Monday, June 12, 2017

Have switched to the ESP8266 -12F for a couple reasons, the exposed ADC and deep-sleep access control.
Using the DHT11 with a one minute sampling rate with 2 AA NiMH freshly recharged batteries, I get 4823 samples with a voltage degradation curve below.  At 2.61V, the reporting to thingspeak still occur, but temp and humidity are reported at value 0.

This just gives a rough idea of battery capacity, and as deep sleep interval grows, the sample count should decrease based on the ~20uA current required by the device.

In terms of range to an AP, I am measuring about 140feet max with two walls between devices.  Using the Android "Wifi Analyser" app this seems to be in the range of -80-85dBm.

When out of range, the wifi manager will boot into an AP host mode for configuration of all parameters.  So if configuration needs to be changed, either device need to be moved out of range or reset when AP  is off.  A problem with this solution is that if the AP host is down for any reason when the ESP8266 tries to connect, the ESP8266 will stay in the AP host mode until properly configured, or  the ESP8266 device is rebooted when the AP is available.

Friday, July 1, 2016

Bee hive monitor

Using an ESP8266 v01 a single-wire DHT11 monitors temperature and humidity and logs to thingspeak channel:

The hive is about 100' from wifi access so am using an open-mesh AP that simply has a more powerful 20mW antenna and appears a reliable connection.  Android app "Wifi analyzer" is good at measuring signal strength.

There are lots of online examples of logging ESP8266 data onto a thingspeak channel, so the only apparent novelty is using a $6 solar-cell battery charger.  It is built for recharging 5V USB cell phones, but the battery is a 3.7V li-ion.  At first I was planning an LDO to get the 3.3V directly from the 5V, but the unit appears to turn off power if significant current (i.e.1A) is not being drawn. Currently the direct battery connection is driving the 3.3V ESP8266.  Current 20 second sensor updates consume more power than solar recharge, so a USB recharge is needed weekly.  Need to figure if simply lowering sample rate fixes this, or if special low-power states are needed.

ESP module was $1.78, sensor about $1, extra bypass cap, and a pullup resistor, so total unit cost <$9.

"Weather protection" is a milk carton.  Power is delivered from left and sensors go into top of hive for center air vent location.

Solar cell charger is under an old window for protection, even though it cuts down recharge rate.

ESP8266 module with DHT11 sensor.

Thursday, June 23, 2016

The automatic chicken coop door was made in April 2016 since I did not want to open the door every morning and close the door every evening.  Not only could I forget, but I did not want the hassle with other daily stuff.

A few folks have made it with the car radio antenna extruding either vertical or horizontal.  I chose vertical to avoid risking debris clogging the action.  If you have chickens you know they are messy.

I used two 12V AC adapters to each power the "retract=open" and "extend=close" and used a single AC timer to have on during the time of day I wanted the door open and it turned off at night when allowing the other 12V AC adapter to close the door.  Others found concerns with significant leakage, but the $18 radio antenna seemed to have a 5-second on or off period after which it would turn off the motor, so my configuration was not leaking measurable AC power when the door was open or closed.

I liked the spare aluminum tracks, but as usual, the more true and plumb you are the better.  Next time I would replace the vinyl door with plywood, but even the vinyl kept the raccoons out.

A concern for reliability could mitigated with periodic oiling, and I had no serious concerns during a few months usage, until the raccoon got the chickens.

Video at coop-door

Thursday, April 28, 2016

First post!  This is much better than maintaining on, but you if you like these topics, you may find some interesting things over there.