Setting PA Speaker Amplifiers

[Vitaliy]

Hi, we just purchased a Behringer X32 (sweet mixer!) for our youth sanctuary in our church, and I decided to re-do the PA system while we were at it. I've always remembered the speakers being very noisy (hum, etc.) when nothing was playing through the system (everything was muted). Trying to adjust the PA levels, I'm now beginning to think that who ever chose the speakers/amplifiers specified something that was way to powerful for that room. If I crank the FOH amplifiers down to the minimum setting (~10%) I get about 90-95db SPL using a cheap meter that I have, when the pink noise is approaching the clipping point of the mixer. Now there is no noise coming out of the speakers when everything is muted, but the amplifiers are cranked ALL THE WAY down. The subs are also barely on, and I'm considering only using one of them now. Same goes for our monitors. This doesn't seem right, can someone let me know if I'm doing anything wrong?

Our setup:
- 2x JBL PRX518s subs
- 2x JBL MRX515 for FOH speakers
- 2x JBL MRX512m for front monitors
- 2x JBL side monitors (will update with model number, don't remember but they are matched correctly with the amps)
- Amps for the FOH Speakers: QSC PLX3002
- Amps for the monitors: QSC RMX1450

Our youth room fits about 90-120 people. I can provide the dimensions of the room if needed.

 

[cmckeeman]

the way i have been taught to set amp levels is play noise at ALMOST clipping level then set the amps so that it is ALMOST painfull, this will make it so that the sound will clip before hurting ears. i'm not sure how right this is but if the sound is at a nice level with the levels just slightly jumping into yellow then that is the best way to set it i have found, as to how to set the sub amp you will want to Eq the room with an RTA to set it with everything else

 

[DuckJordan]

Actually you want your pink noise to be hitting the signal meter pretty consistently and just under clipping. Your faders should be at 0db and you should put the gain on your amplifiers so its louder than you would normally run the system. Then you would have a little headroom for the songs that need a louder volume and have your gain comfortably set. You are actually under powering your rig by quite a bit. the MRX515s recommend 700watts for continuous use for 100hrs. Your amps are putting out 550watts so you are unlikely to over power them. You may hear hissing from unbalanced cables leading to the amplifier or from an input to the mixer. Or you may be on dirty power. I'd check those solutions before messing up your system far enough that an engineer needs to come out and fix them.

 

[Vitaliy]

That's pretty much how I was doing it. I had pink noise going out of the mixer so that the output was at the last spot before clipping. Then I used my meter to get 90-95dB SPL which was already too loud for our application. The amplifiers are far from clipping, because the input knobs are turned all the way down (turning it down another step essentially mutes the speakers).

As for the noise, all cables are balanced although I can't say much about our power condition. The noise did go away when I turned down the controls on the amps. I do remember that on the old mixer the full signal headroom on the output was never completely used. I'm guessing that is why you could hear noise. As for messing up the system, the system was in much worse condition prior to me. Everything was all over the place and it took me a long time just to get it back together. I don't think we've ever had an engineer come look at it.

I know that the speakers are being under-powered, but my question is: is the system too much for that particular room? This includes our speakers and monitors. Since I had to turn down the amps so much in order to get a normal sound level?

 

[DuckJordan]

probably, I'd have to see the room in person to make a more solid recommendation, What do you use these for? Is it just background music? is it services rock concerts? I'd just put your mixer levels down lower and put your amps just under full you can severely damage speakers by under powering them and sending the amps into clipping to get more volume. So get it as loud as you can without clipping and call it a day. (you may need to stay in the lower 1/3rd of your mixers main outputs to keep a decent level.

 

[FMEng]

A common mis-conception is that the gain control on the power amplifier somehow restricts the amount of power delivered to the speakers. That is for the most part untrue. The amplifier can deliver full power to the speaker with the gain set at 1%, provided that the console's output or the amp's input stage is not clipping.

Let's toss out some numbers from the spec sheets: With the gain wide open, your power amp reaches full power when 1 v RMS is applied to its input. The Behringer's maximum output is a 11 v RMS before it clips, and the amplifier's input stage can also handle 11 v RMS before it clips. (That they happen to be the same makes this a bit simpler.) Now some math:

20 log (1 v/11 v) = -21 dB So, with the amp gain wide open, the amplifier reaches full output when the console meter is showing -21 dB. The sound is clean because the console has headroom and we haven't quite overdriven the amp. (Since this is a digital console, we know that full scale on the meter is 0 dB. Analog console meters have a different scale, relative to a voltage.)

