Difference between revisions of "Orchard validation matrix"

From Studio Kousagi Wiki
Jump to: navigation, search
(→‎Validation Matrix)
(→‎Power Consumption)
 
(6 intermediate revisions by the same user not shown)
Line 12: Line 12:
 
| WS2812B driver || OK ||  || confirmed CPU has GPIO performance to operate on-board LEDs; extended chain testing TBD
 
| WS2812B driver || OK ||  || confirmed CPU has GPIO performance to operate on-board LEDs; extended chain testing TBD
 
|-
 
|-
−
| I2C || OK || ||  
+
| I2C || OK || OK ||  
 
|-
 
|-
−
| SPI || OK || ||
+
| SPI || OK || OK ||
 
|-
 
|-
 
| OLED || OK  ||  ||  
 
| OLED || OK  ||  ||  
Line 20: Line 20:
 
| UART || OK ||  || Works well enough for chibiOS
 
| UART || OK ||  || Works well enough for chibiOS
 
|-
 
|-
−
| Integrated radio || ||  ||  
+
| Integrated radio || OK ||  ||  
 
|-
 
|-
−
| BLE || ||  ||  
+
| BLE || OK ||  ||  
 
|-
 
|-
 
| Accelerometer || OK  ||  ||  
 
| Accelerometer || OK  ||  ||  
 
|-
 
|-
−
| Microphone || ECO1 / OK || || Schematic symbol mirrored; requires board respin. Tested OK on EVT1b
+
| Microphone || ECO1 / OK || OK || Schematic symbol mirrored; requires board respin. Tested OK on EVT1b
 
|-
 
|-
−
| MPR121 || OK || ||  
+
| MPR121 || OK || OK ||  
 
|-
 
|-
−
| Jog dial + buttons || OK || || OK for direct touch on dielectric. Not yet tested for capacitive projection through ABS bezel.
+
| Jog dial + buttons || OK || OK ||
 
|-
 
|-
−
| GPIO expander || OK || ||  
+
| GPIO expander || OK || OK ||  
 
|-
 
|-
−
| Fancy battery charger || OK || ||  
+
| Fancy battery charger || OK || OK ||  
 
|-
 
|-
 
| Shipping state power || OK || OK || immeasurable leakage when in "ship mode" e.g. soft battery disconnect state
 
| Shipping state power || OK || OK || immeasurable leakage when in "ship mode" e.g. soft battery disconnect state
 
|-
 
|-
−
| Gas gauge || OK/ECO3 || || gas gauge itself works fine, talks on I2C; but topological position is incorrect and requires ECO3 to be effective
+
| Gas gauge || OK/ECO3 || OK || gas gauge itself works fine, talks on I2C; but topological position is incorrect and requires ECO3 to be effective
 
|-
 
|-
 
| USB charging port || OK ||  ||  
 
| USB charging port || OK ||  ||  
Line 48: Line 48:
 
| Cheesy battery level option || N/A ||  || not populated in EVT1
 
| Cheesy battery level option || N/A ||  || not populated in EVT1
 
|-
 
|-
−
| On-state power characterization || ||  ||  
+
| On-state power characterization || OK ||  ||  
 
|-
 
|-
−
| Standby power characterization || ||  ||  
+
| Standby power characterization || OK ||  ||  
 
|-
 
|-
 
|}
 
|}
Line 56: Line 56:
 
=Power Consumption=
 
=Power Consumption=
  
−
As measured while on battery, ~3.9V through test.
+
As measured while on battery, ~3.9V through test. ~27 degrees C ambient temperature, 70% RH
  
 
* Powered on, 16 LEDs running wave-rainbow at default brightness: 78-79mA
 
* Powered on, 16 LEDs running wave-rainbow at default brightness: 78-79mA
Line 86: Line 86:
 
* FBE-FEI: 21.4mA
 
* FBE-FEI: 21.4mA
 
* Radio stop: 3.0 mA
 
* Radio stop: 3.0 mA
−
* CPU halt: (not executed due to bugs)
+
* CPU to VLLS0: 0.6mA
 +
 
 +
Ship mode:
 +
* 0.09mA - 0.130mA leakage (no hibernate instruction to gas gauge)
 +
 
 +
No difference when hibernation instruction is given to gas gauge.
 +
 
 +
Conclusions:
 +
* VLLS0 stop on Cortex-M0+ should be around 0.0003mA
 +
* So, there is about 0.5mA leakage in the system. This is probably about right, given that the accelerometer, captouch, BLE are still powered up but simply in standby, and also the LDO will have some draw as well. My guess is that if there is stray resistance it's accounting for < 50% of the current leakage budget.
 +
 
