A few weeks ago I bought some of the Parts Express Buyout onkyo drivers to fool around with. I built an MTM using 2 of the 4" "NSB" onkyo/pioneer drivers (PE# 269-570), and an Onkyo tweeter with a neo magnet (PE# 269-776).
For now they are taped into a cardboard box, but hopefully I'll have enclosures built this weekend.
Anyway... an explanation of my development process. After asking a bunch of dumb questions here and getting good feedback, I started to design an excel spreadsheet that would let me play with various crossover values. Here are the CLIO responses of the raw drivers:
As you can see, neither driver exactly has an ideal curve, nor does there seem to be an ideal crossover point.
One of the suggestions that I latched on to was one by Sreten that suggested that I choose individual low-pass and high-pass filter frequencies. That's what prompted me to design this spreadsheet. It allows me to do exactly that, and gives decent visual results.
I wound up settling on crossover points similar to his original suggestion. I'm running the woofers with a 1400hz low-pass 2nd order electrical, and I'm running the tweeter with a 4500hz first order electrical high pass. The approximate result can be seen here:
I wouldn't pay much attention to the red "summed" line... I don't trust it, but you can see what the two filtered responses look like individually.
Here's the crossover I constructed:
I don't have any measurement equipment, so I have no idea what the frequency response looks like. it sounds like the upper midrange is lacking, but since this driver is so harsh in the upper mid, I'm not sure that it's a bad thing. I think the problem is that the paired woofers are quite a bit more efficient than the tweeter, and since the tweeter has a bump at the very high frequencies, everything inbetween is getting lost. I may toss some resistors in line with the mids to bring them down a few dB, and remove the phase plug on the tweeters which will provide a much smoother high frequency response. I'm also going to fool with different component values so that I can begin to hear their individual effects on crossover frequencies & slopes.
But overall, I'm quite happy.
Now I'm going to move on to the vifa/jbl buyout mids, which should be a bit easier to work with.
For now they are taped into a cardboard box, but hopefully I'll have enclosures built this weekend.
Anyway... an explanation of my development process. After asking a bunch of dumb questions here and getting good feedback, I started to design an excel spreadsheet that would let me play with various crossover values. Here are the CLIO responses of the raw drivers:
An externally hosted image should be here but it was not working when we last tested it.
As you can see, neither driver exactly has an ideal curve, nor does there seem to be an ideal crossover point.
One of the suggestions that I latched on to was one by Sreten that suggested that I choose individual low-pass and high-pass filter frequencies. That's what prompted me to design this spreadsheet. It allows me to do exactly that, and gives decent visual results.
I wound up settling on crossover points similar to his original suggestion. I'm running the woofers with a 1400hz low-pass 2nd order electrical, and I'm running the tweeter with a 4500hz first order electrical high pass. The approximate result can be seen here:
An externally hosted image should be here but it was not working when we last tested it.
I wouldn't pay much attention to the red "summed" line... I don't trust it, but you can see what the two filtered responses look like individually.
Here's the crossover I constructed:
An externally hosted image should be here but it was not working when we last tested it.
I don't have any measurement equipment, so I have no idea what the frequency response looks like. it sounds like the upper midrange is lacking, but since this driver is so harsh in the upper mid, I'm not sure that it's a bad thing. I think the problem is that the paired woofers are quite a bit more efficient than the tweeter, and since the tweeter has a bump at the very high frequencies, everything inbetween is getting lost. I may toss some resistors in line with the mids to bring them down a few dB, and remove the phase plug on the tweeters which will provide a much smoother high frequency response. I'm also going to fool with different component values so that I can begin to hear their individual effects on crossover frequencies & slopes.
But overall, I'm quite happy.
Now I'm going to move on to the vifa/jbl buyout mids, which should be a bit easier to work with.
Hi!
Cool stuff, how much do all the drivers and other parts cost altogether?
Now, I'm no expert so please don't shoot me down... but, I think your lacking midrange could be due to phase difference between the two drivers - you probably ought to have both drivers 2nd order, or just the tweeter 2nd order and bass driver full range (as I have currently) connected with inverse polarity.
