Battery Lifespan

Good afternoon.

I am slowly designing a clock that will be controlled by an Arduino Uno and operated by a series of solenoids. To avoid trailing cables, I plan to power it with batteries. I have no idea yet what current it will draw, so plan to incorporate as much "power" as reasonable into the space for the batteries.

Assuming the same circuit and operating conditions, and aiming for a 9v supply, which of the following configurations is likely to last the longest, at least in theory? I know that different cell types discharge differently, but do not know how this affects things in practice.

6 alkaline D batteries (1.5v 12,000mha) in series to give 9v

6 alkaline C batteries (1.5v 8,000mha) in series to give 9v

6 lithium AA batteries (1.5v 1,200mha) in series to give 9v

6 alkaline PP3 batteries (9v 600mha) in parallel to give 9v

6 lithium PP3 batteries (9v 1,200mha) in parallel to give 9v

Any guidance would be appreciated.

James.

Without knowledge of the load that is a difficult question to answer. The more mAh the longer runtime normally.

Are you planning to use rechargeable or single use batteries?

If you plan to use rechargeables that are fixed into the clock then a battery management system is recommended if using li-on.

18650 3.7v cells are worth looking at too.

Edited By Brian Sweeting on 02/04/2018 15:23:46

More info about the current drawn by the solenoids would help - I guess you're not planning to use a starting solenoid out of a car, nor the titchy actuators used in model aircraft? Taking a cell or battery outside it's comfort zone will reduce it's lifetime - sometimes drastically. It's important to understand how much current will be drawn, how often, and for what period. For example, a solenoid ticking every second to move a light mechanism is very different from one striking hours on a large bell.

Generally the battery with the biggest maH figure will give you the longest life. But not if you exceed it's shelf-life or abuse it with too much or too little.

Steer clear of PP3 for power applications. Alkaline cells are often a good compromise not least because they're cheaper than lithium. Other types are available.

Dave

Solenoids are extremely inefficient. extremely. If you like the design to use a pull/push linear action use a tiny motor, gearbox and eccentric.
Arduinos seem to use little power compared to a laptop but it mounts up over 24 hours substantially. On some home control forums where people have analysed home control / security based on such a controller run off battery and solar they have been surprised at the size of solar array needed to keep it going. You need to drop the voltage closer to what it actually uses, 3.7v, and slow the clock, use sleep modes etc .Lots on the web about it.

Re a solenoid drive clock driven by batteries - solenoids in general use a lot of current the bigger they are- big ones may not be the best choice of mechanism for a battery driven clock, depending on what you plan to drive with the solenoids. Rechargeable LiPo with battery management , as used for RC equipment and phones , might work but these batteries usually do not last more than a few years.

There's a reason most common wall clocks these days are made with an Asian AA battery powered quartz crystal regulated circuit with a very compact brief-pulse solenoid escapement. These inexpensive clocks work for a year or more on one AA dry cell battery and last for several years, despite costing just a few dollars / pounds complete with face/dial etc.

Just food for thought, your mileage may vary.....

Important factors are the peak current and temperature range. If your clock is to operate indoors in a normal living environment then the batteries will be happiest, but if it is outside and using conventional alkaline batteries the peak current capability will be eroded at low temperatures (like freezing and below). This suggests that for an LR6 size alkaline cell effective capacity at freezing point is reduced to 50% relative to room temperature - this is a chemistry characteristic so other sizes will perform similarly. As Dave suggests, avoid PP3.

If you really want long life from a primary battery, especially at low temperatures, look at lithium thionyl chloride cells - these are the type used for example in remote reading meters where you have to operate a radio as well as potentially a gas valve, over a 10 - 15 year lifetime. They give 3.6V from a single cell and a D size can have a capacity of 17,000 mAh. Need careful handling, and I'm not sure series connection is advisable.

Coming back to current, often the peak current demand can determine the effective battery life - the electronics may work fine but you could find that the solenoids don't work because the battery voltage sags when you try to draw current. This can be managed (but it isn't straightforward) by using a supercapacitor to provide the current peak.

There is lots of information on battery performance on the web.

The amount of power extractable from any cell will be a maximum when the load is the same as the internal resistance of the cell.

