Mesh Potato - Solar + Battery

[Steve Song]

Hi all,

A question for the solar experts on the list. If we wanted to set up
a remote solar-powered MP with say a 10W solar panel, what sort of
battery would you recommend in combination for long term use?

Thanks... Steve

[Corinna Elektra Aichele]

Hi Steve -

> A question for the solar experts on the list. If we wanted to set up
> a remote solar-powered MP with say a 10W solar panel, what sort of
> battery would you recommend in combination for long term use?

voltage regulated lead-acid batteries, like Panasonic LC-R series. You'll find
these also as power backup in alarm systems.

As a side note: In Germany insurance companies demand that these backup
batteries are replaced every two years. So the companies that perform this
service are a good source of those batteries in mint condition.

You will also need a charge regulator and a low voltage disconnect circuit.

Cheers,
Elektra

[Don Onwunumah]

Steve,

Here is a link to an excellent resource for planning a solar powered
deployment, if a little pessimistic.

http://www.ubnt.com/wiki/Planning_a_wisp_solar_powered_tower

From the guide, Assuming you are powering a 4w device, this equates to
96W per day,

"You will need a battery bank that will run your equipment for a set
amount of days without sunlight, and still not be discharged more than
50%. To work this out, we take the daily power usage, multiply it by
the number of days , and then multiply again by 2 so we factor in the
50% discharge limit."

You can reduce this somewhat to a day with a battery 50% discharged or
less, so if an NS2 requires 4w, it's requirement for the day is 96...
So it's 96 * 2 / 12.. Which equates to a 16AH battery, or an 8Ah if
you want the battery depleted after a day.

A 10w solar panel can deliver ~0.8 to a 12v battery, if the battery is
rated at 16Ah, you will need 16 hrs to charge the battery fully, but
thats if you are lucky, for a 8Ah battery, that's 8hrs, but again if
you are extremely lucky.

Please don't take this as gospel, it is just an assumption not knowing
your exact scenario. As a rule of thumb it is always best to start
with your equipment requirements and go from there as the guide
explains, and always assume that Solar panels do not always meet
specifications.

What are the final power requirements of the MP if I may ask?

Don

On Wednesday, August 11, 2010, Steve Song

> --
> You received this message because you are subscribed to the Google Groups "village-telco-dev" group.
> To post to this group, send email to village-...@googlegroups.com.
> To unsubscribe from this group, send email to village-telco-...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/village-telco-dev?hl=en.
>
>

[David Carman]

Hi Steve

The battery is just a reservoir, so its size is determined by the time
during which the solar panel is not able to power the MP. If the solar
panel cannot recharge the battery fully on *most* days, it's not going
to stay up.

From the testing for AfrikaBurns, I figure that a panel can provide a
continuous feed of 15-20% of its rated power output in a Cape Town
climate. Obviously the panel wattage is halved because of nighttime,
but it also has the sun striking it at an angle most of the time.
Solar panels are surprisingly effective in cloudy conditions.

If an MP consumes an average of 2.5W, a 10W panel is only going to
keep the MP up continuously if sunlight is optimal or the MP is turned
off at night. I suggest a 20W panel for cloudy or temperate locations
to ensure that the battery is fully charged by the afternoon of each
day.

After that, the battery reserve calculation is trivial - it only needs
to keep the MP up overnight with a little more reserve in case of an
especially dark day. If a lead-acid battery discharges more than 50%
of its capacity, it ages considerably more.

So 2.5W at 12V is 200mA. We want to keep the MP up for 20 hours of
zero power input - we need 200mA x 20 hours = 4AH. A battery that can
discharge 4AH without being damaged is 8AH.

So I recommend a minimum 20W panel with minimum 8AH battery.

David

[Vickram Crishna]

On Thu, Aug 12, 2010 at 10:40 AM, David Carman <tid...@gmail.com> wrote:

Hi Steve

On Wed, Aug 11, 2010 at 11:43 PM, Steve Song

<steve...@shuttleworthfoundation.org> wrote:

> A question for the solar experts on the list.  If we wanted to set up
> a remote solar-powered MP with say a 10W solar panel, what sort of
> battery would you recommend in combination for long term use?

