在机器人学领域，机械臂的避障轨迹规划是一个复杂的任务，涉及到多个学科和计算方法。其中一个主要挑战是如何在保证避障的同时，规划出一条最优或近似最优的路径。在这种背景下，RRT系列算法提供了一种有效的解决方案。RRT，即Rapidly-exploring Random Tree（快速探索随机树），是一种基于概率的路径搜索算法，被广泛应用于复杂高维空间的轨迹规划问题中。其核心思想是利用随机采样的方式，不断扩展树状结构来探索整个空间，直到找到目标点。 RRT算法通过随机采样状态空间并以这种方式构建出一棵搜索树，树的节点代表了机械臂可以到达的配置，而树的枝干代表了从一个配置到另一个配置的运动。随着树的不断扩展，算法逐渐接近目标位置。为了更好地处理避障问题，RRT算法经常被加以改进，比如RRT*算法通过引入重连接（rewiring）和最佳优先（best-first）搜索，能够找到更加平滑和短路径的解。而RRT-connect算法则强调通过双向搜索来加快找到路径的速度。这些改进使得RRT算法在具有障碍物的环境中也能找到一条合理的避障路径。 在Matlab环境下进行算法实现和机械臂模拟，可以提供一个直观且强大的平台来测试和优化这些算法。Matlab是一种广泛使用的高性能数学计算软件，它提供了丰富的数学函数库以及可视化工具，非常适合用于算法的快速原型设计、测试和展示。在Matlab中，用户可以定义机械臂的运动学和动力学模型，以及环境的几何模型。然后可以使用Matlab内置的函数或自定义的算法来实现RRT系列算法。用户可以利用Matlab的可视化功能，观察机械臂的运动轨迹，从而分析和评估算法性能。 机械臂轨迹规划是一个综合性问题，不仅涉及到算法的实现，还包含机械臂本体的设计、控制系统的设计以及环境感知与建模等众多方面。在实际应用中，需要综合考虑机械臂的尺寸、重量、关节类型、运动范围等因素，这些都会对轨迹规划产生重要影响。同时，环境中的障碍物分布、动态障碍物的预测等也是规划过程中必须考虑的问题。因此，一个完整的机械臂避障轨迹规划解决方案需要跨学科的知识和技术支持。 在Matlab中，可以通过模块化的方式来构建机械臂避障轨迹规划系统。例如，可以将系统分为轨迹规划模块、控制模块、环境感知模块和用户交互模块等。每个模块完成不同的功能，它们相互配合，共同完成复杂的轨迹规划任务。用户可以通过Matlab界面进行参数设置、算法选择和模拟运行等操作，同时获得包括模拟动画在内的直观结果。 RRT系列算法在机械臂避障轨迹规划方面提供了强有力的工具，Matlab则为算法的实现、测试和验证提供了便捷的平台。通过结合先进的算法和强大的软件工具，工程师和研究人员可以开发出高效的轨迹规划系统，推动机械臂技术的进步。

在这篇文章中，我们将学习如何使用ESP32-CAM开发板来构建数码相机。按下重置按钮后，开发板将拍摄一张图像，并将其存储到microSD卡中，然后返回深度睡眠状态。我们使用EEPROM来存储并获取图像编号。 ESP32-CAM板已包含本草图所需的相机模块，复位开关和microSD卡插槽。除此之外，您还需要一个microSD卡，一个5V电源以及一个USB到串行转换器来上传草图。

无用户上限支持网络环境:支持在局域网内进行加解密操作。支持多用户同时映像浏览解密. 右键加密，加密操作简易:支持文件拖放加密。主顺序中加密。 能快速的加密大量文件 一流加密速度:使用原创的高速加密引擎. 轻松打造满足您个人需要的加密文件夹多种加密设置:加入各种自定义设置。 稳定性更强，性能优化稳定:顺序全部模块进行优化重构。平安更有保障。 可以满足不同层次用户的需要三种加密强度

内容概要：本文详细介绍了使用SIMPLIS进行开关电源仿真的方法和技术细节，尤其是针对多相Buck电路的设计与优化。文中通过具体案例展示了不同控制模式（如COT模式、PWM模式）、软启动策略以及电流均衡算法的应用。同时探讨了如何利用SIMPLIS自带的各种模型和工具来进行高效的电源设计仿真，包括但不限于相位交错控制、动态电流平衡、软启动曲线优化等方面的内容。 适合人群：从事电源设计的专业人士，尤其是对多相Buck电路感兴趣的工程师。 使用场景及目标：帮助工程师更好地理解和掌握SIMPLIS这一强大工具，在实际工作中能够快速搭建并优化复杂的电源系统，提高工作效率的同时确保设计方案的可靠性和高效性。 其他说明：文章不仅提供了理论指导，还有大量的实例代码片段供读者参考实践，使得整个学习过程更加直观易懂。此外，作者分享了许多个人经验教训，有助于初学者避开常见误区。

电源Simplis开关电源及电路仿真案例。 单 多相控制buck仿真电路。 4 8 phase COT D-CAP+ 架构仿真模型， 1-8phase PWM buck仿真模型， 峰值电流模式，D-CAP3模式等，仅供学习，参数可调，可二次开发。 支持Loadline，ZCD，TLVR，softstart等。 电源Simplis开关电源及电路仿真案例详细解析 在现代电子设备中，开关电源技术是维持设备稳定运行的关键之一。开关电源通过快速切换电路的开启和关闭状态，实现对电压的转换和调节。随着技术的发展，开关电源设计和仿真技术也日益成熟，为工程师提供了精确模拟电源性能的工具。本文档介绍的Simplis开关电源及电路仿真案例，涵盖了一系列先进的电源仿真技术和模型，具体包括单相与多相控制的buck仿真电路，以及4到8相的COT（Constant On-Time，固定导通时间）控制模式下的D-CAP+架构仿真模型。 buck转换器是一种常用的开关电源转换器，其工作原理是通过调整开关元件的导通时间来降低输入电压，并输出稳定的直流电压。在这个仿真案例中，我们不仅可以模拟单相buck电路，还能进行多相控制仿真。多相控制能够更有效地管理电流和热量，提高电源转换效率，尤其适用于高功率需求的场合。 接下来，我们来探讨COT控制模式下的D-CAP+架构。COT控制是一种开关电源的控制策略，它通过固定开关的导通时间来控制输出电压。D-CAP+架构则是COT控制模式下的一种衍生，它不仅能够提供快速的负载响应，还能保持良好的稳定性和低噪声特性。4到8相的架构仿真模型能够模拟在多种负载条件下电源的行为，这对于电源系统的设计和优化至关重要。 此外，案例中还提到了1-8相PWM buck仿真模型。PWM（脉冲宽度调制）是一种通过改变开关元件脉冲宽度来控制输出电压的技术。通过调整PWM信号的占空比，可以实现对电源输出电压的精细控制。而多相PWM buck模型可以进一步提升电源的性能，尤其在高电流应用中效果显著。 本案例中还特别强调了峰值电流模式和D-CAP3模式的仿真。峰值电流模式控制是一种电流模式控制方式，它通过监测开关元件的峰值电流来控制开关的导通时间，这种模式下电源系统响应快，稳定性高。D-CAP3模式则是最新的一种控制模式，它在D-CAP+基础上进一步优化，提供了更好的性能。 除了上述的技术点，案例还提到支持多种高级功能，例如Loadline（负载线）、ZCD（零电流检测）、TLVR（温度变化率）和softstart（软启动）。这些功能的加入，不仅使电源设计更加灵活，还大大提高了电源的适应性和可靠性。例如，softstart功能可以减少启动时电流冲击，保护电源不受损害；Loadline功能可以优化电压响应，保持在负载变动时的稳定输出。 电源Simplis开关电源及电路仿真案例提供了一个深入学习和实践先进电源控制策略和技术的平台。该平台不仅包含了多种控制模式和架构的仿真模型，还允许用户调整各种参数，进行二次开发，以满足不同设计需求。这些仿真模型和功能的集合，无疑为电源工程师提供了一套全面的分析和设计工具，从而能够更加高效地完成高质量电源设计。

VCLZip Native Delphi Zip/UnZip Component! (VCLZip Lite: Version 2.23 April 14th, 2002) (VCLZip Pro: Version 3.10 Buid 1 - November 25th, 2007) IMPORTANT: If installing the registered version, please be sure to always re-install/rebuild the components (VCLZip and VCLUnZip) to the component pallette (or rebuild the design time package) so that the ThisVersion property and any other new properties will be properly updated. If your application still does not run without the IDE, open up VCLZip's package, click on options and look at the Directories/Conditionals tab. If KPDEMO is defined, remove it and recompile the package. ***IMPORTANT: Please remember do not install these components into a package by the name of either VCLZip or VCLUnZip. You will receive an error if you do. PLEASE TAKE A LOOK AT THE "WHAT's NEW IN THIS VERSION" LINK IN THE HELP FILE AS IT HAS CONVENIENT LINKS TO ALL OF THE NEW TOPICS. ==================== Version 3.10 Build 1 - Several bug fixes. - Added support for Delphi 2006, 2007 - Added support for BCB 2006, 2007 - Improved memory performance when working with archives containing extremely high number of compressed files. ==================== Version 3.06 Build 2 Made Delphi 2005 compatible Other assorted fixes ==================== Version 3.05 Build 1 Fixed a lot of incompatabilities between VCLZip and WinZip Other assorted fixes ==================== Version 3.04 Build 1 New ZLib methods for optimized compression and decompression of single entities of data in standard ZLib format, without the overhead of the PKZip format. This is excellent for compression of data to be sent across the net, compressing web pages (http compliant compression), blobs, etc. - ZLibCompressStream - ZLibDecompressStream - ZLibCompressBuffer - ZLibDecompressBuffer - ZLibCompressString - ZLibDecompressString Overloaded TStream Methods for Delphi 4,5, BCB 4, and 5 - UnZipToStream - UnZipToStreamByIndex - ZipFromStream Special OnGetNextTStream Event for Delphi 4,5, BCB 4, and 5 - Allows zipping multiple TStreams in one process - More efficient than calling ZipFromStream multiple times Capability to use the latest version of ZLib 1.2.1. - VCLZip currently uses 1.4.1 by default. - By defining ZLIB121, VCLZip will use the latest version of ZLib which is included with the registered version. Some optimization improvements which should show some improvement in zipping and unzipping speed when using TkpStreams with D4, D5, BCB4, and BCB5. ============ Version 3.03 (VCLZip Pro) - Please test your application thoroughly with this new version of VCLZip Pro. While it has been tested and has even been used