I've got a Jensen 15" woofer and three Jensen tweeters someone had pulled from an Emerson console and sold online. The woofer itself has Emerson on it which likely means it was designed specifically for Emerson by Jensen or maybe just that Emerson asked for their name to be put on it.
I used the cabinet from a cheap 1950 Silvertone console that wasn't worth restoring (except the V-M Tri-O-Matic record changer which will go in my Magnavox CR-198C console) as the cabinet for the speaker. The only good thing about the cabinet is the piece of plywood I used for the speaker baffle.
The stock driver configuration of the Emerson console had the woofer in one spot and the three tweeters arranged in a triangle shape at an angle from the woofer. I arranged them as seen here. The extra hole was when I had a pair of dome tweeters on either side of the top tweeter. Didn't like how it sounded as it wasn't very good up close to the speaker given the distance between the two domes.
So I found a ribbon tweeter I had and used that which improved the sound.
I made a 6dB/octave crossover for the speaker as that's what was originally used except the woofer didn't have any inductor in series with it. The reason for the added tweeter crossed over so high is I didn't like how the cone tweeters didn't reproduce the treble up to 16kHz that good.
The plan is to eventually have a new proper open back cabinet built and make it to where the ribbon tweeter mounts above the top cone tweeter.
Cannot decide if I want to paint it black like the baffle is or have it properly finished like the rest of the cabinet is.
Should I keep the 6dB/octave crossover or switch to a 12 dB/octave crossover as in will it make the speaker sound better or is it best to keep the stock crossover slope?
EDIT:
Here's a 12dB/octave crossover I've calculated. The .16mH, .32mH and .12mH inductors I could only get .15mH, .13mH and .33mH, however if I go with this crossover I will measure those and see how close they are to the calculated values, although if the inductors have a 5% or higher tolerance (website doesn't specify) they will be within the calculated value.
I used the cabinet from a cheap 1950 Silvertone console that wasn't worth restoring (except the V-M Tri-O-Matic record changer which will go in my Magnavox CR-198C console) as the cabinet for the speaker. The only good thing about the cabinet is the piece of plywood I used for the speaker baffle.
The stock driver configuration of the Emerson console had the woofer in one spot and the three tweeters arranged in a triangle shape at an angle from the woofer. I arranged them as seen here. The extra hole was when I had a pair of dome tweeters on either side of the top tweeter. Didn't like how it sounded as it wasn't very good up close to the speaker given the distance between the two domes.
So I found a ribbon tweeter I had and used that which improved the sound.
I made a 6dB/octave crossover for the speaker as that's what was originally used except the woofer didn't have any inductor in series with it. The reason for the added tweeter crossed over so high is I didn't like how the cone tweeters didn't reproduce the treble up to 16kHz that good.
The plan is to eventually have a new proper open back cabinet built and make it to where the ribbon tweeter mounts above the top cone tweeter.
Cannot decide if I want to paint it black like the baffle is or have it properly finished like the rest of the cabinet is.
Should I keep the 6dB/octave crossover or switch to a 12 dB/octave crossover as in will it make the speaker sound better or is it best to keep the stock crossover slope?
EDIT:
Here's a 12dB/octave crossover I've calculated. The .16mH, .32mH and .12mH inductors I could only get .15mH, .13mH and .33mH, however if I go with this crossover I will measure those and see how close they are to the calculated values, although if the inductors have a 5% or higher tolerance (website doesn't specify) they will be within the calculated value.
Last edited:
I'm concerned that you have each of the two inductors attached directly to a steel driver chassis and, to boot, right next to the magnet assembly.
Such close proximity to both is sure to alter the inductance of the inductor.
The ribbon tweeter would certainly benefit from having an LC filter rather than just a C.
Such close proximity to both is sure to alter the inductance of the inductor.
The ribbon tweeter would certainly benefit from having an LC filter rather than just a C.
Yes that may indeed affect the inductance.
In the past I've measured an inductor where it would be mounted on a speaker and the inductance either didn't change enough to matter or I compensated for the inductance change by altering the inductor value, but I forget which I did as it was a good while ago.
I might get one of these boards and build the crossover on it.
https://www.parts-express.com/Crossover-PC-Board-3-Way-12-dB-Large-5-x-9-260-134?quantity=1
What I don't understand is given a 12 dB/octave crossover inverts the phase, why do crossover boards like this one label the + terminals for the woofer midrange and tweeter like they do instead of labeling those the - terminals?
