Article: How to choose a power amp

[bongomania]

Originally Posted by cosmonautsix

Pros vs. cons of using a pedal pre-amp (tone hammer or BDDI) and external power amp vs. head/cab or combo? I'm seriously thinking of just picking up a poweramp off CL to drive a 4x10 cab...

Check out the article on preamps in my FAQ. It covers what you need to know.

[Bob Lee (QSC)]

Originally Posted by tekdiver500ft

Sort of. In order to get the best sound quality from a PA or mixer, you want everything prior to the power stage to be at unity gain.

Not necessarily. For example, setting up a crossover or loudspeaker processor often results in different gain settings among the various frequency bands.

So, what I do is turn the input attenuators all the way down, set all channel gain to the appropriate level (as hot as possible with no clipping/feedback), then run all channel strips, subgroups, and the master up to unity. At this point, I turn the attenuators up until I am just slightly louder than my max for the evening. This gives me the best sound the rig can provide, IME. If running a pre/power combo just for bass, then I leave the attenuators wide open.

What piece of gear has input attenuators?

[jnewmark]

Originally Posted by bongomania

That's the speaker impedance, not the output impedance of the device going into the amp.

Now, it is true that whatever the output impedance of the ipod IS, it will affect how well the amp is driven. Voltage isn't the only factor. But all you need is for the input impedance of the power amp to be at least 10x the output impedance of the ipod, which is a good possibility.

Funny. I googled this topic, and most answers were that the Ipod headphone out does not have enough juice, and using the direct out of the dock looses the volume control. Apparently, some sort of preamp is needed to boost the signal, even though some guy measured the output at 2.8 V. But, running it into a typical home stereo outfit is'nt good either as it causes all kinds of distortion !

[rpsands]

Originally Posted by bongomania

That's a very misleading argument, because the gain knobs have nothing directly to do with that limitation. Yes, if you drive the amp at maximum gain you will run out of headroom, and hit a limit of how much wattage is possible to produce with that amp. But turning down the gain knobs doesn't control that limit in a direct linear (or even logarithmic) way. For every increment you turn down the gain knob, you could just turn up the master volume of the preamp, or send a stronger signal into the preamp. You could argue that once the input stage of the power amp is driven into hard clipping, and cannot increase any further, then sending a stronger input would not compensate for turning down the attenuators; but that would be such an outlandishly unrealistic and inappropriate use of the power amp as to render the argument nonsensical. In other words, even though the gain knobs have a definite impact on the output levels, they do not limit the output in any practical sense.

Functionally, they do. You can't get all 1500 watts out of a 1500 watt amp if the attenuators are set at -30db.

It's totally an appropriate use of those knobs to keep the amp from exploding your speakers in the event of an input spike.

[bongomania]

Most of the posts I've seen saying an ipod "doesn't have enough juice" were not from people who had done any measurement or calculation. They seemed to be assuming, using lines like "it was designed for headphones so it won't work".

I've seen a wide range of claims about the output impedance of the ipod headphone jack, but most agree it's somewhere between 5 and 10 ohms.

Power amps have a wide range of input impedances, but for example a QSC PLX is 6 Kohms (unbalanced input) or 12 Kohms (balanced input). That's about 1,000 x the output impedance of the ipod.

IOW, impedance should not be a problem.

So that brings us back to voltage. If the only power amp somebody has on hand happens to have a very high input sensitivity rating, and they play low-volume music from the ipod, that will be too quiet. It might seem like only a tiny difference between 1V and 1.5V, but if you think about it as a percentage it's easy to see why a few hundred millivolts makes a big difference.

[bongomania]

Originally Posted by rpsands

Functionally, they do. You can't get all 1500 watts out of a 1500 watt amp if the attenuators are set at -30db.

It's totally an appropriate use of those knobs to keep the amp from exploding your speakers in the event of an input spike.

I already explained why you are wrong. Try reading it again.

Edit: Also, good luck trying to use the knobs "in the event of an input spike".

[thumbknuckle]

Originally Posted by bongomania

So that brings us back to voltage. If the only power amp somebody has on hand happens to have a very high input sensitivity rating, and they play low-volume music from the ipod, that will be too quiet. It might seem like only a tiny difference between 1V and 1.5V, but if you think about it as a percentage it's easy to see why a few hundred millivolts makes a big difference.

Not that big: 3.5 dB.

[wcriley]

3.5 dB is a big deal when you multiply it by the 30+ dB of gain in the power amp.

[thumbknuckle]

Originally Posted by wcriley

3.5 dB is a big deal when you multiply it by the 30+ dB of gain in the power amp.

3.5dB times 30dB gain is precisely 33.5dB.

[rpsands]

Originally Posted by bongomania

I already explained why you are wrong. Try reading it again.

Edit: Also, good luck trying to use the knobs "in the event of an input spike".

I read what you wrote and it doesn't make any sense. You say it's outlandish to expect that once you hard-clip the input no more signal will get through but that's what I've experienced in practice.

Most power amps start to sound obviously crappy by the time you're overdriving the input, so using more input signal and turning down the output attenuator will both functionally limit the amount of signal that can get to your speakers and also give you audible cues that you're getting too loud.

[bongomania]

Originally Posted by thumbknuckle

Not that big: 3.5 dB.

I said percentage, not logarithm. The actual results are materially quite significant, even if you choose to use a small number to represent it.

[bongomania]

Originally Posted by rpsands

I read what you wrote and it doesn't make any sense. You say it's outlandish to expect that once you hard-clip the input no more signal will get through but that's what I've experienced in practice.

