I spent last night soldering a simple LPF filter together, attached the schematic and hope that xtronic don't mind.
Could someone explain what kind of filter this is? I understand that there are two simple RC filters here. R 6 and C 12 coupled with a pot seem to sweep the value of R (first filter) 50HZ - 150 HZ cut off points
There then seems to be a second duplicate RC circuit through R 5 and C 11.
Does this mean it becomes a 2nd order filter?
Also, is the first opamp just there as a buffer?
Any light shone on this for a noob would be much appreciated
Stu
Could someone explain what kind of filter this is? I understand that there are two simple RC filters here. R 6 and C 12 coupled with a pot seem to sweep the value of R (first filter) 50HZ - 150 HZ cut off points
There then seems to be a second duplicate RC circuit through R 5 and C 11.
Does this mean it becomes a 2nd order filter?
Also, is the first opamp just there as a buffer?
Any light shone on this for a noob would be much appreciated
Stu
The first section, around IC3A, is a second-order Sallen and Key highpass. Judging by the RC values used, it is intended as a subsonic filter.
The two RC sections you mention together indeed form a tuneable second-order low pass, but with a very smooth (gradual, weak, not sharp) transition from passband to stopband. Due to loading effects, the poles of the two RC-sections move away from each other, so you get a response that is even more gradual than a second-order Linkwitz-Riley filter.
What's the purpose of this thing?
The two RC sections you mention together indeed form a tuneable second-order low pass, but with a very smooth (gradual, weak, not sharp) transition from passband to stopband. Due to loading effects, the poles of the two RC-sections move away from each other, so you get a response that is even more gradual than a second-order Linkwitz-Riley filter.
What's the purpose of this thing?
Thats a very simple crossover.
It just 2nd order.
I used a 4 th order on my system.
It just 2nd order.
I used a 4 th order on my system.
An externally hosted image should be here but it was not working when we last tested it.
Thanks Andrew, Yes I've read the Elliot Sound Products website over and over whilst trying to absorb the info at a noob pace.
I've already built my 2nd order RCRC low pass filter so I will try it but i'm already thinking I will be making a new filter very soon!
My bookshelfs are Kef reference 101 which are apparently flat frequency response down to 90Hz, then they drop to -10db at 47Hz.
Not quite sure what I would want my sub filter to do to compliment my speakers, where would the frequencies be best to cross over?
And do you guys make a 'sweepable' cut off point so that the sub can be adjusted to ones taste?
Stu
I've already built my 2nd order RCRC low pass filter so I will try it but i'm already thinking I will be making a new filter very soon!
My bookshelfs are Kef reference 101 which are apparently flat frequency response down to 90Hz, then they drop to -10db at 47Hz.
Not quite sure what I would want my sub filter to do to compliment my speakers, where would the frequencies be best to cross over?
And do you guys make a 'sweepable' cut off point so that the sub can be adjusted to ones taste?
Stu
Are those closed boxes or bass reflex speakers (or something else)?
If they are closed boxes with approximately a second-order Butterworth high-pass response, you could:
1. Make a second-order Butterworth low-pass with the same cut-off frequency for the subwoofer. Depending on the placement of the subwoofer and the other speakers, this may give you a 0 to 3 dB bump around the crossover frequency.
2. If the path lengths from the bookshelf speakers and subwoofer to your ears are nearly the same, which would give a 3 dB bump if you take approach 1, you could instead decide to put second-order Butterworth high-pass filters in the path to your bookshelf speakers and make a fourth-order Linkwitz-Riley low-pass for the subwoofer.
If they are closed boxes with approximately a second-order Butterworth high-pass response, you could:
1. Make a second-order Butterworth low-pass with the same cut-off frequency for the subwoofer. Depending on the placement of the subwoofer and the other speakers, this may give you a 0 to 3 dB bump around the crossover frequency.
2. If the path lengths from the bookshelf speakers and subwoofer to your ears are nearly the same, which would give a 3 dB bump if you take approach 1, you could instead decide to put second-order Butterworth high-pass filters in the path to your bookshelf speakers and make a fourth-order Linkwitz-Riley low-pass for the subwoofer.
Thanks, the kef reference 101 are closed boxes. I have been reading for days about filters and speakers, learning but at a very slow pace. There seems to be so much science. Were you able to guess the second order Butterworth High pass response from the info I gave you (flat to 90Hz, -10dB @ 47Hz)?
Would an ideal sub woofer partner to these speakers then require a bode plot that is symmetrical to this? The bode plot for the frequency response is included in this link:
http://kef.com/uploads/files/en/museum_pdf/70s/Reference_Series_Model_101.pdf
I may be wrong, but would a second order filter drop 12dB over an octave? Which means we are almost accurate if we drop 10dB from 90Hz to 47Hz?
Would I then want the opposite effect for my LPF to match the speakers?