So I recommend a minimum 20W panel with minimum 8AH battery.

Further, I recommend using, literally, a 'battery' ie, multiple units, not a single one. It depends a lot on where you are, but if your local market has access to the inexpensive Panasonic knock-offs from a certain Eastern economic superpower in the Asian region, it is easily possible (almost a near certainty) that you will get upwards of 2 years usability. The downside is your local traders may be buying super cheap stuff, which has very poor reliability, but the upper grade stuff is actually just a few cents more expensive, if available. The smaller units are a whole lot cheaper than the big ones (in India, one 6V battery can retail for IRS 80-90, whereas the 12V battery, which looks a lot like (but isn't) a motorcycle battery, costs IRS 450. Using two 6V batteries in series costs less than half as much (but check the Ah rating, too). 

--
Vickram
http://communicall.wordpress.com

[Peter Jansen]

Hi Steve,

Solar calculations are quite simple, and the general rules have been around for awhile now.

You can get the 'available solar energy' from most Met Offices, but here is a site that gives you the values you need,

http://www.energymatters.com.au/climate-data/?q=cape+town&find=Search

Also on the site the 'stand alone power system calculator' will do the calculation you need.

Basically a 10W panel will output 10 W with an solar irradiation of 1 kWh/m^2.

The battery size required generally the capacity used is that which will run the system for 5 (10 if you don't want it to fail) days of no sun, any thing less than 5 will tend to discharge the battery in winter some time.

So a 2.5 W load on all day uses 2.5 * 24 = 60 Wh/day, Min solar for cape town in 2.44 kWh/m^2/day, so a 10 W panel gives 10 * 2.44 = 24.4 Wh/day, so in winter you don't have enough solar power, in winter the battery will go flat sometime, but might be enough on 'average'.

Now 60 Wh/day is 60 / 12 = 5 Ah/day * 5 days = 25 Ah * 50 % (min discharge) = 50 Ah * 90 % (some losses) = 55 Ahr.

The solar panel needs to re-charge the battery within 2 days of being flat (after 5 days of no sun).

Now the panel angle, basically you want it at right angles to the suns rays during winter to catch the most of the 2.44 kWh/m^2. Most of the available energy occurs when the sun is overhead (because of the atmosphere etc). The power drops off with the cosine of the angle, so at 30 Deg off angle for example the output is 86 % of the right angle value. So when the sun is +/- 30 Deg of right angles to the panel during the day you get more that 86% of the available power. In the tropics you want some angle off horizontal to let the water run off and keep it a bit clean.

Hope that Helps,

Peter Jansen

I don't have any association with energy matters, but the calculator is very useful.

Here is the example calculation from the energy matters site,

Solar Irradiation Information

 

Your location:	Cape Town, South Africa
Using data from Capetown/Df Malan weather station
Min. solar irradiation per day	2.44 kWh / m2
Average solar irradiation per day	5.20 kWh / m2
Max. solar irradiation per day	7.89 kWh / m2

 

Energy

 

Energy usage (per day)	0.06 kWh
Power from other sources (wind, hydro, etc)	0 Ah a day at 12 V

 

Solar System

 

System voltage	12 V
Current required (factoring loss)	5 Ah at 12 V
Solar Panels	2 × Suntech Solar Panel 20Watt 12Volt Multi-crystalline)
Solar Charge Controller	1 × Morningstar Sunguard 12Volt 4.5Amp Regulator
Days of battery backup	5 days
Battery depth of discharge	50 %
Battery bank required (factoring loss)	54 Ah 1 × Fullriver Sealed Lead Acid AGM Battery 12Volt 60Ah HGL

 

[Steve Song]

Thanks everyone for the comprehensive replies!

Cheers... Steve

--
You received this message because you are subscribed to the Google Groups "village-telco-dev" group.
To post to this group, send email to village-...@googlegroups.com.
To unsubscribe from this group, send email to village-telco-...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/village-telco-dev?hl=en.