in at least two production applications for several months now prior to initial release, there are so many combinations of property settings, environment differences, and ways to use VCLZip that you should always test VCLZip completely in your application before deploying. *** New Zip64 capabilities, properties, methods and events: - Uncompressed, Compressed, and Archive file sizes can be up to 2^63-1 bytes in length. - You can compress up to 2147483647 files into an archive. This is compatible with PKZip's Zip64 format. - If a file does not extend beyond any of the original limitations (filesizes of 4 gig or 65535 files) then no Zip64 format information is included in the archive. - property isZip64 - tells you when you are working with a zip file that is using Zip64 format. Much faster processing due to linking to Zlib object files for compression and decompression routines. Blocked Zip Files (spanned zip archives split onto hard drive) - Now completely compatible with PKZip and WinZip split archives file naming format. - For backwards compatability you can tell VCLZip to use the old VCLZip filenaming format by using the BlockMode property. - New method OnFileNameForSplitPart called just before each split filepart is created. VCLZip supplies a default implementation of this method so for most purposes you won't need your own. - method DefaultFileNameForSplitPart - VCLZip calls this internally if you don't define your own OnFileNameForSplitPart. You can also call it from your own OnFileNameForSplitPart if you wish to add some processing to the default behavior. - property BlockMode - determines whether VCLZip uses PKZip/WinZip standard naming convention or VCLZip classic method. - method DefaultGetNextDisk - VCLZip calls this internally if you don't define your own OnGetNextDisk. You can also call it from your own OnGetNextDisk event if you wish to add some processing to the default behavior. - Properties for controlling which files are zipped... - IncludeHiddenFiles - default False; - IncludeSysFiles: - default False; - IncludeReadOnlyFiles: - default True; - IncludeArchiveFiles: - default True; - Event OnGetNextStream - Allows you to zip from multiple streams when using the ZipFromStream method. This improves performance since repeated calls to ZipFromStream causes the archive to be updated on each subsequent call. - property ThisBuild - Tells you the current build. See also ThisVersion - property OnHandleMessage - Handles interactive messages with VCLZip. There is a default, so you don't need to define your own unless you wish to eliminate interactive messages and handle them on your own. This is helpful if you are using VCLZip as a service or on a webserver for instance. ******** Upgrading existing applications that use VCLZip 2.X ********** For the most part, existing applications will work as-is. Just install VCLZip 3.X and recompile your code. Here are some things to be aware of though... 1) If your app currently creates mmBlock archives (spanned directly to hard drive) and you define your own OnGetNextDisk in VCLZip 2.X, you should move your code from this event that handles mmBlock events to the new event OnFileNameForSplitPart. However, if you simply rely on VCLZip's default OnGetNextDisk then you don't have to worry about this. 2) If your app creates mmBlock archives, the default naming convention has changed to match the PKZip/WinZip standard. If you wish to keep the same naming convention then set BlockMode := mbClassic. 3) OnGetNextDisk and OnPrepareNextDisk events are called for the 1st disk now. VCLZip 2.X only calls these events starting with the 2nd disk. 4) properties CompressedSize[Index], UncompressedSize[Index], ZipSize are now Int64 types. 5) Delphi 4, Delphi 5, BCB 4, and BCB5 are all capable of using the Zip64 format. However they use the TkpHugeStream decendants which act just like TStreams except they handle files/stream sizes larger than 2gig. There is a TkpHugeFileStream and a TkpHugeMemoryStream which should handle 99% of all necessary actions. If you currently work with VCLZip 2.X with TBlobStreams or some other type of streams, you can either define your own TkpBlobStream for instance which inherits from TkpHugeStream, or use the TkpHugeStream.CopyFrom(TStream, Count) and the TkpHugeStream.GetStream: TStream methods to give VCLZip your stream and get it back. Ofcourse when using regular TStream decendants in D4,4,BCB4,and 5, you cannot create Zip64 archives. If you use Delphi 6, 7, or BCB 6, you don't have to worry about any of this as the normal TSTream is used by VCLZip and handles large file/stream sizes. ============ Version 2.23 (VCLZip Lite) Added the OEMConvert property. Filenames stored in a PKZip compatible archive normally go through an OEM conversion to make them ascii compatible. When opening the zip file the conversion is undone. If you do not plan on having other zip utilities opening up your archives this conversion process is not really necessary. Setting this property to False will eliminate this process. The default value for this property is True for normal PKZip compatability. Added OnEncrypt and OnDecrypt events. These allow you to replace the standard pkzip encryption with your own. Data is passed to these events a buffer at a time. Use this with care as this is still somewhat experimental and I'm not sure how useful it is yet. You must make all changes within the buffer sent in to you. Treat the entire file as a stream. Byte for byte replacement only. No additional keys can be saved. Added OnRecursingFile event. Sometimes when using wildcards and recursing directories, there was no reporting of progress. This will be fired each time a file matches as the file list is being built while recursing directories. Added the EncryptBeforeCompress boolean property. The default for this property is False and if left like this VCLZip will