+荐课 提问/讨论 评论 收藏 分享
  • 课程目录
  • 相关资源
  • 课程笔记

Welcome to the PowerManagement Framework tutorial for the CC3220 wireless MCU. The objective ofthis tutorial is to understand various low powermodes for the CC3220 device, and to familiarize the user withthe Power Management Framework. After this tutorial, youwill be able to understand the various low power modessupported by the CC3220 MCU, work with the Power ManagementFramework within TI Drivers, configure the peripheralsusing TI Drivers, initialize and enable thePower Management Structure, specify I/O parking for lowpower modes, and finally, debug the PowerAware application. The CC3220 device supportsthe following power modes-- sleep mode-- which is a standardlow power mode for the ARM Cortex-M4 processor, lowpower deep sleep mode-- which I'll be referring to youas LPDS for this or tutorial. In this mode, the functionalclock and peripheral clocks are gated-- to slow clock continuesrunning in the background. Lastly we have hibernate,which is an extremely low power mode that shuts down boththe application and network processors. The slow clock continues torun in this mode as well. This diagramshowcases the features of each of these power modes,along with the relative current consumption and wake-up time. The arrows is on theright depict two things. One, that thecurrent consumption decreases from sleepto hibernate, and also conversely, that thewake-up time increases from sleep to hibernate. You can see how theretention of the RAM, CPU, and peripheral Contexts areaffected by the various power modes and what wake-upsources can be used. Let's go through eachcities a little more detail. In sleep mode, theprocessor clock is stopped, whichhalts code execution. The peripheralconfigurations are retained althoughtheir clocks are gated, which results in the stoppingof any ongoing peripheral transaction if the sleepclock is not enabled. The contents of SRAM andthe slow clock counter are also retainedin the sleep mode. The wake-up time is veryshort in sleep mode, and any interruptcan wake up the MCU. Something to note here isthat the debugger disconnects when the MCU enters sleep mode. Like sleep mode, low power deepsleep mode gates the processor clock while the slowclock counter is retained. Unlike sleep mode,the peripheral clocks are also gated, resultingin configuration loss. However, the SRAM canbe retained either fully or partially. The 256 kilobytes of SRAMis divided into four blocks of 64 kilobytes each. Any number of blocks can beretained in low power deep sleep, but a partialblock cannot be retained. This means the peripheralconfiguration can be retained in RAM by the softwarebefore entering LPDS and can be restoredby retrieving from RAM after waking up. The wake-up time isaround 3 milliseconds executing software overhead. In LPDS, not every interruptcan wake up the MCU. Only a specified set ofprogrammable wake sources are available for this mode. The availableawake sources are-- an interrupt from the networkprocessor, LPDS wake timer-- the duration of which can bespecified in the application. Also, one of a subsetof GPIOs can be set as an external wake-up source. Like sleep mode, thedebugger gets disconnected when the MCU enters LPDS. LPDS only affectsthe application MCU and not the networkprocessor, which essentially means any networkingconfiguration, including the connection to an AP, andall internet socket connections, will be retained acrossentry to and exit from LPDS. With correctparking of the pins, current consumption canbe as low as 1 milliamp with the devicemaintaining the connections to the AP and the internet. This mode is ideal forscenarios where the application remains idle for amajority of the time, but still needs to remainconnected to the network so that it can performsome action upon receiving a signal from the network. Hibernate, unlikeLPDS and sleep, affects both the appsand NWP processors, resulting in extremelylow current consumption. It essentiallyshuts down the chip with only a small numberof registers retained. The slow clock counter iskept alive across this mode. Network configurationsare also lost. An interrupt from theslow clock counter can wake the systemfrom hibernate. Also, a specified set of GPIOscan be set as a wake-up source simultaneously. Though the current consumptionis extremely low in this case, it has a relatively largewake-up time than LPDS. This mode is ideal forthe cases where the device will be inactivefor most of the time and will only have to performsome action periodically, or based on some GPIO activity. So far, we havediscussed the low power modes and their features. In the next slides, we'lldiscuss, from the software perspective, how do youuse low power functionality in an application. The Power ManagementFramework, which will be referred to as PMModule, or PM Framework in this tutorial, is partof the TI Drivers library, and is responsiblefor exercising the various low power modes. Though it can be usedalone without exercising the peripheral driversfrom this library, it is highly recommended to useTI Drivers for the peripherals as well. The reason