[Power] $30 resin panel in my hands from eBay

[Richard Smith]

On Thu, Jun 20, 2013 at 2:16 AM, Mike Lee <curiouslee at gmail.com> wrote:

> Hi,
>
> In case your last 2 days are a quieter ramp down rather than a frenzy...

:)  I've cc:ed power@ so that this is archived and can be referenced later.

> I ordered this $29.99 12 volt, 10 watt resin panel from eBay in the hopes
> that it might be a serviceable and affordable replacement for the original
> 10 watt GP unregulated solar panel for G1G1 XO-1s.
>
> http://www.ebay.com/itm/10W-Solar-panel-10-Watt-12-V-Garden-Fountain-pond-Battery-Charger-Diode-cable-/251190599766?ssPageName=ADME:L:OU:US:3160

> It arrived today (Wednesday) from Xiamen Fujian, China by China Post in only
> 7 business days and one weekend.

Sweet find. Finally!  At $3/W its absolutely a replacement.  It won't
be quite as good though as that looks like its a poly-Si panel.  So
its going to have a lot more derate when it heats up to NOC.  Ask the
supplier if they can provide you with a VI curve.   Also ask if they
have a 15W version.

> I'm hoping to install the supplied diode on the positive lead and solder on

Not necessary for the XO.  The XO input already has diodes in the
path.  Additional ones will only result in more loss.

> an XO power cord to do a quick 2-point measure test today or Friday with an
> XO-1 running 12.1.0. Let me know if there might be any gotchas I should be
> aware of. I will run Power Log Collector.

Use 2 XOs one with the old GP panel and 1 with this new panel so you
can do a relative compare.  Do a crossover test as well where half-way
through you swap the panels so that each XO gets run from each panel.

You are using the olpc-panelpwr-log test right?  Not just
olpc-pwr-log?  panelpwr is what you want for panel testing as it puts
the XO into the lowest possible power mode that you can still log data
from.

Oh and make sure you discharge the batteries in the XOs down to the
same level. My suggestion would be that you run each XO on battery
until powerd shuts the system down.  Then charge each XO from the wall
for 10 minutes.  Just so they have a bit of juice to run on.  Then put
them out in the sun for 30 minutes, swap panels and then let them run
until they charge up or your sun dies.   A voltmeter on the panel
output while connected would be nice too especially if you have one
that can log readings.

Ideally you would run multiple runs of each XO+panel combination with
a battery discharge cycle in between each run so the starting point
was the same for each test.  But thats a much longer test.  You want
to see 2 things.  How does it perform in the bulk charge zone when the
XO is trying to draw more power than the panel can provide (about 1 to
1.5 hours) and then what happens as the battery charges and the
required power draw drops below the output of the panel.  This is
where the XO-1 gets into problems with unregulated panels as once you
cross into that zone the voltage on the input to the XO will begin to
rise and if it crosses the 18V threshold then the XO-1 can go into
overvoltage mode.  I say 18V because thats the worst case but typical
is about 20V.

This also can happen when the panel is put in the sun unconnected to
the XO.  In that case the load on the panel is zero and the output
voltage will be at it maximum value.  If thats > 18V (20V typical)
then when you connect up to the XO it might think its overvoltage and
shut off the input.  This is why the GP panels for XO-1 have a 18V
limiter in them.

> With a successful test, my hope is that for the ongoing contributor's
> program granting XO-1s and the steady trickle of donated units, this could
> be a solution we can point to for the tinkerer or suitcase deployment
> voluntourist.

I look forward to the test results and hopefully you have made a great
find that small deployments can use to get solar panels in qty < 100.

--
Richard A. Smith
One Laptop per Child