BUT, I would get a 3dB bump at approx 90Hz - I kind of understand this I think - would a 3dB bump be extremely undesirable or is it something that some people can live with?
Surely commercial sub woofers that simply have one control for cut off and one for volume must encounter these issues and worse all the time,
Stu
Would an ideal sub woofer partner to these speakers then require a bode plot that is symmetrical to this? The bode plot for the frequency response is included in this link:
http://kef.com/uploads/files/en/museum_pdf/70s/Reference_Series_Model_101.pdf
I may be wrong, but would a second order filter drop 12dB over an octave? Which means we are almost accurate if we drop 10dB from 90Hz to 47Hz?
Would I then want the opposite effect for my LPF to match the speakers?
BUT, I would get a 3dB bump at approx 90Hz - I kind of understand this I think - would a 3dB bump be extremely undesirable or is it something that some people can live with?
Surely commercial sub woofers that simply have one control for cut off and one for volume must encounter these issues and worse all the time,
Stu
Once you get signal below about 200hz, there are myriad anomalies in response caused by room reflections and phase issues. Don't worry at all about a 3db peak at the xover freq. There will be much more variance in response caused by the shape and size of your room that will be far more significant. In fact, if you happen to have a room mode at the xover freq, the 3db peak may help to flatten it! Most rooms have one in the 60 - 80hz range, and it easily can be 6 or more db. This has nothing to do with quality of speakers or xover, but is simple physics (actually, complex acoustics).
If you happen to end up with an objectionable peak at the xover freq, simply move it up or down by adjusting the component values.
Not only the room, but the physical relationship of the sub and the mains can have additive/destructive effects on response. Those can sometimes be managed by moving the sub around.
This is why so many subwoofer plate amps have built-in adjustable xovers, phase adjustments, and parametric equalizers. They sure come in handy, but usually are still not enough to completely overcome all the humps and valleys of response. A pair of bass traps for the corners of your room can also be a big help to smoothing everything out.
Peace,
Tom E
If you happen to end up with an objectionable peak at the xover freq, simply move it up or down by adjusting the component values.
Not only the room, but the physical relationship of the sub and the mains can have additive/destructive effects on response. Those can sometimes be managed by moving the sub around.
This is why so many subwoofer plate amps have built-in adjustable xovers, phase adjustments, and parametric equalizers. They sure come in handy, but usually are still not enough to completely overcome all the humps and valleys of response. A pair of bass traps for the corners of your room can also be a big help to smoothing everything out.
Peace,
Tom E
Yes. Most bookshelf speakers are either closed box or bass reflex types. Theoretically, the closed-box speakers behave as second-order high-pass filters for low frequencies, while the bass reflex types behave as fourth-order high-pass filters for low frequencies. A second-order filter rolls of (asymptotically) at a rate of 12 dB/octave and a fourth-order filter at a rate of 24 dB/octave. Hence, with the info you gave, a closed box seemed likely. As KEF is a good brand, the speakers were not likely to have a huge bump in their response near the resonance frequency. A second-order high-pass with little or no bump must at least be close to second-order Butterworth.
The graph in the document you referred to is indeed close to a second-order Butterworth response with a cut-off frequency of about 80 Hz. The 90 Hz specified by Kef is the -2 dB point, but normally we are interested in the half power point (-3.010299... dB, usually rounded to -3 dB).
I don't think there is an ideal choice. With an 80 Hz second-order Butterworth low-pass for the subwoofer the total amount of power that gets radiated into your room (in all directions) will be flat around the crossover frequency, but if you measure the on-axis response in an anechoic chamber you will see a 3 dB bump if the path lengths from the bookshelf speakers and the subwoofer to the microphone are equal.
With an extra second-order 80 Hz Butterworth high-pass in the path towards the bookshelf speaker, you can make the total response of the bookshelf speaker plus extra filter fourth-order Linkwitz-Riley. (By definition, cascading two n-th-order Butterworth high-pass filters with equal cut-off frequencies gives you a 2n-th-order Linkwitz-Riley high-pass filter. The same applies to low-pass filters.)
If you then use a fourth-order Linkwitz-Riley low-pass for the subwoofer and ensure that the path lenghts are equal, you get a flat on-axis response around the crossover frequency, but a 3 dB dip in the total power radiated into your room. It is up to you to determine what you find least annoying.
I have only limited practical experience with subwoofers. I once made a crossover for a subwoofer for a local radio station, used to augment the bass response of a pair of NS-10M studio monitor speakers. These also had an approximately second-order Butterworth response, I think it was with 85 Hz cut-off.
As the path lengths from the subwoofer and the other loudspeakers to the ears of the listeners were quite different anyway, I chose the Butterworth approach, with the gain of the subwoofer adjusted until it sounded right to me. For what it's worth, the users of the studio liked it.
Last edited:
Very helpful Marcel, thanks.