I'll also add these terminals.
https://www.parts-express.com/Dayton-Audio-BPA-38G-HD-Binding-Post-Pair-Gold-091-1245?quantity=1
In the past I've measured an inductor where it would be mounted on a speaker and the inductance either didn't change enough to matter or I compensated for the inductance change by altering the inductor value, but I forget which I did as it was a good while ago.
I might get one of these boards and build the crossover on it.
https://www.parts-express.com/Crossover-PC-Board-3-Way-12-dB-Large-5-x-9-260-134?quantity=1
What I don't understand is given a 12 dB/octave crossover inverts the phase, why do crossover boards like this one label the + terminals for the woofer midrange and tweeter like they do instead of labeling those the - terminals?
I'll also add these terminals.
https://www.parts-express.com/Dayton-Audio-BPA-38G-HD-Binding-Post-Pair-Gold-091-1245?quantity=1
Last edited:
Here's a VituixCAD model simulating the situation in an idealized manner. Driver D2 is connected with inverted polarity. The filters are of the 2nd-order Linkwitz–Riley topology.
To get the correct summation, both the magnitude and phase responses must match up appropriately. The above model produces the following results, consisting of a perfectly flat summed response.
If we connect both drivers with positive polarity, then we get the following results. This connection produces a deep null at the crossover frequency.
To get the correct summation, both the magnitude and phase responses must match up appropriately. The above model produces the following results, consisting of a perfectly flat summed response.
If we connect both drivers with positive polarity, then we get the following results. This connection produces a deep null at the crossover frequency.
Adding on to the previous post, I wanted to explore what a Bessel filter topology could produce. The VituixCAD model that I came up with is as follows. Not that the low-pass filter's cut-off frequency was computed using 1000/sqrt(2), and the high-pass filter's cut-off frequency was computed using 1000*sqrt(2), where sqrt(2) = 1.4142.
The summed response is quite flat, with just a 0.2dB peak at the crossover frequency.
The key to this behaviour is that the filtered phase responses of Drivers D1 and D2 are almost in phase through the crossover region.
The summed response is quite flat, with just a 0.2dB peak at the crossover frequency.
The key to this behaviour is that the filtered phase responses of Drivers D1 and D2 are almost in phase through the crossover region.
Since you asked me, here's my answer.
In the (large) attachment, note that the midrange driver's positive terminal is connected to the negative input terminal - that's reverse polarity.
Attachments
I've just realised that you said "why" and not "what". Silly old me!
The phase shift in a second order crossover is 180 degrees and it is worth while to ensure that the bass and midrange drivers are acoustically in phase (by anti-phase connection) to avoid a dip in response around the frequency of crossover.
The phase shift in a second order crossover is 180 degrees and it is worth while to ensure that the bass and midrange drivers are acoustically in phase (by anti-phase connection) to avoid a dip in response around the frequency of crossover.
So the woofer with its 12dB crossover has its phase shifted by 180 degrees, right?
The midrange with its bandpass has a 180 degree phase shift due to the high pass portion of the bandpass, but does it also have a 180 degree phase reversal due to the low pass of the bandpass filter? If so why is that?
The midrange with its bandpass has a 180 degree phase shift due to the high pass portion of the bandpass, but does it also have a 180 degree phase reversal due to the low pass of the bandpass filter? If so why is that?
The second order (12 dB/octave) crossover shifts the phase of each driver by 90 degrees such that both drivers end up 180 degrees out of phase.
At the crossover frequency both driver cones will be moving in opposite directions and will cancel each other out, leading to a large dip in the frequency response around the crossover point.
Reversing the electrical polarity of the midrange (+ and - connections are reversed) will reduce the severity of the dip. That's about it explanation wise!
At the crossover frequency both driver cones will be moving in opposite directions and will cancel each other out, leading to a large dip in the frequency response around the crossover point.
Reversing the electrical polarity of the midrange (+ and - connections are reversed) will reduce the severity of the dip. That's about it explanation wise!
Where I get confused is the bass driver is 180 degrees out of phase due to the 12dB low pass and the midrange is 180 degrees out of phase due to the high pass portion of the bandpass.
So given both crossover sections connect to the input terminals of the speaker, wouldn't that make them actually in phase with each other given each driver is phase shifted by 180 degrees due to the individual crossovers?
