You are basically there and have done a good job. I will fill in some details tomorrow.
Last edited:
Pano I must say I'm impressed by your level of support on a thread that is 10 years old!
One thing that surprised me when I first did the test was the -9db appearing in the dsp, for your -12db file.
It’s explained later in the thread about rms and peak values of the file but I still don't fully grasp the effect.
The 0dbFS files I played, did read at 0db in the dsp input, that’s why without knowing more I limited myself to max -10db output when measuring (with a noise cancelling headset on for protection).
But these power numbers are already good to take decisions, maybe refining them could change things a bit but I assume only by very few dbs.
When I tried to calculate with a more precise impedance, at each FR, it was too much guessing reading each driver official spec graph.
Maybe it’s 6.5 ohm, or 6.7 or 7, but the end result doesn't change much anyway.
Now if in real it goes near 5 ohms at this point, instead of the specified 8ohms, that’s a bigger impact on watts.
And then bandpass max power, is the max power delivered at the precise highest point of the response?
Like for example at 300Hz with a woofer band-passed between 200 & 400Hz (assuming no secondary EQ here).
I guess I could check myself with REW and play/measure every 10Hz
You have moved right in on an important point. Where is 0dB? It could be in two or more places, which leads to some confusion. I choose Full Scale as 0dB as it is an absolute point in most of the digital audio files we use. It's call 0dBFS and is a standard reference. But a sine wave can not have an RMS value of 0dBFS, can it? Otherwise it would not be a sine wave, it would just be full scale DC.
The highest RMS value a sine wave can have is 3 Decibles below Full Scale. That's expressed as -3dBFS.But another valid point where you can set 0dB - is the maximum sine wave value. So you will often see 0dB set at the maximum value of the sine wave. That's an RMS value of 3dB below Full Scale. That might be what you are seeing in your values. The peaks of the sine wave will be at 0dBFS, while the RMS value of the sine is -3dBS. Why did I choose the -12dBFS value for the test signals? (that's 9dB below the maximum level for a sine). I chose it because it's a level that is a little stronger then most music, but not loud enough to cause problems. It's also a handy value for figuring out what "power" amplifier you need.
Don't worry too much about impedance and power in this case. Half of the point of this long thread is to encourage people to stop thinking about power, and start thinking about voltage. That vast majority of amplifiers are voltage sources, so it's output voltage that is important. That's why this test measures voltage. Unfortunately we are stuck with the early 20th century way of measuring and thinking about amplifiers, with the results stated in watts. We often have to deal with it and do a little reverse math to figure out what the output voltage of an amplifier might be.
More in the next post.
Again, don't worry about power, think about voltage. In a multi-way system, things become more complex. What frequency do you choose to test each section (voix)? If your source is not linear, you need to be careful. Why would your source not be linear? Tone controls, EQ, DSP, crossovers. All of those things change the frequency response on purpose. That makes determining the maximum voltage needed a little more difficult.
For me, the most straight forward way of measuring the maximum voltage needed in each section of a multi-way system was to turn off all EQ DSP (but not crossovers) and use a tone placed in the center of the bandwidth. That will tell me the voltage output of the amplifier in that section High, Mid, Low, Sub, etc.
From there, I look at my DSP, EQ, etc to see how much boost I might have. Let's say I have a 10dB boost at 70 Hz, then I will know that anything around 70Hz will be at a voltage about 3.2X higher than what I measured. To have enough headroom for that boosted frequency to stay out of clipping the amp will need over 3X the voltage I measured/calculated.
There is another way to measure the maximum voltage needed, and that would be to leave in place all DSP and EQ a use slow sine sweep, while you look for the peak voltage coming from the amp. However, there is an inherent DANGER with this for tweeters and compression drivers. A -12dB signal might be much louder than any music signal they would ever get. I speak from experience frying tweeters with sine sweeps.
You need to be careful with loud signals in the high frequency range.
Last edited:
0.330 volts at 120Hz as measured with a Fluke 73 multimeter at my maximum expected volume level.
I have 8 ohm speakers with a 12" wide range guitar driver as the main. My room is 12'x20'x9' and partially open. I don't listen very loudly most of the time, but I could comfortably listen in the adjacent room at the test level.
I have 8 ohm speakers with a 12" wide range guitar driver as the main. My room is 12'x20'x9' and partially open. I don't listen very loudly most of the time, but I could comfortably listen in the adjacent room at the test level.
That was actually quite a bit louder than my normal listening level. My normal loudest listening level clocks in at 0.080 volts. My normal all day long listening level measures 0.051 volts. Pretty amazing isn't it?
Last edited:
It is an Eminence GA-SC64, rated 100.5 DB / watt and it is running unfiltered. I've been thinking of pairing it with a headphone amp...