behave like normal. If set to True, VCLZip will encrypt each buffer prior to compressing it instead of afterwards. This will cause files to not be decryptable by normal zip utilities thereby adding a bit of extra security. Bugs Fixed: IMPORTANT!!! Behavior of freeing the ArchiveStream (compressed stream) has been modified. VCLZip will now no longer try to free ArchiveStream, you must free it yourself. This was due to a problem where it would be freed automatically if there was a problem with the ArchiveStream when trying to open it as a zip file (possibly corrupt). Best practice is that ArchiveStream should always point toward a TMemoryStream that you create anyway. Modified the SFX code (the code used to create the SFX stub distributed with VCLZip) so that it handles filenames that have been run through an OEM Conversion. The SFX was losing accented characters. This modification means that if you are creating zip files to be used as SFX's you will want to leave the OEMConvert property mentioned above, set to it's default value of True. Modified so that when cursor is changed to hourglass by VCLZip, previous cursor is saved correctly instead of just changing it back to default cursor. Now saves Central Directory Extra Fields correctly. Fixed the SFX code so that it works properly if you use Copy /B to concatenate a zip file to the stub. Due to a Delphi strange behavior sometimes path names for directory only entries would become corrupted. Removed reference to QConsts, replaced with RTLConsts. Sometimes a GPF would result if a corrupt zip file was opened. Using a wildcard in pathname added to FilesList did not work. Using '*.*' as a wildcard in files added to FilesList now is the same as using '*'. VCLZip will now check for CancelTheOperation during initial building of the fileslist instead of just during compression processing. Added a final call to OnTotalPercentDone with 100% because this didn't always happen. Attributes were not getting set correctly for directory-only entries. Fixed a problem that was not allowing ZipComment's to be added correctly to spanned or blocked zip files. Not the same fix as in 2.22. Directories (directory-only entries) were not being restored properly unless DoAll was True. You were unable to delete a directory from which files were recursively zipped until exiting your application. ============ Version 2.22 Now Delphi 6 compatible. New event called {link=93,OnRecursingFile} which gets called as VCLZip recurses directories searching for files that match a wildcard that is entered in the FilesList. This gets called each time a file matches the wildcard. Fixed a bug which kept diskettes from being labeled when creating spanned zip files on WIN31. Fixed a bug which sometimes did not allow zip comments to be added to blocked zip sets. Fixed a bug which caused VCLZip to not properly handle the IncompleteZip exception on spanned zip sets unless you called ReadZip prior to calling UnZip. Version 2.21 (Changes are shown in the build stages as they were implemented) Pre-Release Build 5: When working with temporary files, VCLZip will now rename, instead of copy, the temp file if the destination is on the same drive. This will speed up the adding of files to an existing zip file when the resulting zip file is very large. Pre-Release Build 4: New event called OnPrepareNextDisk which is an event that will allow you, when creating spanned zip files across diskettes, to do things like format a diskette that has just been inserted, or to add or delete files from the diskette before continuing with the zipping process. Fixed a problem that was causing the CancelTheOperation Method to not work properly. Pre-Release Build 3: Fixed bug which caused VCLZip to miscalculate space needed for zfc file if wildcards are put into the FilesList. Fixed bug so you could have FilePercentDone without needing TotalPercentDone when creating spanned zip files Fixed so relative_offset set correctly for spanned zips. Side effect of removing needless write of header. Added code to read local fileheaders if exception thrown when reading a central fileheader. Fixed problem where directories couldn't be created from directory entries because the fullpath wasn't known yet. Result of having moved this code to earlier. Fixed typo in creation of LOC header values which could cause error if reading local headers. Changed so Zip Comment starting position is calculated based on end of central record instead of end of file. Pre-Release Build 2: IMPORTANT: Changed default for FileOpenMode back to fmShareDenyNone as it had been for all but version 2.20. Fixed a problem where drivepart (i.e. C:\) was not being stripped when saving relative paths. Added a BufferedStreamSize property which can increase the speed of creating zips to floppy (and other slow media) dramatically. The new default for this should increase the speed by as much as 3 times, but you can now tweak this especially for your application! Added an ImproperZip property which gets set when VCLZip detects an inconsistency with the zip. This can be useful for detecting when VCLZip was able to open the zip in spite of an inconsistency found. There was no way to know this in the past. Fixed a problem where zip comments in zfc files were not being read correctly. Added a setZipSignatures procedure which allows you to modify the signatures of your zip file. This will cause other zip utilities to not be able to recognize or read your zip files created with VCLZip. Useful if you want to add further security to your zip files. Pre-Release Build 1: Some zip files would not open correctly, throwing an incomplete zip file exception due to an erroneous "extra field length" identifier in headers of some compressed files. These zip files are rare, but a very few people seemed to have several of them. This problem would not affect zip files created by VCLZip, and this problem should only occur in VCLZip 2.20, not in any previous version. If you had Range Checking turned on, VCLZip would get a range check error when using a wildcard that ended with a * as in 'somefile.