49, R50, R51, R52, R53, R54, R55, R56 49.9 Resistor 0402 8 Panasonic ERJ-2RKF49R9X R18, R1009, R1010 0Ω Resistor 0402 3 Panasonic ERJ-2GE0R00X R1015, R1018 187Ω Resistor 0402 2 Panasonic ERJ-2RKF1870X Rterm3 50 Resistor 0603 1 Panasonic ERJ-3EKF49R9V T1 ADT4-6T Transformer CD542 1 Minicircuits ADT4-6T T2 ADTL1-4-75+ 1:1 Transmission Line Transformer CD542 1 Minicircuits ADTL1-4-75+ T3 TC4-1W Center-Tapped Transformer AT224 1 Minicircuits TC4-1W+ U7 74AHC1G04 Inverter SOT23-5 1 TI SN74AHC1G04DBVR U2, U3 24FC128-I/SN EEPROM SO8 2 Microchip Tech 24FC128-I/SN U4 XC2C64A6VQG44C CPLD VQFP44 1 Xilinx XC2C64A-5VQG44C U6 ISL55210 Hi Speed Ultra Low Distortion Differential Amplifier TQFN16 1 Intersil ISL55210IRTZ U1 ISLA112P50 500MSPS, 12-bit ADC QFN 10X10 72 1 Intersil ISLA112P50IRZ Intersil Corporation reserves the right to make changes in circuit design, software and/or specifications at any time without notice. Accordingly, the reader is cautioned to verify that the Application Note or Technical Brief is current before proceeding. For information regarding Intersil Corporation and its products, see www.intersil.com 14 AN1725.0 March 8, 2012 C13 C15 C20 C22 C25 C26 C42 C43 C44 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF 0.1uF C12 C36 DNP C37 DNP SDO CSB OVDD SDIO C35 DNP C27 0.1uF SCLK C24 0.1uF pin36 OVDD pin27 OVDD C14 0.1uF C45 pin56 pin70 pin23 OUTFMT pin22 nap_sleep_normal output_mode pin16 clkdivn pin15 Vcm pin71 pin24 AVDD pin19 AVDD AVDD pin12 AVDD pin6 pin1 AVDD AVDD ISLA112P50/55210EV1Z Schematics C38 DNP 0.1uF Bead L14 Bead C9 DNP C10 DNP 2 3 4 5 SCL SDA AVDD GND GND C11 DNP 6 7 8 9 10 Vinm Vinp GND AVDD OVDD DGND 56 55 ORP ORN D13P D13N D12P D12N D11P D11N DGND 64 63 62 61 60 59 58 57 65 OVSS ORP ORN D11P D11N D10P D10N D9P D9N SDIO SCLK CSB SDO 69 68 67 66 D8P D8N D7P D7N D6P D6N DNC DNC DNC DNC AVDD AVSS AVSS CLKOUTP CLKOUTN VINN VINP RLVDS OVSS AVSS AVDD DNC DNC VCM CLKDIV DNC DNC AVDD CLKP CLKN Vcm clkdivn 11 12 13 14 15 16 17 18 AVDD OVDD OVSS L13 C8 DNP SCL SDA D5P D5N D4P D4N D3P D3N D2P D2N 54 53 52 51 50 49 D10P D10N D9P D9N D8P D8N 48 47 CLKOUTP CLKOUTN 46 10K 45 DGND 44 43 42 41 40 39 38 37 R46 D7P D7N D6P D6N D5P D5N D4P D4N OVDD Bead SYNC_RES_N SYNC_RES_P DNC DNC D0N D0P D1N D1P Bead L3 RESETN OVSS OVDD L4 SD0 U1 ISLA112P50 