This is one reason why alkaline cells are not the best choice of cell for a typical quartz wall clock.

If the current were miniscule the lower energy lithium ion cells may work better than alkaline cells (zinc-carbon cells are likely just as good as the alkaline cells forvlow power duty.

D cells are clearly going to beat C size cells of the same chemistry (shelf life not considered)

Clearly 6 AA cells will not compete with larger format 1.5V cells.

Regardless, you need to convert these silly mha units ( presumably you mean mAh?) to mWh initially. Younmight then be comparing ‘apples with apples’ and not ‘apples with oranges’

6 PP3 Lithium batteries are close to the same energy content as the 6 C cells and may well beat them, especially if the delivery voltage needs to remain high. Storage and operating conditions will also affect the energy recovery for any of the cells/batteries quoted. Manufacture’s claims are always under specific conditions and usually flatter their capacity compared to real life scenarios.

A lot more info required, I’m afraid, to give any sensible response. Even the battery specs might need consulting, too.

Late thought, but it may be best to provide two different battery types; one to power the Arduino (6V @ 50mA continuous), the other to drive the solenoids (9V @ ?mA intermittent).

Apart from letting you pick the best battery for each purpose, a dual power arrangement will protect the Arduino from the solenoids. When these are fired, you don't want the Arduino to brown-out due to a temporary voltage drop.

It might also be useful to de-stress the solenoid battery by using it to charge a large capacitor rather than connecting it directly to the solenoids. A capacitor is much better than a primary cell at delivering a short heavy pulse of current. Not a free lunch unfortunately - capacitors need to be big enough to hold the necessary charge and have sufficient time to recharge between each firing. Have you identified what sort of solenoids you intend to use yet?

Dave

Another problem where solenoids are involved is the spike of voltage you can get when you switch off the power. This is because the solenoid works by magnetism, and this is derived from electricity in a coil. Turn off the coil and the magnetism collapses, generating electricity in the now-turned-off coil, so up shoots the voltage.

There are several ways to reduce or avoid these effects, but you do need to consider what will work best and waste least, to get good battery life.

Cheers, Tim

PS you might try using a weight on a string instead of a solenoid, and rely on a tiny battery to do the accurate switching ...

Thank you for all of your replies, which I need to digest properly yet. A couple of quick replies, though.

Bazyle.

Someone else made this observation as well. I have considered a low-RPM motor like the one below. It claims to draw 30ma at 1.5v giving 25rpm. With a cam and microswitch, it would only need 1 revolution per operation.

Not yet, Dave. I was thinking of making my own, following the details in this clock (page 12), which is what set me off in this general direction in the first place. **LINK**

Kind regards,

James.

Those far more expert than I may wish to comment on the following.

These are purely comments, not arguing for, or against:

Alkaline primary cells seem to have an unfortunate habit of producing corrosion, particularly at the negative terminal. (The corrosion produced by a failed Zinc/Carbon primary cell is even more damaging, if left in situ)

I usually coat the terminals with petroleum jelly, (probably blocking a vent !) before installation to prevent this.

In a multi pack, series connected, I have known one cell to fail, or to go reverse polarity, without apparently affecting the others, except to drop the voltage appreciably.

Secondary cells (rechargeables), have a finite life, in terms of the number of charge/discharge cycles.

The sort of rechargeables in this context are likely to have a fully charged voltage of 1.2 rather than the 1.5 of primary cells, so for 9 volts, you have a choice of seven cells delivering 8.4 volts, or eight, delivering 9.6 volts.

Howard

Interesting link, he's made something work, and using wood with its friction too. Just shows I can overthink the complications. Even now I'm thinking his ratchet is wood-on-wood which has got to be high friction and he could have put some polished brass on the pawl working face to reduce friction.

You still have the problem of an arduino to power rather than a single IC clock counter.

I'm famously bad at maths so the usual health warning applies. Corrections and comments gratefully received!

The Lazy Clock has winding details for a solenoid. It uses 200 feet of 30AWG wire which the internet tells me will have a DC resistance of about 28Ω

With a 3V battery and ohms law I=V/R that's a maximum of about 100mA.