*'. Under certain circumstances, drive information would not be stripped from path information if zipping recursively (including subdirectories) "Retrying" to zip a file that could not be opened using the OnSkippingFile event would not always work correctly. Creating spanned zip set to floppy should be faster now due to removing a needless header write to disk for each file. VCLZip would not compile correctly with MAKESMALL defined. Added code to make VCLZip work with BCB5. Haven't tested this yet though since I don't have BCB5 myself yet. Added readonly boolean ImproperZip property which will be set to True when some sort of problem is found when opening the zip file, even if recoverable. This property will be enhanced and refined in the future. If KeepZipOpen is set to True, when putting in the wrong disk in a spanned zip set, VCLZip would not always properly close the file on the old diskette before trying to open the file on the next diskette. Added ECantWriteUCF exception which will be thrown if VCLZip runs out of room to write the uncompressed file when unzipping. Timestamp was not being set properly when unzipping readonly files. Moved setting of the timestamp to before the attributes get set. ============ Version 2.20 Changes have been made in the following areas: --Performance There are a few code optimizations that should speed up the zipping process slightly. --Spanned Zip Files A new feature, turned on with the SaveZipInfoOnFirstDisk allows VCLZip to create and read spanned zip files starting with the first disk instead of the normally required last disk of the spanned disk set by saving a Zip Configuration File on the first disk. This feature can be used even if creating the spanned zip file directly to your hard drive. A new property, SaveOnFirstDisk, allows you to save room on the first disk when creating a spanned zip file, to allow room for other files, such as setup programs, data files, or a Zip Configuration File. Spanned zip files can now be directed toward disks greater than 2 gig in size as long as you are using Delphi 5 or BCB 4. --UnZipping The new Selected indexed property offers another way to flag files to be unzipped. Files that have the Selected property set to True can be unzipped using the UnZipSelected method. The Selected property will be cleared (set to False) for each file as it is unzipped, but you can also call the ClearSelected method to clear them all. At anytime the NumSelected property can be checked to see how many files have been selected. Also, the UnZipToBufferByIndex and UnZipToStreamByIndex methods allow you to unzip files specified by their index instead of by name or wildcard. The BufferLength property allows buffered output (buffer smaller than the total uncompressed filesize) when unzipping directly to memory (see UnZipToBuffer and UnZipToBufferByIndex). This will cause the OnGetNextBuffer Event to be called everytime BufferLength bytes have been output by VCLZip. Modified to work in all ways with zip files that have "extra fields" in their headers. These tend to be quite rare, but they do show up from time to time. --Zipping Added a property called FileOpenMode which allows you to define the file open mode for files when they are opened to be zipped. Added a Retry parameter to the OnSkippingFile Event that can be used to re-attempt to open a file for zipping that is open by another process. This gives the chance to close the file and continue with the zipping process rather than having to start over again. Added a ENotEnoughRoom exception which will be thrown if there is not enough room to write to the archive, i.e. out of disk space. The new OnUpdate Event gets fired when updating or freshening an existing archive. It is triggered for each file that already exists in the archive as it is either replaced or kept in the updated archive. The AddDirEntriesOnRecurse will cause separate directory entries to be included in archives when doing recursive zips through subdirectories. --Integrity Checking A new method, CheckArchive, will perform an integrity check on all files in an archive. This is much faster than using FileIsOK on each file if testing all files in an archive with VERY MANY files. Further improved checking for corrupted zip files when opening zip files. --Encryption The following new properties and methods allow lower level work with password encrypted archives: DecryptHeader Gets the decryption header for a particular compressed file in an archive GetDecryptHeaderPtr Same as DecryptHeader but easier to use in BCB. DecryptHeaderByte Method Tests a password against the decryption header found in the DecryptHeader property. GetDecryptHeaderByteByPtr Same as DecryptHeaderByte but easier to use in BCB. --Self Extracting Executables Changes were made to the ZIPSFX32.BIN stub itself: - Modified to work with zip files containing "extra fields" in their headers. - Modified to change mouse cursor to an hour glass during processing. - Check for correct file size is now done automatically - Now uses the end of central and central headers to find the first local header. - Added a progress meter - Better checking for corrupted zip files. - Added an information window that can optionally be shown when the sfx is initially started up. - Added an AutoRun option to make the sfx stub run automatially when double clicked with no other interaction from the user. For the new modified sfx stub, ZIPSFX32.BIN, instead of using kpSFXOpt, you should now use the TSfxConfig component to set the options for the sfx stub. The new sfx can be found in the sfx\ subdirectory as usual and is called ZIPSFX32.BIN and the original sfx can be