GUI7001 Software User Manual - Hitachi Kokusai Electric America
Contents
1. Figure 22 Symbols Measurement Spectrogram This is a three dimensional measurement representing the Power Density Z axis vs Frequency X axis vs Time Y axis The viewing angle can be rotate to any position Spectrogram Power Spectral Density dB Hz Time 5 3 0 3 6 9 Frequency MHz Figure 23 Spectrogram Measurement Spectrogram Contour That s the Spectrogram in the X Frequency Y Time view where the color gives the Power Density Spectrogram Contour 0 9 Normalized Time 5 to cn mm j o0 bo 0 1 9 6 3 0 3 G g Frequency MHz Figure 24 Spectrogram Contour Measurement Time Domain This measurement shows both input and feedback IF signals at the same time in a time domain The information bar will show two additional pieces of information e Error Shows the Mean Squared Error between the Input and Feedback signals Time Domain Input Feedback Normalized Amplitude 35 40 45 50 Time us Figure 25 Time Domain Measurement 3 4 Calibration Every AT7001 unit presents a internal linear distortion on feedback path This distortion is compensated during linear pre correction However is not present at the mask filter It generates an error that may be avoided using the calibration process 1 Set the output power do 0 dBm 2 Make an external loop connect the RFoutput on After Filter Feedback Sample in2. Center 563 00 MHz i Span 20 00 MHz Res BW 30 kHz FVBW 100 Hz sweep 5 20 s 1001 pts Figure 8 Crest Factor Reduction 3 2 Linear Pre Correction This is the process is used to compensate for linear distortions for example not constant frequency response and group delay inside the channel A Mask Filter is used at a transmitter output to limit the signal bandwidth and comply with an emission mask The sharper it is the greater the group delay distortion A small amount of attenuation is also introduced at channel edges Linear pre correction is used to compensate for these effects FP When the Linear Pre Correction screen is opened the software read and display the pre correction curves stored on the non volatile memory The next figure shows the linear pre correction main window Linear Industries Inc 2 C Enable Pre Correction 6 Automatic Linear Pre Correction Pre Correction Properties Frequency Response Magnitude dB Automatic Algorithm 0 Frequency MHz Group Delay Group Delay ns Figure 9 Linear Pre Correction Window S a Reset Download an all pass filter Used to bypass the linear correction Enable Linear Pre Correction Used to bypass the linear correction or not without downloading an all pass filter Run Single iteration algorithm to compensate for linear distortions Save Button used to save the result on the non volatile memory Back Close curre
3. 6 dB but for very few time or 0 07 of time the amplifier will drive a power above 4 kW It is clear that the amplifier compressed the high amplitudes because the red line is below the blue one Use this measurement to evaluate how much an amplifier has compressed or clipped a signal The information bar will show the additional information e PAPR Stands for Peak to Average Power Ratio and represents the ratio between the maximum peak power found and the average power Shown for both Input and Feedback signals Complementary Cumulative Distribution Function Instantaneous Power Average Power Ratio dB Figure 15 CCDF Measurement Constellation the constellation is a diagram of the sampled baseband modulating symbols where the Cartesian plane is plotted by its real vs imaginary part The real part In Phase of those sampled symbols carry the transmitted information while the imaginary part Quadrature carry the necessary information to generate the vestigial side band For 8VSB modulation the eight regions represents the eight possible symbols and the seven dashed lines indicates the decision threshold This measurement reflects the signal quality Use it to realize non linear linear and amplitude distortions The information bar will show the additional information e MER Modulation Error Rate It is the mean squared error in dB considering the perfect symbol as the reference signal Constella