So given both crossover sections connect to the input terminals of the speaker, wouldn't that make them actually in phase with each other given each driver is phase shifted by 180 degrees due to the individual crossovers?
I always like to know why something is the way it is.
The part that confuses me is both the woofer and midrange have a phase inversion and both sections of the crossover connect to the + and - input of the crossover. So why does one have to be inverted if both have the same phase shift?
I can understand reversing the phase of the midrange or woofer if there's only a low pass on the woofer.
The part that confuses me is both the woofer and midrange have a phase inversion and both sections of the crossover connect to the + and - input of the crossover. So why does one have to be inverted if both have the same phase shift?
I can understand reversing the phase of the midrange or woofer if there's only a low pass on the woofer.
Parts have been ordered.
Will update when the parts get here.
I would absolutely love to ditch the ribbon tweeter, however I know the stock tweeters do not play good up to 16kHz.
The intent is to use the terminals with push on connectors. That way I can easily reverse the phase of any driver easily.
Will update when the parts get here.
I would absolutely love to ditch the ribbon tweeter, however I know the stock tweeters do not play good up to 16kHz.
The intent is to use the terminals with push on connectors. That way I can easily reverse the phase of any driver easily.
Here's the crossover with the actual values of the parts I was able to get without having to do things such as use several caps in parallel and use larger inductors and unwind them to make the right value.
I'm pretty sure the tolerances of the caps would more than make up for the .02uF difference for the 2.35uF cap and the .01uF difference for the 1.32uF cap and the tolerance of the inductors (tolerance was not mentioned on the webpage for the inductors) more than makes up for the 0.1mH difference on the .12mH and .32mH inductors.
Now had this been a proper speaker build for true stereo HI-FI reproduction, I would have made the crossovers with tighter tolerance parts and adjusted the frequencies until I got a standard cap value and altered the inductors to be the proper value.
I had intended to measure the inductors to see how close to the listed value they were, but I forgot to do it.
Here's the crossover board.
Here's it in the speaker.
The speaker sounds a bit better than it did with the 6dB/octave crossover, however the Sony preamp I am using (temporarily replacing my Schiit Saga S preamp which I use at work until I find something better for there) messes with the sound some and the upper treble doesn't seem as good as it was with the Schiit preamp which I ran in passive mode.
I'm pretty sure the tolerances of the caps would more than make up for the .02uF difference for the 2.35uF cap and the .01uF difference for the 1.32uF cap and the tolerance of the inductors (tolerance was not mentioned on the webpage for the inductors) more than makes up for the 0.1mH difference on the .12mH and .32mH inductors.
Now had this been a proper speaker build for true stereo HI-FI reproduction, I would have made the crossovers with tighter tolerance parts and adjusted the frequencies until I got a standard cap value and altered the inductors to be the proper value.
I had intended to measure the inductors to see how close to the listed value they were, but I forgot to do it.
Here's the crossover board.
Here's it in the speaker.
The speaker sounds a bit better than it did with the 6dB/octave crossover, however the Sony preamp I am using (temporarily replacing my Schiit Saga S preamp which I use at work until I find something better for there) messes with the sound some and the upper treble doesn't seem as good as it was with the Schiit preamp which I ran in passive mode.
I'm unclear regarding which standard values of components you chose to physically populate your board.
For example, the 2.33 uF shown in your schematic is not a standard value, the nearest being 2.2 uF.
A tolerance of +/- 5% gives you a fair amount of leeway when choosing which standard value of capacitor to substitute for the theoretical one.
For the 2.33uF I used a 2uF and .33uF in parallel.
For the 1.33uF I used a 1uF and .33uF in parallel.
I used to try and go for the exact value using several caps until I got the DATS V3 and could select a frequency for a cap value I could make up using one or two caps and just alter the inductance value instead since I can now measure inductance.
For the 1.33uF I used a 1uF and .33uF in parallel.
I used to try and go for the exact value using several caps until I got the DATS V3 and could select a frequency for a cap value I could make up using one or two caps and just alter the inductance value instead since I can now measure inductance.
The physical Z offset of the drivers figures into all of this also. Most textbook crossovers assume the acoustic origins of all drivers are in the same plane (or at the same point), but this is rarely the case in real life on a flat baffle with conventional mounting.