found in the same subdirectory except it is now called ORGSFX32.bin. Just rename it if you prefer that one (use KPSFXOPT instead of TSfxConfig with the old stub). --Miscellaneous The installation is now easier, atleast for first time installers of the source code. The .DPK files for Delphi and .CPP files for BCB are now included. Now these files simply have to be compiled and that's it. There is a separate option in the installation for installing to the different versions of Delphi and BCB. Added a property called FlushFilesOnClose which will cause all files opened for write by VCLZip to have their disk buffers flushed to disk when closed. Added the capability to delete Selected files from an archive using the DeleteEntries Method. The behavior of the OnInCompleteZip Event has been greatly improved. You can now use this event to ask the user to insert the last disk of a spanned disk set rather than having to handle this situation from outside VCLZip. The register procedures were changed so that the components now get installed to the "VCLZip" tab on the palette. I found that for all but Delphi 1 I had to actually manually move the components to the "VCLZip" tab. You may find that you have to do this too if you have already installed VCLZip before. The components now use new bitmaps in place of the old ones on the component palette. Separated many compiler defines into a new file called KPDEFS.INC. ==================================== Version 2.18: 1) Thanks to the hard work of a fellow registered user, added the capability to remove all dependencies on the Dialogs, Forms, Controls, and FileCtrl units by defining the conditional MAKESMALL, which results in a smaller footprint. This can be quite useful when putting VCLZip into a DLL for instance. In order to make this work, go into your Project | Options and select the Directories/Conditionals tab and enter MAKESMALL in the conditional defines text box. In Delphi you can add this conditinal define to the project options of your application that uses VCLZip and then do a "build all". In BCB you will have to add this to the project options of the package that contains VCLZip and then rebuild the package. If you define MAKESMALL, the only things you lose are: a) ZIP file open dialog box that appears when the ZipName is set to "?" b) Select Directory dialog box that appears when the DestDir is set to "?" c) Changing the cursor to an hour glass during some operations. d) No long filename support in Delphi 1 2) Made VCLZip completely BCB4 compatible. 3) Added some exception handling to KPUNZIPP and KPINFLT, mainly to handle unexpected situations when wrong passwords are entered. This fixes the problem with PRP, the password recovery program. 4) For Borland C++ Builder, changed any COMP types to double, getting rid of the compiler warnings for unsupported comp type. This affects the OnStartZipInfo and OnStartUnZipInfo events, so you'll have to change the comp parameter to double in these events if you use them (in both your header files and in the CPP files). 5) Modified OnStartUnZip event so that FName (the filename of the file that is about to be unzipped along with complete path) is now a VAR parameter and can be modified. This allows you to change the path and name of a file that is about to be unzipped. This is especially helpfull in applications like Install Programs. NOTE: You will need to change your current code to add the VAR to the event definition and implementation if you already use this event in your application. (In BCB, add a & just before the parameter instead of VAR) 6) Moved many type definitions to VCLUNZIP.PAS so that kpZipObj won't have to be included in your USES list. 7) Fixed bug that caused GPF when setting Zip Comment to '' (empty string). 8) Moved strings in VCLZip/VCLUnZip into a string table, making the code size a little smaller as well as making it much easier to localize string information. However you have the option of not using the new string table, for whatever reason, by defining NO_RES in your project options (in the conditional defines text box on the Directories/Conditionals tab). 9) Removed the need for several files. No longer included are kpstrm.res, kpstrm.rc, kpsconst.res, kpsconst.rc, kpstres.pas, and for Delphi 1, kpdrvs.pas. In some cases the need for these files was eliminated and in other cases just rolled into the newly included kpzcnst.rc, kpzcnst.pas, and kpzcnst.res. Definining NO_RES in your project options will elimiate the need for these new files but will make your code size slightly larger and you won't be able to localize your application without changing VCLZip source code. 10) Modified the OnFilePercentDone and OnTotalPercentDone progress events to work better when creating spanned disk sets and blocked zip sets. They no longer report 100% when the compressed file still has to be copied to disk. 11) Added the ReplaceReadOnly property. Setting this to true will allow files with the ReadOnly attribute to be replaced during the unzip process. 12) Added the ifNewer and ifOlder options to the OverwriteMode property. (This had somehow made it into the help file but not into VCLUnZip) 13) Added the SFXToZip method which will convert an SFX file to a regular zip file. The header pointers will be properly adjusted during the conversion. 14) Fixed a problem where the OnGetNextDisk event would always revert to the DefaultGetNextDisk method instead of what you entered into the Object Inspector each time your project was re-opened. 15) Fixed a bug that caused CRC errors when unzipping files from spanned disk sets if they were STORED (no compression) and spanned across disks. 