I've made the filter from the first post in this thread which I believe you said may be a 'weak' roll off. I'm pleased it worked but have not been able to test it with my stereo system yet. I've made my class D bass amp quite modular in that I can very easily swap out the filter I made with a new one.
I will make a 2nd order Butterworth filter, I presume I can just type 80Hz into this tool:
Sallen-Key Low-pass Filter Design Tool
to get the R and C values for a second order Sallen Key filter to build around my NE5532 opamp? Would I be more interested in a damping factor of 1 or Quality of 0.707?
Am reluctant as yet to opamp a second order high pass to my Kef speakers and put a 4th order low pass filter for the sub as I have a home built ECL86 (Based on baby Huey) amp to drive the kefs that is just beautiful to listen to. I just crave more bass,
Stu
I've made the filter from the first post in this thread which I believe you said may be a 'weak' roll off. I'm pleased it worked but have not been able to test it with my stereo system yet. I've made my class D bass amp quite modular in that I can very easily swap out the filter I made with a new one.
I will make a 2nd order Butterworth filter, I presume I can just type 80Hz into this tool:
Sallen-Key Low-pass Filter Design Tool
to get the R and C values for a second order Sallen Key filter to build around my NE5532 opamp? Would I be more interested in a damping factor of 1 or Quality of 0.707?
Am reluctant as yet to opamp a second order high pass to my Kef speakers and put a 4th order low pass filter for the sub as I have a home built ECL86 (Based on baby Huey) amp to drive the kefs that is just beautiful to listen to. I just crave more bass,
Stu
Butterworth is Q= 1/sqrt(2) = 0.7071
Measure your sealed box speaker.
You can find the Qbox value. There are many sites showing how to do this and the calculations.
If Q is not exactly 0.7071, you can use a Linkwitz Transform to convert it to 1/sqrt(2).
Then your electrical 2pole Butterworth will give your speaker the acoustic LR4
Match this to a low bass LR4 low pass for the complete Low bass to mid crossover.
Measure your sealed box speaker.
You can find the Qbox value. There are many sites showing how to do this and the calculations.
If Q is not exactly 0.7071, you can use a Linkwitz Transform to convert it to 1/sqrt(2).
Then your electrical 2pole Butterworth will give your speaker the acoustic LR4
Match this to a low bass LR4 low pass for the complete Low bass to mid crossover.
A second-order Butterworth filter always has a Q of 0.5*sqrt(2) ~= 0.7071, otherwise it isn't a Butterworth filter.
It is advisable to choose the C1/C2 ratio only slightly larger than the smallest possible value, which is 2 for Q = 0.5*sqrt(2). For example, you could take C1 = 1 uF and C2 = 470 nF, or C1 = 470 nF and C2 = 220 nF.
Actually I meant only slightly larger than or equal to; there is nothing wrong with using a ratio of exactly two. The advantage of using a C1/C2 ratio close to the minimum possible value is that it prevents extreme resistance ratios and it makes the Q factor less sensitive to resistor tolerances.
Marcel, thanks for the help - I will make a filter with
R1 = 8.2kΩ
R2 = 4.7kΩ
C1 = 0.47uF
C2 = 0.22uF
Would it be wise to precede this with the first high pass filter in my first post? this would then give me a summing opamp stage, the highpass filter and volume?
obviously I would remove R6 and start the low pass section there,
Stu
@Andrew, thank you for your comments too, I will try the linkwitz transform at a later date. From what I understand there are soo many other factors to consider
R1 = 8.2kΩ
R2 = 4.7kΩ
C1 = 0.47uF
C2 = 0.22uF
Would it be wise to precede this with the first high pass filter in my first post? this would then give me a summing opamp stage, the highpass filter and volume?
obviously I would remove R6 and start the low pass section there,
Stu
@Andrew, thank you for your comments too, I will try the linkwitz transform at a later date. From what I understand there are soo many other factors to consider
Look up Linkwitz Reilly active filters.
They will explain how the work better than me.
The components in the LR circuit are scaled to give a certain Q.
With the 4 stages you also get 360 degree phase shift through both upper and lower passes. This keeps upper/lower driver in phase.
I just ran a quick simulation and found that the Q of the high-pass section is quite high, which means that you get a substantial response peak. Presumably R1 and R3 got swapped; with 8.2 kohm between pins 1 and 2 of IC3A and 15 kohm from pin 2 to ground, you do get a decent high-pass response. If this circuit is meant for a specific subwoofer model, it could be that the circuit is correct and just compensates for a roll-off elsewhere.
Do you often play warped gramophone records or use other sources that are likely to have a large subsonic content? If so, then having a subsonic filter somewhere may be a good idea. If, on the other hand, there are no subsonics in the first place, you don't need to filter them off.
It is certainly good to drive the Sallen and Key low-pass from an impedance much smaller than the resistors in the filter network.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.