In Frequency Measurements Part 1, I described why you might need to include frequency as part of a complete data acquisition or data logger measurement. In this part I’ll explain how a data acquisition or logger system typically makes a frequency measurement and point out the pitfalls of each approach.
Analog Frequency Measurement
The measurement of frequency using analog techniques commonly uses an integration approach known as a frequency-to-voltage converter (FVC). A pulse stream (frequency) is applied to the device, and the output is a DC value that is proportional to the applied frequency. The FVC’s analog output makes it suitable for connecting to a wide range to data loggers, and the voltage can be easily scaled by the logger into meaningful engineering units, like RPM or gallon per minute. FVCs suffer some disadvantages as well. They tend to drop out below a threshold input frequency range making very low frequency measurements impossible. They are also designed and specified to operate over a specific frequency range because certain component values that operate well over one range won’t work very well over another. Over a 500 to 100,000 Hz range you may find six or more amplifiers, each one covering a small slice of this spectrum. Finally, response time may be impaired since low frequency response is a characteristic of the analog integrator that’s at the heart of this approach.
Digital Frequency Measurement
Digital frequency measurement techniques resolve all the problems of the analog approach, but not without creating some of their own in the process. The typical digital approach gates the pulse stream from which frequency is to be derived with a signal of a precise and known frequency. The gated result is then applied to a counter that updates after each gating cycle to determine the pulse stream’s frequency. For example, let’s say that our gating frequency is exactly 1 Hz. If the counter reads 1,000 after a gating cycle we’d be correct to surmise that the frequency of the pulse stream is 1 kHz.
The problem with this approach is that a very low pulse stream frequency requires a very low gating signal frequency, and so the update rate of the counter may be very long. An example may help. Assume that you have a flow meter that outputs 50 pulses per gallon. Assume also that your flow rate is 1/2 gallon per minute. This translates to an output frequency of 25 pulses per minute, or 0.42 Hz. If you apply a gating frequency of 1 Hz like before, you’ll get terrible resolution that’s virtually useless. You’ll need to decrease the frequency by at least a factor of 10 for 0.1 Hz resolution, and perhaps even lower depending upon the application. Since the counter cannot update until the gating signal completes one cycle, you get a usable reading of flow every 10 seconds or more. There must be a better way.
In fact, there is a better way called a ratio metric measurement that’s used by the DI-149 starter kit. I’ll discuss that in my third and final installment.