Hand Crank Low Battery Recovery: Difference between revisions

From OLPC
Jump to navigation Jump to search
(Rolled back editing even farther)
No edit summary
 
(3 intermediate revisions by 3 users not shown)
Line 8: Line 8:
===Nature of Power Generated by Human Powered Generators===
===Nature of Power Generated by Human Powered Generators===


The power generated from a human powered generator such as the yo-yo design currently being considered is periodic in nature. A high voltage (+20V) at reasonable currents is provided when the generator is pulled and for several seconds afterwards. There is usually a similar amount of time between pulls during which no power is supplied.
The power generated from a human powered generator such as the [[Peripherals/YoYo|yo-yo design currently being considered]] is periodic in nature. A high voltage (+20V) at reasonable currents is provided when the generator is pulled and for several seconds afterwards. There is usually a similar amount of time between pulls during which no power is supplied.


If a battery is in good condition (not extremely discharged), there should be no problem recharging the battery with power from a human powered generator.
If a battery is in good condition (not extremely discharged), there should be no problem recharging the battery with power from a human powered generator.
Line 23: Line 23:


==Solutions==
==Solutions==

===XO-1.75===
Late in production of the [[XO-1.5]] (starting with the [[XO_1.5_C5|C5]] motherboard), the battery charger was changed the Intersil ISL9519. This battery charger uses a much more aggressive method of trickle charging, which rarely requires more than a couple of seconds to restore a battery to the normal charging regimen. This charger design is also used in [[XO-1.75]], and appears to solve this problem.

===Earlier Thoughts===


One thought is to redesign the 3VPCU rail to be more independent of VIN.
One thought is to redesign the 3VPCU rail to be more independent of VIN.


Is it possible for the EC to detect this condition ?
Is it possible for the EC to detect this condition ?

Is it possible to store more of the power from one pull of the generator, in a capacitor which would then supply the EC and trickle charger for X seconds or minutes? The cap could either be on the motherboard, or on the generator. The user would still have a slow job -- they'd have to pull the generator perhaps once a minute. But that's better than having to pull it 20 times a minute for trickle charging to work.

:This is certainly an option, but it is better to dump the charge directly into the battery. It can take it, and no expensive power storage device is required.

[[Category:Battery & Power]]

Latest revision as of 02:57, 19 September 2011

This page describes the problems with recovering from an extremely discharged battery when using a human powered generator.

These problems were first discussed in the March 9-10 XO design review meeting in Cambridge.

Problem Description

The problem only occurs in the case of trying to recharge a battery which has been extremely discharged using only a human powered generator.

Nature of Power Generated by Human Powered Generators

The power generated from a human powered generator such as the yo-yo design currently being considered is periodic in nature. A high voltage (+20V) at reasonable currents is provided when the generator is pulled and for several seconds afterwards. There is usually a similar amount of time between pulls during which no power is supplied.

If a battery is in good condition (not extremely discharged), there should be no problem recharging the battery with power from a human powered generator.

Recovery from a Low Battery Condition

If a laptop attempts to charge an extremely discharged battery (here defined as a battery whose output voltage is less than or equal to 4.8V), it will pull the VIN power rail down to a voltage below that necessary for operation of the laptop. As this causes erratic laptop operation, the embedded controller (EC) is programmed to instead trickle charge batteries which are extremely discharged (including those discharged so low that the chip inside the battery pack is no longer powered).

This trickle charging consists of providing a small current to the battery for a several hundred milliseconds, every couple of seconds. The EC does this continuously whenever it has a valid mains input (ACIN on GPIAD2) and it cannot communicate with the battery pack.

Interaction

The problem is that while the power provided by the yo-yo will eventually trickle charge the battery, it is very inefficient. We need to develop a way to dump more of the power from the hand generator into the battery.

Solutions

XO-1.75

Late in production of the XO-1.5 (starting with the C5 motherboard), the battery charger was changed the Intersil ISL9519. This battery charger uses a much more aggressive method of trickle charging, which rarely requires more than a couple of seconds to restore a battery to the normal charging regimen. This charger design is also used in XO-1.75, and appears to solve this problem.

Earlier Thoughts

One thought is to redesign the 3VPCU rail to be more independent of VIN.

Is it possible for the EC to detect this condition ?

Is it possible to store more of the power from one pull of the generator, in a capacitor which would then supply the EC and trickle charger for X seconds or minutes? The cap could either be on the motherboard, or on the generator. The user would still have a slow job -- they'd have to pull the generator perhaps once a minute. But that's better than having to pull it 20 times a minute for trickle charging to work.

This is certainly an option, but it is better to dump the charge directly into the battery. It can take it, and no expensive power storage device is required.