Comments on Gains, Measuring Power, DMM, Oscope, and Clamp

[thechris]

==Forward==
In engineering, there is a need to know answers, but a bigger need to know if you have the right question. This document applies this reasoning to the DMM method.

==Physics==
In physics, at least regarding electrical circuits, there are two applicable laws. Ohm's law and Power law.

Ohms Law:
Current = Voltage / Resistance

Power Law:
Power = Voltage * Current

==Algebra==
Using algebra, one can come up with the formula:
Power = Voltage * Voltage / Resistance

and also:
Voltage = Square Root ( Power * Resistance)

==The DMM Method==
The amp manufacturer lists a power output for the amplifier. the Speaker manufacturer lists a nominal impedance* for the speaker, which we can use as the resistance in the formula.

So if we had a 1600 Watt (@1ohm) amplifier, and a 1ohm speaker, we would expect to measure a voltage of:
Voltage = Square Root (1600 * 1) = 40V

now all we need is some voltage measuring device. Such a device does exist, and is the Digital Multi Meter.

==Setting Gains with the DMM==
This is generally a bad idea. when people set up their audio systems they do so to maximize their happiness. and this has nothing to do with the formulas above. Lets take 2 examples of typical users.

==Hi-Fi user==
This user wants things to sound good. there is no reason he needs to turn the gains up to the maximum, undistorted levels given by the DMM. Now the speaker might be too loud.

==Street User==
This user wants to safely maximize the average SPL of the audio system, while limiting distortion to a reasonable level, maybe less then 25%. Depending on speakers, it may be better for this user to use higher gains then the DMM method calls for.

==Benchmarking==
For some reason, people really want to know the output power of the amp. The DMM method really doesn't set the strict requirements for testing this. It can be used, but without some control, it will be inaccurate (but really, who cares...)

==Power?==
Firstly, the DMM method assumed the amp manufacturer was telling the truth about amplifier power. amp power is specified at low levels of distortion. if you just crank the amp to the max, you might measure more power, but this isn't really a fair comparison.

basically, without knowing if the signal is undistorted you will not be able to compare apples to apples when looking at power ratings.

==Voltage?==
Is the DMM really accurate? true RMS meters should be more accurate. There are a slew of various derating factors and such for correcting DMMs, usually these stem from familiarity with some of the calculations but a lack of real understanding.

==Resistance?==
So a 1ohm resistance? thats funny, the DMM says its 0.7ohm. hmm. Should i use that rating? no. The speaker has a property known as impedance. impedance and resistance are rated in Ohms, but impedance changes with frequency. playing a 60hz tone may give a 12ohm impedance while a 200hz tone may be closer to 1ohm. So impedance isn't known accurately either.

==Ah==
The lack of stringent testing conditions -- using a resistive load of known value, measuring distortion, ect... are reasons an amp might be misrated by users of the DMM method.

==Useful==
Sure, the DMM is a useful tool. if you measure 40V on your trusty DMM, then you know the amp probably would put out 1600W or less, which is valueable to know if the speaker handels 1600W and you want to be safe.

==Oscilloscope==
This is the other new fad on the forums. The Oscilloscope is a nice and extremely useful tool that allows you to view the waveform being played. it also acts as a DMM. This allows you to deterimine if an amp is clipping.

==But?==
its expensive and really more suited to competitive SPL users. if your amp has an unregulated power supply, the maximum gain for clipping will change based upon the battery/ect... in which case you might have been better with just a DMM.

Also, most handhelds do not have a spectrum analyzer for determining distortion. the low distortion levels used for rating amps will not be visible on the 64 pixel screens, and waves with less then 1% distortion look very much like sine waves.

==Clamp Meters==
This is the other nifty tool. this is a big Ammeter. it measures current. this way you can find:
Power = Voltage * Current

or:
Resistance = Voltage / Current

==But==
again, this is a tool more suited for SPL competitors. End users really don't care about the speaker impedance -- they care about how loud and clean the speaker plays.

for benchmarking, this method will be likely to give a low reading -- since the impedance is different at the testing frequency, less current will flow. the amp could put out more power, but isn't being allowed to.

==Time?==
There is another issue that has been glazed over. Time. The wave is AC, so power is different at each instant in time. People use RMS values to get around this issue.

==Phase==
But sadly, the speaker is reactive. this means the current and voltage aren't in phase. they will both be sine waves, but not "in phase". for example, one might look like sin(t) and the other cos(t).

==So?==
if you do the math, it turns out that any reactive load will actually draw a negative power (return energy to the source) for a certain period of time. purely reactive loads draw no power. This is just another bump in the road. this bump is overly technical as well.

==Conclusion==
the DMM/Oscope/Clamp meter are all tools. But just because you see a formula doesn't mean you either understand it nor should apply it blindly to your application.

