Hello,
I am attempting to design a power meter to be used with the output of a 100W per channel amplifier. I have seen a few rectifier circuits on line but none seem to do the trick. First, into 8 ohms I need to attenuate the signal from 28.3 V to 5.0 V to be fed to the Arduino analog input. Using a voltage divider circuit this is simple, but because the signal is effectively AC, the Arduino doesn't know what to make of this.
Second, rectifying with a simple diode (1N4148) doesn't seem to do well because of the voltage drop, and the fact that it is not a full-wave rectifier, although perhaps this doesn't matter.
In addition, I've seen circuits that use a 1.2V input from a headphone jack that somehow work fine with no rectification, so I don't really understand how this can be OK.
I've considered using an op amp to attenuate and rectify the signal based on the Texas Instruments LM3915 data sheet, but am wondering if there's an easier yet still effective way of doing this.
Thanks.
Do the rectification before the voltage divider. It will reduce the impact of the diode drop. In practice, the diode drop is not a significant problem. Please post links to the circuits you say don't "do the trick" ( and didn't say why).
I'd probably use a [u]Peak Detector[/u]. (There are full-wave versions).
I've made several sound activated effects (including a VU meter effect) and I always use a (half-wave) peak detector.
Second, rectifying with a simple diode (1N4148) doesn't seem to do well because of the voltage drop
If you put the diode before the voltage divider you can compensate for the ~0.7V drop out of 28V. (And 28V RMS is almost 40V peak.) But, if you also want to measure "normal everyday wattages" you might not want that 0.7V drop.
and the fact that it is not a full-wave rectifier, although perhaps this doesn't matter.
I'd say it doesn't matter since most audio waveforms are symmetrical. If you wanted to make a super-accurate peak-wattage meter you'd want the full waveform.
Even with a voltage divider it would be a good idea to add an [u]over-voltage protection circuit[/u].
In addition, I've seen circuits that use a 1.2V input from a headphone jack that somehow work fine with no rectification, so I don't really understand how this can be OK.
You bias the Arduino input at 2.5V and "isolate" the bias with a capacitor. That allows a 5V peak-to-peak audio waveform. There is a bias schematic attached to [u]this[/u]. (The associated code subtracts-out the bias and finds the positive or negative peak.)
I didn't use a peak detector because it's supposed to be a very-simple example.
P.S
If you have a bridge amplifier (no actual speaker ground) you'll have to design for that.
And be careful when testing with speakers connected. A "100W" speaker is designed to be used with an amplifier hitting 100W (or less) on the peaks while averaging 10-20W (or less), and that's if you trust the specs... You can fry it with continuous 100W test tones. The tweeter and midrange can be fried at even lower power!