Battery Lifespan

[Martin KyteParticipant @martinkyte99762]

Or just build a sundial.

[Neil WyattModerator @neilwyatt]

Posted by SillyOldDuffer on 03/04/2018 10:30:55:

Nothing wrong with putting an Arduino Uno to sleep provided James' clock doesn't need it to be awake for some other reason. The other downside is learning to program sleep modes; they aren't the best place for a beginner to start.

The Uno is my favourite Arduino; I brook no criticism – it is the best of all possible Microcontroller boards!

Even so I have to admit that power efficiency isn't a top Uno virtue. It might suit James to use a Nano or one of the other alternatives. A raw Nano consumes about 20mA compared with a Uno's 50mA. And of course a Nano also has sleep modes, and it's possible to take other measures to reduce power consumption such as chopping out the LEDs (2mA). Might also be fun to see how much you can reduce the chip's input voltage below 5V before the Arduino crashes. Reducing voltage will also reduce the current consumption. (You can bypass the on-board regulator by applying power to the Vin pin.)

Dave

Edited By SillyOldDuffer on 03/04/2018 10:33:10

Or use a raw AVR chip and get the average consumption down to less than the battery self-discharge by using deep sleep modes

[John HaineParticipant @johnhaine32865]

LTD Stirling engine on top of a radiator?

[duncan webster 1Participant @duncanwebster1]

Posted by not done it yet on 03/04/2018 15:18:23:

Posted by duncan webster on 03/04/2018 14:56:41:

Do solar panels work off interior lighting? If so rig one up to charge the clock battery, or put the clock near a window.

About as well as a wind turbine on a car roof!

Actually, there is a solar charged clock in Buckinham Palace. As fa as I know, it is still there. I know one of the designers involved with the project.

Well according to **LINK**

a turbine with an area of 1 sq ft attached to a car doing 30 mph would generate about 50watts, so unless the OP's clock is particularly heavy on the electricity he will be well in!

[Bob Brown 1Participant @bobbrown1]

You can get solar powered watches which use solar power to keep the battery charged.

**LINK**

[James AlfordParticipant @jamesalford67616]

Thank you, again, for the ideas and suggestions. I shall have a look at the Nano, as suggested. I have the Uno simply because it was a present, rather than by active choice.

I am going to try a practical test to get some idea of likely battery life span. I have the Uno wired up to a LCD clock display and LEDs flashing instead of solenoids operating: as an interim I plan to buy some small three volt relays to replace the LEDs as I hope that these will approximate more to the current draw of a small, three volt solenoid. I shall then run it from six C cells and see how long it lasts. I can also use my multi-meter to measure current usage.

I have plenty of time to test it as I am still working on the actual clock mechanism's design.

Regards,

James.

[James AlfordParticipant @jamesalford67616]

Good morning.

Last night, I set up my Uno and clock circuitry to run on six alkaline C cells. I measured the current draw: 60 milliamps when only the LCD was running, rising to 150 milliamps when a LED triggered. Disconnecting the LCD, which is only used to set the time, reduced the draw by a couple of milliamps.

By my estimation, this will give me a running time of about five and a half days at the outside for just the resting circuit. Add in the additional drain for all of the LEDs (currently one per second, minute, hour, am or pm indicator, day of the weeks indicator and two for the quarter and hourly strikes and chimes), and I shall be lucky to see a few days before the batteries go flat. I used the LEDs, which will be replaced in time with the solenoids, to test my progamming, Oh, well, time will tell.

Somehow, I think that I shall need to use an external power supply after all, reserving a small nine volt battery as a back up, should the mains become disconnected.

Regards,

James.

Edited By James Alford on 05/04/2018 07:25:05

[John HaineParticipant @johnhaine32865]

James, one of your problems is that you are supplying the UNO board with 9v which it drops down to 5 v using a linear regulator (assuming your cells are in series). Almost half the energy of the batteries is being wasted as heat in the regulator. So actually all the battery voltage of 6 cells above about 6v which is probably the drop out voltage of the regulator is doing nothing for you. If you want to benefit from the extra cells you need to have a switched mode regulator, a buck converter in the jargon, that effectively transforms the voltage down and the current up, so reducing your current drain from 60 down to 40 mA.

[James AlfordParticipant @jamesalford67616]

John.

Thank you for the suggestion. My understanding is that the Uno needs 9v as an optimum input if using an unregulated battery supply. Is the buck convertor an alternative way of stabilising the voltage?

[John HaineParticipant @johnhaine32865]

Yes, a buck converter is an alternative to the standard linear regulator. To apply to the Arduino, you use one of the other power inputs that bypasses the on-board regulator.

