jra wrote:Since you have a regulator for the ESP8266 anyway, I would consider choosing one that has an Enable pin and use that to toggle power to the ESP8266 and sensors. There are currently two (as of yet untested) ESP8266 breakout boards on Osh Park that use this design, see http://www.forum-raspberrypi.de/Thread- ... ekt?page=2 for details. The page is in German but google translate turns it into understandable English.
Thanks for the reference, I haven't read a translation yet, but if I am understanding you, it sounds like a reasonable approach. I actually thought about such an approach when I got hold of the MCP1725 that I mentioned in an earlier post. That regulator has a shutdown pin (although it is active low) and the RTC could potential be attached there with some kind of inverter. What I would gain would be no need for the MOSFET, but what I would lose is the quiescent current that the regulator would pull all the time it is not on. The data sheet says .1 microamp typical. Using the transistor, theoretically, means no load at all on the battery until it is pressed into action. Admittedly, I don't know what that quiescent output means in battery life for 6 months-1 year (I'm too tired or lazy to try and figure that out),
I have read your posts with a great attention since I am working on a coin cells based project.
My project will run on two CR2032 during 1 year and some months, using deep sleep mode.
As the deepsleep current is still arount 30uA, your RTC solution would extend my barrety life to 4 years !
Now I have one question: how do your circuit shut down?
I understand that the RTC will trigger the power source as its alarm will be ringed. at the end of the operations (say after 3s), what causes the power shut down of the ESP?
DrG wrote:
Thanks very much for the comments. As a result, I have done some homework and redesign using a TC1262 LDO (500 mA) and a LP0701 P-channel MOSFET.
For the regulators, I obtained the TC1262, MCP1825s, and MCP1725. All three are 500 mA and they have similar specifications in my view. The MCP1725 has a “power good” output that may have some uses. For now I am using the TC1262. All three have a 6 volt input limit and if I move from 4X AA 1.2 NiMH batteries to 1.5v alkaline, I will use 3X AA to keep the input voltage away from the limit.
Learning about the MOSFET replacement for the PNP has been challenging and some of it is starting to make sense. I chose the LP0701. Key specifications for me: Id, -500 mA; Vgs~0.7v.
I have not put capacitors in the schematic but will be using them before and after the regulator. I will also test a higher value resistor for R1 (2.2k-4.7k).
The new version of the circuit appears above. It has been bread boarded and is testing just fine. But it is a limited test – I am simply turning on an LED for 5 sec on an alarm occurring every minute. After I do some more coding with the RTC, I will start to move it to an ESP8266 and move forward.
Again, I very much appreciate the feedback and please do comment further if the mood strikes you.
Cheers,
DrG
Start and upload data every 10 minutes:
- RTC Alarm set to 10 min
- when 10 min -> alarm ON -> boot ESP8266 -> check for WIFI OK -> wait 2 sec -> transmit Data -> reset RTC + 10 min -> Full shutdown.
Nothing can consume less than a powered off device.
This is how is looking live data upload, with a small Li-ion Battery , now in day 2 of running:
https://thingspeak.com/channels/58350
When started not fully charged. And a, don't even think it's a 2000mAh Battery, as printed on the label )
And this is the board, with RTC, LI-ion Charger and DHT22 Sensor, on top of a ESP8266 NeEXTEVO v3 board:
Available on Tindie: https://www.tindie.com/products/nEXT_EVO1/universal-ac-mains-dimmer-mpdmv41/