I used my speakers FR on the bass and 1st order on the tweeter for a while, and when I added an inductor to the tweeter and adjusted the values I found a more open and pleasing midrange. I suspect one reason for this *could* be the better phase coherance at the crossover/overlap region.
Just an idea anyway. Thanks for sharing your project with us.
Cool stuff, how much do all the drivers and other parts cost altogether?
Now, I'm no expert so please don't shoot me down... but, I think your lacking midrange could be due to phase difference between the two drivers - you probably ought to have both drivers 2nd order, or just the tweeter 2nd order and bass driver full range (as I have currently) connected with inverse polarity.
I used my speakers FR on the bass and 1st order on the tweeter for a while, and when I added an inductor to the tweeter and adjusted the values I found a more open and pleasing midrange. I suspect one reason for this *could* be the better phase coherance at the crossover/overlap region.
Just an idea anyway. Thanks for sharing your project with us.
Jim85IROC said:I don't have any measurement equipment, so I have no idea what the frequency response looks like. it sounds like the upper midrange is lacking, but since this driver is so harsh in the upper mid, I'm not sure that it's a bad thing. I think the problem is that the paired woofers are quite a bit more efficient than the tweeter, and since the tweeter has a bump at the very high frequencies, everything inbetween is getting lost. I may toss some resistors in line with the mids to bring them down a few dB, and remove the phase plug on the tweeters which will provide a much smoother high frequency response. I'm also going to fool with different component values so that I can begin to hear their individual effects on crossover frequencies & slopes.
But overall, I'm quite happy.
Glad to be of some help, from what I remember one mid/bass unit
crossed over as a I suggested would ~ match the tweeter level.
I don't think I clouded the issue with baffle step correction.
So time to
. Build a 2.5 way. One of your bass units should be rolled in 1st order much lower, the other driver keep the crossover
the same so you need new values for the higher impedance.
The frequency of the 0.5 unit is generally held to be 130 /
baffle width, both in metres, I think its nearer 100 / width.
Try tweaking that. Certainly you should try inverting the phase
of the tweeter also, one connection is likely to sound better than
the other but there are cases when it doesn't matter so much.
Also you can consider a TMM for a 2.5 way. If you build a
MTM 2.5 way which way up it is will also affect the sound.
sreten.edit : P.S. which Vifa JBL buyout ? one of them is a nightmare !
SimontY:
The drivers cost $.87 each, and the crossover parts are a few bucks... maybe $5 total.
The phase difference could certainly be a part of it, and next time I get a chance to play, I'm going to try to reverse the polarity of the tweeter to see how it effects the sound. Since my spreadsheet is rather crude, it doesn't take phase issues into account.
I don't want to run the mid full range because it sounds terrible that way. There is a ton of breakup above 7khz that I want to filter out.
sreten:
I do have intentions to build a 2.5 way as well, but I want to tweak the MTM first. Since I'm doing this strictly as a learning experience, I'm going to try out as many combinations as I can.
The Vifa that I just bought is the 4.5" jbl/vifa jobber. PE # 299-495. Other than a massive dip above 1khz, it doesn't look all that bad. Maybe it'll give me an education on notch filters! I'm going to be using 2 of them in an isobarac sub enclosure that I designed to mate to my tangband full range speakers. I figured that at $6 each, I'd get 2 more to play with.
How much do you think keeping the MTM arrangement for the 2.5 way would effect the sound? I'd rather do this with one baffle if I can help it, and the tweeter is so small that the woofers are still spaced less than 6" apart, center to center. That's closer than a pair of 7" woofers in a typical 2.5 way system.
The drivers cost $.87 each, and the crossover parts are a few bucks... maybe $5 total.
The phase difference could certainly be a part of it, and next time I get a chance to play, I'm going to try to reverse the polarity of the tweeter to see how it effects the sound. Since my spreadsheet is rather crude, it doesn't take phase issues into account.
