Comment by wolfgang42 | Hacker News Reader

Comment by wolfgang42 | Hacker News Reader

From #diamondbattery tweets: https://twitter.com/search?q=%23diamondbattery

Could be used in underground seismic monitors https://twitter.com/Keminoes/status/803297734694465541

Car keys that require batteries. If we don't have a spare battery with us and our keys die, we're stranded. MP3 players, too https://twitter.com/tysongeisler/status/803438384052129796

To power a low-power-mode Arduino for gas detection (CO, CO2, Low O2, etc), fire detection, radiation detection, etc. https://twitter.com/noalear/status/803389146433626112

Could diamond batteries be used to power medical nanobots? https://twitter.com/weirbe/status/803418222972125184

Power clothes that contain sensors, as well as clothes that electrically regulate the temperature. https://twitter.com/CIMCloudOne/status/803440800759574528

how about in cell phones to give a little charge to the battery while the phone is not in use. https://twitter.com/W_Haas/status/803352403512872961

Is there the potential to power watches as a fair amount of waste is generated from depleted watch batteries each year. https://twitter.com/Merrett72/status/803297788163289088

> To power a low-power-mode Arduino for gas detection (CO, CO2, Low O2, etc), fire detection, radiation detection, etc. https://twitter.com/noalear/status/803389146433626112

Except you wouldn't use an Arduino; AVR processors are notorious power hogs when you start counting the milli (and micro) watts. Run-of-the-mill TI MSP430 microcontrollers, aside from being arguably more powerful and better designed at similar price points, consume less than 1 microamp in idle mode.

> AVR processors are notorious power hogs

> MSP430 microcontrollers ... consume less than 1 microamp in idle mode.

As someone who's done extremely low-power development on both platforms, this is an extremely disingenuous comparison. Of course MSP430s consume very little current in idle mode. So do AVRs. People just don't usually run them at low frequencies or in idle mode, because Arduino is targeted at novice users, not people who know how to do low-level power management.

I've been running an ATTiny off the same pair of AA batteries, blinking an LED for 50ms every 5 seconds for 4 years. Just turn all peripherals off, maximum clock prescaler division, sleep mode with wake on watchdog interrupt. There's functionally no way to get better performance out of an MSP430 in real-world applications.

While I prefer the MSP430 instruction set, beyond that there's no obvious all-around advantage over the AVR family.

> maximum clock prescaler division

How do you calculate the tradeoff between the low power of a lower clock frequency and the shorter active time with a higher clock frequency?

If you're computation bound, unless you have a very current-limited power supply, you can just max out the clock when you wake up and turn it back down before sleeping.

In my case I had to be awake for some time anyway, so I just used the lowest clock frequency I could.

Just to back this point up, I have an atmel 328p (auduino duo) thats been running on single tiny solar panel, and 5 farads of super capacitor for about 4 years.

It transmits a temperature reading every 30 seconds over rfm12b radio.

(obviously the panel keeps the charge topped up, but it easily lasts through the night and a very cloudy day. It even works indoors. )

>I've been running an ATTiny off the same pair of AA batteries, blinking an LED for 50ms every 5 seconds for 4 years.

Is this just for fun or is it part of a larger project?

For fun. My parents wanted a fake security system with a blinking light at their house to ward off burglars, so I soldered an LED and a resistor to an ATTiny and then soldered on a double AA pack. It's still going off the first set of batteries.

I've used both. The MSP430 has a few advantages over the AVR family when it comes to low power.

1) There are specific low power designs. They use lower voltages, which allows you to squeeze more energy out of each battery, and they use non-volatile RAM.

2) There are a lot of MSP430 variants, and they have integrated all sorts of features that would have to be external on an AVR series microcontroller. E.g., a real time clock, or an LCD driver.

3) The MSP430 series has variants with almost any quantity of digital or analog IO you could wish for. This means you can directly connect the MSP430 to things that you would need a breakout board for on an AVR.

So even if the AVR series and the MSP series are on relatively equal ground when it comes to power per clock (which they aren't), the MSP430 is a much better chip for low power draw.

Thread a few weeks ago on the MSP430:

https://news.ycombinator.com/item?id=12952100

tl;dr (I hunted up the datasheets); they'll run (i.e. processing numbers, not idling) at 120 µA/MHz, which is miniscule. However, there are some new ARM-Thumb designs which will run at 35 µA/MHz:

http://www.atmel.com/products/microcontrollers/arm/sam-l.asp...