--
Steve Song
Telecommunications Fellow, Shuttleworth Foundation

email:   steve...@shuttleworthfoundation.org
work:    +27 21 970 1220
mobile:  +27 83 482 2088
skype:   steve_l_song
blog:    http://manypossibilities.net
twitter: stevesong

[David Rowe]

A little off topic but I have been playing with Lithium battery powered
potatoes lately, see attached photo. Paul GS gave me this idea - they
have about 4 times the energy for a given size and weight as lead acid.
Rather that solar power, I just needed a way to power a potato for a few
hours, e.g. for testing or portable demos.

I am using model aircraft batteries I bought from hobbyking.com in Hong
Kong. The battery I chose is 1AH at 14.5V or 14Wh, cost $15, and
weights 115g. This tiny battery will run the MP for perhaps 6 hours,
plenty for my application. The charger is at the top, which runs off
12V. I have velcro-ed one battery to my MP 1.3 prototype on the left,
so with a small telephone my MP is now much more portable.

These batteries pack quite a punch, when I accidentally touched the
flying leads together there was a scary loud "crack", an impressive
spark and lots of smoke. These tiny batteries can supply > 50A into
model aircraft motors.

The MP has a microcontroller with A/D and D/A - it would be possible to
modify the MP to have a built in solar charge controller at very little
extra cost. Battery charge state could be read out via Wifi, automatic
watchdog could be added. Nice project if some one is interested.

Cheers,

David

[Don Onwunumah]

David,

Excellent note on lithium - I have considered them versus lead acid, it's great to hear you have actually used them.

w.r.t your comment:

The MP has a microcontroller with A/D and D/A - it would be possible to
modify the MP to have a built in solar charge controller at very little
extra cost.  Battery charge state could be read out via Wifi, automatic
watchdog could be added. Nice project if some one is interested.

I am very interested in this - as I am planning to deploy the MP in locations where remote monitoring of power supply would be beneficial. Not having an actual MP yet, I have tried looking for the specifications of the AR2317 - but can only get a basic block diagram from Atheros, and this seems to be linked to the RF switch. It would also be great to get the H/W specs of the MP's somewhere in the site (unless I missed it).

Can you possibly expand on where one starts with designing a solar controller ( I assume you interface into the DAC/DAC bus  - with some control interface to the processor). I can get up to speed reasonably quickly but only of if I have enough data sheets!

Thanks in advance,

Don

[David Rowe]

Hi Don,

Thanks for your interest in this project.

The Mesh Potato circuit diagram in on the wiki. We have an Ar2317
processor for Wifi and an Atmel microcontroller (uC) that is currently
used as some glue logic for the FXS port. It's very similar to an
Arduino type microcontroller. It has plenty of analog and digital I/O
that is not being used.

There is no decent data available on the AR2317 - the good people at
Atheros require us to pay USD$100,000 to get the data, even for a
project with huge good will like the Village Telco. We have some how
managed to develop the Mesh Potato despite this.

Yes, some sort of control bus is required between the AR2317 and the uC.
Something could be knocked up a spare GPIO and the CPLD. Perhaps make
the uC look like another SPI port. One side of the AR2317 serial port
is already connected to the uC.

Elektra might have some ideas on appropriate algorithms for solar charge
controllers.

Cheers,

David

> To unsubscribe from this group, send email to village-telco-dev
> +unsub...@googlegroups.com.

[Roey Rosenblith]

Hello David and Don, 

I've been lurking on this list for a while, but here in Uganda, my company Village Energy (www.villagenergyuganda.com) which works on solar solutions is trying to put together something very similar for village pay phones. We would be very interested in collaborating on such a project (as we are already currently working on it). 

From a cost perspective it would seem like Lead Acid would still be the way to go. First they are very readily available throughout Africa and people in rural village already have distribution systems setup for them. One would think that mobility would not be a big deal hear since these phones and mesh potatoes would be fairly stationary once their in place. I've cced here our electrical engineer Eric Thorne, who is currently working on this project, we would be happy to have a discussion with you guys though. 

