Here is a report on the chlorinator, courtesy of Dhaval's D-Lab group: chlorine_final_report.pdf

Dhaval adds: yeah so the 5V is beacuse we were thinking of using bicycle dynamos which run at max 6. the theoretical cut off is minimum 1.4 but we found out in practice (using non standard temperature, concentrations etc) that it uses about 3.5 V. at 5 V using our surface area of electrodes and seperation of electrodes (which control the current) it ran at 500 mA so thats 2.5W.

our approach was entirely experimental. we played with area of electrodes, separation, voltage, current, concentration of NaCl and volume of reaction cell to get the reaction to work with a bike dynamo under 15 mins. and then we used a 300 bucks chlorine meter to see if it produces enoough chlorine

if you want to use other things, i would suggest: play with all the above variables so that you get X mg of chlorine within T mins.

to get X mg, what happens is that lets say you want to purify 10 L of water. then what you should do is have at least 5mg per liter of chlorine in the water that people drink. so initially, you make your chlorine mixture in your little reaction mixture of Y ml and then you pour that in 10L, so that means that your reaction mixture should have at least (5mg/L x 10 L / Y/1000 ml) of chlrone. these values are in the 100-500 mg/L and normal pool chlorine meter only measure up to mg/L so that means to test your reactant you have to dilute your to be tested mixture by at least 100-500 times (which happen to also increase error) if you use non-300 bucks pool side tester

be wary: even if you iniditally put in 5mg/L of chlorine, within 30 mins, your water has to contain under 3mg/L (the taste threshold of chlorine), aka about 2 mg will have to be consumed in bacterial killing.