I've looked around the site for suggestions, but think this question is a unique one and would love some conversation.

I need a method to measure the average power in to a transducer.

This method must

  1. Be able to measure the power of complex input waveforms (audio input, in this case)
  2. Be able to measure the power of a signal oscillating around a DC voltage
  3. Be able to measure the power of a signal oscillating around 0 V (AC)
  4. Be able to measure the power of a signal which oscillates around some DC value, and not fail when the oscillation is of such a magnitude that portions of the signal fall below 0 V

The general parameters of the waveforms I'm measuring are

  • Audio, with and without DC bias
  • Chirps, with and without DC bias
  • Single frequencies, with and without DC bias

Before you ask, the DC bias is absolutely necessary! The physics of the devices for which I must determine the average power input require that I test signals with and without a bias.

In sum, I need a method of measurement which can accept voltage and current frequencies from 0-25kHz, is minimally invasive, is accurate to within a hundredth of a Watt (tenth might be okay too), and can handle voltages both with and without a DC bias that have peak to peak values from 0.1 V - 50 V.

An idea I've been given is that I could use transformers to remove the bias, measure the bias and the alternating component, and add them to obtain the average power.

Thanks in advance for any advice you can give!

  • \$\begingroup\$ Something like this plus some software (and a PC) could do the job. Can't find a "better" one right now, but there are some which are even cheaper and might still fit; your 25kHz requirement allows for the most simple of those devices. - 0.01W however may be a problem. \$\endgroup\$ – JimmyB Apr 14 '16 at 15:27
  • \$\begingroup\$ 16MHz sampling rate! That seems very impressive. This does look like a possibility; I might have to condition the inputs still, however. The ideal solution seems to be a DAQ which collects at 50kHz (at least, accurate amplitude representation), is fairly cheap, and has a large input voltage range. \$\endgroup\$ – xcnmoore Apr 14 '16 at 17:32
  • \$\begingroup\$ Please avoid using the expression "complex input waveforms", because "complex" has a specific meaning in signal processing, and no waveform you can measure will be "complex". :) \$\endgroup\$ – pipe Apr 14 '16 at 17:39
  • \$\begingroup\$ My bad, complicated! \$\endgroup\$ – xcnmoore Apr 14 '16 at 17:41

I think that by sampling the signal and integrating by SW you can achieve the results you want. Here comes the questions: What ADC resolution and what sampling rate. What quality of filter will you provide to concentrate on your required BW. Audio filters can be complicated because of their very narrow BW. The waveform type (with, without DC, etc.) is no issue for an integrating algorithm. It is only and issue for the analog front-end that has to condition the signal to the ADC range. Another limitations I see are the achievable accuracy, how to calibrate your HW/SW, and how to adapt the instrument to the range you specified, be it via mechanical/electronic attenuators and with or without auto-ranging.

Each option you want to add will be another complication of the SW. So, where to draw the line to keep it simple? Not easy to say.

  • \$\begingroup\$ I guess that I'd prefer to use a cheap ADC of some sort, like Arduino. I think they are usually 10 bit with 0-5V input range. The sampling rate is about 10kHz on Arduino, which is fine for my purposes; I'm most interested in an average power. Also, I didn't make this clear, but I'm interested in power over all frequencies, so I don't think there is any reason to isolate any! Or would 10kHz be a problem with not being able to capture those higher frequencies? \$\endgroup\$ – xcnmoore Apr 14 '16 at 17:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.