Does That Generator Really Work (and how do you know?)!?!?!?!

With my experience working for small scale alternative energy companies like Tremont Electric (Kinetic Energy Harvesting) and Voltaic Systems (Solar Generation), I always get two questions when walking around with prototypes:

1. Do those things actually work?

2. How do you know how well they work?

The first question is easy.  Yes, they work.  However since they are small scale energy producers, a customer needs to have the right expectations.  If you buy a product with a solar panel the size of your palm to exclusively power your heavily-used 4G smartphone, you will find yourself with a large bill and an even larger sad face.  However if you need a small power source on a camping trip where plugging in isn't an option, then you've got a product worth checking out.

The second question is more interesting.  A product might specify the power output of the energy source, such as “Our 5x10 panel is capable of generating 5W.”  However this not what you will see going to your target device.  In most small-scale energy producing devices like these, there isn't enough power to reliably charge a device directly from the energy source.  Instead, the energy goes into a battery that charges over time and can deliver a stable supply to the phone.  For a generic case, see the block diagram below:

As with any power transfer system, each step costs efficiency.  Next we’ll examine efficiency losses step by step.:

Energy source:

We will assume for the purposes of this article  that this is perfectly efficient.  I don't think it's fair to ding the manufacturer because solar panels struggle to produce 19% efficiency.

Input Power Conditioning:

This is the step that takes the power from the energy source and converts it into something acceptable to charge the battery that is used.  There is a lot of variation here depending on the type of energy generation, the size of the source, output impedance, etc...  But I'll throw out an efficiency of 85% to have a number to work with.

Battery:

This is the battery that is designed in the device and slowly charges whenever energy is present.  On top of internal resistance losses, batteries also have coulombic efficiency losses.  It is reasonable to expect an overall efficiency of 80% for this stage.

Output Power Conditioning:

Some switchers are specifically designed to take Li-Ion voltages and create a 5V USB standard output.  I'm sure you can find about 10 datasheets of parts from TI, Linear Tech, etc... that will quote exact efficiencies, but it is safe to assign a 75% efficiency over the entire voltage rage of a Li-Ion battery (2.9-4.2V).

Sum it all up, you're looking at a total efficiency of about 50%.  Ouch.  So your 5W panel can be reasonably expected to give you 2.5W to your target device.  But all of those numbers are just thrown together for the general case.  How can you, the user, KNOW what you're getting in terms of efficiencies?  Measure it in your application!

The best way to to measure any efficiency is to look at power in vs. power out.  In this case, current and voltage can be measured from the energy source to the electronics when charging.  Then the output current and voltage can be measured to your cell phone.  Once you've got the test setup prepared, you just need to log the data and save it for later review.  If only someone would make an easy, well-explained video on how to log a couple of channels with a lot memory.  Thanks Jeremy!

Anyone want to try it on a device of theirs?  I would love to see some results from real use cases in the comments!  User data is always the best!