ragile fancontrol hld.md

Feature Name

Fan control policy for ragile device.

High Level Design Document

Rev 0.1

Table of Content

• Feature Name
  • High Level Design Document
  • Table of Content
    • Revision
    • Scope
    • Definitions/Abbreviations
    • Overview
    • Requirements
    • Design
      • Platform capabilities
      • Platform restrictions
      • Policy
      • Emergency policy
    • SAI API
    • Configuration and management
      • CLI/YANG model Enhancements
      • Config DB Enhancements
    • Warmboot and Fastboot Design Impact
    • Restrictions/Limitations
    • Testing Requirements/Design
      • Unit Test cases
      • System Test cases
    • Open/Action items - if any

Revision

Rev	Date	Author	Change Description
0.1	09/22/2021	Ragile Team	Initial version

Scope

This document gives the details of fan control design for Ragile device.

Definitions/Abbreviations

Definitions/Abbreviation	Description
MAC	Medium access control chip
BOARD	Motherboard
CPU	Central processing unit
INLET	Air inlet
OUTLET	Air onlet
CPLD	Complex programmable logic device
FPGA	Field-programmable gate array

Definitions/Abbreviation	Description
INLET_T	Temperature detection point of INLET
OUTLET_T	Temperature detection point of OUTLET
CPU_T	Temperature detection point of CPU
BOARD_T	Temperature detection point of BOARD
MAC_T	Temperature detection point of MAC

Overview

In order to ensure the stable of the networking switch at an appropriate temperature, this document provides an structure of fan control based on the temperature points.

Requirements

The functional requirements include:

• A stable method of obtaining four temperature points. Like etc.
• A method to control Fans, like CPLD or FPAG etc.
• An common platform API that encapsulates above content.

Design

Platform capabilities

• Fan speed is described by 0 ~ 255(0 ~ 0xff) levels, 0 means stopped and 255 means maximum, the default level is 96(0x60)
• The temperature detection points involved are CPU_T, INLET_T, OUTLET_T, BOARD_T, MAC_T.
• Support fan redundancy.

Platform restrictions

• For safety reasons, level 0 is not allowed, the minimum is limit to 51(0x33).
• Fans with opposite directions are not allowed.

Policy

• When keep level 96(0x60).

• When and device temperature is growing, calculate fan speed following formula:

• When and device temperature is cooling down, there is two policy:
  • When use formula above.
  • Otherwise keep the original speed.

Definitions	Description
	Inlet temperature
	Minimum allowable temperature
	Maximum allowable temperature
	Speed of fan
	Minimum speed of fan
	Maximum speed of fan
	Slope of fan speed and temperature
	Inlet temperature measured last time
	Inlet temperature measured current
	Fuse that determine whether to trigger fan control

Emergency policy

• When the device status fails three times in a row, set level as 187(0xbb) until it back to normal. Then restart the control policy.

• There is two way to determine device status:

  1. Error reading temperature point.
  2. or

• When

  or or or or

  enter the warning alram state, print corresponding log, turn state LED to amber and adjust all Fans to full speed.

• When

  or and

  enter critical alarm state, print corresponding log, turn SYS_LED to red. When any one of the following two conditions is met, reset the machine.

  • 
  • and and and

• To avoid jitter in the measurement, warning state and crital state need to be verified twice.

Definitions	Description
	Outlet temperature
	MAC temperature
	BOARD temperature
	CPU temperature
	Error temperature low threshold, default is -50C°
	Error temperature high threshold default is 50C°
	MAC warning temperature threshold
	MAC critical temperature threshold
	OUTLET warning temperature threshold
	OUTLET critical temperature threshold
	INLET warning temperature threshold
	INLET critical temperature threshold
	CPU warning temperature threshold
	CPU critical temperature threshold
	BOARD warning temperature threshold
	BOARD critical temperature threshold

SAI API

NA

Configuration and management

NA

CLI/YANG model Enhancements

NA

Config DB Enhancements

NA

Warmboot and Fastboot Design Impact

NA

Restrictions/Limitations

NA

Testing Requirements/Design

NA

Unit Test cases

Run command show platform fan status to check current fan speed, alias FAN_SPEED

• Pluging out one or more fans, check FAN_SPEED is it running at full speed after
• Heating up chips to warning threshold, check FAN_SPEED is it running at full speed.
• Heating up chips to critical threshold, check system is it reset.

System Test cases

NA

Open/Action items - if any

NA