* The DMM should be used to determine the manufacturer's "maximum undistorted" gain for your audio system. this can be used as a reference to help you set gains based upon the more relevent factors like the way the speakers sound.

* The DMM+Clamp should be used to determine speaker impedance (magitude) for purposes of SPL competition. creative use of these two can also help tune a box.

* The Oscope can be used like a DMM and can be used to get a better "maximum gain" for your amps, but it isn't perfect, and using amps with a regulated PSU will give more accurate results. (note, i am not telling you to buy only amps with regulated PSUs. just that this method works better with them)

==Epilog==
I may include more info here later. for instance, there are ways to use a good DMM to determine clipping. this thread might be a good sticky as well.

* for the DMM method, the prefered test signal is a sine wave. this is because a sine wave is an ideal signal for testing. it has a known peak to rms and peak to average ratio. further, its derivitive and integral are also sinusoidal which means filters will not change these ratios.

[gotslap?]

well put...im tired of seeing so many people say this amp puts out x amount of power because they saw it bench tested...a clamp meter and DMM used at the exact same time will tell you everything...voltage, impedence, and power

[Saynkuck123]

^^^ x2

[profundus-sanus]

Why isn't this a sticky yet?!?!

Also you should put in how to measure clipping with a dmm

[thechris]

sure, the most basic ways (with varying degrees of accuracy) involve either:
1.) creation of bandstop filter (see "twin-t" network).
2.) multiple tests.

in a linear, undistorted system, bumping up the volume setting by 1 notch should give you 1 or 2 dB increase of signals.

so if you were to test the gain at 10 locations, and at each location, try two different volume settings, you would expect to see the ratio of volume 26 vs volume 25 stay constant.

as you increase the gain to moderate/severe clipping, the percentage increase from volume setting 26 vs volume setting 25 will be lower.

this is a basic graphical means to determine clipping.


beyond this, you would set up a bandstop filter of some sort. a twin T gives something like 9dB of attenuation at the 3rd harmonic, which is already at -9dB at maximum clipping levels. but attenuations of the fundamental of 40-80dB are possible, so there is a decent amount of contrast.

In this scheme, a tone is played and the external filter largely blocks it out. as distortion begins, harmonics are generated, and these are not filtered strongly, and can be measured.

the downsides are:
1.) need to build filter
2.) need to choose a tone that works well with filter
3.) small signals are measured.
4.) some of the main tone will bleed through, and the gain knob isn't linear, meaning its best to set the gain, ramp through volume settings (which have known gain factors), then adjust the gain and try again, until the desired volume setting is the point of clipping.


there are other methods as well. for instance, if you disable filtering and play a square wave, the output will not be able to increase much once clipping sets in. (before the square wave could increase in amplitude. once clipping sets in, it cannot increase). clearly there is a concern with testing with full power square waves when a speaker is attached!

there are also waves that are asymetric that will add a DC offset when clipped, but normally have no DC component.

methods that test with wave shapes other then sinusoids should not be filtered! and as always, your mileage may vary.

[tymoto21]

you can delete my post if im totaly wrong on this but looking at the ac chart in my class the formula is (v^2*cos theta)/r. because amps are ac wouldnt that affect power?

[thechris]

i don't frequent this forum as much.

not sure what theta is in this case. I suspect it is the phase difference of voltage and current, in which case the cosine would represent "power factor"

"""
==Phase==
But sadly, the speaker is reactive. this means the current and voltage aren't in phase. they will both be sine waves, but not "in phase". for example, one might look like sin(t) and the other cos(t).
"""

and thats what I mean in this passage. the voltage and current (peaks) might not occur and the same time. as a result the average power might be less then the simple multiplication of RMS voltage and RMS current. eg, measuring the voltage will give some value, say 1V rms. and current would give 1A, rms. this would seem to imply 1W on average. but it could be possible that the average is 0. this happens if the energy isn't used, but rather is just transfered to some energy storage mechanism (moving mass, springs, ect...) only to be returned later.

[Fallen]

I often wonder if it matters so much what an amp puts out into a reactive load vs what is can put out into a resistive load. It's easy to say it'll put out x amount of power into y sub at z frequency in w box. There is just too many variables, the box effects the impedance of the driver more than anything else.

IMO what is important is that is can source the voltage and supply the current without any voltage drops. I think the only fair way to say what an amp puts out would be a resistive load. Often times when I design a box the impedance around tuning is over 20 ohms! This would be a huge drop in power, but it is also where the enclosure is most efficient, it'll play with the same output(or even if it was -3dB, that only half power, but the resistance didn't double, it went up by a factor of 20).

I just don't think using a sub in an enclosure its a fair measure of output power. It might be useful to say how much power is going to the sub. But because the subs impedance varies greatly, whilst output remains constant, the power into the sub will vary too even with the same voltage in.

At least thats my take on it.