16) Added the OnZipComplete and OnUnZipComplete events. If defined, these will fire at the very end of a zip or unzip operation (after all files have been processed, not after each file). These events will rarely be used since, normally you will be able to do the same thing at the point that the call to Zip or UnZip returns, but these events can be useful when using VCLZip in threads where in certain circumstances the return from the Zip or UnZip methods are not seen. 17) Creation of SFX files has never been easier!!! The addition of the MakeNewSFX method allows you to create Self Extracting Executables without the need to create a zip file first. The files that you specify in the FilesList property will be zipped, using all the normal VCLZip property settings, and the SFX will be created, all in one step! In addition, you can create configurable SFX files using this method, and you can do this especially easy by adding the new unit kpSFXOpt to your application's USES list and using the new 32bit SFX stub that is now distributed with VCLZip. This allows you to easily set things like SFX Dialog caption, default target extraction directory, file to launch after extraction, etc. 18) Fixed a memory leak that only affects applications using VCLZip that are compiled with Delphi 2, and that use wildcard specifications in the FilesList property. Version 2.17a: 1) Fixed a bug that was keeping VCLZip from reading truncated zip files or sfx files that did not have their headers adjusted. 2) Fixed a bug that was causing a directory to be created on the C drive when doing integrity checking with the FileIsOK property. 3) Added {$V-} to kpZipObj.PAS 4) Moved two AssignTo methods to public instead of private in kpZipObj.PAS Version 2.17: 1) Added Memory zipping and unzipping capabilities through the UnZipToBuffer and ZipFromBuffer methods. See the documentation for these methods in the Help File for more information. 2) New FileIsOK Property allows you to check for the integrity of individual files within an archive without actually unzipping the file. 3) Fixed a bug that kept checking of volume labels from working on WIN31 when working with spanned disk sets. 4) Removed all references to ChDirectory so that VCLZip will be more thread safe allowing separate instances of VCLZip in separate threads to be performing zip/unzip operations at the same time. 5) A new public property PreserveStubs allows you to make modifications to sfx archives and have the archive remain an SFX rather than revert back to a normal zip file. 6) Added a default OnGetNextDisk event. If one is not defined, then the default event will be called when the situation arises that a new disk is needed when zipping or unzipping a spanned or blocked zip archive. 7) Added more power to the wildcard capabilities. Now you can qualify the * wildcard character, for instance: * would satisfy any number of contiguous characters as long as they are all a thru e. * would satisfy any number of contiguous characters as long as none of them were a thru e. This allows you to do things like include files in specific direcories into your ExcludeList. For instance: VCLZip1.ExcludeList.Add('c:\test\*.txt') would exclude the zipping of all .txt files in the test directory but not in any subdirectories. 8) Fixed other minor bugs and made other code enhancements. Version 2.16: ***Please be aware that if you currently use the OnSkippingFile event in any of your applications, version 2.16 will require a small modification as this event has an added parameter and one of the current parameters is used a little differently when being called by the zip operation. Please see the help file for more information. 1) The OnSkippingFile Event has been changed slightly, adding a parameter for the filename. 2) OnSkippingFile is now called when a file to be zipped is skipped because it is locked by another application. See the Help File for more information. 3) Fixed a bug with the Exclude and NoCompressList where they were ignoring entries with anything before the extention (i.e. 'somefile.*' as opposed to '*.zip') if you were saving directory information. 4) Fixed a bug that caused an error if you added a wildcard with a non-existent directory to the FilesList. 5) A few other minor bug fixes. Modifications for 2.15 include: 1) PackLevel can now be set to 0 (zero) which means no compression at all (STORED only). 2) New property ExcludeList is a new stringlist that you can add filenames and wildcards to in order to specify files that you do not wish to be included in an archive. 3) New property NoCompressList is a new stringlist that you can add filenames and wildcards to in order to specify files that you wish to be STORED with a PackLevel of 0 (zero), no compression. 4) All compiler warnings and hints were removed. Modifications for 2.14 include: 1) Delphi 4 compatability. 2) Added ability to use complex wildcards when specifying which files are to be zipped. This includes wildcard characters not only in the filename but also in the pathname. This allows you to specify directories using wildcards, for instance: VCLZip1.FilesList.add('c:\test\w*\mycode*.pas'); would get all PAS files beginning with mycode in subdirectories under TEST that begin with the letter w. Wilcards may be much more complex than this. Please see the help file for more information. 3) Added the ability to override the RECURSE property setting when specifying files to be zipped. By adding the following characters to the beginning of the filenames being added, you can override whatever the current setting is for the RECURSE property: '>' will force recursion into subdirectories '|' will force NO-recursion For instance: VCLZip1.FilesList.add('>c:\windows\*.ini'); will get all .ini files in and below the windows directory reguardless of what the recurse property setting is. and: VCLZip1.FilesList.add('|c:\windows\sys*\*.dll'); will get all .dll files in subdirectories of the windows directories that start with 'sys' but will not recurse into any directories below the sys* directories. 