I have built a 3-way speaker consisting of two 8" Peerless woofers, two Vifa NE123W-08 midranges, and a Satori TW29RN-8 tweeter. The woofers are in their own bass bins, with mids and tweeter in separate enclosures. The woofer is crossed actively, the mids and tweeters passively with this professionally designed xover. The designer is no longer available to assist me, so I ask for your suggestions on where to start tweaking this system.
The speakers are rather dull and congested, with lifeless upper midrange. The soundstage is adequate, but voices seem to stay inside the enclosure and are somewhat muffled. When I first purchased them, I ran the midranges full range by themselves, with the same active xover on the woofers. Even without a tweeter, the system was lively and present, perhaps a bit too bright and raggedy on top. Now it seems that the upper midrange is missing. I've tried reducing the tweeter pad, but that doesn't change much other than making music sound etched and screechy. I think the problem lies in the upper response of the midrange, but I don't know what to try. I have also tried swapping tweeter leads in case the phase was inverted, but, as expected, that makes it worse. The xover components are very high quality; no corners cut there. The mids/tweeter amp is superb, the wiring is also high quality.
The midrange drivers are wired in parallel. I am wired to take action if given some direction.
Peace,
Tom E
The speakers are rather dull and congested, with lifeless upper midrange. The soundstage is adequate, but voices seem to stay inside the enclosure and are somewhat muffled. When I first purchased them, I ran the midranges full range by themselves, with the same active xover on the woofers. Even without a tweeter, the system was lively and present, perhaps a bit too bright and raggedy on top. Now it seems that the upper midrange is missing. I've tried reducing the tweeter pad, but that doesn't change much other than making music sound etched and screechy. I think the problem lies in the upper response of the midrange, but I don't know what to try. I have also tried swapping tweeter leads in case the phase was inverted, but, as expected, that makes it worse. The xover components are very high quality; no corners cut there. The mids/tweeter amp is superb, the wiring is also high quality.
The midrange drivers are wired in parallel. I am wired to take action if given some direction.
Peace,
Tom E
I'm thinking a place to start is with global EQ, rather than touching the crossover. This will help in some ways, give you information in others and will prevent you creating new problems at the same time.
I'm not thinking the crossover is optimum, but it is probably OK... and improving on it will take some work.
I'm not thinking the crossover is optimum, but it is probably OK... and improving on it will take some work.
Have you measured the frequency response of axis? It can be very helpful to see what's happening at 15, 30 and 45 degrees horizontally off axis. Are you listening on the design axis of the speaker? You can determine this by measuring the frequency response vertically on/off axis.
Make sure the polarity on the mids is the same, especially if they are 2x per cab. Then try flipping all of them to the opposite polarity. It could be that the absolute phase is out at the xover point.
Later,
Wolf
Later,
Wolf
Thanks for comments so far.
I am not at all interested in band-aid solutions such as EQ, global or otherwise. These are excellent drivers, housed in carefully constructed enclosures, driven by superb amplifiers.
I have not measured anything, on or off axis. I do not need a measurement to determine that something is wrong. I have other speakers, cheaper drivers in flimsier enclosures that sound much clearer.
I am certain that the midranges are all in the same phase. I have already tried swapping between woofers and mids, and mids and tweeter, but it only makes the sound worse, so I suspect, and the simulation graph shows, that there is a deep null at the xover point. I guess that indicates that phase is not the problem.
Pete, can you give a little more info about your comment? I know what you mean by order, but what do you mean by lower? If this is 3rd order, do you mean 2nd instead? Can you provide component values or give me some hints? That's going to change A LOT of the speaker characteristics, but I'm willing to try it with some guidance. The xovers are external and hard-wired, so somewhat easy to work on.
Again, thanks to all. I have put a lot of time and more than few bucks into these speakers, and I am getting frustrated that my SEAS co-axials in the kitchen, driven by a cheap Class D amp, have more apparent resolution than this tweaked, bi-amped, complex, expensive main system.
Peace,
Tom E
I am not at all interested in band-aid solutions such as EQ, global or otherwise. These are excellent drivers, housed in carefully constructed enclosures, driven by superb amplifiers.
I have not measured anything, on or off axis. I do not need a measurement to determine that something is wrong. I have other speakers, cheaper drivers in flimsier enclosures that sound much clearer.
I am certain that the midranges are all in the same phase. I have already tried swapping between woofers and mids, and mids and tweeter, but it only makes the sound worse, so I suspect, and the simulation graph shows, that there is a deep null at the xover point. I guess that indicates that phase is not the problem.
Pete, can you give a little more info about your comment? I know what you mean by order, but what do you mean by lower? If this is 3rd order, do you mean 2nd instead? Can you provide component values or give me some hints? That's going to change A LOT of the speaker characteristics, but I'm willing to try it with some guidance. The xovers are external and hard-wired, so somewhat easy to work on.
Again, thanks to all. I have put a lot of time and more than few bucks into these speakers, and I am getting frustrated that my SEAS co-axials in the kitchen, driven by a cheap Class D amp, have more apparent resolution than this tweaked, bi-amped, complex, expensive main system.
Peace,
Tom E
Tom, the frequency response of this driver is flat enough, and the breakup peak far enough away from the crossover point, that a lower order crossover should suffice. My starting point for these would be 1st order, but you may prefer 2nd. I usually make the tweeter xover one order higher & reverse polarity for best phase alignment, YMMV. As to parts values, they're easy enough to calculate using the impedance values from the spec sheet, fortunately these drivers have fairly flat impedance, looks like Zmin 7.26 ohms is the value you should use for your calcs. (I usually run my own impedance plots with ARTA to get higher resolution than most spec sheets).