A buck converter circuit using a standard chip from TI that costs £1.20 from RS is described here. The description isn't very good because it omits the key point, which is that an inductor is in effect used to drop the voltage which can give very high efficiency. You can probably also buy ready-made converters if you look around the web.

These are key components in modern electronics – for example in a mobile phone the LiPO battery gives around 4V but several of the ICs run on ~1.8V, so the 2.2 V difference has to be dropped as efficiently as possible to maximise battery life. This means that a lot of development has gone into making the chips as efficient and cheap as possible, and simple to use.

[SillyOldDufferModerator @sillyoldduffer]

Yes, 'buck converter' is a better, but more complicated/expensive way of regulating power.

A downside of the Arduino Uno is that it's designed to 'keep it simple' for beginners, and that includes simplistic advice about how best to power it. Sensible and safe. Actually there are a number of possibilities.

It's a 5V board. You can power it with a USB cable from a computer or a wall-wart charger. (USB is 5V). Or feed it 5V from an external stabilised power supply through the Vin Pin. Or feed it 6 to 20Vdc rubbish through the power socket.

As John says the Uno's power socket input uses a linear regulator to burn off excess volts, which is very inefficient if the device is being powered by a battery. To avoid overheating and voltage drop-out problems Arduino recommend keeping the power socket input voltage between 7 and 12V well within the absolute range which is 6V to 20V.

However, whenever I battery power a Arduino I use a 6V accumulator like this example. Although below the 7V recommendation it works. I have to keep an eye on it though: one day I might plug in a Uno and find its regulator blips out on only 6V causing crashes etc.

The more advanced Arduino M0 uses a switch mode regulator that doesn't burn excess volts. It only requires 20mA while being 3 or 4 times faster than a Uno, with massively more memory and several other tasty features. Don't get too excited though, the price for the goodies is that it uses 3.3V logic and the input-output pins are comparatively under-powered and delicate. In other words it's much easier to blow up!

Something wrong if turning on a LED causes the current to rise to 160mA. I suspect a short-circuit. The maximum current provided by a single output pin is 40mA (20mA recommended) and a small LED might draw between 2 and 10mA. Are you switching the LED with a transistor? If so don't forget that, unlike a solenoid, the LED will need a dropper resistor to limit the current. (Try 330 – 1500 ohms)

Dave

[Ady1Participant @ady1]

I had forgotten about my old watch

Seiko solar watches seem to go on forever kinda thing nowadays, they use some sort of capacitor battery and top it up with a solar face

You would need to research it to get the details

It's a very reliable system, I got my second hand ebay one 10 years ago

Edited By Ady1 on 05/04/2018 10:58:32

[magpieParticipant @magpie]

In an early version of my Electro-Mechanical Fibre optic Clock, I also intended to use solenoids to operate the cylinders. However, I found them to be rather inefficient and abandoned that idea in favour of ratchet wheels and cams to operate the main cylinders. The drive on the final version of my clock is a synchronous motor which has worked very well for a number of years now, and it keeps perfect time. A series of post about it can be found in the clock section, at the top of page 6. It may give you a few ideas. Good luck with your project.

Edited By magpie on 05/04/2018 12:09:51

[James AlfordParticipant @jamesalford67616]

John and Dave: Thank you for the advice on buck convertors. I shall look into this.

John: have removed two batteries, running it now with six volts. This has reduced the current draw to between 35ma and 75ma. I shall check again at the weekend, though. My everyday multimeter has a small dial and is hard to read at the best of times. It is most useful for estimating than measuring. I have a large AVO tucked away and shall check with that.

Dave: I have a number of LEDs fitted, only one flashing at a time. One is triggered through a transistor, the rest straight from the pins. They all have resistors fitted, but I think only 100 ohm. Perhaps I should be increasing this, at least whilst using LEDs to test the circuits.

Magpie: I had a quick look at your LED clock thread. It looks really impressive and I shall read it in detail. Thank you.

Regards,

James.

 

Edited By James Alford on 06/04/2018 07:10:50

[I.M. OUTAHEREParticipant @i-m-outahere]

May be a strange question but is his clock going to display the seconds ? If it is then you could significantly reduce the poewr usage just by eliminating this by using the movement to only show minutes like a digital clock does . I wonder if some form of capacitor discharge circuit to drive the solenoid my be more efficient ?

[James AlfordParticipant @jamesalford67616]

XD35q
Not a daft question at all. Ideally I should like to display the seconds. However, I am happy to not do so if it saves significant battery life and allows the device to run for several months on one set of batteries. To get a realistic idea of potential running time I really need to wire a solenoid in the circuit to test it properly.

