Page MenuHomePhabricator

Updated 856 Days AgoPublic

The power utility in Haiti is extremely unreliable. We are building a prototype energy storage device to provide energy for low-power equipment such as Wi-Fi APs.

If successful, we can build many of these devices in country to provide reliable power to blanket the UCCC campus in Wi-Fi without depending on the local grid.

An enterprise Wi-Fi system is deployed on the UCCC campus fed by local fiber. There are three switches with an estimated power draw of 22W each. If we can power this network equipment for 7 hours a day, we'd be in good shape and provide more value than the local utility.


Power outages occur nearly daily and may last up to 20+ hours. The quality of grid power is very poor and can damage sensitive equipment. People frequently hack their own unmetered power connections to existing wires which can result in a forced blackout.

Hooking up solar to interface with the grid is complex and dangerous. Therefore, we're looking at options which keep the system 100% off grid.


@reid was inspired to create this project by the $200 solar self sufficiency blog post on Hackernoon in July 2017. The post details a solar setup costing ~$230 which can provide about 350Wh/day of solar energy.

Since this system is based around a cheap lead-acid battery, we could use the cheap inverters and combine them with locally available solar panels and car batteries. Twelve volt car batteries would be readily available, provide roughly 45AH of energy and could be wired in parallel to provide more energy storage.

On January 16, 2018, I contacted UCI about this idea and they confirmed it's needed, beneficial, and batteries are indeed available. Unfortunately, solar panels are no longer available for free, but some panels may be locally available for purchase.

Prototype system

We will construct a prototype system consisting of the Hackernoon parts from Amazon.

We'll assemble them with the Hackernoon post instructions and measure the endurance by using similar network equipment.

Questions for us

  • T9: Scheduled power on and off Can we build a device which kills and restores DC power on a schedule? The controller cannot do this by itself.
  • What about the prototype is painful to setup? We should make the system as easy as possible to install.
  • T10: Document setup and installation Think about ways which don't involve English writing. Same for repair!
  • How does the controller work? Is the quality OK? Would we get better quality making one ourselves?
  • Are there cheaper or better parts available outside of Amazon?
  • Optimal panel placement? Use Google Earth Pro to investigate the location of the building and it's relation to the sun. These's a nice guide for determining azimuth which is intended for setting up Wi-Fi antennas.
  • How will we secure the panels in country? We may need to ask if flat roof access is available. It may be better to mount on a pole outside on near the roof.
  • How will we secure the batteries? Perhaps create instructions to build wooden storage boxes which can be locked. These can be assembled before our arrival.

Upcoming trips (deadline)

The best available trip to install the system is coming up:

Upcoming trip to visit Haiti.

Mar 23 2018, 5:00 AM

Production system

A real system would use car batteries sourced in Haiti. It may also use solar panels in Haiti.

We prefer to use car batteries already in Haiti. This avoids shipping heavy batteries and makes the system easy to maintain since replacements are always available. A typical car battery can hold about 45AH of energy. We'd prefer AGM (absorbent glass mat) or gel batteries but can use any available battery.

Once we better understand the current draw of the equipment with the prototype, we can get a good idea of what panels and batteries we'd need in a real installation. Our prototype battery can hold 396Wh (12V * 33AH), but a regular car battery can hold 540Wh (12V * 45AH). We'd expect a theoretical production of 4.7 sun-hours/day (based on Miami, FL sunlight which is the most southern data I can find) to yield (4.7 * 100W panel) = 470Wh/day. Actual production we'd place around 350Wh/day, much less than theoretical.

This may be enough for something small, but we might consider using 200W of panels (theoretical 940Wh/day) with 2 car batteries wired in parallel for 90AH (1080Wh) of storage. This would easily power a light equipment (25W) for an entire day.

Questions for our partners

@reid asked our partners the following questions on January 19:

  • How much would automotive AGM (absorbent glass mat), gel, or regular batteries cost you? Any example of cost would be helpful.
  • Do you have 12V solar panels available for purchase? 100W and up. I can very quickly and easily get 100W panels for $120 here in California, but costs will be cheaper if I search around or order in bulk.
  • Perhaps you folks have a good idea of your typical energy production in Haiti from your home solar setup? My estimate of 4.7 sun-hours/day may not be very accurate.
Last Author
Last Edited
Jan 19 2018, 9:50 PM

Event Timeline

reid added a subscriber: reid.
reid created this document.
reid edited the content of this document. (Show Details)
reid edited the content of this document. (Show Details)Jan 19 2018, 9:06 PM
reid edited the content of this document. (Show Details)Jan 19 2018, 9:08 PM
reid edited the content of this document. (Show Details)Jan 19 2018, 9:11 PM
reid edited the content of this document. (Show Details)Jan 19 2018, 9:33 PM
reid edited the content of this document. (Show Details)Jan 19 2018, 9:37 PM
reid edited the content of this document. (Show Details)Jan 19 2018, 9:50 PM