XO1.5 1.2V power sequencing ECO

From OLPC

Revision as of 05:45, 9 August 2009 by Wad (Talk | contribs)
(diff) ← Older revision | Current revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This describes a change to early prototypes of the XO-1.5 motherboard.

Contents

Scope

This engineering change order fixes a problem with the sequencing of power supply turn-on to the VX855 chip, which prevent power up. Boards having this problem will occasionally fail to turn on, indicated by the EC flashing the red battery charge LED. Laptops needing this ECO will also fail to resume properly after a suspend (the light are on, but nobody is home).

These power issues are exacerbated by suspend/resume testing.

There is a similar ECO for problems when initially turning on the laptop.

This only applies to XO-1.5 A-phase (both A1 and A2) motherboards.

Problem Description

When powering up, the A1 prototypes bring up all run-time power supplies simultaneously, as partially shown in Figure 1:

  • Green - VCCP (1.05V supply for processor bus)
  • Magenta - +1.2V (VDD supply to VX855)
  • Yellow - VCORE (to processor)
  • Cyan - 1.8V_GOOD (This signal has a pullup to the +3.3V supply, and shown both the timing to +1.8V being good simultaneous with +1.2V and VCCP, as well as a later turn-on of the +3.3V supply, which enable the VCORE turn-on.)

The VX855, while relatively tolerant of run-time power sequencing, does specify that +1.2V (VDD) be supplied before the others. It is likely that other supplies such as +1.8V and +3.3V are currently supplied to the chip before +1.2V.

To correct this, the VX855 SUSB (MAIN_ON) signal is run solely to the +1.2V supply, and the +1.2V_GD (power good) signal is used to drive the MAIN_ON net turning on the remaining supplies. The +1.2V quick drain circuit is disabled by this ECO, the addition of an inverting transistor is required to fix it. The resulting timing is shown in Figure 2:

  • Green - VCCP
  • Magenta - 1.8V_GOOD (requires +3.3V supply)
  • Yellow - VCORE
  • Cyan - +1.2V (Now happens 8 mS before the others -- we can adjust the +1.2V_GD time constant by removing/changing PC16.)

A second problem, not visible in the above traces, is oscillation on the SYSOK signal, especially during resume. It is to correct this problem that I recommend removing U3, and wiring pins 2 and 4 together.

Procedure

  1. Remove R234 (right next to RTC crystal).
  2. Lift PU7, pin 6 (near PL6) and tie it to PRE_MAIN_ON (pad of R234 closest to VX855).
  3. Remove U3 and tie U3, pin 2 to U3, pin 4.
  4. Tie the MAIN_ON (U3, pin 1) signal to +1.2V_GD (D12, cathode, marked by several parallel lines at one end).
  5. Remove PC16 and PC17

At one point, the removal of PR67 was also recommended. This was incorrect. For optimal suspend/resume performance, leave PR67 present.

Results

Experience with the A2 boards indicates that this is a required ECO for proper suspend/resume operation.

Done to:

  • Isabel - which went from frequent red LEDs to never showing one.
  • Graham - where it fixed problems with resume leading to the flashing red light
  • CJB - which went from occasional flashing red LEDs to none.
  • All A2 motherboards
Personal tools
  • Log in
  • Login with OpenID
About OLPC
About the laptop
About the tablet
Projects
OLPC wiki
Toolbox