4. Compensation unchecked There differences between IOT and Solid State amplifiers regarding its memory non linearities As a rule of thumb Type A achieves better results for IOT amplifiers and Type B for Solid State 7 Click on Run The whole process consists of single iteration After a few seconds the results will be applied automatically Both algorithms works independently gt Crest Factor Reduction will be automatically bypassed when the algorithm is running 8 Once the correction performed satisfied results click on Save push button to save the curves on the non volatile memory gt After running the algorithm the presence of unbalanced shoulders means that there are memory effects left 9 If Crest Factor Reduction option is enabled make it active on AT7001 front panel and set a reduction level When the amplifier operates above the saturation level i e Clipping the signal peaks the CFR may reduce intermodulation However be aware there is a trade off between intermodulation reduction and MER using CFR The more the reduction the worst the MER The next Figure shows typical results using the Memoryless and the Memory Pre Correction Usually the Memory correction achieves better results than the Mermoryless correction Ret 8 dBm Htten 10 dB Peak Center 563 MHz hes BH 3 kHz VEH 360 Hz sweep 1 565 5 441 pts Figure 6 Memoryless vs Memory Comparison A There is no guarantee to achieve bett
5. amplifier gain in dB vs normalized input amplitude in dB Full Scale The blue points show the Instantaneous Transfer Function as it has a dynamic behavior Their spread characteristic is caused by noise and memory effects The red line is the Memoryless Transfer Function that best fits the blue points This is a powerful tool to evaluate a power amplifier transfer function and thus its linearity Check how spread the blue points are in order to realize how much memory effects the amplifier has Always perform this measurement using Feedback Before Filter sample The information bar will show the additional information e Peak to peak amplitude distortion Shows the Memoryless Transfer Function peak to peak value AMAM Instantaneous Transfer Function l Memoryless Transfer Function Gain dB 30 25 20 15 10 5 0 Normalized Power dBFS Figure 13 AM AM Measurement AM PM AM PM stands for Amplitude Modulation Phase Modulation it represents the phase deviation in vs normalized input amplitude in dB Full Scale The blue points show the Instantaneous Transfer Function as it has a dynamic behavior Their spread characteristic is caused by noise and memory effects The red line is the Memoryless Transfer Function that best fits the blue points This is a powerful tool to evaluate a power amplifier transfer function and thus its linearity Check how spread the blue points are in order to realize
6. information e Peak to peak group delay distortion self explained Group Delay Group Delay ns Frequency MHz Figure 20 Group Delay Measurement gt Crest Factor Reduction will be automatically bypassed when this measurement is being performed Power Spectrum Density This is a spectral analysis using the Power spectrum Density estimation of the IF signal The in band signal average power is normalized to be the O dB point The information bar will show the additional information e Shoulders attenuation 3 25 MHz shows the shoulder attenuation 3 25 MHz from channel center frequency relative to the in band average power with a 500 kHz equivalent resolution bandwidth Power Spectral Density es i b co oo co 40 Power frequency sample dBrHz 50 r 4 F J 0 J B 9 Frequency MHz Figure 21 Power Spectrum Density Measurement Symbols This measurement represents the purely real demodulated symbols over time Figure 22 shows the demodulated symbols after a mask filter without linear pre corrections and then with Straight indicates a high MER signal The information bar will show the additional information e MER Modulation Error Rate It is the mean squared error in dB considering the perfect symbol as the reference signal alas ae ee weg t ow Nes an eS Amplitude 0 500 1000 1 500 2000 2500 3000 3500 Symbol Number
7. on measurements like Frequency Response Group Delay MER Constellation etc
8. window Measurement Window This Figure shows the last active measurement File gt Export Use this option to export the measurement to one of the following formats e emf gt Enhanced Metafile e eps gt Encapsulated Postscript e jpeg gt Joint Photographic Experts Group e pdf gt Portable Document Format Information Bar This bar shows additional information regarding the last measurement Amplitude Modulus Shows the normalized complex envelope modulus for the Input and a Feedback signal where I nj O in is the modulus for a complex signal with real part Jin and imaginary part n It is useful to evaluate the signal compression caused by the amplifier non linearity as it gets close to the saturation region Perform this measurement using the Feedback Before Filter sample Figure 12 shows a signal that has been compressed at the saturation level and expanded at the turn on region typical in class AB amplifiers The information bar will show two additional important pieces of information e Error Shows the Mean Squared Error between the Input and Feedback signals e Peak Compression shows in dB how much the maximum peak found was compressed Amplitude Modulus Normalized Amplitude 10 Time us Figure 12 Amplitude Modulus Measurement AM AM AM AM stands for Amplitude Modulation Amplitude Modulation it represents the normalized output inout amplitude ratio