Roey

Roey Rosenblith
Director, Village Energy (U) Ltd.
Mob: +256-75-710-9902
Landline: +256-75-436-3749
www.villageenergyuganda.com
Kondogoro Zone
Kulambiro Kisaasi
Nakawa Division
Kampala, Uganda
__________________________________
The contents of this e-mail message are privileged and/or confidential and intended strictly for use by . If you are not the intended recipient, any review, dissemination, copying, distribution or other use of the contents of this message or any attachment by you is strictly prohibited. If you receive this communication in error, please notify us immediately by return e-mail or by telephone (+256-075-710-9902), and please delete this message and all attachments from your system. Thank you.

[Eric]

Hi everyone,

This sounds like a really exciting project! As Roey said, we're
leaning towards using flooded lead-acid batteries for our application
because the capacity and voltage requirements make Li-ion seem
prohibitively expensive ... the system is completely solar-powered,
and the battery should be able to run a near-700 mAh load at 12 V for
24 hours (the 12 V FLAC battery also has a convenient voltage).
Putting Li-ion cells in series and parallel to get to the electrical
requirements right would also make charging more difficult. Thoughts?

As for charging cells with uCs, this is something that I'm probably
going to be working on for the next few months at least. I've got a
little experience with Atmel uCs and managed to prototype a very
simple charge controller that from time to time sampled the battery
voltage and controlled switches to isolate the battery from charging
and discharging sources when over- and under-voltages were
respectively reached. In between samples, it would go to sleep by
using an external 32.768 kHz crystal. I wanted to try a watchdog timer
but didn't get to learn how to properly program the fuses before I had
to shift gears. Not sure if this is useful for anyone but I'd love to
share what I've got.

As for charging algorithms, I think most of it has to do with
switching between different charge regimes at the right time (fast
charging, trickle charging) to keep the battery within its safe zone.

Best,

Eric

> > >         village-telco-...@googlegroups.com<village-telco-dev%2Bunsubscr i...@googlegroups.com>
> > .

> > >         For more options, visit this group at
> > >        http://groups.google.com/group/village-telco-dev?hl=en.
>
> > > --
> > > You received this message because you are subscribed to the Google
> > > Groups "village-telco-dev" group.
> > > To post to this group, send email to
> > > village-...@googlegroups.com.
> > > To unsubscribe from this group, send email to village-telco-dev
> > > +unsub...@googlegroups.com.
> > > For more options, visit this group at
> > >http://groups.google.com/group/village-telco-dev?hl=en.
>
> > --
> > You received this message because you are subscribed to the Google Groups
> > "village-telco-dev" group.
> > To post to this group, send email to village-...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > village-telco-...@googlegroups.com<village-telco-dev%2Bunsubscr i...@googlegroups.com>
> > .

[elektra]

Hi all -

having run a solar-powered home for 18 years, I'd like to suggest to use
voltage regulated lead-acid batteries (VRLA). 7 Ampere hours is
sufficient to power a Mesh-Potato for approximately 30 hours. However
discharge cycles going below 80% should generally be avoided in order to
have the batteries last long, so I'd recommend about twice the capacity.
So far we haven't pulled all registers to save power on the MP yet. The
FXS module consumes considerable power once the driver is loaded. Maybe
we can save some power here.

Floating lead-acid batteries from motorcycles or cars have properties
that are not really well suited for such an application. They can only
take a very limited amount of charge-discharging cycles.

A simple shunt charge regulator will do. We can do pulse-width
modulation of the incoming current by means of a single MOS-FET
transistor. I have been building charge-regulators like this. They were
shunting the solar panel when the battery voltage was exceeding 13.8
Volt and started to charge again at 13.4 Volt. Depending on the state of
charge the length of the pulses varies. Once the battery is charged the
battery receives a single short charge pulse every now and then. If the
battery is under load the pulse length increases.

We need to have a low-voltage disconnect, too. This can again be done by
a single transistor, or we can put the Atheros SoC and the FXS chip in
suspend mode, if this is supported.