Just a thought - 2x 87dB midwoofers in parallel should give 93dB, same as the tweeter - yet it seems there is ~ 6dB padding on the tweeters???
This seems the easiest thing to try - just short out the 3.3 ohm resistors & see how it sounds
This seems the easiest thing to try - just short out the 3.3 ohm resistors & see how it sounds
Are the frequency plots in Leap (attached) measurements of the real system (i.e. not simulations)?
The page three plot is a classic case of tweeter phase needing to be reversed relative to the mids. When reversed the result should be the plot on page 2.
The page three plot is a classic case of tweeter phase needing to be reversed relative to the mids. When reversed the result should be the plot on page 2.
Couldn't disagree more. Acoustics dominates the drivers. EQ (a target) is always necessary.
But you do need measurements to see what exactly is wrong.
Without measurements it's like working in the dark.
The Madisound design is for a single woofer and tweeter on page 4...
If I know Madisound, for parallel twin woofers, they would have written NE123W X 2 on page 4 for parallel woofers. Something is in error here on the crossover design. As PeteMck says, you need more level on the tweeter.
Or less level on the midbasses. Which will need a fairly straightforward bass crossover adjustment when wiring series. Once you figure out the error.
Last edited:
Steve, note that the heading on the other pages DOES indicate "NE123W x 2". Only the page that shows the xover schematic does not refer to two mids. I also began to suspect that something was amiss with the schematic, and the actual design might be for only a single mid driver instead of two in parallel.
I have already tried to reduce the tweeter pad in gradual steps, but that seems to make the speaker sound bright and etched, but does not give better upper midrange. I suspect the tweeter might not be the problem, but perhaps the mids roll off too soon or too sharply.
But, based on advice given here, I will try again to fuss with the tweeter pad.
Pete, you advise to short 3.3 ohm resistors, and I understand your reasoning for that. BOTH of them in each xover, or which one?
Peace,
Tom E
I have already tried to reduce the tweeter pad in gradual steps, but that seems to make the speaker sound bright and etched, but does not give better upper midrange. I suspect the tweeter might not be the problem, but perhaps the mids roll off too soon or too sharply.
But, based on advice given here, I will try again to fuss with the tweeter pad.
Pete, you advise to short 3.3 ohm resistors, and I understand your reasoning for that. BOTH of them in each xover, or which one?
Peace,
Tom E
Maybe I should disconnect one of the mids to see if the damn thing plays better? Then it would be a simple matter of adjusting the existing midrange xover components for four ohms instead of eight?
I just ran this circuit through my simulator. I was struggling to find close equivalents, but certain things stood out.
The bass circuit looks like a 4 ohm circuit and worked well enough though I got a -6dB point nearer 4kHz with two parallel 8 ohm drivers.
The tweeter circuit was reasonable and phase matched, but too low in level based on the specs I saw.
Removing resistors is one way to get it louder, but I was concerned about too low an impedance.
Turns out if you remove the 3.3R in front of the tweeter filter you get a well-behaved 4dB gain and about a flat 6 ohms. Further removing the attenuator after the reactive components drops load impedance to 4 ohms, changes the slope, and makes the filter uncomfortably bright at the top end. TBH, it's a bit hard to see what 47R achieves. Might as well not be there IMO.
I have a feeling the bass filter is OK. The tweeter filter needs a rework IMO.
The bass circuit looks like a 4 ohm circuit and worked well enough though I got a -6dB point nearer 4kHz with two parallel 8 ohm drivers.
The tweeter circuit was reasonable and phase matched, but too low in level based on the specs I saw.
Removing resistors is one way to get it louder, but I was concerned about too low an impedance.
Turns out if you remove the 3.3R in front of the tweeter filter you get a well-behaved 4dB gain and about a flat 6 ohms. Further removing the attenuator after the reactive components drops load impedance to 4 ohms, changes the slope, and makes the filter uncomfortably bright at the top end. TBH, it's a bit hard to see what 47R achieves. Might as well not be there IMO.
I have a feeling the bass filter is OK. The tweeter filter needs a rework IMO.
Thanks so much, Steve.
Attached is the simulated impedance curve provided by Madisound. It was provided separately. I think they use a general template for their xover designs and plug in different values. I have gotten several from them, and they always put a resistor parallel to the tweeter. I'm not sure what purpose it serves.
Attached is the simulated impedance curve provided by Madisound. It was provided separately. I think they use a general template for their xover designs and plug in different values. I have gotten several from them, and they always put a resistor parallel to the tweeter. I'm not sure what purpose it serves.
re:' BOTH of them in each xover, or which one?': try Steve's suggestion first ' remove the 3.3R in front of the tweeter filter" & tell us how it sounds to you.
"47R Might as well not be there IMO"
In this case it's too high to do anything much
"47R Might as well not be there IMO"
In this case it's too high to do anything much
+1, I happily listen to the final result but I have to measure my way there...
+1, I had my doubts before I got a DSP but I see no way back now when I finally got it. I don't use it for room correction (an area I still doubt) but its perfect to make a decent crossover turn excellent (ruler flat with any sane target line). I can also use it to extend bass on closed speakers, just add a Linkwitz transform. It also makes it easy to make your whole sound chain sound anyway you want. Too "forward", reprogram. Too "dark", reprogram. Too "warm", reprogram. I constantly believe I found the perfect balance but still find room for improvements. Just tweak, reprogram and enjoy the new sound, I love it!