The Lazy Clock pulses once per minute, far slower than a pendulum clock. Assuming each pulse lasts a full second, that's:

24hrs x 60minutes x 1seconds = 1440 pulses per day

Assuming each pulse consumes 100mAseconds, that's 1440 x 100 = 144000 mAseconds per day

8000mA Hours of charge is 8000 x 60minutes x 60 seconds = 28,800,000 mA seconds.

So the C type batteries connected to give 3V might be expected to last 28800000 / 144000 = 200 days

Actually I'd expect much better than that because the coil is an inductor (tending to make the resistance considerably higher while the current is rising) . On the other hand, a pulse longer than 1 second might be needed to make the clock tick.

Challenge for friends of the Imperial system: how many turns will 200 feet of 30AWG wire make on a 13/32" diameter bobbin? The inductance of the solenoid can be calculated from the number of turns to give a more refined estimate.

Dave

James,

All other things being equal, then capaicity in mAh decides how long they will last.

If you put two batteries in series then their voltages add and total capacity in mAh is the same as a single cell.

If you put two batteries in parallel their capacities add and the voltage remains the same.

Cells in parallel tend to have higher self-discharge rates due to voltage imbalance (but not some chemistries that self-balance well like Lithium-ion).

So the answer is simple, the largest possible cells in series - the six d-cells.

using alkaline batteries the duty they have to perform is largely relevant as all formats will respond similarly, except PP3 which has a poorer performance at higher discharge rates.

The secret to long life with arduino is to sleep the processor as much as possible, and keep all outputs set low and high impedance as much as possible. Try an d keep pulses to the solenoids as short as is possible consistent with reliability. Consider lengthening the pulses as the voltage drops with flat batteries to extend useful life - arduino can measure battery voltage for you.

.

I've just received an eMail from ebay Seller: roberts9713

He has 'Super Capacitors' at what appear to be very reasonable prices.

MichaelG.

Would you be better off using an SLA battery , lithium ion or lipo pack to run the clock ? Rather than changing batteries when they go flat you could add a discrete charging socket to the clock and use a plug in charger to recharge the battery / batteries .

Don't try "matching the cell impedance" to get the most energy from the battery! If you match the impedance, half the available energy is dissipated in the internal resistance. Example: 2V cell, 1 ohm ESR, 1 AH capacity. Load with a 1 ohm resistor, 1 amp flows, 1 watt dissipated in the load for 1 hour = 1 watt hour. Compare with a 9 ohm load, 1/5 amp flows for 5 hours, dissipating 9/25 = 0.36 watts, which over 5 hours is 1.8 WH. The available energy from the cell is 2 WH (2 volts x 1 amp-hour). (This is, amongst other reasons, why my 1 kW workshop heater isn't "matched" to the mains!)

Solenoids are very inefficient as has been remarked. When you power them up the inductance is quite low so they draw a high current quickly, but once fully drawn in much less current is needed to keep it that way because the magnetic circuit has low reluctance. So the ideal way to drive them is with lots of current to start but reduce it as they pull in, only supplying enough to keep them drawn in. I've seen guidance on the web somewhere for this. You may also be able to recover some energy from them when they release, if you have a way to store it.

Supercaps might be useful, but they have limited working voltage, and need careful management - if you just charge them through a resistor they give little overall energy benefit because as much energy is dissipated in the charging resistor as available to the load.

There has to be a reason why the reciprocating electric motor, using solenoids instead of cylinders to drive cranks, never caught on!

Don’t compare maximum power with maximum usable energy. The two are not the same. I specifically mentioned power, which will be the maximum available from cells. That is why one can start a car engine with a lead acid accumulator. Probably not important in this application.

Without actually knowing the design details we can only guess, mind! Obviously not Big Ben sized, but it is the single reason why pp3 size might not be appropriate.

I second that with regards to sleep the pocessor. The other trick would be to use a higher voltage to pull the solonoid in and a lower voltage to maintain. You would obviously keep the pulse width as short as possible.

regards Martin

To add to n-d-I-y's comments, remember that energy is how much possible work you can get from a battery (etc) and power is how quickly you can get it. In short, energy is stored work, and power is the rate of doing work.

Cheers, Tim