Let's Make Robots!

step-up or joule thief or.. ?

I`m trying to convert the wireless doorbell outside from battery to a renewable source such as solar or wind. My problem lies in that the battery required is a 9v.

My first thought was to produce a high enough voltage to charge the 9v but that might require more solar panels than I have. The 2nd idea was to use a lower voltage battery or super-cap as replacement and stepping up the voltage to 9v. I haven`t taken any measurements of current usage but I don`t imagine it could be more than a couple hundred mA while pushing the button and much lower in idle.

The joule thief circuit is very simple and can make 3v out of 0.5. Would simply adding more turns to the coil make a higher voltage? Or should I look at a capacitor charge circuit instead? I haven`t given much useful information but maybe someone who has had more experience with solar power or any kind of simple voltage boosters could give some advice on where to start looking?

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

The 'Joule Thief' is horribly, horribly innefficient. Due to its design it also run regardless of output current draw, so if you're using it to step up battery voltage the batteries will constantly be loaded even if the output circuit is off. They're also very noisy.

My first recommendation would be to check out the doorbell circuit itself - does it actually need 9V to function well? Try replacing the 9V battery with various lower voltages, and see if/how performance is affected. There may even by a voltage regulator on board the circuit to step down the input voltage, which you can bypass, replace or remove to allow the circuit to run on lower voltages.

Once you've determined the optimal input voltage for the doorbell circuit (best compromise between performance and low voltage), there are 3 main ways to feed the device as you've already identified: direct charging & power, inductive step up converter, capacitive step up converter.
If the solar cells/wind generator are easily able to produce the required voltage and current to charge the batteries, and the batteries are likewise able to power the circuit, then there's nothing wrong with the direct approach. A few blocking diodes will ensure that the power from the batteries isn't dumped back into the solar cells/generator, but we can get into circuit design later. You can also use a combination direct/step up design that uses a booster at low cupply voltages, but switches to direct when the supply is high enough.
Inductive step up converters can be made to self-start at quite low voltages, and they're typically very compact. They're a little noisier and less efficient than the capacitive types, but they can convert quite large amounts of power if designed correctly (i.e. good current capability). I've built a number of small 2-3 transistor boost converters which will start cranking out 5-9V with an input as low as 0.56V, although they're not very efficient until the voltage rises further. With a good selection of parts you can make a small, clean and efficient converter, especially if you use a dedicated controller IC like one of the many Maxim boost converter chips.
Capacitive step up converters are quiet, light, clean and very efficient. The downside is that they can be quite large, have fairly limited current output, and require an oscillator to drive them. If you increase the speed of the driving oscillator you can reduce the size of your capacitors and still get the same output, but the oscillator itself will draw more power at higher frequencies. The oscillator circuit also typically operates at a higher starting voltage than for an inductive step up converter, so you'll need a higher supply voltage before anything is produced at the output. I'm not sure if there are any good capacitive step up converter ICs out there, but it's worth a look.

The next step really is to characterise the circuit you need to power - minimum input voltage, maximum current draw, 'resting' current draw, etc. Naturally you'll also need to figure out the specifics of your power supply device in a similar way, whether you use solar or wind.

Hi TF, thanks for the response. Lots of good information there!

I didn`t have a 9v handy to test the transmitter so I stacked up some AAs instead. It appears it works as low as 4v but I need 7-8v to get sufficient range. In idle there is no current draw and only 4mA when holding the button down.

With such a low current draw I`m wondering if I can`t just use a 1F 5.5v super cap charged up with a solar panel and boosted with a charge pump. I found this schematic here for an easy charge pump. If it can turn 3v into 5.4 I guess it could make 8v out of 5. I think the only parts I don`t have are the special diodes.

Do charge pumps suffer from the same problem as the joule thief circuit in that current gets drained even when not in use? I might have to check the button wiring and put it in between the cap and charge pump in that case.

I've actually built a bigger but very similar version of that circuit before; you can use just about any diode but the germanium diodes are extra low voltage so they make for better efficiency. You can keep adding capacitor 'stages' until you get the voltage you want - the one I built was spitting out over 2kV (multimeter doesn't register any higher than that) from a 3V supply =D
Make sure you get the 74HC14 and not the 74HCT14, the 'HCT' version requires a higher operating voltage to work.

Super caps are great for solar powered applications, but their higher self-discharge rates mean that you'll probably need to collect enough sun every day for continuous operation, whereas a rechargeable battery will be able to hold onto its charge for a few days even with no sun at all. It really depends on how good your solar cells are at generating power in bad weather.

The charge pump will keep running even when the output isn't being used, but it won't draw as much current.
If you want to be really clever you can use one of the 6 inverter channels on your 74HC14 (if you use a circuit like the one you posted) to make a basic voltage detector. When the detector senses that the charge pump has 'filled' the output capacitor it automatically turns off the charge pump circuit. When the voltage drops again (as the transmitter circuit uses up the stored power) the detector will 'unlock' the charge pump and it'll start charging the output capacitor up again.

I had some 74HC14 to use to clean up encoder signals so I used 1 and made a 3 stage charge pump. I tried the super cap at first but you are right, they lose charge very fast and I couldn`t see it lasting even the night let alone a couple of rainy days. With 3 stages though I get a nice 8v out of 2 AAA rechargables.

It only draws a few mA while idle but I like the idea of using an inverter to turn itself on and off. Is it a simple case of feeding the final voltage back into an input and then the output somewhere into the RC oscillator?

Yep, a diode and a few resistors is all you need to correctly bias the input for the 'trigger' voltage you want, and then the output feeds back into the oscillator via another diode to make it stall whenever the trigger voltage is exceeded. I can sketch you up a simple trigger circuit which I used recently for a similar purpose if you'd like.

I ended up not bothering with adding the trigger to stop the charging. The doorbell was refitted this afternoon and works fine so far with the solar cells perched on the roof. I will have to come up with a more robust way of fixing them on there or I can see them getting blown away in a typhoon.

The weather is supposed to be cloudy and rainy for the next couple of days so I am looking forward to seeing how the batteries hold up and if the water washes the cells off the roof.

Thanks for all your help TF! A few pictures for your trouble :)



Lookin good ezekiel! Did you put those solar cells through a laminator or something?
The "standard" joule thief may be inefficient, however there are other switching circuits that can be easily made. The kit at www.thejoulethief.com uses a kind of switching transistor that handles the pulsing of the coil to drive the LED