Limiter Setting: Pink Vs. Sine

[0x15e]

I plan to spend some time this weekend setting up my DR260's limiter to protect my freshly repaired subs.

It's been recommended that I run a 60 or 80Hz sine wave through it, then measure output voltage from the amp. I have a couple of questions about that.

First, is there any advantage to using pink noise instead of sine waves? I have a TrueRMS multimeter so measuring it wouldn't be a problem.

Second, what level should I use for the test signal? 0db peaks or what?

[SoundInMotionDJ]

I use sine waves. I have 1000Hz, and 60Hz.

1000Hz is something of a "standard" used to get the power ratings for amps. It is also good for setting the limiters for tops.
60Hz is handy for setting the limiter for subs. The rule of thumb I go by is the test tone should be 1/3 octave away from any crossover frequencies.

When setting the limiter, I run the signal from the source -> mixer -> effects -> limiter as large as I think it will ever be during the performance. Then I engage the limiter and measure the resulting output from the amps. I adjust the limiter until the amps are producing the desired "limited" signal.

--Stan Graves

[wallywally]

Using a sine wave is the safest way. If you set the limiter using pink noise and then run a sine wave the voltage will be higher. Musical instruments will give you a voltage somewhere between sine and pink depending on the type of instrument.

[BrentEvans]

Using a sine wave is the safest way. If you set the limiter using pink noise and then run a sine wave the voltage will be higher. Musical instruments will give you a voltage somewhere between sine and pink depending on the type of instrument.

To expound on this a bit...

Music program and pink noise are both variable-state electrical signals. You said you had a True RMS multimeter, which is good, but even so, it needs a steady-state electrical signal to read, it can simply do so at any frequency. Sine waves provide just that.. and therefore represent the true limit points you want to achieve.

[Bill Fitzmaurice]

it needs a steady-state electrical signal to read, it can simply do so at any frequency. Sine waves provide just that.. and therefore represent the true limit points you want to achieve.

If you measure a pink noise source the voltage won't be constant, it will continuously flutter about. To a great extent that flutter will be stabilized when reading a limited signal, as the limiter reduces the signal peaks. But it's still a lot easier to get a good reading with a sine wave source. Technically the pink noise is closer to actual program material, so the reading is more 'accurate', but if the reading keeps jumping about to the extent that you can't tell at what voltage it's actually clamped then you're just spinning your wheels.

[0x15e]

Thanks for the info. It's been a lot of help in understanding what's going on.

I did some experimenting with it last night and I think I'm starting to figure it out. The meter didn't really have a problem with pink but it was definitely more responsive to the 60Hz sine wave (and easier to measure).

The program material is DJs playing drum & bass and dubstep, which half the time has extended long waves at low frequencies, so I think the sine will get the job done just fine.

[Captain D]

I checked the accuracy of my Fluke 73 multimeter in Volts AC mode using sine waves and found the following; its accurate at lower frequencies (20 to 640 Hz) but gets steadily more inaccurate as frequency increases.

Sine waves were generated using Realtime Analyzer (http://www.ymec.com , a very useful program!)- fed through mixer to the power amp (bypassing the DRPA). Final voltage for signal was set at 4.00 volts (Peak) and measured via a calibrated Tektronix TAS220 oscilloscope. The scope and the Fluke 73 were connected to the signal simultaneously. Signal voltage on the scope was checked each time the frequency was changed but it showed no noticeable change over the range tested.


Actual Peak Voltage-------- square root of 2-------- Actual RMS voltage
4.00 1.4142 2.828

Tested Points
Frequency (Hz)-------- RMS Volts observed-------- % of Actual

20-------- 2.773-------- 98.0
40-------- 2.828-------- 100.0
80-------- 2.839-------- 100.4
160-------- 2.840-------- 100.4
320-------- 2.838-------- 100.3
640-------- 2.822-------- 99.8
1280-------- 2.756-------- 97.4
2560-------- 2.531-------- 89.5
3000-------- 2.434-------- 86.1
4000-------- 2.199-------- 77.7
5120-------- 1.947-------- 68.8
6000-------- 1.767-------- 62.5
7000-------- 1.583-------- 56.0
8000-------- 1.421-------- 50.2
9000-------- 1.276-------- 45.1
10240-------- 1.122-------- 39.7

(The data fit a 4th order polynomial pretty well - from 320 Hz and above, anyway)

Results:

At 20 Hz the meter dips a bit in accuracy, but 98% of actual isn't too far off. For setting subs using any freq. in the range 40 to 160 would be fine....but it seems prudent to pick a frequency a good bit away from your system's high and low pass settings to avoid any attenuation from the filters. Stan Graves' rule of limiting using frequencies 1/3 octave away from any crossover frequencies seems a great place to start.

For setting limits on tops any freq. in the range 80 to 640 Hz should keep your limit setting accurate, again depending on where your high-pass filter is set. For my meter 1000Hz would show rms voltage about 98.6% -- pretty good for government work.

These are just the results of an older Fluke 73, a non-True RMS meter. YMMV depending on the meter you've got. I imagine a True- RMS meter would fair much better but I don't have one to test.

The upshot of all this is that the error could have a serious impact when setting limiters. For example, if you set an output limit of 30Vac using a 8000 Hz test signal, your system would hit 60Vac when it dished real music with substantial content in the range 125 - 640 Hz, blowing just about any driver we deal with here. An unlikely, near-worst case scenario for sure....but you get the point.

The data would also indicate that its not good to use pink noise (or white noise for that matter) when setting limiters for power amp output. (Don't know if anyone does that - but its been discussed.) Error in the higher frequency ranges could likely cause pink noise to show a much lower RMS reading than actual.

Again, YMMV, but I thought some might find this interesting.


Cheers,

Captain D

[0x15e]

Awesome. More info / data is always appreciated.

I ended up testing with 60Hz for lows and 1000Hz for highs. I'm far less worried about the highs since the amp is more or less matched to them in the first place and I don't have to go anywhere near clipping on the amp for them to be loud enough. The subs performed really well at the last party using the new settings.

[Dan30]

Sorry to bump an old thread.

I ran two tests. The first was a 100hz sine wave through my RMX2450 which topped out at 69 volts any higher and the clip lights started fluttering. When I ran the pink noise test I could only get to about 18-20 volts before the power section started clipping. Is this normal?

[0x15e]

That's normal. I've found that a sine wave will max out the amp far quicker than pink noise or program material.

My last go at setting this all up was done using a combination of the three, but mostly program material. I set the limiters on the tops using pink but then used several different tracks of program material (some house, dubstep, and drum & bass) to really get things tuned for the subs. This has been the best setup yet.

[el_ingeniero]

That's normal. I've found that a sine wave will max out the amp far quicker than pink noise or program material.

My last go at setting this all up was done using a combination of the three, but mostly program material. I set the limiters on the tops using pink but then used several different tracks of program material (some house, dubstep, and drum & bass) to really get things tuned for the subs. This has been the best setup yet.

Hmm, that goes against the conclusions the poster above came to:

1) don't use pink noise for setting limiters.
2) instead use a sine wave at least 1/3 of an octave away from any crossover frequency.

So if you have a low pass at 40 Hz and a high pass of 100 Hz on your subs, 60 or 65Hz is a good place. For tops, if your highpass is set at 120 Hz then somewhere in the range 180 to 1000 Hz is fine.

Pretty much that's the instructions in the plans for what to use as a signal, if I recall correctly.