I don't want to run the mid full range because it sounds terrible that way. There is a ton of breakup above 7khz that I want to filter out.
sreten:
I do have intentions to build a 2.5 way as well, but I want to tweak the MTM first. Since I'm doing this strictly as a learning experience, I'm going to try out as many combinations as I can.
The Vifa that I just bought is the 4.5" jbl/vifa jobber. PE # 299-495. Other than a massive dip above 1khz, it doesn't look all that bad. Maybe it'll give me an education on notch filters! I'm going to be using 2 of them in an isobarac sub enclosure that I designed to mate to my tangband full range speakers. I figured that at $6 each, I'd get 2 more to play with.
How much do you think keeping the MTM arrangement for the 2.5 way would effect the sound? I'd rather do this with one baffle if I can help it, and the tweeter is so small that the woofers are still spaced less than 6" apart, center to center. That's closer than a pair of 7" woofers in a typical 2.5 way system.
Hi Streten,
I think your comments are a little more meaningful than mine were, lol.
In your opinion, is it reasonable to make a '2.5-way' with MTM positioned drivers? I can't see why it wouldnt be, and I am in an experimental sort of mood (currently using electronic BSC and 2-way MTM layout) Apologies for threadjacking, I can just never help myself!!
Ooh, one reason it'd work in some cases (mine) - if your seating position fits either the top or bottom mid better than it does the tweeter...
Thankyou
I think your comments are a little more meaningful than mine were, lol.
In your opinion, is it reasonable to make a '2.5-way' with MTM positioned drivers? I can't see why it wouldnt be, and I am in an experimental sort of mood
(currently using electronic BSC and 2-way MTM layout) Apologies for threadjacking, I can just never help myself!! Ooh, one reason it'd work in some cases (mine) - if your seating position fits either the top or bottom mid better than it does the tweeter...
Thankyou
Well, I wouldn't call this "good" sound. It's tolerable considering what the drivers cost. These mids offer minimal detail, and the highs are very bright.
I'm sure not replacing any of my other speakers with this little experiment, although I do plan to a/b it with my tangband full range speakers when I get those done.
However, I am very happy with my results considering that this is my very first crossover design. As I tweak it, I hope to be able to improve upon it. My only wish is that I was able to measure the frequency response so that I could correlate what I hear to what the response looks like.
I'm sure not replacing any of my other speakers with this little experiment, although I do plan to a/b it with my tangband full range speakers when I get those done.
However, I am very happy with my results considering that this is my very first crossover design. As I tweak it, I hope to be able to improve upon it. My only wish is that I was able to measure the frequency response so that I could correlate what I hear to what the response looks like.
Nothing "wrong" with a MTM 2.5 way, usually the top unit
is the midrange so T&M are inverted compared to TMM.
Same issues as a TM versus MT.
IMO if you want to tweak your MTM 2 way then you still
need to implement BSC by cutting the midrange back by
6DB by oversizing the series inductor and possibly adding
a parallel resistor to it.
Again you need to be 3dB down at the baffle frequency,
6dB in the midrange and then c/o as you currently do.
IMO 2.5 way is more elegant, but a BSC'd 2 way MTM
can take more juice in the midrange.
The vifa/JBL is a nightmare as a bass/mid unit, in a sub it will be
fine. It is however entirely unsuited to an isobaric sub and will
work much better reflexed in parallel. We are hardly in subwoofer
territory, 2 litres per driver tuned to 80Hz is about as good as it
gets, unless you drop the port tuning and add active bass boost,
going over 2 litres per driver is not sensible.
sreten.
is the midrange so T&M are inverted compared to TMM.
Same issues as a TM versus MT.
IMO if you want to tweak your MTM 2 way then you still
need to implement BSC by cutting the midrange back by
6DB by oversizing the series inductor and possibly adding
a parallel resistor to it.
Again you need to be 3dB down at the baffle frequency,
6dB in the midrange and then c/o as you currently do.
IMO 2.5 way is more elegant, but a BSC'd 2 way MTM
can take more juice in the midrange.