9. 3 Bypass the Non Linear and Linear Pre Corrections 4 Open the Signal Measurements window and Measure Frequency Response between Input and Feedback After Filter The result is going to be the internal linear distortion to be calibrated Measurement Software E Frequency Response Signal Selection Input Feedback Before Filter Feedback After Fitter Measurements _ Persist Frequency Response v Magnitude dB 0 Frequency MHz Peak to peak frequency response distortion 0 76 dB Figure 26 Internal feedback linear distortion 5 At the main screen click on Advanced and the on Calibration The Calibration window will open 6 Press Run the calculate the distortion The inverse response will appear compared to the last Frequency Response measurement Next Figure shows the results Feedback Path Calibration E Frequency Response Calibration Reset Magnitude dB 0 Frequency MHz Group Delay Non Yolatile Memory Group Delay ns 0 Frequency MHz READ THE USER MANUAL before performing this operation Figure 27 Calibration window 7 Press Save on the Non Volatile Memory Panel Close the Calibration window 8 Back to the measurement window measure the frequency response again Now with the compensation taking place a flat response will appear The calibration will take effect on Linear Precorrection and
10. D Athlon 64 Windows Server 2008 AMD Opteron AMD Sempron To install MCR perform the following steps 1 Install the MATLAB COMPONENT RUNTIME Execute the MCRinstaller exe file located in the MATLAB Component Runtime folder 2 Install the USB Driver Execute the CDM 2 04 06 exe file located in the USB Driver folder This is a Virtual COM Port driver Microsoft WHQL certified and compatible with the following operating systems e Windows Server 2008 e Windows Server 2008 x64 e Windows Vista e Windows Vista x64 e Windows XP e Windows XP x64 e Windows 2000 e Windows Server 2003 e Windows Server 2003 x64 ie MATLAB R Compiler Runtime 8 InstallShield Wizard Welcome to the InstallShield Wizard for MATL AB __ mattas r compiler Runtime 7 8 Compiler Runtime The InstallShield R Wizard will allow you to modify repair or remove MATLAB R Compiler Runtime 7 8 To continue click Next The MathWorks Figure 1 MATLAB Component Runtime Installation Window 3 Operation Before running the software be certain that the AT7001 is on and USB cable is connected between the exciter and PC gt The exciter USB Port is located at the rear panel Run the executable file GUI7001vXXX exe located in the GUI7001 folder where XXX indicates the version The main screen will appear as follows Linear Industries Inc Graphic User Interface E Communication Settings Port Software Ver
11. Linear Industries Inc GUI7001 Software User Manual For the AT7001 Digital Exciter f f POPA oe s ah hee Po DIGITAL EXCITER sge sne atata Rev E October 2009 Revision History REV DESCRIPTION DATE BY A Initial Version 2008 10 08 Henry Douglas B Udated to new non linear correction algorithm 2009 02 02 Henry Douglas C Crest Factor Reduction Type A and Type B Models 2009 07 27 Henry Douglas D New layout CCDF description correction 2009 09 09 Henry Douglas E Connection message changed 2009 10 19 Henry Douglas How to Contact Linear LINEAR INDUSTRIES INC NORTH AMERICA DIVISION 2531 Technology Drive Suite 310 Elgin IL 60124 Web site www linear tv com Phone 847 428 5793 Toll Free 847 428 5793 Fax 847 428 5785 About this document i gt Indicates a hint A Indicates important information 1 Introduction GUI7001 is the software developed by Linear Industries Inc and is used to perform non linear and linear pre correction and signal measurements with the AT7001 ATSC 8VSB digital exciter 2 Installation system Requirements for Windows Table 1 Minimum System Requirements Operating Systems Processors Disk Space RAM Windows XP Intel Pentium 466 MB 1024 MB Pentium 4 and Service Pack 2 or 3 above 2048 MB recommended Windows Server 2003 Intel Celeron Service Pack 2 or R2 Intel Xeon Intel Core Windows Vista Service Pack 1 AM