being, that allthe drivers are power aware, and register themselves withthe Power Management Module. This simplifies the managementof peripheral configuration, where the MCU iscycling through LPDS. TI Drivers are a collection ofeasy to use interface drivers for most of the peripherals. The interface is similar withTI Drivers for other MCUs from Texas Instruments. It sits atop of driverlib,which is a low level peripheral driver library. Use of generic RTOS APIs makethe drivers re entrant, which means they can work with anyRTOS or even without an RTOS, provided a portinglayer is provided. The CC3220 SDK providesthe porting layer for TI RTOS, Free RTOS,and non-OS environments. These drivers arepower aware and are supported for the followingtool chains in the SDK-- Code Composer Studio,which is a free IDE from Texas Instruments, IAREmbedded Workbench, and GCC. Like previouslymentioned, PM Framework is integrated with TI Drivers. Each device driver registersthemselves with the PM Module when opening the interface. The application needs toinitialize the PM Module at the start, thoughit can be enabled or disabled in the run time. Typically, upon entering LPDS,the peripheral clocks are gated and the peripheralconfigurations will be lost. The PM Module takes careof enabling the clock after coming out of LPDS. It also restores the peripheralconfiguration to the state that it was beforeentering LPDS. If the pins are not parkedproperly before entering LPDS it can lead to current leakage. Therefore, it is necessaryto park the pins. PM Module takescare of pin parking as per the parkingstructure, which is specified wheninitializing the PM Module. Although the PMModule takes care of managing the peripheralconfigurations and pin parking, it still allows forregistration of hook functions for various power events likeentering LPDS, exiting LPDS, and entering Hibernate. If working in an RTOSenvironment, and PM Framework is enabled, PowerManagement checks with all the registereddrivers to check if the MCU can be put into LPDS. Only after confirming thatnone of the peripherals is currently activedoes it allow the apps MCU to enter LPDS. Also, it makes sure that ithas sufficient time to cycle through an LPDS cycle. So if a task is about to beactive again very shortly, Sleep load is exercisedinstead of LPDS. If working in anon-OS environment, the Power Management Functionmust be called explicitly by the application,after which, it makes similar checks todecide if the app's MCU can enter LPDS. Unlike LPDS, afunction call needs to be made by the applicationto put the SOC into Hibernate. Regardless of whetheror not the PM Framework is enabled, onceits API is invoked, the system will enter Hibernate. The duration of time thedevice will remain in Hibernate is passed as anargument to a function. The device will enterHibernate if the duration is above a minimumamount, and will remain so until either thetimer has elapsed or there is anexternal trigger event. The GPIO wake-up sourcecan be specified in the PM Configuration Structure. I/Os can be retainedduring Hibernate, which are also specified inthe PM Configuration Structure. The developer needs to makesure of the following things when using Hibernate mode-- stop the network processorbefore entering Hibernate, disable the retention afterwaking from Hibernate, the retention for SFlash pins,which are specified as group 1, is already disabledby the boot loader. For other groups,the retention must be disabled by theapplication, but only after reconfiguring the pins. A sample PowerManagement Structure and a brief explanationfor each of these fields is provided below. The pin parkingstructure for LPDS can be specified in PMConfiguration Structure. It is extremely importantto park the pins appropriately to getlower current consumption. Following are the possibleparking states for any pin. I sample parkingstructure is as follows-- I/Os can be retained inhibernate mode as groups, individual retention forthe I/Os is not supported. Shown here are the four groupswhich consume most of the I/Os. Again, make sure todisable retention after waking up from Hibernate. The debugger willdisconnect no matter which low power the device enters. It is possible toreconnect the debugger once the device comes out ofLPDS and stays in active mode. Using settings which keep thedebug interface alive in LPDS and do not park JTag pinsare necessary if it is wished to reconnect to the debugger. To reconnect the debuggerafter waking up from LPDS, the application needs to makesure the app's MCU does not enter LPDS again immediately. The following canbe used to do so-- disable the power policyand LPDS exit hook function, or put a usercontrolled while loop which can be brokenafter reconnecting back from the debugger. Most of the settingsshown thus far are used in thefollowing SDK examples. Idle Profile--this exercises LPDS and is an RTOS-based examplewith different RTOSs being used per CCS and IAR. Idle Profile Nonos exercisesLPDS and is a nonon example. Sensor Profile is an applicationthat exercises Hibernate mode. All of these examples
课程介绍 共计1课时,12分32秒


