[Peripherals] Charging the XO off a bicycle dynamo

[Richard A. Smith]

Holden Bonwit wrote:
> Updates will now occur on the wiki at the following section: 
> http://wiki.laptop.org/go/Battery_and_power#Practical_trials
> 
> Please contribute if you can, by posting there, or replying to me.

> The XO charging light and icon come on, but there is insufficient
> current draw available to actually charge. To provide more power, I
> hooked up a second dynamo to the input of the electronics above.
> Again, the charge light came on. I wasn't able to measure current
> draw,

You can use the laptop to measure current draw.  Rather than do your 
tests with the laptop off do them with a half charged battery and the 
laptop on.  Then you can use either OpenFirmware or my power logging 
script to read the current sensor inside the battery.

By using a half charged battery the charging system will still try to 
the max amount of power it can to charge the battery but the laptop will 
still run and allow you measure things.  So the result will be the same 
but 5 to 7 watts of your net power is lost in running the XO.

There are currently 2 tools of reading the current sensor.  OpenFirmware 
has a 'watch-battery' command that will print out the current battery 
info.  Running this while on battery will show that its running at 6.x 
volts and between minus .8 and 1 amps.  The exact numbers will be 
different depending on the current battery voltage. If you hook up your 
charger and start pedaling you will see the current draw decrease.  If 
you can really get 6W going then the current should switch to a positive 
number.  Which means that you are charging the battery.

The other tool is olpc-logbat (stable builds) and my new power logging 
script available here:

http://dev.laptop.org/~rsmith/olpc-pwr_log

This script show the same info that 'watch-battery' but it also has 
logging and uses the accumulated current register in the battery to 
compute how much mAh you have delivered into or drawn out of the battery 
since you stated the script. At the end of the day thats the number you 
care about.

Providing only 6 watts you will be able to just barely charge the 
battery.  But its will help you to see what you are doing.

If you can get the current draw to be around positive 1.5 amps then you 
will have reached the maximum power draw of the onboard DC/DC converter. 
   This means that when the laptop is off you will charge at the max rate.

> but the 8712 was heating up quite a bit. My thought is that the
> AC was not in phase, and so there might have been destructive
> interference. I did not pedal for more than 20 seconds for fear of
> letting smoke out somewhere along the line.

Destructive AC interference would not result in your linear regulator 
getting hot.  Out of phase AC would _decrease_ your net effective 
voltage after your diode and reduce the amount of power available since 
the dynamos [1] are working against each other.

You age gong to need one diode per dynamo to isolate each diode from 
each other.  With a diode per dynamo the phase of the generator does not 
matter. In fact in half-wave rectified (single diode) system you will 
get maximum power transfer when they are completely out of phase rather 
than in phase.

Hmm... I also just realized that if you have a half-wave rectified setup 
then you were really only able to deliver a max of half the rated output 
due to 50% of you output wave not delivering any power.  If you can get 
a full-wave setup (bridge rectifier) then it would be better.

The reason your regulator got hot is because it was working harder.  It 
was working harder _because_ you were able to keep the voltage above 12V 
and _because_ you had more power available. It was doing its job. 
Linear regulators dissipate more wasted power (ie heat) when the voltage 
differential between the input and output grows.

Keep plugging.

[1] I'll add that what you have can't be a dynamo.  Dynamos by 
definition generate a DC output.  If you have AC output its an 
alternator but the common terminology seems to be to call this a dynamo 
anyway.

-- 
Richard Smith  <richard at laptop.org>
One Laptop Per Child