Hi all,
I'm interested in the topic above: how is it - in theory - possible to plan/calculate the RIGHT gain structure in order to get a 3-way active system work correctly when increasing and decreasing volume on the preamp ?
Let's assume all the factors which might need correction:
- different amplifiers (100W for lows, 50W for mids, 20W for highs)
- simple line-level 3-way crossover
- different speaker sensitivity
I chose the most "difficult" way by intention to learn finally how the heck we're going to bring levels together so that even in such a different setup the whole frequency range is getting louder or quieter equally on all 3 bands when turning the volume knob at preamp side line-level.
The most simple setup would be of course:
- take 3 identical amps (6 for stereo)
- speakers have about the same sensitivity
.. and voilá !
But this is just not always the case and I would like to understand the fundamentals behind this all. DIY room-stereo, no PA system here, no DSP.
What I know:
- bass needs tons more power than mids or a ribbon crossed high enough to
move adequate amount of air for the same SPL (to stay linear at all vol levels).
What I don't know is:
- how much is the difference on amp side (power need) in a general setup ? Approximately.
- how does this all relates to amp gains (there're differences of course) and speaker sensis.
- let's design such a system in theory through an example of .. dunno.. 90dB bass, 95dB mid and 99dB sensitivity for highs. (speaker sensitivity).
I've googled a lot but maybe my fault - no big results. On most forums I read one should set all amps gains 'til clipping - wtf. Definitiely wrong, what if I use a tube amp for mids with less negative feedback which doesn't clip that sudden (distorts slowly instead). So I need some kind of real explanation behind, relating with power, decibels, etc.
Some ppl also said: for DIY amps we can adjust gain via amount of feedback too. I definitely won't do it just for the sake of volume equalization, changing feedback also changes speaker control (damping) and the overall tone of the sound so this is a no-go.
I also read: gains can be adjusted either on line-level after the input, OR on the crossover itself (actually line level too but already split), or changing amp input sensitivity (not sure if this is a proper way of it).
So there're a lot of confusions here in other communities as well. Can someone explain the basics ? (Again: no PA stuff with plate amps and tons of potmeters). I might need a simple block diagram with the signal path and the logic, where to correct what - if needed. And with speakers of different sensitivities and different amp designs we need to correct for sure.
Reason for this: I plan to scrap my old passive Butterworth XO from my 3-way DIY boxes and try to understand how to match amps needed to drive bass-mid-high sections of different speakers to achieve equally changing SPL levels when changing volume.
I'm interested in the topic above: how is it - in theory - possible to plan/calculate the RIGHT gain structure in order to get a 3-way active system work correctly when increasing and decreasing volume on the preamp ?
Let's assume all the factors which might need correction:
- different amplifiers (100W for lows, 50W for mids, 20W for highs)
- simple line-level 3-way crossover
- different speaker sensitivity
I chose the most "difficult" way by intention to learn finally how the heck we're going to bring levels together so that even in such a different setup the whole frequency range is getting louder or quieter equally on all 3 bands when turning the volume knob at preamp side line-level.
The most simple setup would be of course:
- take 3 identical amps (6 for stereo)
- speakers have about the same sensitivity
.. and voilá !
But this is just not always the case and I would like to understand the fundamentals behind this all. DIY room-stereo, no PA system here, no DSP.
What I know:
- bass needs tons more power than mids or a ribbon crossed high enough to
move adequate amount of air for the same SPL (to stay linear at all vol levels).
What I don't know is:
- how much is the difference on amp side (power need) in a general setup ? Approximately.
- how does this all relates to amp gains (there're differences of course) and speaker sensis.
- let's design such a system in theory through an example of .. dunno.. 90dB bass, 95dB mid and 99dB sensitivity for highs. (speaker sensitivity).
I've googled a lot but maybe my fault - no big results. On most forums I read one should set all amps gains 'til clipping - wtf. Definitiely wrong, what if I use a tube amp for mids with less negative feedback which doesn't clip that sudden (distorts slowly instead). So I need some kind of real explanation behind, relating with power, decibels, etc.
Some ppl also said: for DIY amps we can adjust gain via amount of feedback too. I definitely won't do it just for the sake of volume equalization, changing feedback also changes speaker control (damping) and the overall tone of the sound so this is a no-go.