SimpleLink Wi-Fi 电源管理 框架

欢迎学习适用于 CC3220 无线 MCU 的 电源管理框架教程。 本教程的 目标是了解 CC3220 器件的 各种低功耗模式, 并使用户熟悉 电源管理框架。 在学习本教程之后, 您将能够了解 CC3220 MCU 支持的各种 低功耗模式, 在 TI 驱动程序内 使用电源管理框架, 使用 TI 驱动程序 配置外设,初始化 并启用电源 管理结构, 为低功耗模式 指定 I/O 设置, 最后调试功率 感知应用。


我们在这里所谈论的 “未使用的运放” 不是指在芯片储藏箱或防静电袋中的运放;而是指在同一个封装里面的多个运放中未被使用的部分。最近论坛中的一个提问促使我来研究这个问题,在处理这个问题时,我无意中看到一篇由我同事Todd Toporski发表的好文章(点击此处,查看原文)。他非常出色地概括了关于这个问题的几个重要方面及其原因。这里,我总结一下并加入了一些自己的想法。 最好将未使用的运放连接为一个带...
maylove 模拟与混合信号
       在过去只使用电池的时候,单电源完全就是个行业术语。如果市场营销人员将“单电源”标签贴在放大器上,所带动的兴趣和销量将会体现出这类产品的吸引力。现在仍是如此,但您的确需要知道这个词的真正含义。    ...         在过去只使用电池的时候,单电源完全就是个行业术语。如果市场营销人员将“单电源”...
qwqwqw2088 模拟与混合信号
一、零线和地线区别 零线是在供电端(发电厂、变电站、变压器)接地,或在入户前重复接地,是工作接地线,是输电线路的一部分,电流经电厂→火线→负载→零线反回电厂。 地线在用户端接地,和用电电器的金属外壳或人体可触部位连接,使机壳与大地等电位,保护人体不触电。零线不与输电线路构成回路,正常情况下没有电流。注意两者的区别: “ 1.零线和地线这两个是不同的概念,不是一回事,千万别互换或混接。...
qwqwqw2088 模拟与混合信号
msp430f5438 RTC日历怎么校准以及怎么跟现实时间同步?
一个项目要日历,不知道msp430f5438 日历准不准以及怎么跟现实时间同步以及怎么校准。如果给5438断电的话是不是内部的原先对日历校准的数据丢了,是不是上电又要校准这样用好象不方便?我这样理解对不对,初次使用MSP430,希望得到大家的帮助!...
lyzjhzdz 微控制器 MCU



2019年12月09日 10:18:18



2019年11月17日 10:00:16



2019年09月10日 08:22:24



2019年08月04日 09:06:27



2019年06月17日 01:15:54



2019年06月05日 16:05:16



2019年04月11日 17:21:52



2019年03月12日 19:30:21


很有用的课程啊 啊 啊啊

2019年03月12日 12:18:12


About Us 关于我们 客户服务 联系方式 器件索引 网站地图 最新文章 手机版

站点相关: EEWORLD首页 EE大学堂 论坛 下载中心 Datasheet 活动专区 博客

北京市海淀区中关村大街18号B座15层1530室 电话:(010)82350740 邮编:100190

电子工程世界版权所有 京B2-20211791 京ICP备10001474号-1 电信业务审批[2006]字第258号函 京公网安备 11010802033920号 Copyright © 2005-2023, Inc. All rights reserved