What's new

Can you answer a really silly crossover question?? (1 Viewer)

Chris PC

Senior HTF Member
Joined
May 12, 2001
Messages
3,975
Being unhappy with the current bass management and associated crossovers in receivers, I wanted to clarify that I understood how receiver crossovers worked. I understand crossovers in speakers, but I wanted to make sure that understood the way they work in receivers.
In the case of both HIGH PASS and LOW PASS, does a crossover entail rolling off the frequencies BEFORE the "center frequency of 100 hz"?
For example, we'll discuss the crossover between the front Left and Right main speakers and the Subwoofer.
If the crossover between the subwoofer and the front main speakers is 100 hz, does that mean that the signal sent to the subwoofer has low frequencies are rolled off starting at say 80 hz and the signal sent to the front main speakers is rolled off at say 120 hz, each which whatever respective slope is applied?
What I mean is, the crossovers in receivers are not steep brick walls, but more likely a shallow 6 dB per octave blending with the frequencies sent to each dropping off well before that crossover frequency? For instance, at the crossover frequency, the sound is being reproduced by both subwoofer and mains somewhat equally?
 

Jeremy Anderson

Screenwriter
Joined
Nov 23, 1999
Messages
1,049
That sounds accurate to me, although 6db/octave seems a bit shallow. My Onkyo receiver slopes at about 12db/octave, and my old JVC was 18db/octave, both crossed at 80hz. I prefer the 12db/octave of my Onkyo because it maintains directionality in upper bass a bit better than my JVC did.
This is why you can hear phase problems between the sub and speakers primarily in the crossover range, because if the phase is reversed, the sub cancels out the speakers.
 

DaleB

Stunt Coordinator
Joined
Jul 23, 2000
Messages
103
It has always been my understanding that 100 hz in this case, is the point where it starts to roll off.
Imagining a bell curve, with the peak being 100 Hz. The DEGREE of roll-off (the 'shape of the curve') is controlled by the type or configuration of the particular filter being used.
Imagining also a straight horizontal line intersecting the peak, would be the uniform level of all other frequencies.
The curved slopes representing the high & low pass responses of the filters.
That's how I visual a typical receiver or processor crossing over the lower frequencies, below 100Hz, to the subwoofer, and those above 100Hz to the other speakers. Now, how well the particular speakers or sub 'respond' to that could be another topic.
Ideally, WHERE they crossover, should be continuous in level of amplitude.
Maybe my observation is oversimplified, or I missed something entirely!
I also imagine when designing a filter, the response of the components results in calculations to ensure there is no roll-off starting before 100 Hz.
 

Brian Bunge

Senior HTF Member
Joined
Sep 11, 2000
Messages
3,716
If I'm not mistaken, the "crossover point" is actually 3dB down from reference. In other words, the point at which the high pass and low pass filters intersect one another is 3dB down (this is assuming you're looking at a graphical representation of the two filters). And in this case that point is at 100Hz. You will definitely still get output from your mains below 100Hz and from your sub above 100Hz. With a 12dB/octave crossover this would mean that at one octave down (50Hz) you will still have output from your mains attenuated by 12dB. At one octave up (200Hz) you will still have output from your sub attenuated by 12dB.
Brian
------------------
 

DaleB

Stunt Coordinator
Joined
Jul 23, 2000
Messages
103
Wouldn't there be some kind of compensation to avoid a dip in response, by extending the 'actual' frequency where the cutoff starts?
 

Gregory S

Stunt Coordinator
Joined
Jan 3, 2001
Messages
53
Brian,
You are correct. The specified crossover frequency is down 3 db. So, if the specified frquency is 100 Hz, the response is down 3 db at 100 Hz and therefore the actual start of the roll off is slightly higher (assuming a high pass filter).
Dale,
The roll off prior to the specified crossover frequency is a nature of the beast. However, if you want the start of the roll off to actually occur, say at 100 Hz, move the specified crossover point lower in frequency (90 Hz or so). I really don't think you would notice anything, but you could do it, assuming you are designing your own circuit.
Greg
 