I also read: gains can be adjusted either on line-level after the input, OR on the crossover itself (actually line level too but already split), or changing amp input sensitivity (not sure if this is a proper way of it).
So there're a lot of confusions here in other communities as well. Can someone explain the basics ? (Again: no PA stuff with plate amps and tons of potmeters). I might need a simple block diagram with the signal path and the logic, where to correct what - if needed. And with speakers of different sensitivities and different amp designs we need to correct for sure.
Reason for this: I plan to scrap my old passive Butterworth XO from my 3-way DIY boxes and try to understand how to match amps needed to drive bass-mid-high sections of different speakers to achieve equally changing SPL levels when changing volume.
To the amps: I don't have them yet but don't want to have 6 identical modules consuming space and heating like crazy where I need decent ones for bass only. For the highs I would use much smaller ones which implies less power into 8 Ohms but with the same amplification curve to be able to follow bass volume changes. (Or vice-versa). I think you understand.
Most active crossovers e.g. Behringer CX3400 or Citronix cx34 (both three way stereo, lower end of the market) allow you to set up the low and high crossover frequency (with a 24db/octave ramp) and also the individual gain on each driver (low, mid, high). I would plan to set these at your normal listening level so that you get the integration you want, possibly using a calibrated microphone to get a frequency/db plot using pink noise.
Other digital active crossovers also allow the addition of filters/equalisers to smooth out particular dips/peaks in the trace.
But I can see your dilemma - if you set the system up to sound right at a particular volume but have different power amps for each driver pair what happens to the balance if you turn up/down the volume. each driver pair will get a different differential in power ?
Looks like you really need a 3D map of input volts (volume) vs three channel separate gain vs crossover/filters
I guess the alternative is to use something like the minidsp that may give you more control.
Look forward to hearing the replies.
Other digital active crossovers also allow the addition of filters/equalisers to smooth out particular dips/peaks in the trace.
But I can see your dilemma - if you set the system up to sound right at a particular volume but have different power amps for each driver pair what happens to the balance if you turn up/down the volume. each driver pair will get a different differential in power ?
Looks like you really need a 3D map of input volts (volume) vs three channel separate gain vs crossover/filters
I guess the alternative is to use something like the minidsp that may give you more control.
Look forward to hearing the replies.
In My oppinon, and experience (open to correction).
Different powers don't matter (more power just gives you more headroom before clipping).
Different gains on the amplifiers don't matter (though you are best to optimize these to the drivers if you can). You can just attenuate the higher gain amp more, or use it for a less sensitive driver.
Say you have amps with 20db and 30db of gain. regardless of input voltage one is always going to have 10 db more gain than the other. If that is too much attenuate it to the level it needs to be.
Really all you need to do is have the right attenuation in the crossover for each output and you should get consistent levels throughout the volume range regardless of the amps the crossover is driving.
Tony.
Different powers don't matter (more power just gives you more headroom before clipping).
Different gains on the amplifiers don't matter (though you are best to optimize these to the drivers if you can). You can just attenuate the higher gain amp more, or use it for a less sensitive driver.
Say you have amps with 20db and 30db of gain. regardless of input voltage one is always going to have 10 db more gain than the other. If that is too much attenuate it to the level it needs to be.
Really all you need to do is have the right attenuation in the crossover for each output and you should get consistent levels throughout the volume range regardless of the amps the crossover is driving.
Tony.
Last edited:
Yes, this is why you have got no results googling the problem......there isn't one Setting Sound System Level Controls
Last edited:
My opinion, and to my knowledge, it is indeed as written by wintermute.
a (good) amplifier has a fixed gain, independent of input voltage and power. Speakers are also assumed to have a constant FR over a "wide" power range: with 10dB more power applied, the whole FR will lift 10dB.
as such, with your example units "90dB bass, 95dB mid and 99dB high", we can take any speaker as reference and add/subtract gain to the other amplifiers. Say we take the bass as reference, the mid amp needs 5dB less gain, the high will need 9dB less gain.
Say we have a 100W amp on the bass, how much power is needed on the tweeter? with this calculator What is a decibel (dB)? we set 100W as reference, subtract 9dB from that, to obtain a value of 12.58W. So applying 100W to bass and 12.58W to the high will attain similar SPL of both drivers.
