Solar Power for Silar’s Orphanage: A Case Study


Sora Edwards-Thro, a college freshman, coordinated, funded, purchased, and installed a solar system in Port au Prince Haiti during her 2014 Christmas break. I was involved as a consultant, along with many other helpers, that she engaged, often via phone, or skype. Nick Doiron gave her a big assist by loading XSCE schoolserver software on a new 8 watt intel NUC (with an i3 processor), and carrying it in his suitcase to Silar’s location in Haiti. Sean Collins also went to Silar’s, in December, and collected the information about connected electrical devices.

Once installed, the solar panels, controller, intel NUC, and Netgear 3g modem became a system which monitors, and reports essential information about the new power system both locally, and to the internet cloud, where the funders can check up on their investment. (For additional informationūüôā




  • $660 — 4x Trojan T-105 6 volt batteries
  • $730 — 2x 300 Watt solar panels
  • $176 —¬†30 amp charge controller MMPT 3250RN (chinese Epsolar)
  • $145 — 10 GB per month for 12 months 3G internet access
  • $470 — intel NUC, 1.5TB hard disk, 8GB ram, USB ethernet dongle
  • $20 — TP-Link WR841n wifi
  • $50 — Netgear MBR1210 3G modem
  • $200 — miscellaneous tools, materials

Totals: $2450. It should be mentioned that plane tickets, and additional ground transportation, for multiple people, at least doubled this fairly modest expense.

Planning vs actual performance

When we started planning the solar system, we tried to ask ourselves about the expected loads. We came up with the following:

  • 600WHr –Charge 20 XO laptops once a day
  • 300WHr –School server 24hrs per day
  • 600WHr –Provide 10 rooms with lighting 3 hours per day at 20 watts per light
  • 300WHr –Provide power for a Church service 3-4 hours ¬†(lighting and sound system).
  • 250WHr –Power for DC refrigerator

We decided pretty quickly that we could not support a refrigerator. After the installation, the graphs showed that some of our estimated requirements were high.


The first day on the above graph is Sunday, and the -100 watt draw was during the church service.  We felt good that we could supply energy so that the gasoline generator would no longer provide noise, pollution, and a drain on the pocketbook. During the middle few days, the loads were not connected to the controller, and were not registered on the wattsinout bottom graph (but the dip in voltage after the fourth day indicates the drawing of power). During the last two days, you can see the impact of the nightly lighting.

Design Decisions

There was some lively debate about how to keep the low voltage lights from being subjected to 110vac. Finally we chose to abandon the wires inside the walls of the orphanage, and string up new low voltage wiring to each of the rooms.  And then there was also lots of discussion about whether to design a 12V or 24 system. When you are in the business of distributing power you want the highest voltage possible, and the largest wire size.

We chose #10 AWG stranded wire, purchased in the US, and after much soul searching, to run our whole system at 24 volts. ¬†The compact flourescent bulbs were available in either 12 or 24 volts. We wanted the DC system to be able to function across the whole of Silar’s two properties ( the longest distance was about 150 Meters).

One of the biggest disappointments so far was that the solar panel output seems to be about 50% of their advertised output (that has been true in every case for me). Of course, there may be some opportunity to improve  the situation, because the panels are currently lying horizontal.


This was not the first set of batteries that Unleashkids installed at Silar’s. In March of 2013, a 15Amp smart charger, and two Trojan T-105 batteries was installed at Silar’s¬†to power the first school server installed on XO-1.75 hardware.

During a followup visit in late 2013, we checked the specific gravity of the batteries, and monitored the charging current  when grid power was on. The specific gravity was extremely low, indicating a completely discharged battery, and the inability to accept charging current indicated the same thing. At that point we developed the hypothesis that the grid power was not on long enough during the day to match the loads that were being drawn.

So we installed an AC Powerline recorder. The data showed that there were periods as long as 3 days in a row when the power never came on. The average grid power was 7 hours per day, usually in the early morning hours. There was really no way to determine whether the 110vac inverter, that was part of the system, had been connected to the AC power distribution system of the orphanage, and used to power room lighting, laptops, tablets, cell phones etc (my suspicion).

2 Responses to Solar Power for Silar’s Orphanage: A Case Study

  1. Pingback: A Solar System for Silar’s Orphanage, Port Au Prince Haiti | OLPC School Server

  2. Kim says:

    An admirable project for anyone, let alone a student…

    Estimating power output is messy territory, and panels rarely achieve their full “rated” level outside laboratory conditions. Variables that affect production include their inclination with respect to the sun, temperature (they start to lose efficiency at 80F, and losses become appreciable in full sun above 90F), shading from trees/clouds/rain (of course), and ambient humidity (moisture in the air scatters and absorbs solar energy). Direct current declines with distance, even with good thick wires, and batteries aren’t entirely efficient either, further taxing the output of such a system.

    Still, congratulations!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.