Edited By James Alford on 06/04/2018 12:35:10

[John HaineParticipant @johnhaine32865]

This may be a retrograde suggestion, but the most efficient way to display seconds would be to use a quartz mechanism with the stepper motor wires brought out rather than driven by the internal oscillator and divider. These little stepper motors only need a minuscule current at around a volt for a short period to drive them on, so consume almost no energy or current. Then the seconds display is just a conventional hand. You could even use the rest of the train to drive the minute and hour hands…..

[James AlfordParticipant @jamesalford67616]

Posted by John Haine on 06/04/2018 15:16:35:

This may be a retrograde suggestion, but the most efficient way to display seconds would be to use a quartz mechanism with the stepper motor wires brought out rather than driven by the internal oscillator and divider. These little stepper motors only need a minuscule current at around a volt for a short period to drive them on, so consume almost no energy or current. Then the seconds display is just a conventional hand. You could even use the rest of the train to drive the minute and hour hands…..

True, but it would not be the design that I hope to create.

I tested the current draw again today with an AVO with a large, clear dial. It showed 45ma, rising to 55ma when an LED flashed.

James.

[James AlfordParticipant @jamesalford67616]

An update on battery life.

For what it is worth, using four alkaline 1.5v C cells to power the Uno, a LCD clock display and LEDs flashing each second, minute and hour, along with a flashing LED for quarterly and hourly chimes, the system became unstable after seven days. The chime LEDs are disabled automatically at night.

Whilst I know that this could be improved by following the suggestions given here, I think that I have concluded that I shall need to use a transformer for the final device, reserving a battery for back-up, should the mains go off.

Thank you for your advice.

James.

[SillyOldDufferModerator @sillyoldduffer]

Posted by James Alford on 12/04/2018 07:17:21:

An update on battery life.

For what it is worth, using four alkaline 1.5v C cells to power the Uno, a LCD clock display and LEDs flashing each second, minute and hour, along with a flashing LED for quarterly and hourly chimes, the system became unstable after seven days. The chime LEDs are disabled automatically at night.

Whilst I know that this could be improved by following the suggestions given here, I think that I have concluded that I shall need to use a transformer for the final device, reserving a battery for back-up, should the mains go off.

Thank you for your advice.

James.

Thank you for trying the experiment and letting us know the result.

As the low-power members of the Arduino family consume about half as much power as Uno, you would still only get a fortnight out of four C cells – the answer is still a transformer.

Looking more deeply into earlier suggestions of putting the Arduino into sleep mode, I don't think that's a good idea in your project as it stands. In sleep mode pretty much everything is turned off including the Arduino's clock oscillator. The low-power timer used to wake the Arduino up periodically is an RC oscillator – not accurate. The solution is not to use the Arduino itself as the timekeeper, but to connect it to a Real Time Clock module like this, as described here. Then, you can put the Arduino in power saving mode waking up periodically to check the time, and only updating the display and striking as needed before dozing off again. RTC chips are ultra-low power devices, typically lasting 2 years on a button cell.

If I were you, I'd build the clock as originally conceived and may be return to these more advanced possibilities later if it suits you. One of the nice things about microcontrollers is that you can come back and improve the way they work (accuracy and power-saving) without having to mess with a clock or model's mechanicals.

Dave

[Neil WyattModerator @neilwyatt]

It's worth looking at this excellent tutorial on redcuing arduino power consumption on the Sparkfun website.

I've had an RTC module running on a 2032 lithium button cell for nearly 2 years and going strong, so taht's agood idea of Daves.

Neil

[James AlfordParticipant @jamesalford67616]

Dave,

I have also come to the conclusion that I am better off with a transformer and worrying about saving power at a later stage. I should have to reduce consumption dramatically to make much practical improvement, especially once the solenoids are in the circuit instead of LEDs.

I am already using an RTC to keep the time, the RTC module having its own battery. The Uno interrogates the RTC for timings, with the Uno controlling the functions at intervals determined by the RTC.

 

Neil,

I shall have a look at the Sparkfun article. I have used a few ideas from that site, but not on this aspect.

The circuit that I have built uses a LED display to show the time, with a couple of press buttons wired in to advance and retard the time. This will be used to set up the time whenever required, rather than plugging the final clock into a laptop and setting the time via the IDE. I shall arrange for the display to switch off when not required. Ultimately, I should like the finished device not to require a computer to be viable.

 

Again: thank you all for the help and suggestions.

Regards,

James.

 

Edited By James Alford on 13/04/2018 07:15:38

[Author]

Posts