Most power amps start to sound obviously crappy by the time you're overdriving the input, so using more input signal and turning down the output attenuator will both functionally limit the amount of signal that can get to your speakers and also give you audible cues that you're getting too loud.

Let me try this again, because I think we are coming at this from two totally different angles.

If you have a 1500 W amp, and you turn the gain knobs down to -30 dB, how big is the maximum wattage peak the amp can produce?

The true answer is "it depends on the input signal".

I did not say either of the things you think I said. I did not say that using the knobs was inappropriate, and I did not say that "it's outlandish to expect that once you hard-clip the input no more signal will get through". This is specifically why I said "read it again", although now I'm sorry I took that tone. So honestly I do apologize if I came off too snippy there. But really, really: I did not say either of those things. Maybe you could re-read those parts and see if a new meaning comes out?

What I did say, or at least what I was trying to say, is that there is no way to know how much wattage your amp is putting out just by looking at the position of the gain knobs (above zero). Turn the knobs down 50%. Are you getting 50% less wattage? NO! You don't know how much. Do a calculation based on the maximum voltage before hard clipping, minus the gain knob setting, times the gain of the amp. Do you know how much wattage is being put out? NO! Because the amp will put out much more than its rated wattage when the power section is driven into hard clipping. How much more? We don't know!

And very few end-user musicians have the equipment and knowledge necessary to make useful measurements of the wattage output of their amp.

So if a typical end user cannot know how much wattage is being put out at any setting of the gain knobs, then in what way do those knobs practically control the wattage?

My point was not that the wattage isn't ever reduced by using the gain knobs; clearly it can be. My point is that we don't know by how much, or how much wattage is really coming out, especially when dealing with transient peaks, and the gain knobs are no help to us in that regard.

[Bob Lee (QSC)]

3.5 dB is not a great audible difference, but it could spell the difference between an amp being able to reach full power and only about 44% of full power.

[rpsands]

I'm not sure but I think on some power amps it might actually be a gain attenuator - Bob, could you explain how the volume knobs work on say, a PLX amp?

[tekdiver500ft]

Originally Posted by Bob Lee (QSC)

Not necessarily. For example, setting up a crossover or loudspeaker processor often results in different gain settings among the various frequency bands.

I read back my post, and I wasn't entirely clear, so I understand the questions. Let me clarify: I was talking about running through the board. Best sound comes when input gain, channel slider, subgroup (if used), and main volume are all at unity. This is why many boards have attenuators on the back end. It allows you to run everything at unity and still keep volume manageable without materially affecting sound quality.

Originally Posted by Bob Lee (QSC)

What piece of gear has input attenuators?

Again, my clarity was lacking, so thanks for asking, thus allowing me to be more clear. The power amps have input attenuators, at least the ones I use. So, I turn them all the way down so nothing gets too loud while I'm setting the board. Once everything is at or near unity (+/- 5 dB on my boards is close enough to not change sound quality, so I'll often set it all at -5 dB so I have somewhere to go if I need to get louder), I then turn up the attenuators until I achieve slightly more volume than I think I'll need, again, allowing me somewhere to go should I have to.

I hope that clears up the confusion.

[Bob Lee (QSC)]

It's just a gain control (think of it as a stage with a gain variable anywhere between 0 and 1), situated in between the input stage (fixed gain) and output stage (also fixed gain).

Voltage gain is multiplicative, so input gain × gain control × output gain = overall channel gain.

They're often erroneously called "input attenuators," but if they actually were you could use them to prevent overdriving the input stage. But since pro amps generally have balanced inputs, they have balanced input stages and no input attenuators.

[Bob Lee (QSC)]

Originally Posted by tekdiver500ft

I read back my post, and I wasn't entirely clear, so I understand the questions. Let me clarify: I was talking about running through the board. Best sound comes when input gain, channel slider, subgroup (if used), and main volume are all at unity. This is why many boards have attenuators on the back end. It allows you to run everything at unity and still keep volume manageable without materially affecting sound quality.

The main point of a mixer is that you can combine signals in variable amounts (through manipulating the gain of the individual channels) to get the desired blend. Unity gain with mic inputs or passive instrument inputs would be particularly disappointing, because they tend to be of very low levels.

The power amps have input attenuators, at least the ones I use.

If they are pro amps with balanced inputs, they probably don't; they would have gain controls instead that are not part of the input.

[bongomania]

Originally Posted by Bob Lee (QSC)

It's just a gain control (think of it as a stage with a gain variable anywhere between 0 and 1), situated in between the input stage (fixed gain) and output stage (also fixed gain).

Voltage gain is multiplicative, so input gain × gain control × output gain = overall channel gain.

They're often erroneously called "input attenuators," but if they actually were you could use them to prevent overdriving the input stage. But since pro amps generally have balanced inputs, they have balanced input stages and no input attenuators.

...And this is true for most other brands as well. They use all kinds of different labels around the knobs, but it's the same setup really.

[Bob Lee (QSC)]

Originally Posted by bongomania

...And this is true for most other brands as well. They use all kinds of different labels around the knobs, but it's the same setup really.

Exactly right.

[tekdiver500ft]

It's always been my understanding (and it may be a terminology issue, I don't know) that if unity is as high as the control will go, then it is an attenuator. A gain control increases input voltage. So, on my power amps, since all the way clockwise is unity, and everything else is marked as "-X dB" they are (according to what I was taught in my audio engineering classes) attenuators. I've never opened up the amps to see actual circuits, so I don't know if they are attenuators or gain controls in actual design.



*edit*

Oops, took too long, it's already been discussed. I type too slowly...