Consider for a second the Internet of Things. A vast network of connected devices, programmable matter, and wearable electronics can only mean one thing: there’s going to be a ton of batteries. While changing the battery in a smoke detector may seem tolerable, changing the batteries in a thousand sensor nodes is untenable. The solution to this problem is self-contained sensor nodes, and right now the best power source for mobile devices is probably solar.
For his Hackaday Prize entry, [Shantam Raj] is building a self-contained sensor node. It’s a Bluetooth device for the Internet side of this Thing, but the real trick to this device is solar energy harvesting and low power capabilities through optimized firmware.
Basically, this system is a low-power SoC with Bluetooth. The power from this device comes from a small solar cell coupled with a very efficient power supply and some new, interesting supercapacitors from Murata. These supercaps are extremely small, have high storage capacity, low ESR, and fast charging and discharging. The test board (seen in the video below) provides a proof of concept, but this device has a problem: there’s a single ‘sanity check’/power LED on the board that consumes 4 mA. The microcontroller, when running the optimized firmware, only consumes 1 mA. Yes, the LED thrown into the prototype that only serves as an indication the device is on is the biggest power sink in the entire system.
This project is great, and it’s exactly what we’re looking for in The Hackaday Prize. If the Internet of Things ever happens as it was envisioned, we’re going to be buried under a mountain of coin cell lithium batteries. Some sort of energy harvesting scheme is the only way around this, and we’re happy to see someone is working on the problem.
Neat!. Cool to see him making a PCB for the energy harvesting power circuit and all; ‘got my vote for the best product!
pulse the led once every 10 seconds when powered on initially and once an hour or a day after that just to have physical feedback that it is working (at least while debugging).
Or use some sort of LCD display. Those take almost nothing in terms of power. Not one of the character modules, but a raw LCD, such as a 7-segment calculator display.
or series the led with a microswitch to press when you need to view it.
e-ink display might be the perfect one here. No power draw unless the text needs to be changed (if i remember correctly)
eink takes too much power when changing tho.
Never mind I just watched another video from the same person and the device is in low power most of the time and only goes active when the available power reserves collected is enough to make it active for a few seconds. Some interesting chips for sure.
We have yet to see an IoT BEAM Bot.
I’ve been looking forward to seeing more energy harvesting of motion/vibrations. A wireless pedometer would be battery-free in my utopia.
TV remotes, wristwatches, LED keychains, ect…
Nice job on the solar!
What is that youtube video supposed to show? A blinking light?
HaD, can you contact entry submitters and ask them to shape up their content a bit before you feature them?
HaD reader(s) (in this case sdfdsfd), could you please test all links in the article (in this case one) before assuming the submitter has no up to shape content prepared?
Sorry – couldn’t resist ;)
… This is the like the only idea I had, and was unsure I could pull it off.
Meh. Did this many years ago to demo BLE. 1mA sounds a lot.
22mF ??? WTF is that? Super cap??
as stated on his/its had.io page:
“AVX Supercapacitor 22mF BZ054B223ZS Best Cap series supercapacitor from AVX. Very low leakage caps.”
This seems to be an appropriate URL: http://www.avx.com/products/supercapacitors/bestcapreg-bz-series/
OK 22,000uF , so now we have pF, nF/MMFD, uF/MFD and mF – nothing at all confusing there !!!
not really confusing just unfamiliar – it’s just properly applied metric units.
I’d have used 0,022F or as you did 22kµF (ok that last one is just awful!).
But where did you find MMFD and MFD? I’ve never heard of these deprecated (Wikipedia) units before – please keep those away from me!
‘In the beginning’ there was MFD. There was also nF which isn’t used much any more. Today we might express nF as 0.022uF which is 22millimicro (22muF) and just as weird as 22kuF. MMFD was even less common. Capacitors were also color coded.
MFD is still used in electrics rather than electronics –
https://www.umatechnology.org/wp-content/uploads/2016/07/50-5-370-440-capacitor-HCKS500D050R440Z.jpg
Use a simple e-ink indicator, requiring no power to maintain its indication state.
Not very useful for telling you if there is power available then is it?
If the design consumer consistently very low current and is powered by a small solar cell, why use high end supercaps that put a lot of cost and effort into their ability be be charged and discharged quickly. Seems like a waste to me. Put those caps in a drone or something will utilize the attributes you are paying a premium for. Just my 2 cents but it sounds like the supercaps are in there for more of a buzzword than their being a need for that tech.
Sounds like a great idea.
You could even call it…
e…
en…
Water, sea, lake…..
Ocean!
EnOcean !
You’re late BTW.
Yes, I have seen them in some buildings (i think there are multiple companies making them). Also seen rooms super heated or super cooled because the light was not turned on for a long time and the thermostat died.
EnOcean is OSHW now?!
Probably not. But there is no info you could not find by checking the datasheets of energy harvester ICs. Just because something is open source does not mean it brings in any new information or benefit.
Instead of having the LED connected and running how about a Momentary SMD switch to test the board or a dip for toggle.
http://www.ebay.com.au/itm/140pcs-14-Types-Tactile-Switches-Push-Button-Momentary-Tact-Mini-SMD-Pushbutton-/252610114683?hash=item3ad0bc747b:g:Z50AAOSwjDZYlW5O
CRAZY
http://www.ebay.com.au/itm/10-x-Red-1-Bit-1P-Position-2-54mm-Pitch-Toggle-Switch-PCB-DIP-Ping-Dial-Code-/371637071752?hash=item56874baf88:g:MYMAAOSwQaJXRGWi
Even crazier
Great. I wish to get 1$ everytime when I hear about “10 years lasting IoT device”. And then you need to attach some sensor to this extra-ultra-low-energy marvel…
to be fair electronic water meters (ultrasonic) are rated at 16 years on one battery, have nice big LCD screens AND can be read wirelessly.
I think the ultrasonic sensor burns most power, a bit of googling shows me models with huge batteries (3.6V 19Ah). They seem to require close to 1mW of average power. That is still a lot, but you probably care more about the service cost of battery replacement than the battery cost itself.
I never got why people build a super low energy node and then power it from something intermittent as ambient solar, with a reserve of a few day’s worth of energy. What if you leave the building for a few days and there is no light inside and your thermostat gets stuck?
You can easily make a temperature sensor that will last 10 year on a pair of AA batteries. By that time whatever technology and system was around that node will probably get replaced with something cheaper and better.
Just make sure you put good lithium AAs in there, not the junk cells packaged with almost every remote control.
It is true lithium has better shelf life. But good quality alkalines can be 5-7 years. Whatever charge is left after this is the manufacturers have different opinions, but I would say we can still count on 50%.
So you could go for the simple option to have 5-7 years of operation with regular alkaline or go for the much more expensive lithium, depending on the application.
BQ25570 and BQ25505 are small power harvesting marvells! They can be sampled thru TI, however arn’t breadboad friendly. Nothing to scare “generic smd to dip adapter” pcb can be purchased in Ali. Probaly this especific application needs the super ultra low leak and ESR. Probably for simpler applications 4F super capacitors ( http://s.click.aliexpress.com/e/VrrJyjM ) can do the job.