ELAR advice document 2
Version 1: please send comments to David Nathan
Unless you have mains power available, or use only very low powered devices for which you can bring enough alkaline batteries, you will need an accumulator (battery/batteries) and strategies for replenishing it (/them). Lead-acid batteries, such as those used in cars, are currently the best accumulator solution as a form of "appropriate technology" for remote or poor locations. The disadvantage of lead-acid batteries is that they are heavy. However, this is currently the price we pay for a relatively high capacity, flexible, robust, locally-manageable, low-tech and low-cost power supply.
We recommend the following planning strategies for powering in the field:
We only tested one replenishment option – charging from the mains, using a 5 amp car battery charger (available in-country, UK cost GBP25). It took about 12 hours to fully charge the battery; we would recommend that a higher-powered (e.g. 15 amp) charger should be considered in situations where mains power is available only for a short time (e.g. town visits are short, or when using a generator that runs sporadically). Charging the battery by fitting it into a car would take from half an hour up to 2 hours depending on the car and how it is used. We did not investigate solar replenishment.
| time | action |
| 11:28 | booted up and macro started |
| 14:38 | last recorded entry |
| time | action/battery voltage |
| Before test | unloaded 12.98V |
| 17:19 | booted up and macro started |
| 03:09 next day | inverter cut off |
| After test | unloaded 11.5V |
Fig 1 shows that the rate of voltage drop varied according to the type of activity on the laptop. However, the voltage drop was reasonably smooth over a range of loads, and the progress of battery discharge (as measured simply by its voltage) was quite predictable. Note that the battery voltage recovers significantly, but gradually, over a period of several minutes, when the battery is unloaded. Therefore, to interpret battery voltage readings, the load status must be taken into account.
The battery was full at the beginning of the test. The test was stopped when the laptop was still running but the battery voltage dropped to less than 12.0V under load (11.96V). This kept a considerable safety margin, both for the laptop and the battery, since it is important not to discharge the battery too low (see XX).
In this test, the battery powered the laptop for 11.5 hours. Extrapolation of the test results to a final battery voltage of 11.8V would (at an average rate of -0.056 volts per hour) yield a total running time of nearly 14 hours.

Fig 1. Battery capacity test, April 2006, 85 Ah battery - Targus transformer - IBM T42P
Although a 12 volt battery is larger and heavier than 3, or even 6, laptop batteries, it is more flexible. Laptop batteries are very expensive, and usually only suit a particular model. The cost of 6 laptop batteries would exceed the price of the laptop itself, while car batteries are cheap, and can be bought in-country as required. Additionally, replenishing (recharging) laptop batteries must typically be done using the laptop itself, so flexible approaches to replenishment are not available. Therefore, laptop batteries are not likely to be a good option for the powering system's accumulator.
Test B showed that the low voltage solution yields a longer running time and is perhaps 30-40% efficient. However, it requires a special laptop transformer that is expensive and which may not be readily available. This method does not offer the flexibility of the inverter, which can be used as a multipurpose power supply since it provides normal mains power up to its rated capacity.
Obviously, the running time available will also depend on the size and type of laptop. Not did we test the laptops' running time in relation to the specific activities carried out on them. As the Dell laptop ran, its moment-to-moment power consumption fluctuated between 2 and 8 amps, largely depending on the level of hard-drive activity. Other settings, such as screen brightness, use of wireless devices etc, will also affect power consumption, so laptop users should make themselves aware of the power implications of various tasks and settings.
The fieldworker should become familiar with how her/his equipment works and how the voltages change with use. It is necessary to carry a digital multimeter to check voltages. Digital multimeters are cheap, light, and make it very easy to check the capacity of a battery, calculate average power consumption rate, and predict the amount of time left to power devices. In addition, they can be used to check all kinds of batteries and to identify a variety of faults with electrical equipment.
Applying these principles to the field powering situation means that there are going to be many alternative ways to provide power, most of them involving concessions, effort, flexibility and creativity on your part. Nevertheless, local situations will provide some opportunities. For example, car batteries and chargers are things that will exist in many places, even in remote and poor communities. People may be able to either provide batteries for you to use, or, perhaps better, they might be pleased to use a battery (or batteries) that you buy locally and ultimately give to them in return for helping you with your replenishment strategy. A possible scenario would be that you bought two batteries and arrange to have them swapped every other day to be charged, either when a generator is running, by fitting into a vehicle, or by transporting to a source of mains power and charging it there. In such a scenario, the batteries (and charger, if relevant) would be useful to people after you leave. According to our tests, if you can arrange to have a battery charged every other day, you can probably use a laptop computer as much as you need – i.e. from 3 to 6 hours a day depending on manner of use.
Author: Tom Castle and David Nathan
Version: First draft
Date: 5 May 2006
Acknowledgements: Thanks to the following people for helpful input: Rob Kennedy (Language Centre, SOAS), Bernard Howard (Linguistics, SOAS).