Cheers,
Elektra

[David Rowe]

Hello,

Some more thoughts:

1/ As Elektra suggests deep cycle batteries are designed for this sort
of job However I imagine they are specialised kit in much of Africa. I
wonder if a regular starter battery would work OK if we are only
discharging to a few % of it's AH capacity? For example my (petrol) car
starter battery is 60AH, a MP would use 2-3AH per day at 12V.

2/ A MP power reduction project would be a nice project for some one.
As well as low power operation of the FXS port there might be some
register that lets us reduce the AR2317 clock speed. We rarely use 100%
CPU, especially when not making calls.

3/ Elektra with the shunt regulator the mosfet switch connects across
the panel? So when the switch closes if shorts the panel or the
duration of the pulse? Does it get hot? A circuit might help if you
have one. Sounds like a cool design.

4/ A sleep timer was something the Air Jaldi people were very interested
in. The MP could be told to go to sleep from midnight to 6am for
example. So the uC would open the power switch (same one used for low
voltage cut out), then close it again hours later to bring the rest of
the MP up again. This can dramatically cut the night-time power
consumption and hence battery capacity needed.

5/ I messed around with a related device (measured battery voltage,
watchdog, sleep timer etc) a few years ago:

http://www.rowetel.com/blog/?p=68

Cheers,

David

[dbur...@jcis.net]

Deep cycle batteries should be easy to find in any city near water, since they are standard kit for just about any marine power system. For example, the backup power supply on the OpenBTS installation in Niue was built from locally purchased boat batteries. There are very few things that can be purchased locally in Niue but deep-cycle batteries were no problem.

Sent from my Verizon Wireless BlackBerry

--
You received this message because you are subscribed to the Google Groups "village-telco-dev" group.
To post to this group, send email to village-...@googlegroups.com.

To unsubscribe from this group, send email to village-telco-...@googlegroups.com.

[Jeff Wishnie]

On Aug 20, 2010, at 2:04 PM, dbur...@jcis.net wrote:

> Deep cycle batteries should be easy to find in any city near water, since they are standard kit for just about any marine power system. For example, the backup power supply on the OpenBTS installation in Niue was built from locally purchased boat batteries. There are very few things that can be purchased locally in Niue but deep-cycle batteries were no problem.
>
> Sent from my Verizon Wireless BlackBerry

In my work with Inveneo we were able to find high quality deep-cycle batteries from solar and UPS dealers in most larger African cities. The most difficult place to source them was Freetown in Sierra Leone, but even there we were able to find a few importers and suppliers.

best,

Jeff

[elektra]

Hi -

> 3/ Elektra with the shunt regulator the mosfet switch connects across
> the panel?

yes. The circuit is very effective and simple. Of course you need a
diode to prevent current from the battery floating to transistor. To
have such a diode is a good idea, anyway.

> So when the switch closes if shorts the panel or the
> duration of the pulse? Does it get hot?

At the power levels we need for the MP, no. A heat sink is not required,
given that the MOS-Fet switches fast enough.

> A circuit might help if you
> have one. Sounds like a cool design.
>

Sure, I'll draw one.

> 4/ A sleep timer was something the Air Jaldi people were very interested
> in. The MP could be told to go to sleep from midnight to 6am for
> example. So the uC would open the power switch (same one used for low
> voltage cut out), then close it again hours later to bring the rest of
> the MP up again. This can dramatically cut the night-time power
> consumption and hence battery capacity needed.
>

I thought about this idea, too. However I would personally prefer a
telephone system that allows to make calls anytime. There are number of
reasons that crossed my mind: Call someone in case of a emergency. Make
a cheap international call to someone in a timezone far away during
their daylight time.

Cheers,
Elektra

[David Carman]

Hi List

A few notes:

Lithium vs Lead-acid: Definitely lead-acid, due to expense,
availability and that methamphetamine addicts steal lithium batteries
for their modified Birch.