Now, let's turn the amp gain down to half way, which is 10 dB less gain. Now, the console has to drive a bit harder for the amp to reach full output. The console meter would now read -21 +10 = -11 dB when the amp reaches full output. The sound is still clean because both units still have headroom.

Now, let's turn the amp way down so that the gain is 30 dB less. The console meter would need to reach -21 + 30 = +9 dB, but our intrepid console cannot go above 0 dB. Now, the console is working full out, but it still can't drive the amplifier to full power. The console's meter is pegged, the console's output is clipping, and the amplifier's input is clipping. Yet, the amplifier's maximum output is 9 dB less than full, or 10% of it's capability, and we hear massive amounts of distortion because there are two stages in full clipping. It sounds bad, and we can't enjoy all of the power the system is capable of.

One theory is that the ideal setup is to have all stages prior to the power amp just below clipping when the amp's output clips. That would happen with the amp gain set 20 dB below full. Since max is 32 dB on its scale, the "right" setting is 32 -20 = 12 dB. With it set that way, the console meters will tell you when you are very near the limits, and this allows you to reach maximum sound level if you need it.

What does this boil down to? Turning the gain down does not impose any restriction upon the complete system, provided neither the console output or the amplifier input run out of headroom. It also means that cranking the amp gain doesn't get us a system that can get any louder. It just means the console's output level would be lower.

Of course, we've ignored signal to noise ratios. The OP did point up to the noise penalty to getting more sound output if we don't actually need it. And, adding more equipment in the chain, like EQ or crossovers makes the math more complex. But, those are discussions for other posts.

 

[museav]

Looking at maximum levels and clipping is important but don't forget to consider how you actually operate everything in normal use. Also consider that your pink noise source probably has a crest factor, or the ratio of the peak to average level, of around 6dB where music may have a crest factor of 10-20dB, thus the average signal level in normal use is likely further below the peak levels than with pink noise.

Looking the at electronics side and using the house system as an example, the analog output of the X32 is a +4dBu nominal and +21dBu maximum level. The maximum input level for the PLX3002 is +22dBu, which means it will not clip until after the X32 output does. However, the input sensitivity or the input level that derives full output for the PLX3002 is just +5.7dBu (with an 8 Ohm load). 21dBu minus 5.7dBu is 15.3 dBu and that means that the amplifer is actually providing its full rated output when the mixer output meters show around -15. Thus to have 0 on the mixer represent the full amplifier output you probably want at least 15dB of attenuation on the amplifier inputs.

That attenuation is to make the maximum mixer output correlate directly to the full amplifier output, however you may also want more headroom in the mixer and/or a lower nominal volume in the room with a reasonable nominal level at the mixer outputs, which might typically average around -17 to -21 on the mixer output meters. In that case you could run the mixer at a lower level or simply add additional attenuation at the amplifier inputs.

So set the amp for around 15dB of attenuation and play some music with average levels around or a bit above -20 on the mixer meters and see how that relates to the desired levels. If it is too loud, add more attenuation at the amp inputs.

Similar with your powered subs, they would apparently have to be set around +30 just to accommodate 0 on the mixer meters representing full output and perhaps lower if you don't need to full output or want additional headroom.

 

[Vitaliy]

Thanks for the information! It actually helped a lot to clear up setting PA gain structure, somewhere where I'm only beginning to venture into.

So, for our system, the "right" way would be to set the FOH amps to 12-15db? So that at the same moment when the mixer clips, so do the amps (their inputs) right? How would I need to set up the amps for our monitor wedges? My biggest problem is that the system is way too loud when set correctly, which is probably the opposite of the issues most people have. Clipping isn't an issue.

As for our subs, they don't have dB indications on their controls. It's a min/max knob, so I'm not sure how I would set the subs up. Currently I do it by ear (play music, and make the system sound as balanced out as possible). How would you do it the right way?

 

[museav]

So, for our system, the "right" way would be to set the FOH amps to 12-15db? So that at the same moment when the mixer clips, so do the amps (their inputs) right? How would I need to set up the amps for our monitor wedges? My biggest problem is that the system is way too loud when set correctly, which is probably the opposite of the issues most people have. Clipping isn't an issue.