12. and persisted with a symbol period multiple The open eyes points indicates the exact moment for the receiver to sample and decide in order to extract the message Eyes wide open are desired This measurement reflects the signal quality Use it to realize non linear linear and amplitude distortions The information bar will show the additional information e MER Modulation Error Rate It is the mean squared error in dB considering the perfect symbol as the reference signal Eye Diagram i MS l A wy h j j Se tay FASSEN CN CN t co iil i si apnyduyy Symbol Figure 18 Eye Diagram Measurement ll show the Frequency Response t to measure linear distortions This measurement w Frequency Response Use i in Iter lly a fi Ica typ der test Ice un existent on your dev lon formati self expla iona lon Il show the add bar wi lon The informat Peak to peak frequency response d ined tort IS Frequency Response Magnitude dB 0 Frequency MHz Figure 19 Frequency Response Measurement gt Crest Factor Reduction will be automatically bypassed when this measurement is being performed Group Delay This measurement will show the Group Delay existent on your device under test typically a filter Use it to measure linear distortions The information bar will show the additional
13. ation algorithm to compensate for non linear distortions A Memory Effects Compensation Check this box to compensate for memory non linear distortions Uncheck to compesante for memoryless distortions only Save Push this button to save the result on the non volatile memory Back Close current window AM AM Figure that shows the AM AM behavior for the pre correction AM PM Figure that shows the AM PM behavior for the pre correction 10 DUT If the Memory Non Linear Pre Correction option is enabled two PA memory models may be selected Type A and Type B try both to compare which one fits better Se SS Memoryless vs Memory Pre Correction Non linearity is present when the AM AM and AM PM transfer functions are not straight lines i e when the DUT Device Under Test has not a constant gain and a constant output phase delay with respect to the input amplitude When such distortions don t change with time they re called memoryless non linearity or static non linearity It means the output signal at the present moment depends only on the input signal at the present moment also However some amplifiers present a dynamic behavior Their AM AM and AM PM behavior change with time in a short time range nano micro seconds It happens because the output signal at the present moment doesn t depend on the input signal at the present moment only but also on its previous values It generates more distortion and a me
14. er results using the Memory Pre Correction It depends on the amplifier to be linearized Figure 7 presents the corresponding curves used to compensate for the non linearities of Figure 6 Those curves holds the inverse DUT characteristic i e the necessary characteristic to compensate for the distortion AM AM AM AM Gain Normalized 0 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 6 0 9 1 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Amplitude Normalized Amplitude o _ o M D i AM PM AM PM Oo co Phase ormalized Memory p m N o nN mm 2p 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 6 0 9 1 l 0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 1 Normalized Amplitude Normalized Amplitude Oo Figure 7a Memoryless curves Figure 7b Memory curves The memory curves Figure 7b are three dimensional plots where the color means amplitude in dB for AM AM and phase in for AM PM A distortion is present when horizontal parallel lines present different colors When Crest Factor Reduction is used intermodulation may be reduced However there is a trade off with three collateral effects e MER reduction e Frequency response distortion e Image frequency rejection Figure 8 represents typical results when Crest Factor Reduction is used Marker 2 A 3 250000000 MHz Input RF PNO Fast 1 Trig Free Run IFGain Low Atten 6 dB SENSE INT ALIGN ALTO 02 11 46 PM Jun 03 2009 10 aman Ref 4 38 dBm