4) The [ and ] characters previously used as special wildcard characters have been changed to < and > since [ and ] are valid filename characters. If you still need to use the previous characters for backward compatability, I can show registered users how to easily modify a couple of constants in the source code in order to go back to the old style. See "Using Wildcards" in the help file for more information. 5) A few bug fixes. Modifications for 2.13 include: 1) New property ResetArchiveBitOnZip causes each file's archive bit to be turned off after being zipped. 2) New Property SkipIfArchiveBitNotSet causes files who's archive bit is not set to be skipped during zipping operations. 3) A few modifications were made to allow more compatibility with BCB 1. 4) Cleaned up the Help File some. 5) KWF file now works for Delphi 1 and Delphi 2 again. Still can't get context sensitive help in Delphi 3. 6) Cleaned up some of the code that was causing compiler warnings and hints. Modifications for 2.12 include: 1) Added a TempPath property to allow the temporary files path to be different from the Windows default. 2) Modified VCLZip so that any temporary files that are created receive a unique temporary filename so as not to clash with any other files in the temporary directory. This also allows working with zip files residing in the temporary directory. 3) Fixed a bug in the relative path feature. 4) Fixed a bug that caused a "list out of bounds" error if a file in the FilesList did not actually exist. Modifications for 2.11 include: 1) Fixed password encryption bug for 16 bit. 2) Fixed "invalid pointer operation" when closing application bug. 3) Fixed path device truncation bug which caused inability to modify existing archives in 16 bit. 4) Fixed inability to cancel during wilcard expansion bug. 5) Added capability to better handle corrupted timestamps. 6) Added capability to open and work with SFX files that were created with the COPY/B method (header files not adjusted). 7) Other small bug fixes. I'm still working on a bug which causes a GPF when continually unzipping the same file thousands to millions of times. This mainly affects programs like the Password Recovery Program (PRP) which uses the brute force method of searching for an archive's password. Modifications for 2.10 include: 1) Capability for 16bit VCLZip to store long file/path names when running on a 32bit OS. 2) New property (Store83Names) which allows you to force DOS 8.3 file and path names to be stored. 3) Better UNC path support. 4) Fixed a bug to allow files to be added to an empty archive. Modifications for 2.03 include: 1) Volume labels now get written correctly to spanned disk sets in Delphi 1 for all versions of Windows. 2) Delphi 1 VCLZip now correctly recognizes when it is running on Windows NT. 3) Fixed a problem with zipping files in the root directory when StorePaths = True. 4) File and Zip Comments are now read correctly from spanned/blocked zip archives. 5) Fixed a buf that was causing "Duplicate Object" errors. Modifications for 2.02 include: 1) Fix for file comments which were supposed to be fixed in version 2.01 but weren't. 2) Fix for stream zipping. Version 2.01 would not create a new archive if using a stream. (The Stream Demo now allows creating new zip files to streams too) 3) A few other minor modifications to further solidify the code. 4) A modification to the Zip Utility Demo which allows unzipping from Blocked zip files as if they were single zip files. 5) Added a read-only, published ThisVersion property which reflects the version of the VCLZip/VCLUnZip that you are currently working with. Modifications for 2.01 include: 1) Fixes for exceptions that were caused when CANCELING a zip or unzip of a spanned zip file. 2) Fix for a possible problem when zipping or unzipping a spanned zip file when one or more of the compressed files resided on more than 2 of the spanned parts. 3) Fix for file comments which were broken in version 2.00. Additional features for version 2.00 include: 1) Modify/Add internal file details (filename, pathname, timestamp, comment) for any file while zipping, in the OnStartZip event. 2) Add an Archive Comment while zipping in the OnStartZipInfo event. 3) Delphi 1 compatiblity for VCLZip. 4) Stream to Stream Zipping - Archives themselves can now be TStreams! 5) New Relative Path Information option. 6) Unzip archives that weren't zipped with the Relative Path option turned on as if they had been by determining how much path information to use with the Rootpath property. 7) Modify timestamps for files in existing archives (you could already modify filenames and pathnames for files in existing archives) 8) The OnBadPassword event now allows you to supply a new password and try the same file again when unzipping. 9) Source code has been cleaned up so that it will compile under Borland C++ Builder with no modifications. Also some bugs were fixed, most importantly: 1) An empty file, that had been compressed into an archive would cause any file added to the archive to cause the archive to approximately double in size. Any archives containing empty files are not corrupted, they are OK. This was simply a fix to the way the archive was processed. 2) After creating an SFX file, you had to close the zip file before you could modify it in any way, otherwise a stream read error was encountered. See the Help file for more information on new features. This zip file is part of a self contained installation program. Just run it and the installation program will begin. Contact vclzip@bigfoot.com for further information Thanks! Kevin Boylan