Application Note 1725 SC0 1 OUTMODE NAPSLP AVDD AVDD SDIO SCLK CSB SDO EP AVSS AVDD OUTFMT 0 GND 15 72 GND 71 AVDD 70 OUTFMT 0.1uF OVDD36 D1N D1P D2N D2P D3N D3P 28 29 30 31 32 33 34 35 25 DGND26 OVDD27 RESETN 22 output_mode nap_sleep_normal23 AVDD24 clk_inp clk_inn AVDD19 20 21 Under DUT. L15 bead L16 bead R16 L17 bead L18 bead D0N D0P DNP R17 DNP anlg_1.8V HI= Gray_Code LO= Unsign Ft=Twos_Comp R15 DNP anlg_1.8V R25 DNP OUTFMT R14 1K HI=2mA LVDS LO= LVCMOS Ft= 3mA LVDS anlg_1.8V R22 DNP output_mode R24 DNP HI= clk_div4 LO= clk_div2 Ft=clk_div1 anlg_1.8V R26 DNP clkdivn R23 DNP HI= Nap LO = Normal Ft= Sleep nap_sleep_normal R28 1K INTERSIL PROPRIETARY AND CONFIDENTIAL. SUBJECT TO NONDISCLOSURE AGR Title Size B FIGURE 38. ISLA112P50/55210EV1Z (1 OF 3) ISLA112P50/ISL55210 Number page1 Revision C AN1725.0 March 8, 2012 ISLA112P50/55210EV1Z Schematics (Continued) ANLG_3.3V 2 R1000 2 1 L1001 1 +anlg_3.3V C1001 + BEAD 4.7uF 0 R1001 VS_AMP C1002 1.0uF 5kohm/DNP C1004 R1003 2.05kohm/DNP C1003 DNP X2Y 0.09uF VS_AMP R1002 20ohm/DNP 34 5 NC NC Rterm1 DNP 4 R1007 0 100ohm Rbypass2 1 2 DNP 2 1 Rbypass4 1 2 DNP Cterm2 2.2pF/DNP 3 Vi+ R1010 0ohm 4 FBR1011 412ohm VCM 13 15 14 Vs+ Vcm R1008 412ohm 1 Fb+ R1009 0ohm 2 Vi- R1005 7.68kohm/DNP Vo+ NC NC Vo- R1021 130ohm C1010 R1013 12 Vinm 29.4ohm 11 L1002 47nH 4.7uF C16 DNP 10 R1014 9 29.4ohm Cdiff 10pF L1003 47nH C1012 C1015 DNP C1016 DNP To ADC Vinp 4.7uF R1015 187ohm 2 C1017 VS_AMP R1016 28.7ohm 0.1uF C1009 X2Y 0.09uF ADT1-1WT Vtest C1019 C1008 DNP R1017 28.7ohm 1 C1018 R1018 187ohm R1016, R1017 default populate as 1uH Inductors U7 1 2 3 0.1uF 2 VS_AMP NC VCC 5 A GND 74AHC1G04 Y 4 Rterm2 50ohm/DNP Pd 1 2 DNP Rterm3 50ohm FIGURE 39. ISLA112P50/55210EV1Z (2 OF 3) 0.1uF Application Note 1725 1 Cterm1 2.2pF/DNP GND 100ohm R1019 422ohm R1020 107ohm C1014 1uF U6 ISL55210 GND 25 R1006 . R1004 10kohm/DNP 8 2 3 R18 ADTL1-4-75 6 Pd 2 4.7uF 16 Vs+ 1 7 . ADT4-6T C1007 1 2 6 INPUT Rbypass3 1 2 DNP EP GND 1 GND 16 2 5 Rbypass1 1 2 DNP 1 0 16 C1005 0.1uf AN1725.0 March 8, 2012 ISLA112P50/55210EV1Z Schematics (Continued) anlg_3.3V Anlg_5V RESETN D13P D13N R37 4.7K D11P D11N 0.1uF R39 R3 1K D5P D5N D4P D4N D3P D3N D2P D2N VCC VCC 0.1uF ID EEPROM U3 1 2 3 4 WP A0 A1 A2 Vss 8 7 6 5 Vcc WP SCL SDA WP_2V SC2 SD2 24FC128-I/SN U4 =value