Welcome to the Blog
Welcome to the EnergyZarr blog… as you can see, my surname is Zarr which very often is confused with Czar or Tsar (Russian) or Zar (German) which translates to “autocratic ruler” – cough… I assure you, the only thing I rule is about 14 square feet of my garage. However, I did think it was a catchy title, so here we are.
I’m an engineer by education as well as by nature and love to discuss topics related to improving our way of life through technology. This blog is dedicated to discussions on energy efficiency and other related topics such as energy harvesting, lower-power technologies, LED lighting, and so forth. I will be making weekly updates (maybe more depending on traffic) and look forward to feedback from readers.
This inaugural blog entry examines units of measure for efficiency in analog semiconductor components. With all the emphasis on making energy-consuming devices more efficient (cars, refrigerators, etc), it makes sense to look at the components that make up these systems. The issue in the past has been that a specification such as power consumption doesn’t tell the entire story.
For example, imagine two chain saws that both advertise they have a 2-horsepower engine with a 40-minute run time (same size tank). However, one saw has vanadium-steel chain blades with a ceramic coating that lowers the friction as you cut through a tree. It also has a revolutionary chain design that more efficiently removes wood as it cuts, making it much faster. The other saw has standard steel blades. If you simply compared the power consumption of each saw, you might miss the fact that the “high-tech” saw will actually do more work with the same fuel since it will cut faster. The engine will use less fuel per cut. So, a better measure of a chainsaw’s efficiency might be cuts per liter of fuel where a “cut” is standardized to mean a 6” diameter branch (or something similar).
In electronics we see the same thing. A good example is analog-to-digital converters (ADCs). An engineer designing equipment with many channels of data conversion will often look at the ADC power specifications along with the speed and resolution. The problem is that not all data converters are equal. There are “marketing bits” and there are “real bits,” also known as the Effective Number of Bits (ENOB). Like our chain saw analogy, the entire story may not be told with simply comparing the 12-bit, 50 MSPS ADC’s power ratings.
A better way to compare ADCs is to consider how much error the converter contributes to a given power consumption. This tells you the real story of how good the underlying technology is for efficiently converting AC analog signals to bits. See equation 1 below – this is a metric that uses the real bits (ENOB) which is a function of both the signal to noise and distortion (SINAD) contributed by the converter. It also normalizes the number of channels and the frequency at which the converter is running. The result is energy (in Joules) per conversion. This is exactly how much energy it costs the system every time a conversion is made. Now comparing various converters becomes much clearer and even allows dissimilar technologies to be compared (e.g. Flash and folding ADC architectures).
There are many more metrics for other technologies such as serial cable drivers / receivers, power supplies and more. Check out my white paper on National’s PowerWise family where I talk more in detail about this. Also, tell me what you think about using metrics such as these to design lower-power systems. I’d love to hear from you.
Till next time…