It isn't clear from the previous graphs - if the absolute polarity of the drivers is reversed, will the off-axis peak and dip reverse vertical position?
It isn't clear from the previous graphs - if the absolute polarity of the drivers is reversed, will the off-axis peak and dip reverse vertical position?
Tom,
Is this the graph you were asking about?

The above graph is of a Butterworth 3rd order crossover (LP and HP) with the both driver connected with the same polarity.
Reversing the polarity of one driver (e.g. tweeter) will product the following response:

NOTE: angles > 0 are defined as "above" the on-axis plane
As you can see, the position of the null and the response 3dB peak have switched with respect to the on-axis plane.
The shape of the response remains the same, it's just flipped.
Judging from the graph, it it's important to consider whether the tweeter is above or below the woofer and where the position axis is located when choosing the relative polarity of the two drivers.
The on-axis frequency response is flat, and the power response is identical, for both normal and reversed polarities.
-Charlie
I'm not currently subscribed to the AES, so I would have to make a trip to the library. I believe that this is Small's original paper on the Butterworth crossover. Is that correct?
You might want to take a look at this Rane app note (available online):
The graphs most recently above are confusing. I was referring to graphs in this link. Anyway, you answered my question - position does not change if both drivers reverse polarity.
The graphs most recently above are confusing. I was referring to graphs in this link. Anyway, you answered my question - position does not change if both drivers reverse polarity.
Sorry, I guess I misread your post.
The graphs above are for the case of both drivers in phase and one driver with reversed phase. Only the relative phase is important, so reversing both gives the same responses as when both are connected with normal polarity, whateve "normal" is.
-Charlie
Progress Report!
Well, I'm finally making some progress...
Here are some pics of the new arrivals around here:
The upper picture is of (left to right): Universal Filter Board, Modular Crossover Board, original "Main Board" concept with on board PS and EQ.
I ordered the Main board and an earlier version of the Modular board in December, and they just showed up at the end of last week because of some delay in the mail service. 😡 As a result, I had the latest order sent via 2-day express service!
The latest boards look very nice. The fab work seems to be top quality and the boards feel quite solid. The larger board is 4x4 inches and the smaller (in hand) board is 2x4 inches.
The next step is to build up some of these boards to test out the design.
-Charlie
Well, I'm finally making some progress...
Here are some pics of the new arrivals around here:


The upper picture is of (left to right): Universal Filter Board, Modular Crossover Board, original "Main Board" concept with on board PS and EQ.
I ordered the Main board and an earlier version of the Modular board in December, and they just showed up at the end of last week because of some delay in the mail service. 😡 As a result, I had the latest order sent via 2-day express service!
The latest boards look very nice. The fab work seems to be top quality and the boards feel quite solid. The larger board is 4x4 inches and the smaller (in hand) board is 2x4 inches.
The next step is to build up some of these boards to test out the design.
-Charlie
board build report
I finally built up one of the Universal Filter boards. Each board has two independent filters that can be configured for up to 3rd order, so you can build asymmetric filters. I built a 1st order HP, 3rd order LP crossover.
I have to say that the boards were really easy to build. Because each component is labeled to match the schematic shown in the attached picture, I just pulled up TI's free FilterPro program (version 2), created the filter sections that I needed, ordered the parts, and then inserted each component on to the board where it should go.
I will be testing these out soon, but here is a pic of the built up board. That should be the last hurdle before these are ready to go. More complete information on the schematic is available in this pdf file:
http://audio.claub.net/temp/universal_filter_board_connections.pdf
-Charlie
I finally built up one of the Universal Filter boards. Each board has two independent filters that can be configured for up to 3rd order, so you can build asymmetric filters. I built a 1st order HP, 3rd order LP crossover.
I have to say that the boards were really easy to build. Because each component is labeled to match the schematic shown in the attached picture, I just pulled up TI's free FilterPro program (version 2), created the filter sections that I needed, ordered the parts, and then inserted each component on to the board where it should go.
I will be testing these out soon, but here is a pic of the built up board. That should be the last hurdle before these are ready to go. More complete information on the schematic is available in this pdf file:
http://audio.claub.net/temp/universal_filter_board_connections.pdf
-Charlie


Modular Crossover Board - initial test results
I finally put together one of the modular crossover filter boards (shown in post #105 in the lower pic as LA-X2F2). This design has a state variable filter that provides simultaneous, complementary second order low and high pass functions with Q and corner frequency adjustable via on-board cermet trimmer potentiometers. The potentiometers are "pluggable" - they plug in to sockets that use something similar to what is used for DIP ICs. This allows the end user to remove and adjust the pots for excellent matching and precision. You can also adjust them on the fly while they are in the board if you like.
Here are some quick tests that I ran with the board connected to a basic LM7815/7915 based power supply, with both the PS and MXB board out in the open on my desk. I set Fc = 1k Hz, with Q=0.7 initially, and then jacked up to Q=3. Then I changed to Fc = 500 Hz, Q = 0.7 and finally I took some quick distortion measurements with those settings. In the passband, the distortion level is around -95 dB (just under 0.002%). I used an LME49740 IC for these tests. For fun, I substituted an OnSemi MC33274 IC and the distortion rose to about 0.5% (about 200 times higher!) but this isn't much of a surprise since the OnSemi op-amp datasheet indicates that it has about 100 times higher distortion as the National LME IC!
NOTE: the frequency response measurements were carried out using ARTA's spectrum analyzer with signal averaging. This results in the rather noisy looking responses, but is not indicative of circuit noise.
FREQUENCY RESPONSES:
DISTORTION:
.
I finally put together one of the modular crossover filter boards (shown in post #105 in the lower pic as LA-X2F2). This design has a state variable filter that provides simultaneous, complementary second order low and high pass functions with Q and corner frequency adjustable via on-board cermet trimmer potentiometers. The potentiometers are "pluggable" - they plug in to sockets that use something similar to what is used for DIP ICs. This allows the end user to remove and adjust the pots for excellent matching and precision. You can also adjust them on the fly while they are in the board if you like.
Here are some quick tests that I ran with the board connected to a basic LM7815/7915 based power supply, with both the PS and MXB board out in the open on my desk. I set Fc = 1k Hz, with Q=0.7 initially, and then jacked up to Q=3. Then I changed to Fc = 500 Hz, Q = 0.7 and finally I took some quick distortion measurements with those settings. In the passband, the distortion level is around -95 dB (just under 0.002%). I used an LME49740 IC for these tests. For fun, I substituted an OnSemi MC33274 IC and the distortion rose to about 0.5% (about 200 times higher!) but this isn't much of a surprise since the OnSemi op-amp datasheet indicates that it has about 100 times higher distortion as the National LME IC!
NOTE: the frequency response measurements were carried out using ARTA's spectrum analyzer with signal averaging. This results in the rather noisy looking responses, but is not indicative of circuit noise.
FREQUENCY RESPONSES:



DISTORTION:

.
Last edited:
Universal Filter Board build report (board #2)
I was also able to build up another Universal Filter Board with a new set of parts that came in a few days ago. Like attempt #1, it's designed to implement a third order low pass filter, but this time using the Multiple Feed Back (inverting) topology. After I built the first board, I realized that I had inadvertently switched the layout and that you can't use this board to make a Sallen Key 3rd order filter - thus the switch to MFB topology in this build for the 3rd order section. I also implemented a first order high pass filter using a non-inverting (Sallen Key) topology.
I did a quick test on ARTA for both filters including distortion. The results look pretty good. I used an LF353 dual op amp I had laying around, since I have so far not ordered any better quality dual op amps.
I'm a little puzzled about the difference in distortion levels for each filter, although they do use totally different topologies. If anyone had any ideas why they would be so different, using the same IC, please post your thoughts. The frequency response is right on target, however, and I will remeasure distortion when I get some better ICs.
This board looks like a "go"!
Frequency Response and Distortion: 2nd order LP filter, Q=1, Fc=250 Hz
Frequency Response and Distortion: 1nd order HP filter, Fc=250 Hz
.
I was also able to build up another Universal Filter Board with a new set of parts that came in a few days ago. Like attempt #1, it's designed to implement a third order low pass filter, but this time using the Multiple Feed Back (inverting) topology. After I built the first board, I realized that I had inadvertently switched the layout and that you can't use this board to make a Sallen Key 3rd order filter - thus the switch to MFB topology in this build for the 3rd order section. I also implemented a first order high pass filter using a non-inverting (Sallen Key) topology.
I did a quick test on ARTA for both filters including distortion. The results look pretty good. I used an LF353 dual op amp I had laying around, since I have so far not ordered any better quality dual op amps.
I'm a little puzzled about the difference in distortion levels for each filter, although they do use totally different topologies. If anyone had any ideas why they would be so different, using the same IC, please post your thoughts. The frequency response is right on target, however, and I will remeasure distortion when I get some better ICs.
This board looks like a "go"!
Frequency Response and Distortion: 2nd order LP filter, Q=1, Fc=250 Hz

Frequency Response and Distortion: 1nd order HP filter, Fc=250 Hz

.
The distortion is but not the noise 🙁If the distortion is vanishingly small, then twice the distortion is still vanishingly small. This is one reason why I am using high end op amps, so that the number of op amps is not really of major concern.
Your boards look very good for prototyping though.
Also, you have to keep in mind the big picture. These circuits will be used as crossovers for loudspeakers. Loudspeaker drivers convert an electrical signal in to a mechanical one (e.g. sound waves) with an inherent distortion that is often on the order of 0.1%, or 1% if SPLs are "high" and the driver is reaching Xmax. So chasing distortion in the crossover circuit that is much smaller in magnitude really does not make all that much practical sense. The loudspeaker driver distortion dominates.
-Charlie
This explanation has several times been rebuked. Loudspeaker distortion is DIFFERENT from the distortions in electronics, which are still heard at tiny amounts. Otherwise everybody would use opamps.
This explanation has several times been rebuked. Loudspeaker distortion is DIFFERENT from the distortions in electronics, which are still heard at tiny amounts. Otherwise everybody would use opamps.
I agree with the first part of your post - speaker distortion is a of a different nature.
But there are many reasons why people would or would not use opamps and often they have nothing to do with the actual sound quality.
jan
.....................Loudspeaker drivers convert an electrical signal in to a mechanical one (e.g. sound waves) with an inherent distortion that is often on the order of 0.1%, or 1% if SPLs are "high" .....................
Many quote the high distortion of speakers when delivering high SPLs....................This explanation has several times been rebuked. Loudspeaker distortion is DIFFERENT from the distortions in electronics, which are still heard at tiny amounts...............
It shows clearly on the response chart which some manufacturers publish.
I have also seen a few manufacturers show the distortion plots for two different SPL output levels. These also clearly show that distortion reduces dramatically when output levels are reduced.
I cannot recall seeing distortion figures nor distortion plots for output levels that are say -20dB ref maximum nor for any lower output levels, maybe -40dB ref max.
Example for a speaker that has a maximum output level of 105dB, we sometimes see distortion quoted for 96dB spl and 90dB spl.
what about for 85dB spl (-20dB ref max) and for 65dB spl (-40dB ref max).
I suspect the speaker distortion becomes so low it is difficult to measure accurately, but that is me just guessing.
Who knows better?
- Status
- Not open for further replies.
- Home
- Source & Line
- Analog Line Level
- active crossover board capabilities - I want your input