What stops your Arduino project?

I'm concerned by power as @javagoza is. In fact i want to make an iot device with an arduino, and can't make it because it uses so many mAh!

If one wants to join on a easyEDA project to make a physically compliant UNO  :

- powered directly by a 18650 cell (4.2 - 3.3V) / solar cell combo

- without usb, leds, and other power consumers.

- with an adjustable boost driven by the uno delivering power up to 2A on the Vin pin (for activating shields, especially GSM)

So one can make battery powered iot projects with the UNO, and drop them in remote areas.

I need electronics experts.

Best to everyone.

Perhaps keep in mind that a lot of those mAh are under user control however. A number of people who used to develop projects on the Arduino Uno on the test bench, ended up deploying them on the Arduino Pro Mini 5v to save power out in the field. Optionally deploying them on the Pro Mini 3v3 version reduced power consumption even further.

The Pro Mini did away with the second microcontroller used for the USB UART bridge, reduced the number of LEDs down to pin 13 and provided a jumper to disable the onboard voltage regulator, so pretty much what you have mentioned above.

On top of that the folks at Atmel designed all sorts of power saving features into the AVR family including the ATmega328P chip used in the Uno / Pro Mini allowing you to get current down from mA to μA and those suited for battery-powered projects. ( They even documented it all for you )

The requirement for 2A to drive shields is possibly the more significant issue here as now you are talking 'A' not 'mA' (let alone 'μA'). Once again though, Uno shields probably aren't the best choice for remote battery-powered projects out in the field. Great for the test bench but not necessarily optimised for low power use out in the field.

I saw a project a while ago using a Pro Mini to power up a MKR GSM 1400 once an hour in order to take a ultrasonic sensor reading and transmit it as a SMS txt message over the mobile phone network, which is another way to save lots of power if you don't need to be constantly transmitting over GSM networks - power everything down and go into low power sleep mode when not doing anything and save lots of mAh in the process.

I recall a similar technique was used in one of the Ben Heck Show episodes on here for a battery powered geocache project where an ATtiny was used to power up a Beaglebone Black.
/challenges-projects/element14-presents/benheck/ben-heck-exclusive/w/documents/872/episode-197-ben-heck-s-beaglebone-black-geocache-episode
 

Waveshare have been creating some solar power management modules for remote powering of microcontroller projects, which sounds like they could fit in with what you are trying to achieve. There are both 14500 cell and 18650 cell variants with solar cell charging.

Have you seen Doug's rather impressive solar powered super capacitor project ?
A MSP430 microcontroller powered off capacitors being charged via solar.
/challenges-projects/design-challenges/just-encase/b/blog/posts/just-encase-solar-super-capacitor---project-outline---blog-2

Atmel produced an application note a while back demonstrating their AVR microcontrollers running off a capacitor using some of the 'picopower' mode power-saving features. I think there were also some YouTube videos covering it.

beacon_dave said:

The Pro Mini did away with the second microcontroller used for the USB UART bridge, reduced the number of LEDs down to pin 13 and provided a jumper to disable the onboard voltage regulator, so pretty much what you have mentioned above.

Hello Dave,

Thank you very much for your insights, links, and brainbusters.

Yes it can be a good idea to plug a 3.3V pro mini on an uno sized PCB, for not-reinventing the wheel. I'll give it a try asap.

GSM shield may eat 2A for few ms during the power rush. Then it's about 300mA, for 1mn. This is 5mAh, twice a day. Solar cell can provide this much of the year.

This is for a POC, so finding things on the shelf will help run faster through the process.

Have a nice day.

Sparkfun who make the Pro Mini have some power saving tips here:
https://learn.sparkfun.com/tutorials/reducing-arduino-power-consumption/all

This was the project demo using a Pro Mini powering a MKR GSM 1400 showing the power consumption at idle, sensor reading and SMS messaging stages:
DIY Water Level Indicator using MKR GSM 1400 with and without power saving:
https://www.youtube.com/watch?v=xor1Vk66k2I

and with Pro Mini power switching added:
https://www.youtube.com/watch?v=YVho42UC43Q

Looking at the spec sheet of the GSM module on the MKR GSM and it looks like you can potentially optimise power there as well.    

These were the Waveshare solar power modules I had seen, which may give some ideas for your proof of concept:
Waveshare solar power manager:
https://www.waveshare.com/solar-power-manager.htm

https://www.waveshare.com/product/modules/others/power-relays/solar-power-manager-c.htm 

Old Atmel AVR picoPower Labs demo:
https://www.youtube.com/watch?v=DGtYKFhLxPg

and related application note:
http://ww1.microchip.com/downloads/en/appnotes/doc8349.pdf

Thank you Dave.