在Java程序设计的学习和实践过程中，会涉及到一系列的基础知识点和概念，这些内容对于理解和掌握Java语言至关重要。在给出的深圳大学-Java程序设计-相关资料1中，我们可以提炼出以下几点核心知识点： 1. Java语言的数据类型包括基本数据类型和引用数据类型。基本数据类型有八种，包括四种整型（byte, short, int, long）、两种浮点型（float, double）、一种字符型（char）和一种布尔型（boolean）。 2. 标识符的命名规则要求遵循字母、数字、下划线和美元符号的组合，且不能以数字开头。标识符区分大小写。 3. Java语言使用的是Unicode字符集，支持国际化编程。 4. 在进行数据类型转换时，从低精度到高精度类型转换无需显式转换，而从高精度到低精度类型转换时通常需要强制类型转换。 5. Java数组的声明和初始化有多种形式，但必须在使用之前完成声明和初始化。二维数组的声明需要注意其维度的初始化顺序。 6. Java语言中的控制结构包括顺序结构、选择结构（if-else, switch-case）和循环结构（for, while, do-while）。 7. Java的控制流语句如if-else和switch-case提供了条件分支和多分支选择的结构。 8. Java的循环控制语句如for、while和do-while可以控制代码的重复执行。 9. 运算符在Java语言中用于执行数据运算，包括算术运算符、比较运算符、逻辑运算符、位运算符等。 10. Java中的类是对象的模板，可以包含变量（成员变量）和方法（成员方法）。类可以被声明为public或非public，但一个文件中只能有一个public类。 11. 类的继承、封装和多态是面向对象程序设计的三大特性。通过类的继承可以实现代码的复用。 12. Java的关键字包括数据类型关键字、控制流程关键字、访问控制关键字等，例如int、class、public等。 13. Java中的构造函数用于在创建对象时初始化对象的状态。 这些知识点是Java程序设计中最为基础的内容，对于初学者而言，掌握它们是学习Java语言的基础，也是进一步学习Java高级特性的前提。

GS算法（Gerchberg-Saxton算法）是一种用于从强度信息中恢复相位信息的算法，最初于1972年提出，广泛应用于光学领域。该算法的基本流程包括初始估计、傅里叶变换、频域约束、逆傅里叶变换和空域约束的迭代过程。原始GS算法在空域约束时直接使用目标振幅，容易陷入局部最优解。而Fienup算法通过引入反馈调节量（步长α）改进了约束条件，显著提高了收敛速度。文章还提供了MATLAB代码实现，对比了两种算法的运行结果，展示了Fienup算法在相位恢复和模拟衍射输出上的优势。 GS算法，即Gerchberg-Saxton算法，是一种在1972年被提出，用于在已知强度信息的条件下恢复波前相位信息的计算方法。该算法主要用于光学领域，尤其在光学系统的相位恢复及计算光学领域有广泛应用。GS算法的基本原理是通过迭代的方式逐步接近真实的相位信息，其核心步骤包括对强度信息的傅里叶变换、应用频域约束条件、以及进行逆傅里叶变换来更新空域信息。通过反复迭代，算法可以逐步修正相位信息，最终达到波前恢复的目的。 GS算法的迭代过程首先需要一个合理的初始相位估计值，然后通过傅里叶变换将其转换到频域，在频域中对相位进行调整，使之满足已知的振幅信息。接着，通过逆傅里叶变换将调整后的频域信息转换回空域，再根据空域中的振幅信息进行调整，以此循环往复直至得到满意的结果。然而，GS算法的一个主要问题是其迭代过程可能会被局部最优解所困，导致恢复过程的效率和准确性受限。 为了解决这一问题，后续研究中提出了Fienup算法。Fienup算法是对GS算法的一个重要改进，它通过引入反馈调节量（步长α）来优化频域和空域的约束条件，有效避免了局部最优解的陷阱，大大提高了算法的收敛速度和恢复精度。Fienup算法的提出，为相位恢复问题的解决提供了更为高效和稳定的途径。 文章中提到了MATLAB代码的实现，将GS算法和Fienup算法进行了对比。通过具体的编程实现，可以看到Fienup算法在相位恢复和模拟衍射输出方面相比于原始的GS算法有着明显的优势。MATLAB作为一种广泛使用的数值计算软件，提供了强大的矩阵运算和数据处理能力，这使得算法的验证和实验变得更加方便快捷。代码实现部分可能包括对初始估计的生成、傅里叶变换和逆变换的实现、以及如何在迭代过程中应用频域和空域的约束条件等关键步骤的详细描述。 此外，这篇文章也为读者提供了更加直观的算法效果展示，通过图形化的方式对比了GS算法和Fienup算法在不同迭代次数下的恢复结果，使读者能够更加直观地理解两种算法的性能差异。通过这种直观的展示，研究者和工程师可以更加容易地根据实际需要选择合适的算法进行相位恢复。 光学算法、相位恢复、MATLAB是与GS算法相关的三个关键领域。光学算法涉及到光波传播和相互作用的数学描述；相位恢复则是光学测量和成像中的关键步骤；MATLAB作为一种科学计算软件，为这些复杂算法的实现提供了有效的工具。这些领域之间的交叉融合对于推动光学技术的发展起到了重要作用，特别是对于光学测量和图像处理等领域，精确的相位恢复技术可以带来更为清晰和精确的图像，从而提高光学系统的性能。

### edaSDRAM相位角计算详解 #### 一、引言 在现代电子系统设计中，特别是涉及高速数据传输的应用场景中，如高性能计算、通信设备等，正确配置和优化存储器接口变得尤为重要。本文将详细介绍如何进行SDRAM相位角计算，包括所需的关键参数获取方法以及具体的计算步骤，并通过一个实际案例来加深理解。 #### 二、芯片说明 在本案例中，我们将使用以下两种主要组件： - **FPGA芯片**：EP3C55F484I7 - **SDRAM芯片**：MT48LC16M16A2 这些芯片的选择基于它们在性能和成本之间的良好平衡，适合用于需要大量数据快速交换的应用场景。 #### 三、计算所需参数的获取 ##### 3.1 SDRAM参数获取 为了获取SDRAM的相关参数，需要参考该芯片的数据手册。可以通过访问`datasheet5.com`网站下载MT48LC16M16A2的数据手册。以下是关键参数及其含义： - **TDS (Input Data Setup Time)**：1.5 ns — 输入数据建立时间，即数据必须稳定在时钟边沿之前的时间。 - **TDH (Input Data Hold Time)**：0.8 ns — 输入采样保持时间，即数据必须在时钟边沿之后保持稳定的时间。 - **THZ (Output High Impedance Time)**：5.4 ns — 输出高阻时间，即输出信号变为高阻态的时间。 - **TOH (Output Hold Time)**：3 ns — 输出保持时间，即输出信号在时钟边沿之后必须保持有效的时间。 ##### 3.2 FPGA参数说明 对于FPGA参数的获取，首先需要设置时间分析模式，确保选择“Use Classic Timing Analyzer during compilation”。接下来，根据原理图连接情况，进行编译以获得所需的参数： - **TCOMAX (Maximum Clock-to-Output Delay)**：3.537 ns — TCO中最大的数，即从时钟输入到输出信号稳定的最大延迟。 - **Tsu (Maximum Input Data Setup Time)**：7.952 ns — tsu中最大的数，即输入数据必须在时钟边沿之前稳定的最大时间。 - **Th (Maximum Input Data Hold Time)**：7.4 ns — th中最大的数，负数去掉符号后再比较，即输入数据必须在时钟边沿之后稳定的最大时间。 - **TCOMIN (Minimum Clock-to-Output Delay)**：1.942 ns — tco中的最小值，即从时钟输入到输出信号稳定所需的最小延迟。 这些参数对于后续的相位角计算至关重要。 #### 四、相位角计算公式 根据上述参数，我们可以计算出读写操作中的超前和滞后的相位角，这对于正确配置PLL（Phase-Locked Loop，锁相环）至关重要。具体计算步骤如下： - **读滞后**：Toh - Thmax = 10.4 ns - **写滞后**：Tclk - TCOMAX = 3.433 ns (假设时钟频率为29.5M * 4，Tclk表示时钟周期) - **读超前**：Tcomin - TDH = 1.142 ns - **写超前**：Tclk - THZ - Tsu = -4.882 ns 最终需要计算的数值为：\[ \frac{(-3.433 - 4.882)}{2} = -4.1575 \] ns。 #### 五、设置PLL参数并编译下载 根据上述计算结果，我们需要设置PLL的相位调整值为-4.15 ns。这一步骤可以通过FPGA开发软件中的相应设置完成。具体设置方法依赖于所使用的开发环境，但通常涉及到调整PLL模块中的相位偏移值。完成设置后，重新编译项目并下载到目标FPGA板卡上。 #### 六、总结 通过对SDRAM相位角的精确计算，可以有效地提高数据传输的可靠性和稳定性，从而提升整个系统的性能表现。希望本文提供的步骤和示例能够帮助读者更好地理解和掌握这一过程。如果您在实践中遇到任何问题或有更深入的探讨需求，请随时联系作者：qx7873087@163.com，一起交流学习。

根据提供的文件信息，本文将详细解析SDRAM相位角计算的相关知识点，包括所需参数的获取方式、具体计算方法以及PLL参数设置等内容。 ### 一、SDRAM与FPGA概述 SDRAM（Synchronous Dynamic Random Access Memory）即同步动态随机存取存储器，是一种广泛应用于计算机系统中的内存类型。其主要特点是通过同步时钟信号实现数据访问，能够提供高速的数据传输速率。FPGA（Field-Programmable Gate Array）即现场可编程门阵列，是一种高度可编程的集成电路，常用于实现复杂的数字逻辑功能。 在FPGA设计中，正确配置SDRAM接口对于确保系统稳定性和提高性能至关重要。其中一个关键步骤就是SDRAM相位角的计算。 ### 二、SDRAM参数获取方法 #### 1. SDRAM参数获取 SDRAM的参数主要来自于其数据手册。通常，这些参数包括但不限于： - **TDS**（Input Data Setup Time）: 输入数据建立时间，即数据需要在时钟边沿之前保持稳定的时间。 - **TDH**（Input Data Hold Time）: 输入采样保持时间，即数据需要在时钟边沿之后保持稳定的时间。 - **THZ**（Output High Impedance Time）: 输出高阻时间，即输出端口由正常状态转换为高阻态的时间。 - **TOH**（Output Hold Time）: 输出保持时间，即输出数据需要在时钟边沿之后保持稳定的时间。 这些参数可以从SDRAM芯片的数据手册中获取。例如，MT48LC16M16A2型号的SDRAM，可以在数据手册中找到对应的参数值。 #### 2. FPGA参数说明 FPGA参数的获取需要进行特定的设置，以确保能够准确获取到所需的时序参数。以下是一些关键步骤： - **设置时间分析模式**：在FPGA设计工具中，需要设置使用经典时序分析（Classic Timing Analyzer），以便获取到准确的时序数据。 - **连接SDRAM接口**：根据设计需求，正确连接SDRAM接口至FPGA。需要注意的是，某些参数（如c1）可能需要预先计算得出。 - **编译获取参数**：完成设计并编译后，可以从编译报告中获取到关键参数，如TCOMAX、Tsumax、THmax等。 ### 三、SDRAM相位角计算 #### 公式计算 相位角计算基于FPGA和SDRAM的关键时序参数。计算过程中，主要关注以下几个方面： - **读滞后**（Read Skew）: Toh - THmax - **写滞后**（Write Skew）: Tclk - TCOMAX - **读超前**（Read Advance）: TCOMIN - TDH - **写超前**（Write Advance）: Tclk - THZ - TSUmax 其中： - **Toh** 代表SDRAM的输出保持时间； - **THmax** 代表SDRAM的最大输出高阻时间； - **Tclk** 代表时钟周期； - **TCOMAX** 代表FPGA的最大时钟输出时间； - **TCOMIN** 代表FPGA的最小时钟输出时间； - **TDH** 代表SDRAM的输入采样保持时间； - **THZ** 代表SDRAM的输出高阻时间； - **TSUmax** 代表SDRAM的最大输入数据建立时间。 通过以上公式，我们可以计算出读写操作的滞后和超前时间，进而确定SDRAM相位角的具体数值。 #### 计算实例 以文中提供的数据为例： - 读滞后：Toh - THmax = 10.4ns - 写滞后：Tclk - TCOMAX = 3.433ns (假设时钟频率为29.5MHz * 4) - 读超前：TCOMIN - TDH = 1.142ns - 写超前：Tclk - THZ - TSUmax = -4.882ns 根据上述计算结果，可以得出相位角的最小值（min）为： \[ \text{min} = \frac{-3.433 - 4.882}{2} = -4.1575\text{ns} \] ### 四、PLL参数设置 PLL（Phase-Locked Loop）即锁相环，是一种用于调节频率和相位的电路。在FPGA设计中，通过调整PLL参数，可以实现对SDRAM时序的精确控制。 #### 设置PLL参数 根据上述计算结果，可以将PLL的相位差设置为-4.1575ns（文中简化为-4.15ns）。具体设置步骤如下： - 在FPGA开发软件中打开PLL配置界面。 - 调整PLL参数，使输出时钟相对于输入时钟具有所需的相位偏移。 - 完成设置后，重新编译并下载设计至FPGA板。 通过这种方式，可以确保SDRAM与FPGA之间的数据交换在正确的时序下进行，从而提高系统的整体性能和稳定性。 SDRAM相位角计算是一项重要的FPGA设计任务，涉及到了解和应用SDRAM及FPGA的关键时序参数。通过合理设置PLL参数，可以有效地优化系统性能。