So it is not that hard to setup a system: a Xover at line level, and line level controls at the input of the power amps.
Take care of phase!
a (good) amplifier has a fixed gain, independent of input voltage and power. Speakers are also assumed to have a constant FR over a "wide" power range: with 10dB more power applied, the whole FR will lift 10dB.
as such, with your example units "90dB bass, 95dB mid and 99dB high", we can take any speaker as reference and add/subtract gain to the other amplifiers. Say we take the bass as reference, the mid amp needs 5dB less gain, the high will need 9dB less gain.
Say we have a 100W amp on the bass, how much power is needed on the tweeter? with this calculator What is a decibel (dB)? we set 100W as reference, subtract 9dB from that, to obtain a value of 12.58W. So applying 100W to bass and 12.58W to the high will attain similar SPL of both drivers.
So it is not that hard to setup a system: a Xover at line level, and line level controls at the input of the power amps.
Take care of phase!
Rod Elliott - ESP
The Effect of Amplitude Inaccuracies:
BiAmp (Bi-Amplification - Not Quite Magic, But Close) - Part 2
The Effect of Amplitude Inaccuracies:
BiAmp (Bi-Amplification - Not Quite Magic, But Close) - Part 2
Let’s start with the end result in mind. You know your gain structure is correct when all components in your system clip at the same time.
So first you should find the volume that your preamp clips at, on my emotive umc-1 that’s at 58. This is the max volume on the preamp I will use for this exercise.
Now my amp connected to my 97db jbl woofer has an input sensitivity of .7v, has clipping lights indicator and also a variable input sensitivity (what some folks think is a volume knob on the amp). 200w/8ohms
Ideally the preamp should output more than the .7v that the amp needs, so it should start clipping with the preamp volume at 58. Play pink noise thru the preamp and Turn the input sensitivity till the clipping lights barely come on. You now have the amp gain matched. (You don’t need to have speakers connected).
Now let’s move to the hf section. I had a topping amp with input sensitivity of .46v, 15 Watts and horns with 110db efficiency. We go thru the same exercise with this amp with preamp set to 58.
After we go thru this exercise, everything is gain matched, however the volume levels are not necessarily correct between tweeter and woofer. If you chose the amplifier with correct gains for each section they will be more or less in the ballpark and you will need to fiddle with the input sensitity a bit or drop the levels somewhere else like the crossover by just a few dB and you are all set.
However in real life you may have already have amps that you want to use which have a huge difference in levels. You can then use an attenuator between speaker and amp to reduce the wattage the speaker gets- ( whether it will affect sonically is a diff question).
Remember also why we are doing the gain matching- to maximize head room, higher signal to noise ratio, operate within linear region of the component. If you don’t gain match it’s not necessarily a problem( example I listen at max 45 volume even tho everything was calibrated at 58), unless it manifests as a noise or hiss.
If you pair a low input sensitivity (.25v) with high gain (36db results in higher output power) with highly sensitive speakers (110db) it is a recipe for disaster. This combination will result in a very loud spl even with very low input signal. As a result you will be very close to the noise floor and all kinds of bad things will happen.
So first you should find the volume that your preamp clips at, on my emotive umc-1 that’s at 58. This is the max volume on the preamp I will use for this exercise.
Now my amp connected to my 97db jbl woofer has an input sensitivity of .7v, has clipping lights indicator and also a variable input sensitivity (what some folks think is a volume knob on the amp). 200w/8ohms
Ideally the preamp should output more than the .7v that the amp needs, so it should start clipping with the preamp volume at 58. Play pink noise thru the preamp and Turn the input sensitivity till the clipping lights barely come on. You now have the amp gain matched. (You don’t need to have speakers connected).
Now let’s move to the hf section. I had a topping amp with input sensitivity of .46v, 15 Watts and horns with 110db efficiency. We go thru the same exercise with this amp with preamp set to 58.
After we go thru this exercise, everything is gain matched, however the volume levels are not necessarily correct between tweeter and woofer. If you chose the amplifier with correct gains for each section they will be more or less in the ballpark and you will need to fiddle with the input sensitity a bit or drop the levels somewhere else like the crossover by just a few dB and you are all set.