You probably want at least 15dB or so of attenuation at the amps, but that is just to get 0 on the mixer meters to represent full output for the amp. But there is nothing wrong with adding additional attenuation at the amp inputs to reduce the resulting system level associated with different mixer output levels.

An 'old school' approach used to be to determine the amplifier input attenuation required to get reasonable gain structure and the desired output levels, then insert fixed pads of that attenuation before the amplifier, thus allowing the amp input gains to be turned all the way up. That allowed leaving the amps 'wide open' and not having to worry about someone turning them up.

As for our subs, they don't have dB indications on their controls. It's a min/max knob, so I'm not sure how I would set the subs up. Currently I do it by ear (play music, and make the system sound as balanced out as possible). How would you do it the right way?

You could look at the response on a RTA or FFT analyzer and making some assessments based on that information. You could also use a dual channel analyzer to view the related Transfer Function that might show whether options such as such as delaying the subs or mains are appropriate. That can be extremely useful information in optimizing a system, however in the end it is still what you hear that determines if the results are acceptable or not.

 

[gordonmcleod]

THX did a lot of work optimising noise floor in there systems One thing they did was to adjust the 50% mod level of the input devices for 300mv and then adjust the SPL in the theatre on the amp gain pots
They of course sized amps taking into account their input sensitivity and speakers based on their SPL rateing

 

[DuckJordan]

THX did a lot of work optimising noise floor in there systems One thing they did was to adjust the 50% mod level of the input devices for 300mv and then adjust the SPL in the theatre on the amp gain pots
They of course sized amps taking into account their input sensitivity and speakers based on their SPL rateing

Movie theaters and live theaters are two different breeds, the goal of a movie theater is to replicate audio from one crowing to another where lice audio is about the inteligabilty using live sources.

We have both in our space and it takes a whole different form of knowledge for either.

 

[museav]

Movie theaters and live theaters are two different breeds, the goal of a movie theater is to replicate audio from one crowing to another where lice audio is about the inteligabilty using live sources.

We have both in our space and it takes a whole different form of knowledge for either.

+1, the primary goal for cinemas and screening rooms and particularly with THX is usually to provide not only a specific listening experience but also as close to the same experience as possible in every room. They also don't have to worry about things like gain before feedback and are usually working with controlled acoustical ennvironments, espcially when addressing THX which includes standards for the associated acoustical environment.

 

[gordonmcleod]

+1, the primary goal for cinemas and screening rooms and particularly with THX is usually to provide not only a specific listening experience but also as close to the same experience as possible in every room. They also don't have to worry about things like gain before feedback and are usually working with controlled acoustical ennvironments, espcially when addressing THX which includes standards for the associated acoustical environment.

My point was in respect to the issue of system noise floor and that gain structure of "the building blocks " should not be taken lightly

 

[TimmyP1955]

The common wisdom is to set up the gain structure such that everything clips simultaneously. Unless the system has a LOT of headroom over what you'll need, I disagree with this methodology. Most "weekend warrior" systems will occasionally (if not often) be run hot enough that there will be a little amp clipping. This can slip by relatively unnoticed if it's not too much too often. However if something else in the system clips at the same time, it will sound BAD. You want headroom in everything ahead of the amps so that there is no chance of clipping anyplace else. If you are using system limiters, this this is especially important.

Most boards and processors can put out around +22 dBu. I recommend that the gain structure be set up such that the amps clip with the board and processors (if applicable) no hotter than say +12 (consult your owners manual to figure out where on the meters that is - some have 0 = 0 dBu, some don't). This will give you a 10dB margin between board/processor clipping and amp clipping. Then of course don't forget to set those DriveRack limiters. You'll be able to push the system with minimal sonic degredation.

 

[DuckJordan]

I disagree when your amps clip they start blowing drivers, when your board clips it sounds bad. Is rather let the moron complain about sound quality than hearing about blown drivers.

 

[Vitaliy]

Looking the at electronics side and using the house system as an example, the analog output of the X32 is a +4dBu nominal and +21dBu maximum level. The maximum input level for the PLX3002 is +22dBu, which means it will not clip until after the X32 output does. However, the input sensitivity or the input level that derives full output for the PLX3002 is just +5.7dBu (with an 8 Ohm load). 21dBu minus 5.7dBu is 15.3 dBu and that means that the amplifer is actually providing its full rated output when the mixer output meters show around -15. Thus to have 0 on the mixer represent the full amplifier output you probably want at least 15dB of attenuation on the amplifier inputs.