The vifa/JBL is a nightmare as a bass/mid unit, in a sub it will be
fine. It is however entirely unsuited to an isobaric sub and will
work much better reflexed in parallel. We are hardly in subwoofer
territory, 2 litres per driver tuned to 80Hz is about as good as it
gets, unless you drop the port tuning and add active bass boost,
going over 2 litres per driver is not sensible.
sreten.
Ugh, you make it sound so complicated. 
6db sounds like major overkill to me. I'm only using a little bit right now, too much killed the life totally, god only knows how much in db tho, hehe.
The basic idea though:
Slap an inductor in series with one mid/bass driver (at about 300hz or so), which then becomes the 0.5. Leave the other mid/bass to do everything it was doing already, but connect it in series with a small resistor to balance the sounds out.
Please confirm!!!
Thanks!
Jim85IROC - sorry if this is not relevant to your thread
6db sounds like major overkill to me. I'm only using a little bit right now, too much killed the life totally, god only knows how much in db tho, hehe.
The basic idea though:
Slap an inductor in series with one mid/bass driver (at about 300hz or so), which then becomes the 0.5. Leave the other mid/bass to do everything it was doing already, but connect it in series with a small resistor to balance the sounds out.
Please confirm!!!
Thanks!
Jim85IROC - sorry if this is not relevant to your thread
sreten,
what's so bad about the isobarac arrangement? The box size with the isobarac alignment was very small, which is exactly what I was aiming for with the design. I used an EBS alignment, which yielded a fairly flat response between 45 and 130hz, all in a 4.2 liter enclosure. Since I'm going to be running the sub passively (plus this is $$$$ driven), I don't think active equalization would be realistic.
I posted a question about cone excursion last week, and the resulting information indicates that I should be able to get 90dB of output above 50hz, which will be more than enough for the intended use.
Getting back to BSC... that's definately something I want to tackle, but I think that first I should hone my basic crossover skills before getting involved in BSC. These speakers have a very high qts, so their bass response isn't exactly going to be "lean" to begin with. Here are a couple alignments that I've modeled in WinISD. Box volume is 1 cubic foot (28 liters).
To get any flatter, I need to double the volume, which is already excessive for 2 4" drivers. With a narrow (5") front baffle, I suspect that this bloated response should provide reasonably adequate BSC as long as they stay relatively close to a wall.
SimontY:
Don't worry about if it's relevant or not. We're both learning here, and thankfully sreten has the patience to help us with our questions.
what's so bad about the isobarac arrangement? The box size with the isobarac alignment was very small, which is exactly what I was aiming for with the design. I used an EBS alignment, which yielded a fairly flat response between 45 and 130hz, all in a 4.2 liter enclosure. Since I'm going to be running the sub passively (plus this is $$$$ driven), I don't think active equalization would be realistic.
An externally hosted image should be here but it was not working when we last tested it.
I posted a question about cone excursion last week, and the resulting information indicates that I should be able to get 90dB of output above 50hz, which will be more than enough for the intended use.
Getting back to BSC... that's definately something I want to tackle, but I think that first I should hone my basic crossover skills before getting involved in BSC. These speakers have a very high qts, so their bass response isn't exactly going to be "lean" to begin with. Here are a couple alignments that I've modeled in WinISD. Box volume is 1 cubic foot (28 liters).
An externally hosted image should be here but it was not working when we last tested it.
To get any flatter, I need to double the volume, which is already excessive for 2 4" drivers. With a narrow (5") front baffle, I suspect that this bloated response should provide reasonably adequate BSC as long as they stay relatively close to a wall.
SimontY:
Don't worry about if it's relevant or not. We're both learning here, and thankfully sreten has the patience to help us with our questions.
Yeah 1 cuft is ridiculous for a 4" driver, even for 2,
You should be looking at a sealed box, reflexing will just give
you a boom box. Sealed box 0.3 cuft for 2 drivers Q = 1.
And you'll need BSC, has to be near 6dB or you won't match
the tweeter sensitivity, here you don't have much choice.
Build your subwoofer it its a learning exercise, with the box
volume far better optimisations in terms of efficiency and
extension are possible with suitable drivers.
An isobaric is -6dB in efficiency and maximum output.
No commercial manafacturer has ever sold a reflexed isobaric.