However in real life you may have already have amps that you want to use which have a huge difference in levels. You can then use an attenuator between speaker and amp to reduce the wattage the speaker gets- ( whether it will affect sonically is a diff question).
Remember also why we are doing the gain matching- to maximize head room, higher signal to noise ratio, operate within linear region of the component. If you don’t gain match it’s not necessarily a problem( example I listen at max 45 volume even tho everything was calibrated at 58), unless it manifests as a noise or hiss.
If you pair a low input sensitivity (.25v) with high gain (36db results in higher output power) with highly sensitive speakers (110db) it is a recipe for disaster. This combination will result in a very loud spl even with very low input signal. As a result you will be very close to the noise floor and all kinds of bad things will happen.
@scottjoplin, yes you could attenuate the amp inputs. But imho it has these drawbacks. One, by decreasing the amp input you are moving closer to the noise floor, so lower signal noise ratio.
The other reason is that typically i will adjust my components starting from source moving to speakers. I have several components after my preamp, a minidsp 10x10 for active crossover, another for Dirac room correction and another for bumping up the sub line level voltage. So by the time I am setting the amplifier I have already gain matched everything in the chain before. So ideally I wouldn’t want to change anything that I have setup once already. The minidsp do not have as fine a control over gain mostly like 6db/12db increments too.
At the same time unless you are grossly mismatched in gain your suggestion will work as well. Especially people using regular components designed for the home segment should be ok.
The other reason is that typically i will adjust my components starting from source moving to speakers. I have several components after my preamp, a minidsp 10x10 for active crossover, another for Dirac room correction and another for bumping up the sub line level voltage. So by the time I am setting the amplifier I have already gain matched everything in the chain before. So ideally I wouldn’t want to change anything that I have setup once already. The minidsp do not have as fine a control over gain mostly like 6db/12db increments too.
At the same time unless you are grossly mismatched in gain your suggestion will work as well. Especially people using regular components designed for the home segment should be ok.
**** yeah.. I thought it's not that simple...
My father also told me it's enough to compensate on each amp's volume control.
Let's say I "convert" my 3-way old setup from my childhood, stupid Butterworth with random speakers (regarding sensitivity):
- 86dB bass
- 88 mid
- 91 tweeter
According to him, if I buy 6 identical class AB modules on eBay (or whatever, just an example) and drive each speaker from these, I would need:
- leave the bass on max
- attenuate the mids a bit
- attenuate the tweeters a bit more
And I can put a master volume control into the preamp on signal level (too) which affects then the whole systems's perceived volume.
I might believe this theory (since vol. controls - at least most of them - are also LOG types) so if everything's set fine, I have a coherently behaving system overall.
With 6 identical amp modules (100W 8 Ohm each).
Theory.
But: why the heck shall I use 100W class AB heavy beasts for the tweeters as well ? I would rather use smaller ones. Or same big but Class A maybe. Or whatever.. Class AB, AB mid-high, Class D bass. Just a thought.
This makes the whole thing the most complicated.
(I think somebody has already provided an answer... I read now..)..
Edit: #11 adzy, got your point. I would do the same, at the very beginning once on line-level and then goodbye problem.
Last edited:
You are exactly right, many people would use a low power class A amp for tweeters in your situation. Read the Rane note I linked you too earlier, it's very good and clear
Last edited:
Exactly. Why would I burn let's say 100W heat for the tweeters in a Class AB endstage just because I need 100W for the bass section ? I could go with 50W mids and maybe 10W for the tweeters. (Just assuming). 50W amp with a little "steeper" curve and 10W with an even more steep curve compared to bass - so they walk together as I change master volume.
Whoaa, a bit insane..
How is it done in PA ? Actually, the recipe is there. Do they use the same insane power amplifiers for the mids and tweeters like for the bass drivers ? Can't believe. They also equalize them somehow somewhere so "smaller" amps can drive smaller drivers, staying with the bass at all volume levels nicely.
I'll need to have this confirmed. Otherwise seems legit at first glance. We'll see.
In case you didn't spot it earlier...............this Rane note........Setting Sound System Level Controls
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.