Forum Replies Created
-
Posts
-
Hi,
I think you might have already found your answer in the forum.
But, as I have been using the same resistor divider and a capacitor to measure the battery voltage using the internal 1.1V reference all this while but it is always on pin ADC6 or ADC7. So, if you look at section 23.9.5 of the datasheet (register DIDR0), that might be answer. I have yet to dig the Arduino core on whether they does this when using the pin as an analog pin. Chances is they don’t as this is granular detail in power optimization but I could be wrong.I don’t think it’s hard to use the same code base for the Seeduino LoRaWAN. It’s basically replacing the LMIC portion (SPI based) with the RHF76-052AM portion (UART based). I have never use that RHF76-052AM module, so I’m not sure.
Hi Francisco,
It’s best to use the RadioHead library. The board was heavily tested with this library and if you look at the examples bundled with the RadioHead library, you’ll notice there are guides within the examples for the Mini Ultra Pro. This is for both RFM95W and RFM69HCW radios.Arduino LMIC is only for RFM95W to run LoRaWAN stack on the board.
It looks that problem stems from the AVR runtime library power.h that has been wrongly patched and accidentally removed the
power_usart3_disable()andpower_usart3_enable()functions. Will look into this in January as I’m super busy at the moment unfortunately. But, you can add the necessary macro retrieved from earlier version of the AVR runtime library.Would be great if you had a mini-Mega2560
ATMega2560 is a super expensive chip for an 8-bit MCU.
Also:
‘shop’
‘Arduino Official’
http://www.rocketscream.com/blog/product-category/boards-kits/arduino-and-compatibles/arduino-official/
goes nowhere.No longer selling official Arduino boards. Will remove them, thank you very much for pointing this out.
Are you sure they are empty? flush() could be implemented in Stream class. I could be wrong because I rarely use 32U4.
Try putting some characters into print() and call flush() before sleep. See whether all characters are sent before sleep.Else it is best to simply put a delay by comparing it to millis(). Just estimate the time by calculating how many bytes you are sending.
The 32U4 USB needs to be detach before entering a power down mode and attach back once awake.
So, it supposed to look like this:USBDevice.detach(); // Add your wake up source here, example attachInterrupt() LowPower.powerDown(SLEEP_FOREVER, ADC_OFF, BOD_OFF); USBDevice.attach(); delay(1000); // In order for PC to have enough time to show the connected COM portBut, unfortunately, if you look at the core code and search for USBDevice.detach(), you’ll see that it’s an empty function (not implemented). So, you need to add the detaching the USB portion before sleep:
// Disable USB clock USBCON |= _BV(FRZCLK); // Disable USB PLL PLLCSR &= ~_BV(PLLE); // Disable USB USBCON &= ~_BV(USBE); // Add your wake up source here, example attachInterrupt() LowPower.powerDown(SLEEP_FOREVER, ADC_OFF, BOD_OFF); USBDevice.attach(); // In order for PC to have enough time to show the connected COM port delay(1000); // In order for PC to have enough time to show the connected COM portAlso add some delay in the setup code portion so you have enough time for code uploading.
Hi Jeremy,
Depending on which XBee module you use, the XBee could consuming more than the 250 mA of current during transmission? If yes, you would need an external regulator as the on-board MCP1700 is good up to 250 mA. You can also try to add some bulk capacitor on the 3V3 before trying the external regulator approach. Based on my past experience on using XBee (older models though circa 2009-2012), the TX current can be very high.
No it would not work if put to sleep. That’s how the Time library work without a physical alarm as it depends on the millis() to count time. The synchronization only happens in setup() to let the Time library knows the real calendar time. From there on wards, it depends on millis() to compute the alarm time in the future. millis() doesn’t run in sleep as it depends on timer 0.
You would need to call the sync function again once awake from the 8 s power down period. Best method is using an external RTC with interrupt like DS3231. That way you can put the MCU to sleep forever and once the alarm fires, the interrupt pin will wake the MCU. But, again the millis() is stop during these sleep period. I don’t quite get why you need to maintain the millis().
The time library depends on either the millis() OR an external hardware RTC. In power down sleep mode, the Timer 0 that millis() uses is stopped. Best way is to use an external hardware RTC and tie an external interrupt capable pin such as pin D2 or D3 to wake the processor up from power down sleep.
Hi Thomas,
It looks like you are powering the board from VIN (judging from your comment on PG LED turning ON), hence the mA range. If you want to achieve 20 uA, please power it through the Li-Ion/Pol connector as shown in this guide.You can think VIN as your secondary power (external DC, solar power) that is not always around. PG turns on when a valid and good VIN is applied on VIN or the USB port is connected.
Hi Matthias,
A quick question – do I actually need to read out the Device ID of the Mini Ultra Pro and inserting it into the TTN dashboard?
I mean, couldn’t I make up my own Device ID (or let TTN generate one) and insert that Device ID into the LCIM code that I upload to the Ultra Pro Mini? Or is the actual hardware Device ID necessary (if so, why)?You can used a random number during development, but in real LoRaWAN deployment, it is necessary to have a unique 64-bit ID. So, every single device is unique in the network. For example, every single RN2483/RN2903 module has an unique 64-bit ID too (same ID chip used on the Mini Ultra Pro V2). When doing an OTAA, basically you join the network through a process of exchanging some unique keys with the backend to enable the network to differentiate each devices globally (before assigning the dynamic DevAddr). You might want to read the importance of it here.