Awesome,
Ok I got it, the tracks I found were probably defined like that, -3db RMS.
Agreed on the voltage, I only needed to convert back to watts to be able to compare other amps since that’s the most common spec. Very few list their max voltage.
Yes defining the right points per bandpass was bit troublesome, considering PEQS, crossover and house target effects combined.
So I will re-do with a method closer to yours.
Fortunately the last tune has very little boosts, and cuts only once all of these are summed, even if it’s an OB setup.
So I’ll keep the PEQS that linearise each driver response, but will deactivate the house target.
This will give me a baseline for any future tune changes, I will know what is the power requirement for flat.
Then I will also re-do it with house target ON.
This will give me an idea of the house target effect on power requirements to compare to flat.
And since I realised only now that REW can generate any frequency, I will use new ones for the woofers, at 100Hz and not 160Hz.
That should be about 5db higher, and might explain why the last measured voltage was so low compared to before with your 120Hz sine.
And ok I will also limit all these measurements to -20db as the multimeter is precise enough and the numbers for -30/-20/-10db matched so far > safer for the gears, and me.
¡Muchas gracias Pano!
@grosso de rien et de nada!
You've got it figured out, you should not have any problem, even though once you start putting on layers of EQ, it can get tricky.
You need to be aware of this when you are calculating your needed headroom and the voltages needed for you to play and not clip.
You've got it figured out, you should not have any problem, even though once you start putting on layers of EQ, it can get tricky.
Yes REW has a handy tone generator built in. Be Aware! If REW states a level of -12dB, that is NOT the same as the test tones I posted at the beginning of this thread. The REW tone will be 3dB lower than my tones. Why? Because the level shown by REW is for peak values. For example the setting shown in the screenshot below will generate a tone that has its peaks at -12dBFS. But the RMS value of the tone will be -15dBFS.
You need to be aware of this when you are calculating your needed headroom and the voltages needed for you to play and not clip.
For those of you who want to do your own calculations and decibels, here are two very handy calculators.
https://daycounter.com/Calculators/Decibels-Calculator.phtml
http://www.csgnetwork.com/ohmslaw2.html
Both of these can be saved has simple HTML on your local machine and will run offline. That way you will always have them.
Some basics to remember:
Also very important to remember is Peak value vs RMS. A sine wave has an average (RMS) value 3dB below its peak value. For example if a sine wave has peaks of -1dB, then its RMS value will be -4dB. Recorded music, however, is quite different - having peaks at 0dbFS but and RMS value (per track) that could be -24dB for dynamic classical recordings to -10dB for modern hyper-compressed pop and rock. To reach the same average levels, you'd need to turn up the volume of the classical recording by 14dB. That would mean classical music voltage peaks 5X higher than the compressed rock track.
OK, you probably would not want the classical average level that high, but you see my point.
https://daycounter.com/Calculators/Decibels-Calculator.phtml
http://www.csgnetwork.com/ohmslaw2.html
Both of these can be saved has simple HTML on your local machine and will run offline. That way you will always have them.
Some basics to remember:
- 6dB is 2X voltage 4X power.
- 9dB is 2.83X voltage, 8X power.
- 12db is 4X voltage, 16X power.
- -------------------------------------
- 1 volt into 8 ohms = 1/8 watt
- 1 volt into 4 ohms = 1/4 watt
- 10 volts into 8 ohms = 12.5 watts
- 10 volts into 4 ohms = 25 watts
- 28.3 volts into 8 ohms = 100 watts
Also very important to remember is Peak value vs RMS. A sine wave has an average (RMS) value 3dB below its peak value. For example if a sine wave has peaks of -1dB, then its RMS value will be -4dB. Recorded music, however, is quite different - having peaks at 0dbFS but and RMS value (per track) that could be -24dB for dynamic classical recordings to -10dB for modern hyper-compressed pop and rock. To reach the same average levels, you'd need to turn up the volume of the classical recording by 14dB. That would mean classical music voltage peaks 5X higher than the compressed rock track.
OK, you probably would not want the classical average level that high, but you see my point.
Hi Pano.
I have 3-ways speaker with tri-amp 3-ways active crossver. My room about 4x4x3.8m, measure by oscilloscope at 120hz 1.25Vrms on low amp.
My idea is to measure mid amp at around 1khz and high amp at around 7200hz. My cross-over cut at 400hz & 3500hz.
Do you think it's reasonable to do so?
I have 3-ways speaker with tri-amp 3-ways active crossver. My room about 4x4x3.8m, measure by oscilloscope at 120hz 1.25Vrms on low amp.
My idea is to measure mid amp at around 1khz and high amp at around 7200hz. My cross-over cut at 400hz & 3500hz.
Do you think it's reasonable to do so?