dig_1.8V IO(2) IO(2) IO(2) IO(2) I(2) GND IO(1) VCC IO(1) IO(1) IO(1) R2 4.7K 23 24 TDO 25 26 dig_1.8V SCLK 27 SDO 28 R5 1K 29 R6 1K 30 31 SDIO CSB 32 33 C3 dig_1.8V SPI_CONF R35 1K R1 1K D6P D6N 0.1uF IO(2) TDO GND VCCIO2 IO(2) IO(2) IO(2) IO_GLB_S/R IO_GOE IO_GOE IO_GOE R7 1K TCK TMS TDI IO(1) VCCIO1 I/O(1) I/O(1) GND I/O(1) I/O(1) I/O(GCK) 11 TCK 10 TMS 9 TDI 8 R9 1K 7 VCC 6 R10 1K 5 R11 1K 4 3 R12 1K 2 R13 1K CPLD SPARE1 1 C2 D1P D1N D0P D0N dig_1.8V 0.1uF SPARE EEPROM U2 1 2 3 4 PC3 A0 A1 A2 Vss Vcc WP SCL SDA 4.7K 4.7K D7P D7N C5 0.1uF 8 7 6 R33 5 R32 WP_2V 0 0 R31 R29 dig_1.8V CLKOUTP CLKOUTN C34 0.1uF 10K D8P D8N C33 dig_1.8V 22 21 20 19 18 R4 1K 17 16 R8 1K 15 dig_1.8V 14 R41 1K 13 R40 1K 12 PC1 D10P D10N dig_1.8V C1 33uF dig_1.8V 4.7K D9P D9N dig_1.8V C4 R34 D12P D12N PORn_ExtResetn_fpga 33uF C7 ORP ORN dig_1.8V dig_1.8V VCC SC0 SD0 24FC128-I/SN PC6 PC7 PC0 daughter card detected PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC4 PC5 PC2 WP SPI_master_drive CPLD SPARE1 SCLK_3V CSB_3V MISO_3V MOSI_3V SD0 SC0 SD2 SC2 PORn_ExtResetn_fpga MH1 MH2 MH3 MH4 JTAG connector PLCD Programing J1 DGND GND Tooling Hole Tooling Hole Tooling Hole Tooling Hole 1 3 5 7 9 11 13 2 4 6 8 10 12 14 VCC TMS TCK TDO TDI INTERSIL PROPRIETARY AND CONFIDENTIAL. SUBJECT TO NONDI 2MM HDR 14P SMT Title Size B AN1725.0 March 8, 2012 Date: File: 53475 1879 FIGURE 40. ISLA112P50/55210EV1Z (3 OF 3) ISLA112P50/ISL55210 Number R page3 16-Nov-2011 Sheet of C:\Documents and Settings\GHENDRIC\Desktop\New Drawn By: Fol Application Note 1725 AVDD 33uF anlg_1.8V C31 C_vdd3_anlg Dig_5V IO_GOE VAUX IO(2) IO(2) IO(2) IO(1) IO(1) IO(1) IO(1) IO(GCK) IO(GCK) 17 DGND GND AVDD 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 126 128 130 132 134 136 138 140 142 144 146 148 150 152 154 156 158 160 162 164 166 168 170 172 174 176 178 180 34 35 36 37 38 39 40 41 42 43 44 33uF C6 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39 41 43 45 47 49 51 53 55 57 59 61 63 65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99 101 103 105 107 109 111 113 115 117 119 121 123 125 127 129 131 133 135 137 139 141 143 145 147 149 151 153 155 157 159 161 163 165 167 169 171 173 175 177 179 R30 1K VCC R36 1K R38 1K WP_2V SCLK_3V CSB_3V SPI_master_drive R42 1K MISO_3V MOSI_3V J6 dig_1.8V OVDD OVDD