You'll find near identical performance possible with a single
driver at double the efficiency as impedance is double.
Also for a good match with the TB's you"ll need two V/J's in parallel.
sreten.
You should be looking at a sealed box, reflexing will just give
you a boom box. Sealed box 0.3 cuft for 2 drivers Q = 1.
And you'll need BSC, has to be near 6dB or you won't match
the tweeter sensitivity, here you don't have much choice.
Build your subwoofer it its a learning exercise, with the box
volume far better optimisations in terms of efficiency and
extension are possible with suitable drivers.
An isobaric is -6dB in efficiency and maximum output.
No commercial manafacturer has ever sold a reflexed isobaric.
You'll find near identical performance possible with a single
driver at double the efficiency as impedance is double.
Also for a good match with the TB's you"ll need two V/J's in parallel.
sreten.
I suppose building a sealed box would be smarter... I just was hoping to get usable bass out of these without a sub. But hey... these are "experimentation" speakers, so I'll try a few different configurations.
I'm still thinking about the sub, and it seems that with active equalization, doing a non-isobarac enclosure would make sense, but without having the ability to build the active filter makes it hard to get the low frequency extention without a huge box or going isobarac. I can understand why OEMs don't use an isobarac design. With their resources, building an active filter to compensate for a small enclosure is much less expensive than using a 2nd woofer.
But... I just went back and looked at my design. There was one thing I didn't notice until now... port length.
With a 2" port, I'm going to need a length of 26". That'll be hard to fit into an 8" cube.
With a 1" port I'll need a little over 6". A 1" port should make plenty of port noise!
I think I just bought 4 $6 paper weights.
I'm still thinking about the sub, and it seems that with active equalization, doing a non-isobarac enclosure would make sense, but without having the ability to build the active filter makes it hard to get the low frequency extention without a huge box or going isobarac. I can understand why OEMs don't use an isobarac design. With their resources, building an active filter to compensate for a small enclosure is much less expensive than using a 2nd woofer.
But... I just went back and looked at my design. There was one thing I didn't notice until now... port length.
With a 2" port, I'm going to need a length of 26". That'll be hard to fit into an 8" cube.
With a 1" port I'll need a little over 6". A 1" port should make plenty of port noise!
I think I just bought 4 $6 paper weights.
I didn't plan to bring this thread back from the dead, but now that it's being referenced in another "NSB" thread, I wanted to provide an update.
After Jeff B released his excel crossover program, I plugged all of my data into it to see what things looked like compared to my crude spreadsheet design. Surprisingly, my crossover modeled flat to +/- 4.5dB over the bandwidth range. Better than I thought, but I could also see why the mids were so "shouty". The 6dB/octave cap wasn't sufficiently attenuating the bump at the low end of the tweeter's spectrum.
I added a shunt .15mH inductor on the tweeter circuit, and also added a series 2 ohm resistor with the NSBs to flatten out a lower midrange bump. Of course, the 2 ohm resistor is going to raise the qts, so I may not have actually gained anything... we'll see. It'll be easy enough to remove if necessary.
I haven't listened to the new crossover yet (hopefully tonight), but here is a schematic. According to Jeff B's spreadsheet, this models flat +/- 2dB over the spectrum.
After Jeff B released his excel crossover program, I plugged all of my data into it to see what things looked like compared to my crude spreadsheet design. Surprisingly, my crossover modeled flat to +/- 4.5dB over the bandwidth range. Better than I thought, but I could also see why the mids were so "shouty". The 6dB/octave cap wasn't sufficiently attenuating the bump at the low end of the tweeter's spectrum.
I added a shunt .15mH inductor on the tweeter circuit, and also added a series 2 ohm resistor with the NSBs to flatten out a lower midrange bump. Of course, the 2 ohm resistor is going to raise the qts, so I may not have actually gained anything... we'll see. It'll be easy enough to remove if necessary.
I haven't listened to the new crossover yet (hopefully tonight), but here is a schematic. According to Jeff B's spreadsheet, this models flat +/- 2dB over the spectrum.
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.