Brian Bunge

Senior HTF Member
Joined
Sep 11, 2000
Messages
3,716
Dale,
The summed output of the two drivers naturally boosts the response back up at the crossover frequency, so no compensation should be needed. What I was describing is just a graphical respresentation of where the two independent filters meet. Any added compensation could actually cause a peak in the response at the crossover frequency.
Brian
------------------
 

Chris PC

Senior HTF Member
Joined
May 12, 2001
Messages
3,975
The picture in my mind is also complete, except of course, we don't know the parameters for each receiver design. we don't know what the slope of the crossover is, and where exactly the roll-off begins for high and low pass. You'd have to know the slope, aswell as the crossover frequency, in order to know that.
 

Brian Bunge

Senior HTF Member
Joined
Sep 11, 2000
Messages
3,716
Chris,
I believe most receivers implement a 2nd order (12dB/octave) crossover somewhere between 80-100Hz. It should be available in the specs sheet. My Sherwood Newcastle uses a 2nd order crossover fixed at 100Hz.
Brian
------------------
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Gregory,
Don't forget that in addition to the slight rolloff beginning above the crossover frequency you've got a subwoofer with a low-pass filter that is adding in to make the overall SPL work out correctly.
I posted this on AVS Forum in a discussion of bass management.
Assume crossover is 80Hz, high pass is 12dB/octave and low pass is 24dB/octave. This is the THX curve for crossover frequencies.
Assume the desired volume is 75dB for the test tone and that the test tone is being output from one speaker.
At the crossover frequency, both the subwoofer and the speaker are outputting 72dB (added together in that curious bit of mathematics known as logarithmic arithmetic) this totals out to 75dB.
If we jump one octave up to 160Hz the outputs are as follows: main speaker 74+dB, subwoofer 48dB.
Going up another octave to 320Hz, the outputs are:
Main speaker 75dB (so close that this is accurate enough), subwoofer 24dB (it's inconsequential and adds the slightest fraction of a dB vs the main speaker).
At 640Hz, the subwoofer is contributing zero.
Moving below the crossover frequency....
At 40Hz, the subwoofer output is 74+dB and the speaker output is 60dB.
At 20 Hz, the subwoofer output is 74+dB and the speaker output is 48dB.
I did a much more elegant job of explaining this on AVS Forum where I also got into what shows up in the subwoofer from a frequency based perspective using a bass guitar as my example.
Hope this helps everyone out.
Regards,
------------------
John Kotches
Contributing Writer
Link Removed
 

Gregory S

Stunt Coordinator
Joined
Jan 3, 2001
Messages
53
John,
Thanks for that detailed answer - I'll try to find your post over at the AVS Forum.
John, another question: Can two or more pre-outs be summed in the analog domain by using a summing buffer? Basically, an opamp used as a summer in a voltage follower configuration. So, if two pre-outs were each putting out 1 volt the output of the opamp would be 2 volts.
What I'm thinking is that 2 pre-outs (say, left and right main) can be summed together to form a composite signal to be sent to a subwoofer.
Is that basically what happens when pre-outs are summed together for the subwoofer - the output voltage increases?
How is it generally done when done in the analog domain? I know most of the time it is done digitally.
What is the typical input voltage range to a powered subwoofer? If summing works the way I described, wouldn't you eventually overload the subwoofer if too many pre-outs were summed - you would exceed the input voltage range of the subwoofer's amplifier?
If this can be done in analog, I could potentially make my own bass management system. At least it's fun thinking about!
Thanks for any help you can offer. Have a good weekend!
Greg
 

John Kotches

Senior HTF Member
Joined
Mar 14, 2000
Messages
2,635
Greg,
I've not played with the electrical aspects of crossover construction, so I'm afraid I'm not going to be much help with your questions.
Regards,
------------------
John Kotches
Contributing Writer
Link Removed
 

Users who are viewing this thread

Sign up for our newsletter

and receive essential news, curated deals, and much more







You will only receive emails from us. We will never sell or distribute your email address to third party companies at any time.

Forum statistics

Threads
356,994
Messages
5,127,971
Members
144,226
Latest member
maanw2357
Recent bookmarks
0
Top