> Car keys that require batteries. If we don't have a spare battery with us and our keys die, we're stranded.

There was a great solution for this in use for over a hundred years.

My 2006 has a workable solution as well that doesn't require a traditional key, there's a slot you can put your key fob in to start the car even if the battery in it is dead (only option on trims sold without the Smart Key System) and the driver door has a mechanical lock that can be opened with a key hidden in the fob.

Keyless entry and ignition is a great convenience, especially for my wife since she can keep the fob buried in her bag - and it makes it easy for me to put my daughter in the car without trying to wrangle the fob out of my pocket to unlock the door. But it floors me that so many manufacturers just ignore the use-case where the battery in the fob is dead...

But do they, though? I'm thinking it's just a matter of actually reading the owner's manual.

For instance: there's a key hidden in the fob of my Nissan Leaf that is able to mechanically unlock the driver door, same as your 2006. To start the car, you place it against the "start" button. I'm assuming there's some inductive trickery going on.

Anyway, I'd be surprised if our cars (2006 and 2015) were the exception, rather than the rule.

> But do they, though? I'm thinking it's just a matter of actually reading the owner's manual.

It's in the glove box...

We're on our third keyless entry car. The 2014 Chevy Volt was the worst, but they all had that feature. I had to read the manual to find the trick, but you:

* Open up the compartment on top of the center of the dashboard

* Take out the rubber liner on the bottom of the compartment

* Pop the mechanical key out of the fob

* Poke the mechanical key into an unlabeled hole in the bottom of the above compartment

And bingo, your car will start.

The Kia Sorrento we have with a keyless system is much more sane: The center panel has a really obvious place to push the whole keyfob in, and it will start with a dead fob battery.

The Toyota Prius we used to own also had keyless, and also had a place to insert the whole keyfob to start the car when the battery died. And it was also pretty obvious (and labeled).

It's only the Chevy Volt where it was an absolute mystery what you're supposed to do (unless you read the manual). I don't think I would have ever guessed that you should disassemble the compartment on top of the dashboard and stick the key into an unlabeled hole. Would have been afraid of shorting something out.

On my Ford Fusion, it apparently has a passive RFID chip in the key fob in case the battery dies, and the reader is in the bottom of the cup holder. And a hidden key inside the fob, that you can use to pry off a plastic shroud on the door handle to get access to a hidden cylinder lock on the door.

By the time this technology is out there, car keys would be obsolete. You would nt own a car.

I have a 2012 Lexus that actually tells you "place key on start button" if it thinks the key battery is dead.

What? I had a 2004 BMW that would charge the key battery through induction when you had it in the ignition and the car was running. Battery never died in 10 years owning the car. How have things gotten worse since then?

Key doesn't go in the ignition, kind of hard to charge a battery if it's floating in a purse or pocket.

That's my point, why would I ever buy keyless entry/start? Of all the things that bother me with car ownership I never felt burdened by having to put a key in the ignition. That's the best part!

The only appealing thing to me would be that keyless entry/start usually comes with remote start, which in the northern climates would be nice.

But I'm with you as far as having a physical key goes. Never have I thought, "man, this key stuff is just so tough!" especially since every other door in my life still uses a mechanical key which would likely be attached to the remote key fob. When I get home and turn off the car I'm going to need those keys in my hand anyway to get in my house.

Depends on how you start your car I guess.

My car is also a BMW product, but it was a big toggle switch on bottom of the dashboard for Engine on/off. Pulses red when it's ready to start.

I have a 2010 Prius with keyless entry and start - you just need the key in your pocket or bag. It may sound a bit gimicky, but I really like it. No more hunting for the key, I just leave it in my ruck sack.

If the battery dies there is a hidden key in the fob that can be used to unlock the drivers door. To start the car you can hold the fob on the 'Start' button - there is an RFID chip inside which is read.

A tablet computer that does not need external power would be nice. Even using this in an e-book reader with an e-ink display would be great.

A wireless mesh network that needs no external power for emergency communications would also be nice. Just imagine being able to place nodes on utility poles and not having to worry about paying for electricity.

Reply on news.ycombinator.com