I've come to thinking that any powering device that uses USB ports to power an IOT gizmo is a nonsense. Into the wild power is scarce and if we want to drop and forget the IOT device we need to power it as much as possible from direct battery connection.

Real world IOT is far from mains, WIFI, bluetooth. It leaves you with satellite, GSM, or LoRa way of connect to the cloud. For satellite and GSM you need solar powering. For LoRa you can't go farther than very few miles away from gateways / mains.

RTC / interrupts are a must as IOT devices are to sleep 99% of time.

I would say that 'real world IOT' is more varied than you suggest. A lot of it is near to mains, Wi-Fi, Bluetooth, and a lot of it isn't. (And some of it falls somewhere in-between e.g. generator power / in-car). You therefore need consider power and network connectivity on a case-by-case basis.

The Uno dev board, like many dev boards, has USB primarily for ease of use to facilitate the prototyping of microcontroller projects (originally for educational purposes). It is slightly easier for most to connect an Uno form factor dev board to a PC via USB  than starting out with just a bare chip pushed into breadboard. In a lot of cases though, after programming and debugging, students are often expected to remove the DIL package microcontroller chip and plug it directly into their own circuit board, and then power it appropriately for the intended use case. In some cases it would be nice to have a Uno style dev board which you could snap in half to turn it into a Pro Mini style deployable board without the 'excess baggage'.

Powering via a USB port may or may not be appropriate. In some cases it is just a readily available connector providing a +5v supply. With the Uno however with the USB comes the second 16U2 microcontroller to provide the USB to UART bridge, so if you aren't actually using that then you are likely wasting a significant amount of power. However the same could also be said for the main 328P in that if you aren't using most of its features, or not using them efficiently, then you could also be wasting power. Buck/boosting of power supplies can be inefficient use of limited power as well. In most cases, you have to weigh all this up as part of the larger picture and consider the cost of development time versus that of simply using a higher capacity solar panel/battery power source to start with.

Off-grid applications out in the wild are a very common use case for IOT these days. I've seen cases though where people are still taking a wheelbarrow load of 12V car batteries out to a remote field stations every week just to keep the power going. I think they have been doing this for so long now with data logging applications, that they haven't really considered other options such as low power + solar.

For GSM you still need to be within range of a GSM base station which isn't always the case out in the wild. In theory, LoRa just needs to be in reach of another LoRa node in order to be able to relay the data. In some applications this is practical. However you will likely need synchronisation so that the LoRa nodes all come out of sleep at the same time in order to be ready to receive a transmission and to retransmit it before going back to sleep again.

> (originally for educational art installations purposes)

Arduino project was started for artists use, the educational aspect came just after that.

> (originally for educational art installations purposes)

Arduino project was started for artists use, the educational aspect came just after that.

Thanks for the correction. I was under the impression that it was born more out of teaching physical interaction in the arts / design classes.

"...Arduino was initially designed for artists, designers, and others who want to incorporate physical computing into their designs without having to first become electrical engineers..."

"...The Arduino project was developed in an educational environment and is a very popular educational tool..."

"...Arduino was born to teach Interaction Design, a design discipline that puts prototyping at the centre of its methodology..."

[ Source: Getting Started with Arduino, 4th Edition: Massimo Banzi and Michael Shiloh ]

"...Forged in the crucible of a classroom, the idea of an accessible, affordable electronics development platform was under serious investigation. It would eventually give birth to the Arduino UNO..."

https://blog.arduino.cc/2021/12/09/one-board-to-rule-them-all-history-of-the-arduino-uno/?queryID=8893646e63fe5f75bf458290b61d600d

I was aware that the Processing IDE was successfully being on the software side to teach digital art to non-programmers which then led to a student project the Interaction Design Institute Ivrea in Italy to extend it to physical computing which resulted in the Wiring project with production of the various Wiring dev boards which eventually were used for teaching physical computing. 

The Arduino project appeared to reuse the Wiring IDE but replaced the Wiring ATmega128 boards initially with lower cost ATmega8 based dev boards in kit form, which could be more easily assembled by the students on the physical computing courses.

"...the earliest of Arduino’s self-assembly boards, which were used almost exclusively in the classroom... ...So it’s optimized for education, not for electronic operation!”..."

So lots educational appearances throughout its history. I still think that 'Processing' and 'Wiring' weren't the best of project names to use.