I just want to clarify, 15dB of attenuation means 32-15=17dB on the knobs of the amplifier, right?

Also, did you mean to say 4 Ohm instead of 8? It seems (correct me if I'm wrong) but at 8 Ohm the specs say that the input sensitivity is 1.7 Vrms = about 6.83 dBu (if the calculator I'm using is correct ).

This is the first time I'm looking into amplifier audio levels and dBu vs. dBV vs. Vrms, etc. so I'm a little lost.

On the amplifiers, the 32dB setting (all the way clockwise) means that 0dB of attenuation of the input signal is occurring, right? In order for me to "correctly" set this setting on any amplifier I would want to the mixer to clip at the point right above or at an amplifier's maximum output? If I was to increase attenuation I would never achieve maximum output of the amplifier because the mixer would clip before that, right?

Last week we used the X32 for the first time, and I set the FOH amplifier to the 12dB mark. During our worship portion of the service the mixer would have an average signal of -18. Volume levels seemed good, but I can say that we could probably decrease the input attenuation in the amplifier so that we could use the maximum power of the amplifier without clipping if needed but also keep volume levels at the level we desire.

 

[museav]

I just want to clarify, 15dB of attenuation means 32-15=17dB on the knobs of the amplifier, right?

Yes, although the controls are not actually varying the amplifier gain, which is fixed at 40X or 32dB, QSC apparently labeled the input attenuator controls as though they do vary the gain, thus the 32 value for wide open or 0 attenuation.

Also, did you mean to say 4 Ohm instead of 8? It seems (correct me if I'm wrong) but at 8 Ohm the specs say that the input sensitivity is 1.7 Vrms = about 6.83 dBu (if the calculator I'm using is correct ).

Good catch! I initially noted both the 8 Ohm and 4 Ohm senstivity values and when I edited that I apparently matched the wrong impedance and level.

On the amplifiers, the 32dB setting (all the way clockwise) means that 0dB of attenuation of the input signal is occurring, right?

Correct.

In order for me to "correctly" set this setting on any amplifier I would want to the mixer to clip at the point right above or at an amplifier's maximum output? If I was to increase attenuation I would never achieve maximum output of the amplifier because the mixer would clip before that, right?

"Correctly" is all about what you are trying to achieve and I think that is the cause for some of the differing responses. Some situations may be trying to get the absolute most they can from the system with experienced operators. Other situations may normally operate well below the full system capabilities with inexperienced operators. Yet other situations may have some different combination of goals and operators. That can affect how you might approach some of the gain structure.

Another potential element if you are basing the gain structure on clip inidctaors is what those clip indicators are actually showing, especially for amplifiers. Are they showing actual clipping or are they a warning that you are some certain level below clipping? For an amplifier, does the clip indicator reflect the input, the output, some certain level of distortion regardless of where it occurs or what?

Then there are possible details of the system. For example, if the device before the amplifier has variable analog level outputs then you may choose to attenuate the signal there rather than at the amplifier input, especially if that device is physically near the amplifier and those level controls may be less readily adjusted. For example, an often overlooked fact is that some of the dbx DriveRack products come set with certain maximum analog input and output levels but you can vary those via internal jumpers for the 260 or software for the 4800/4820 in order to set 0dBFS for the internal digital signal to represent different analog levels. Thus rather than applying as much, or any, attenuation at the amplifier input you could potentially set the corresponding maximum analog output of the processor to be lower. The critical point here is that if you are dealing with analog signals then the gain adjustment for digital devices must be to the analog levels after any D/A conversion rather than being digital attenuation before the D/A.

The overall point is that while there are some widely accepted general guidelines and processes for many aspects for system gain structure, there are also often potential application and product specific considerations.

 

[Vitaliy]

Hi, thanks to all for helping me out setting the amplifier. I decided to go with setting the FOH up in the way that the mixer and amplifiers both clip at the same point.

I do have one more question, though. For our monitor amplifier, the QSC RMX1450, only the input sensitivity at 8 ohms is stated. That would be fine, except for the fact that we have two JBL MRX512M monitors in parallel on each channel. Each speaker is rated at 8 ohms, but since we have two of them running in parallel we now have 4 ohms. I can't find information about the sensitivity at 4 ohms anywhere. How should I approach this?