I would like to make something to measure amount of water in flower pots. My idea was a MCU waking up every few hours checking the water level then setting some indicator and falling asleep again. I don't know what I could use for the indicator. It could be something like 7 segmet display or just a few dots. But most of such devices are made of LEDs which consumes too much energy for battery driven operation.
I could use LCD without backlight. They are everywhere and must be consuming so little energy but in our local shop they have nothing like a 7 segment or 4 dots LCD display. Just large displays with many pixels. Besides I think the number from the display will fade if the MCU goes to sleep.
Ideal solution would be some e-Ink with a few LARGE pixels but I am not aware of seeing anything like this.
Do you have any suggestion what I could use?
Surely you can just flash a single LED with a very short duty cycle. This will not use too much energy.
When you are alerted, you can wake up a "whole display". This goes back to sleep shortly afterwards.
Be realistic. A Fire alarm has to be able to alert you and have sufficient backup battery to run for a sensible amount of time. Your flowers are not much different.
When you have responded to the emergency, you will need to replace the batteries. And replace the fire extinguishers.
That is not what I want to do. I have some "self watering" (is it English?) pots. There is about 10l of water under the soil connected with it by wicks. It's hidden it the pot so you cannot see the water directly. It was supplied with a "measuring bar" - one end of the stick should be floating on the water while the other should be over the pot's surface. In reality the bar is stuck most of the time measuring nothing. Instead of getting something working better I would like make an electronic measuring (a bit overkill, I know). So if I needed change the batteries after each time the plants need watering it would be a bit expensive.
David's idea of using a very low duty cycle to flash the LED should greatly reduce the power used by the indicator. It sounds like it would be a simple and inexpensive method.
Have you considered electro-mechanical indicators that use power only to change state? Search for "electro-mechanical vane display". There are some on eBay.
Avionics equipment used to use little "ball indicators" that held their state when unpowered but a quick search didn't find any suppliers.
You could probably invent something with a solenoid and a spring loaded flag.
My general idea was to wake up briefly and sleep in between.
Monitoring your water level only needs to be done every 30 minutes. The power use is trivial.
When it wakes up an LCD display, it goes back to sleep as soon as you have read the display. Tiny amount of power.
Emergencies might need a flashing LED or piezo. The power use is modest but significant.
Smajdalf:
That is not what I want to do... if I needed change the batteries after each time the plants need watering it would be a bit expensive.
I suspect you have not understood what David suggested.
Due to persistence of vision you can flash an LED on for less than one millisecond (1ms) and still clearly see it. If your LED consumes 10mA then the average current is only 10 uA, which will be comparable to running an AVR processor in low power mode. This is less current that an LCD display driver will take.
If you use AA cells, they have a capacity of around 2800mAh, ie in theory, for example, they will provide 2.8A for 1 hour, or 1mA for 2,800 hours (~3 months) or 10uA for 280,000 hours (32 years).
Also considering the operating scenario it is likely you would want to check the plant water level every few days because otherwise the plants will die. Thus the LED will not be flashing for long!
You might also like to consider tapping power from the pump motor supply.
Thanks for your suggestions.
I din't try it but I guess a LED with 0.1% duty cycle will be hardly visible indoors with no chance being noticable in direct sunlight. The electro-mechanical indicators look better I wonder if I will be able to get any.
As Bodmer said, the LED flash would be visible indoors.
Direct sunlight would have no chance. A mechanical indicator will require a lot more energy. It will need to withstand bad weather.
You could easily test the visibility yourself. I think this method will work. It depends on the LED but the flash should be obvservable indoors.
You could construct your own solenoid triggered device.
Personally, I'd buy a solenoid but you can find sites on-line telling you how to build one. Get a solenoid that retracts when energized.
Use a short piece of dowel rod as a flag. Attach a hinged rod to the bottom of a horizontal surface. Make sure the rod swings down freely.
The solenoid is mounted near the free end of the rod such that when the solenoid is energized the plunger retracts and the free end or the rod drops. After refilling the water tank, reset the rod manually.
It would be slightly more complicated to use a spring to replace gravity but you wouldn't be as limited in mounting options.
And, if you want a more elegant solution, find a solenoid activated valve similar to those used in lawn irrigation systems. Use this valve to directly refill your water tank.
bodmer:
A float activated semaphore would be so much simpler
A float activated semaphore is what I already have. It somehow get stuck most the time wich makes it useless. Buying better/finding why it get stuck and fixing it could be cleaner solution. On the other hand inexpensive nearly no power indicator looks like very useful thing for me.
It is offtopic here but another solution would be to send the info wirelessly to some (wall-powered) info center. Is there some way of small non-expepensive (preferably two way) wireless communication? I just need to send a few bytes and don't care much about reliability because I can send the data repeatedly. Wireless would be even better than the indicator because there would be no "ugly electronic" visible in the pot...
I have experimented with various low cost wireless approaches for remote telemetry and got systems working reliably:
Bluetooth - not particularly low power, low range, point to point, needs a hub to display info
433MHz ISM band - low power, low range, needs a hub, some only have AC coupled data input/output
nRF24L01+ - very low power, can be in a network, high data rate, low range, needs a hub
ESP8266 - WiFi, high power but in bursts, can be put into very low power state, very good range
SIM900/800 - high power, uses GPRS cell phone network, small monthly costs involved, can send emails or SMS messages, can be deployed remotely anywhere there is cellphone coverage
My favorite is the ESP8266 as it can use my existing Wifi network and post information on IOT sites like ThingSpeak. I have also configured one setup to send a Tweet on a daily basis to say it is OK and another when the system needs attention. The ESP8266 is also supported by the Arduino IDE via an add on package making it a familiar development system. The main problem is that the ESP8266 must be powered for approx 20s to get a Wifi link and send data and it takes very high current bursts when it transmits so the power supply must support that (1000uF of decoupling works). Despite these limitations it is possible to design around this, for example I have a totally solar powered setup that uses little 4V solar panels charging 2x1F supercapacitors and this reliably sends data during the day time to ThingSpeak whenever the supercapacitors reach 3.6V.
So add you watering system to the Internet of Things and check on it with you mobile phone wherever you are!
There are quite a few gotchas with any setup, for example I had a 433MHz setup running but no one within 50yds of my house could use a keyfob remote to get into their car as the transmissions, though legal, acted as a jammer!
