4 ms·
> If the power fails they'll come on at the wrong time. I was assuming the timer was battery-powered and not prone to power failure, while the pumps were. This
by chime 16y ago
> If the power fails they'll come on at the wrong time.
I was assuming the timer was battery-powered and not prone to power failure, while the pumps were. This could cause the timer to turn on the pumps, not detect that the pumps were off for 40 minutes during the cycle because of power failures, and then shut off the pumps after a predetermined duration. So the timer would start/stop a one hour cycle but during those times, the pumps would only work for 20 minutes. That could cause crops to fail because the pumps didn't irrigate sufficient water.
To fix this, you'd need a timer that would also detect the duration the pumps were working and adjust accordingly, increasing the cost of the system. Plus how does the timer deal with the next cycle schedule to start at 4am-5am when it has been pumping 30-40mins extra before the cycle before it took longer? I was just trying to say that setting up a workable schedule on such a device could be a lot more complex/expensive than a human texting "pump 2 on", "pump 2 off".
- jacquesm 16y agoBut a human texting 'pump 2 on' suffers from the exact same problem, compounded by the delays of receiving the SMS (which in my case sometimes only appear hours or days after sending). So it was meant as being the exact equivalent of what they have now. (I didn't see any evidence of a feedback mechanism in the article but I may have overlooked that).