15. gital pre correction the PA linearity is improved and extended so that the PA can be operated at higher power This means that a lower power lower cost linearized PA can be used in place of a higher power higher cost PA Furthermore the linearized PA operates more efficiently since it is operated closer to saturation The benefits are compounded because a lower power PA operating more efficiently consumes substantially less power than an inefficient higher power PA When the Non Linear Pre Correction screen is opened the software read and display the pre correction curves stored on the non volatile memory The Memory Effects Compensation checkbox is not going to be enabled if this option is not included The Figure 5 shows the non linear pre correction main window Linear Industries Inc DUT T Automatic Non Linear Pre Correction Pre Correction Properties 7 Enable Correction 0 4 0 5 0 6 Mormalized Amplitude 0 4 0 5 0 6 Normalized Amplitude 1 Turn on non linear pre correction at front panel 2 Connect the feedback sample at Before Filter 6 10 0 dBm Figure 5 Non Linear Pre Correction Window 1 Reset Download linear LUTs Look Up Tables Used to bypass the corrections 2 Enable Non Linear Pre Correction Used to bypass the corrections or not without downloading linear LUTs 3 Amplitude Trigger Adjust the amount of peak power used to trigger the signal Capture Run Single iter
16. how much memory effects the amplifier has Always perform this measurement using Feedback Before Filter sample The information bar will show the additional information e Peak to peak amplitude distortion Shows the Memoryless Transfer Function peak to peak value AM PM de A z A ee eT 0 Fi eed Pa ee a 7 Fe 2 oe RES UW Os tt 3 te ype yh E IE gt 1 t d 37 au CF Yat y Ra 2 ww ee T ay Rr 7 i E E Pent Pr x r t oe ee ot A et a t 4 JAS Sie Yet 3 ri 8 REA S os es te ee yr S i 30 25 20 15 10 Normalized Power dBFS Figure 14 AM PM Measurement CCDF CCDF stands for Complementary Cumulative Distribution Function It represents the probability vertical axis of the Instantaneous Power be x dB above the average power where x is the value in the horizontal axis For example suppose an amplifier is driving 1 kW into a dummy load and the Feedback Sample was taken from its output In the measurement below Figure 15 the O dB point has a probability around 40 It means 40 of time the Instantaneous power is above 1 KW or 0 dB and thus 60 of time the Instantaneous power is below the average power 1 KW Now another example the 6 dB point for the Feedback signal red dashed line has a probability of 0 07 It means that 99 93 of time the Instantaneous Power is below 4 kW
17. moryless pre corrector is not able to compensate for these memory effects This pre correction is an automatic procedure It uses a signal taken from a feedback sample located between the power amplifier output and the mask filter input Perform the following steps to run the pre correction 1 Connect the AT7001 RF output to the amplifier to be linearized If an output filter is used run the algorithm with the filter connected to the amplifier because the reflected signal coming from the filter may change the amplifier behavior 2 Set the desired power 3 Take the Before Filter Feedback Sample and make sure is connected the AT 7001 rear panel before running the pre correction E The coupler used for the sample must be directional with isolation greater than 15 dB in order avoid miscalculation created by a reflected signal The coupling port must have 50 O impedance in order to avoid a linear distortion in the feedback path Also the coupling port response must be as flat as 0 3 dB in a 30 MHz range The feedback sample level must be between 10 and 0 dBm You can use the Signal Measurements software to check if the feedback sample is present 4 Press Reset to bypass the last correction 5 Check or uncheck the Memory Effects Compensation checkbox lf Memory Effects Compensation is checked select between Type A or Type B at the DUT menu This option is not available if the memoryless pre correction is selected Memory Effects
18. nning 5 Once the correction achieves satisfying results click on Save push button to save the curves on the non volatile memory Figure 10 presents the corresponding curves used to compensate a mask filter Frequency Response Magnitude dB D Frequency MHz Group Delay Group Delay ns a Frequency MHz Figure 10 Typical Linear Pre Correction Curves 3 3 Signal Measurements This software is used to perform measurements at the signals generated by the AT7001 Digital Exciter There are fourteen measurements avaiable Linear Industries Inc Measurement Software Signal Selection Input Feedback Before Filter O Feedback After Filter Measurements Amplitude Modulus C Concatenate Figure 11 Signal Measurements Window As E Signal Selection Select which signal is going to be measured Input is the signal at the DUT input or the modulator output The Input signal is always the reference Both feedback samples are also available Some measurements allow just one signal selection In order to make a comparison others allow two signals at the same time Measurements This pop up menu shows all fourteen measurements available For each measurement there is an additional option like Concatenate Persist or Average the consecutive measurements Measure Push Button Click once to start measuring Click again to stop Back Close current
