DeepSleep duration: can't exceed approximately 30min

Projects Programming
1

When my DeepSleep length is 30min or more, the code works i.e. it goes to sleep and then wakes up again.

If I go above that then the application goes to sleep but never wakes up.

The bits of code that I am using for this are

#define SLEEP_LENGTH 60*45

and

delay(100);
  ESP.deepSleep(SLEEP_LENGTH * 1000000, WAKE_RF_DEFAULT); // // deepSleep time is defined in microseconds. Multiply seconds by 1e6
  delay(100);

Schematic.png
Schematic.png800×537 46 KB

Currently nothing is connected to VIN and GND2 as I am powering via the USB port. I will add the VIN power part later.

I read somewhere that instead of a straight short to the RST pin that I should include a resister between D0 and RST. This does not seem to have helped.

I am wanting to put the application into deep sleep for 60min.

WeatherStationBMP280-Blynk-Sheets.ino (3.5 KB)

Schematic.png
Schematic.png800×537 46 KB

2

You have:

#define SLEEP_LENGTH 60*45
  ESP.deepSleep(SLEEP_LENGTH * 1000000, WAKE_RF_DEFAULT);

The compiler sees:

  ESP.deepSleep(60 * 45 * 1000000, WAKE_RF_DEFAULT);

What type does deepSleep() take as the first argument? What value are you passing it? Does that value fit into that type?

You could always nap for 1 minute at a time, and do that 60 times. The power consumed on waking 60 times for a new nanoseconds will be trivial.

3

What type does deepSleep() take as the first argument? What value are you passing it? Does that value fit into that type?

Yes I thought of that but I can't seem to work out what the type is. Not sure where to find this in the libraries.

You could always nap for 1 minute at a time, and do that 60 times.

The longest I can go is 30 seconds so I suppose I could do it 120 times instead.

4

Not sure where to find this in the libraries.

Post a link to the library, or libraries. Use the icon to the left of the X2 icon.

60 * 45 * 1,000,000 = 2,700,000,000

says:
Long variables are extended size variables for number storage, and store 32 bits (4 bytes), from -2,147,483,648 to 2,147,483,647.

Since the calculation exceeds the upper end of a long, I suspect that the deepSleep method takes a long.

I guess it takes a signed type, so you can take negative length naps. Any programmer worth a damn would have used an unsigned type, so you could have slept for 4,294,967,295 microseconds (or 71+ minutes).

Of course, any programmer worth a damn would have figured out that sleep for only a few microseconds would have saved any real energy, so would have defined the sleep interval in milliseconds, allow for almost 1200 minute naps. Now, THAT would save energy.

5

Hi,

Yes I did think that might be the issue and did also find it strange that the units had been defined as they have been; but then I am still bumbling my way through all of this so who am I to question? :slight_smile:

This is the library that I am using is GitHub - esp8266/Arduino: ESP8266 core for Arduino

Thanks

6

so who am I to question?

Anyone has the right to question apparently stupid decisions.

This is the library that I am using

Providing a link at that high a level is like saying "I got it from ebay". A more focused link is in order.

7

I see this definition at Arduino/cores/esp8266/Esp.h in esp.h:

 void deepSleep(uint64_t time_us, RFMode mode = RF_DEFAULT);

So I suspect that the compiler is doing the math first, in signed long and then promoting the negative result to uint64_t as some humungous positive number. Try casting. Also turn up warnings in preferences to the max and you'll likely see some overflow warnings.

8

So I suspect that the compiler is doing the math first, in signed long

Yes.

Try casting.

Or declare one of the value as unsigned long:
ESP.deepSleep(SLEEP_LENGTH * 1000000UL, WAKE_RF_DEFAULT);

9

ESP.deepSleep(SLEEP_LENGTH * 1000000UL, WAKE_RF_DEFAULT);

Thanks, that did it :slight_smile:

All working.