Deep-cycle vs standard: True deep-cycle absorbent glass mat (AGM)
batteries used in marine applications are expensive and unnecessary.
Rather use a cheap gel lead-acid battery (often touted as deep-cycle
but not truly so) and look after it with a clever regulator. The best
source in South Africa for such batteries is http://deltecpower.co.za/

A clever regulator needs to:
a. separate the battery from the power consumption, so that it can cut
power when battery charge drops to 50%.
b. Apply a charge of 14.4-15V when the battery needs charging and
apply a charge of 13.5-13.8V when the battery is full.

I have been using a Solsum 6.6f from
http://www.sustainable.co.za/steca-solum-6-6f-solar-charge-controller.html
(see attached picture). It's rated at 6A and requires a minimum input
of 15VDC, which it then regulates to one of the above voltages. The
main components are:
a. ATTINY44V-10SSU uC
b. 2 x FR2407 MOSFETs
c. 2 x 75321S MOSFETs
d. FAJ15 GFX diode

20W solar input requires max 2A and price increases exponentially with
amperage so it would be worthwhile sourcing a 2A circuit with same
function. I haven't found one yet.

The regulator has 3 LEDs that flash different patterns according to
charge state. I am not sure whether the regulator shunts the panel
when the battery is full, but that would certainly not work for when
the regulator is deployed as a UPS downstream from a 220VAC
transformer.

The regulator requires minimum 15VDC input because it cannot step the
voltage up when charging. Solar panels actually deliver 21VDC (20W at
447mA max). The Linksys PSU, as well as many other similar 12V PSUs
actually deliver 16VDC. We should ensure that the PSUs that ship with
the MP do the same, so that the same regulator/battery combination
could be used as UPS and conversion to solar involves nothing more
than connecting a panel.

Optimum lead-acid charge voltages fluctuate wildly with temperature,
so the battery should not be sited in direct sunlight. I suggest
keeping the battery and regulator inside the house close to the phone,
which would mean another wire pair running to the MP to deliver
battery charge information. Battery charge info is nice but is it
necessary? All that it would provide is notice of imminent power-down.
I would rather see a separate regulator/battery unit that can just be
spliced into the DC power supply (solar or transformer) and sit in a
cool place.

David's WiSPCaR is useful for preventing router hangs on remote
routers, but they are not necessarily power-related. I find on WRTs
and Nanos that certain processes hang while others stay up, which
would complicate an onboard monitoring system. I prefer to set a cron
job to cycle remote routers every early morning.

David C

[Eric]

As for using starter batteries for these kinds of applications, Roey
and I asked around and we found out that there have been several solar-
payphone projects that use regular starters instead of deep-cycle
batteries that are still working fine two years after installation.

Even though starters aren't designed for this kind of application, if
you make the capacity high enough and 20-30% depth of discharge, they
should get a couple thousand cycles at least ... which is good for a
year or two.

Eric

[David Rowe]

Hi Eric,

Yes that makes sense. My cheap petrol car starter battery gets
discharged a few % each time I crank the car (200A for 10 seconds) and
lasts for 3+ years.

With this thread I was just pointing out that because the MP consumption
is so low, and starter batteries so common, it's nice to know that a
starter battery will do the job if that's all that is available. I
doubt we will be in a position to mandate battery choice.

In terms of implementation on an MP, here is a suggested plan:

[ ] Implement a prototype charger with a regular Atmel microcontroller.

[ ] Reflash the MP uC and hook up some wires to the prototype charger
components. This will kill the telephony functionality but demo the
charger on MP hardware.

[ ] design a communication path between the AR2317 and the uC. Key
issue is how to make sure telephony software (a FIFO buffer) is not
interrupted when doing charger work.

[ ] implement a communication path, write driver software and test it.
We need a way to send bytes between the AR2317 and the uC. This could
be tricky. There is one spare GPIO on the AR2317. The AR2317 RS232 tx
already goes to the uC. Communications could possibly limited to when
there is not an active phone call. It is also possible to reprogram the
CPLD, e.g. to implement a second SPI device.

[ ] integrate running prototype charger and telephony on MP at the same
time.