19. nt window Frequency Response Figure that shows the pre correction frequency response which is the inverse filter response Group Delay Figure that shows the pre correction group delay which is the opposite filter group delay This pre correction is an automatic procedure It uses a signal taken from a feedback sample located at the mask filter output Perform the following steps to run the pre correction l Connect the AT7001 RF output to the DUT Set the nominal power lf an amplifier is used run the non linear pre correction first and then the linear pre correction This will make linear calculation more accurate Make sure the After Filter Feedback Sample is connected the AT7001 rear panel before running the pre correction gt The coupler used for the sample must be directional with an isolation greater than 15 dB in order avoid miscalculation created by a reflected signal The coupling port must have a 50 Q impedance in order to avoid a linear distortion in the feedback path Also the coupling port response must be as flat as 0 5 dB in a 30 MHz range The feedback sample level must be between 10 and 0 dBm You can use the Signal Measurements software to check if the feedback sample is present 4 Click on Run The whole process consists of single iteration After a few seconds the results will be applied automatically gt Crest Factor Reduction will be automatically bypassed when the algorithm is ru
20. sion Automatic Pre Correction Hardware Version Measurements Software Figure 2 GUI Software Main Window Choose the correct Serial Port number and Click on Connect lf the communication is ok the information bar displays Successfully Connected and Hardware Version Text Box is updated i The Serial Port number may vary between computers All features in this software are optional The prompt screen will show which options are included which are not Linear pre correction INCLUDED Memoryless non linear pre correction INCLUDED et C ProjetosWigital_Exciter matlab GUIw4 00 GUI7001 v400 distrib GUI7001 v400 exe Jof x El Memory non linear pre correction INCLUDED J UDED Succeeed Figure 3 Prompt Window lf an option is included the corresponding button is enabled Linear Industries Inc Advanced Graphic User Interface E Communication Settings Automatic Pre Correction Non Linear Pre Correction Software Version Linear Pre Correction Disconnect Hardware Version Measurements Software Signal Measurements Figure 4 AT7001 connected with all options enabled 3 1 Non Linear Pre Correction This is the process is used to compensate for the power amplifier non linearity in order to reduce the intermodulation at the output signal and thus improve PA efficiency With the automatic non linear di
21. tion Pate 6 bei tiei edi siti pi bik 2a ee ee 2 ee as r F ge 8 FRY 5 25 Be 4 4 E o PF F ge g r v 2 l E E go S F si S Z fiig 10 e tisa O 7 ED Real Part Figure 16 Constellation Measurement Constellation Contour the constellation is a diagram of the sampled baseband modulating symbols where the Cartesian plan is plotted by its real vs imaginary part The real part In Phase of those sampled symbols carry the transmitted information while the imaginary part Quadrature carry the necessary information to generate the vestigial side band For 8VSB modulation the eight regions represents the eight possible symbols and the seven dashed lines indicates the decision threshold The Constellation Contour differs from the Constellation in the sense it shows probability regions for a demodulated symbol to fall in coded in colors A red area means high probability as the blue area means low probability This measurement reflects the signal quality Use it to realize non linear linear and amplitude distortions The information bar will show the additional information e MER Modulation Error Rate It is the mean squared error in dB considering the perfect symbol as the reference signal Constellation Contour Imaginary Part Real Part Figure 17 Constellation Contour Eye Diagram The Eye Diagram is the purely real modulating symbols filtered with a raised cosine filter
Download Pdf Manuals
Related Search