 +
==Battery life estimates==
 +
With 1200mAh battery:
 +
* 16 LEDs on a modest pattern will last 14 hours
 +
* LEDs off, but OLED, BLE on: ~30 hours
 +
* VLLS0 sleep will last 2000 hours (80 days)
 +
* Ship mode will last ~10,000 hours (~1 year)
 +
 
 +
With 4000mAh battery:
 +
* 16 LEDs on a modest pattern - 47 hours (~2 days)
 +
* LEDs off, but OLED, BLE on: 100 hours (~4 days)
 +
* VLLS0 sleep: 6,500 hours (9 months) -- this is about self-discharge rate anyways, so probably half this duration
 +
* Ship mode: 30,000 hours (~3.8 years) -- in reality, self-discharge is faster than this.

Latest revision as of 19:48, 4 June 2015

Validation Matrix

Validation matrix
Subsystem basic status extended status notes
Kinetis-W CPU OK basic SWIO functions OK, OS operational
WS2812B driver OK confirmed CPU has GPIO performance to operate on-board LEDs; extended chain testing TBD
I2C OK OK
SPI OK OK
OLED OK
UART OK Works well enough for chibiOS
Integrated radio OK
BLE OK
Accelerometer OK
Microphone ECO1 / OK OK Schematic symbol mirrored; requires board respin. Tested OK on EVT1b
MPR121 OK OK
Jog dial + buttons OK OK
GPIO expander OK OK
Fancy battery charger OK OK
Shipping state power OK OK immeasurable leakage when in "ship mode" e.g. soft battery disconnect state
Gas gauge OK/ECO3 OK gas gauge itself works fine, talks on I2C; but topological position is incorrect and requires ECO3 to be effective
USB charging port OK
USB charging port OC protect
Cheesy battery charger option N/A not populated in EVT1
Cheesy battery level option N/A not populated in EVT1
On-state power characterization OK
Standby power characterization OK

Power Consumption

As measured while on battery, ~3.9V through test. ~27 degrees C ambient temperature, 70% RH

  • Powered on, 16 LEDs running wave-rainbow at default brightness: 78-79mA
  • Halted (debug mode) 16 LEDs frozen on wave-rainbow: 76.5mA
  • Captouch off: 76.5mA (not measurable reduction)
  • LEDs off: 40mA
  • Accel off: 39.8mA
  • BLE off: 33.1mA
  • OLED off: 58.7mA
  • Boost off: 25.3mA
  • gpiox off: 25.3mA
  • PEE-PBE: 22.7mA
  • PBE-FBE: 22.1mA
  • FBE-FEI: 21.5mA
  • Radio off, CPU halted: 3.0mA

Alt-order:

  • Halted, everything on: 77.5mA
  • Captouch off: 77.4mA
  • LEDs off: 40mA
  • Accel off: 39.9mA
  • BLE off: 33.1mA
  • Boost off: 33.2mA
  • OLED off: 24.7mA
  • gpiox off: 25.3mA or 24.6mA?
  • PEE-PBE: 22.7mA
  • PBE-FBE: 22.1mA
  • FBE-FEI: 21.4mA
  • Radio stop: 3.0 mA
  • CPU to VLLS0: 0.6mA

Ship mode:

  • 0.09mA - 0.130mA leakage (no hibernate instruction to gas gauge)

No difference when hibernation instruction is given to gas gauge.

Conclusions:

  • VLLS0 stop on Cortex-M0+ should be around 0.0003mA
  • So, there is about 0.5mA leakage in the system. This is probably about right, given that the accelerometer, captouch, BLE are still powered up but simply in standby, and also the LDO will have some draw as well. My guess is that if there is stray resistance it's accounting for < 50% of the current leakage budget.

Battery life estimates

With 1200mAh battery:

  • 16 LEDs on a modest pattern will last 14 hours
  • LEDs off, but OLED, BLE on: ~30 hours
  • VLLS0 sleep will last 2000 hours (80 days)
  • Ship mode will last ~10,000 hours (~1 year)

With 4000mAh battery:

  • 16 LEDs on a modest pattern - 47 hours (~2 days)
  • LEDs off, but OLED, BLE on: 100 hours (~4 days)
  • VLLS0 sleep: 6,500 hours (9 months) -- this is about self-discharge rate anyways, so probably half this duration
  • Ship mode: 30,000 hours (~3.8 years) -- in reality, self-discharge is faster than this.
Retrieved from "https://kosagi.com/w/index.php?title=Orchard_validation_matrix&oldid=1710"