Some references:

1) MP01 schematic, note AR2317 GPIOs and uC circuit:

http://villagetelco.svn.sourceforge.net/viewvc/villagetelco/sch

2) uC code for MP01:

http://villagetelco.svn.sourceforge.net/viewvc/villagetelco/david/avr/

Cheers,

David

[Matthew Lloyd]

From New Zealand Red Cross experience, we designed our portable VHF
repeaters (used by Pacific Island Red Cross Societies), with an
external waterproof plug so that any 12VDC source could be used should
the sun not shine or the internal battery die of old age, etc...,
etc...

> > > Elektra- Hide quoted text -
>
> - Show quoted text -

[David Carman]

On Thu, Aug 12, 2010 at 7:10 AM, David Carman <tid...@gmail.com> wrote:
> So I recommend a minimum 20W panel with minimum 8AH battery.

I have been running a new solar-powered Nano2 on a mountain site for
the past few weeks, with a 12AH lead-acid battery reserve.

The 30W solar panel was sourced locally by a third party and appears
to be a "string ribbon" type. Because the 30W spec was way over the
requirement, we sited the panel between rocks so that it would not be
immediately visible, thus reducing the chance of theft (which is what
had happened to our previous 20W amorphous silicon panel at a more
exposed site nearby). The panel gets no morning sun, but is fully
exposed and aimed toward to the late afternoon sun.

We had a rather dark week, in which the panel did not receive
unclouded sunlight for about 5 days, and the regulator cut the load to
the Nano. In cloudy conditions facing directly towards the sun, I
found the amperage on the panel to be only 12mA * 21V = 250mW, way
below the 140mA * 21V = 3000mW from the 10W monocrystalline panel I
had experimented with over the previous equinox.

So it appears that efficiency may vary according to panel type at less
than optimal sun exposure. I shall confirm this by measuring current
on a sunny day soon, to ensure that the new panel is indeed capable of
delivering 30W.

In the meantime, we have replaced the battery with a 60A car battery
to power through cloudy days. This is more expensive and more at risk
of theft, but we are relying on the camouflage for security. I'll also
set up an adjacent node to ping the Nano and notify by email if it
loses power. This security works for batteries as power is lost
immediately after theft, but not for panels.

A recent thread on a private South African WiFi list described the
problem of panel theft. The only proven solution was to fit the panels
with expensive satellite trackers. After a few thefts and subsequent
recoveries, the thieves left the panels alone. Hopefully camouflage
will turn out to be a less expensive but equally effective solution
for us.

David

[Paul Wolstenholme]

On Sep 6, 7:56 pm, David Carman <tidg...@gmail.com> wrote:
> I have been running a new solar-powered Nano2 on a mountain site for
> the past few weeks, with a 12AH lead-acid battery reserve.
>

> ...we sited the panel between rocks so that it would not be
> immediately visible, thus reducing the chance of theft ...

> The panel gets no morning sun, but is fully
> exposed and aimed toward to the late afternoon sun.
>

> ... In cloudy conditions facing directly towards the sun, I

> found the amperage on the panel to be only 12mA * 21V = 250mW, way
> below the 140mA * 21V = 3000mW from the 10W monocrystalline panel I
> had experimented with over the previous equinox.

> ...
> David

If my understanding of solar panel information posted on the internet
is correct, shading a small part of a solar panel can pretty much
negate the benefit of the whole series string. Even if the panel is
facing the sun, nestling it amongst rocks that shade parts of the
panel could result in a significant reduction in output.

Paul W.

[David Carman]

On Tue, Sep 7, 2010 at 1:44 PM, Paul Wolstenholme
<paulwolste...@gmail.com> wrote:
> If my understanding of solar panel information posted on the internet
> is correct, shading a small part of a solar panel can pretty much
> negate the benefit of the whole series string.  Even if the panel is
> facing the sun, nestling it amongst rocks that shade parts of the
> panel could result in a significant reduction in output.

Thanks Paul - I didn't know that.

However the complete or partial shading only happens in the morning -
for most of the afternoon the panel is completely illuminated.

Another reason for nestling it in rocks is to protect from bushfires.

David