Hi Cogitech,
1. Can you post at least the tweeter's specs? It would be interesting to see the attenuation at the tweeter's Fs.
2. Just curious, what do you like about this tweeter? Have you heard a good ribbon tweeter?
3. 1st order parallel is a classy choice. Are you familiar with 1st order series? It has some unique advantages / tradeoffs.
1. Can you post at least the tweeter's specs? It would be interesting to see the attenuation at the tweeter's Fs.
2. Just curious, what do you like about this tweeter? Have you heard a good ribbon tweeter?
3. 1st order parallel is a classy choice. Are you familiar with 1st order series? It has some unique advantages / tradeoffs.
Not to sound like Debbie Downer, but probably not. I take it those xo values from an online calculator or something? That tweeter will likeling ring at Fs (right around 2000hz I suspect). 2nd order would protect it better. Also, if there is a phase mismatch, which is quite likely with a dual woofer arrangement, you could end up with a null at the xo point. The acoustic center offset between the woofer and tweeter are very close to the half wavelength at 6000hz.
I've worked with the larger version of that tweeter (nd28f-6). It sounds great
I've worked with the larger version of that tweeter (nd28f-6). It sounds great
http://www.solen.ca/pdf/dayton/ND16FA-6.pdf
A) A few people at Parts-Express have raved about it being as good as a ribbon, no ringing, etc.
B) I love the form factor. For my application (retro-fit) this is one of the only tweeters that will work. Drill a hole and shove it in
C) The price is right.
I chose this type of crossover because it requires the fewest components, which means 2 things to me; cheap and easy. I am actually not familiar with any crossovers, as I am a newbie, and chose FR for simplicity initially.
Not to sound like Debbie Downer, but probably not. I take it those xo values from an online calculator or something? That tweeter will likeling ring at Fs (right around 2000hz I suspect). 2nd order would protect it better. Also, if there is a phase mismatch, which is quite likely with a dual woofer arrangement, you could end up with a null at the xo point. The acoustic center offset between the woofer and tweeter are very close to the half wavelength at 6000hz.
I've worked with the larger version of that tweeter (nd28f-6). It sounds great
Yep, online calculator. I wanted to aim for a very high crossover because I love the presence and tone of my EL70s and I really only thought some shine at the very high end might be nice.
If I crossover at 6k, how will it ring at Fs (2283hz) ?
Forgive my ignorance, but I don't understand the third-last sentence (last sentence of first paragraph, to be precise). Is that a bad thing, or a good thing?
Last edited:
Hi cogitech, 1st order series uses the same number of parts as 1st order parallel, just wired differently. A lot of people feel 1st order sounds best when it's an option.
An oversimplification but whereas parallel 1st order consists of two "separate modules," (a high pass and low pass, the values of which you might choose more or less independently), series crossover has an interesting property in that the cap and coil work together to split the signal "sort of perfectly." I've not stated it well but that's the basic idea. I think it's even simpler than the crossover you've suggested (same number of parts but potentially more tolerance and potentially easier to dial in).
The trick is that the drivers will have significant overlap as 1st order, and while that has good phase characteristics (which are -totally- audible in my limited experience), the shallow attenuation means you want to cross well away from the tweeter's Fs. And well away from any midrange / treble problems that the fullrangers might have on their top end (probably not a consideration with widebanders but a consideration for woofers of larger size).
I'm definitely no expert but your project is similar to something I'm fiddling with.
Interesting. Thanks for taking the time to explain that to me. It does sound as if series 1st order would be better. 6k crossover point is quite far from the tweeter Fs (2283hz), so I suppose I'm OK there. The EL70s start to roll off very, very gently after about 5k, which is why I chose 6k to bring in the tweeter. I don't think the EL70s have any real "problems" in that upper range... In fact, they are actually really quite nice on their own, just rolled off in a classy, "warm", "vintage" sort of way.
Perhaps an L-pad would be handy for the tweet, so I can choose just how much "air" I want.
The other option is to just stick with the EQ or "Treble" knob and leave well enough alone. I'm just a bit of a tinkerer and need something to tinker with, or at least think about tinkering with.
Last edited:
My two cents:
My two cents: It looks way too risky as far speaker building goes; talking from experience and have old double woofer boxes around that fit this idea.
Take a look at Madisound's HDS kit, they have been there since 1972 and the fellows there designed this.
I've built 10 so far, and using one for my center channel. It will take a couple of months of break in period, but once it get going, it's remarkable. The other 9 are spread out among friends, they think they are great and are spreading the word around.
HDS Home Theater Kit - Parts Only: Madisound Speaker Store
My two cents: It looks way too risky as far speaker building goes; talking from experience and have old double woofer boxes around that fit this idea.
Take a look at Madisound's HDS kit, they have been there since 1972 and the fellows there designed this.
I've built 10 so far, and using one for my center channel. It will take a couple of months of break in period, but once it get going, it's remarkable. The other 9 are spread out among friends, they think they are great and are spreading the word around.
HDS Home Theater Kit - Parts Only: Madisound Speaker Store
Attachments
Hi cogitech,
But with a first-order crossover, the attenuation is only 6db/octave so 6k is actually very _close_ to Fs. Let's say it's an octave and a half (1.5k is two octaves away from 6k). So it will be attenuated 6db at 3k, 12db at 1.5k. So let's call it roughly 9db at Fs.
That's probably not enough attenuation and so for a first order crossover, for that particular tweeter, you probably want a higher crossover point (or alternatively, a steeper slope but there goes all the fun!)
Air in my experience is around 12k. Don't get discouraged, you are on a good and exciting (and affordable) DIY path with first order crossovers. The quality of sound is in keeping with single driver principles, especially the series variety.
Hi ODougbo,
There are lots of ways to deal with double woofers though. What's your concern? Lobing? Are you saying go with something traditionally used in a D'Appolito configuration? If I understand the OP, he just needs some sparkle in the top two octaves so this is still a widebander / supertweeter project in my (not expert!) opinion.
If lobing is a worry, one (perhaps workable) option would be a bipole, with the tweeter on one side, and then you could decide whether it should face forward or backward (or who knows, upward).
There are lots of ways to deal with double woofers though. What's your concern? Lobing? Are you saying go with something traditionally used in a D'Appolito configuration? If I understand the OP, he just needs some sparkle in the top two octaves so this is still a widebander / supertweeter project in my (not expert!) opinion.
If lobing is a worry, one (perhaps workable) option would be a bipole, with the tweeter on one side, and then you could decide whether it should face forward or backward (or who knows, upward).
Someone already sorta answered your question, but here goes. Like they said, Fs is really only a couple octaves away at best. If the slope was 4th order or something, that would be fine. But in this case, at 6db per octave, it isn't much. On top of that, there will be an impedance peak right at Fs that'll reduce the effectiveness of that cap, resulting in even less attenuation. On top of all that, excusion will be increasing as frequency reduces at 12db/oct, yet the cap is only reducing by 6db/oct.
Do you have the tweeter already? If so, place some tones through it at 1/2 watt or so. Every tweeter is a little different, but you should be able to hear what I'm talking about. I recently did this with an SB29RDCN-004. Fs is at 550hz on that tweeter and at a mere 1/4watt or less, it went zingggg. Oh it sounded gross. I've got that thing so the transfer function is -30db at Fs. I felt that was necessary as a design decision.
The AC offset is a little more difficult to explain. The sound comes from the back of the cone on the EL70 and the back of the dome on the tweeter. If both are flush mounted on the baffle, the sound is coming from about 15 or so millimeters difference between the two (15mm is a guess, you'd have to measure to know). Well as the sound waves leave the baffle, they'll be out of sync with each other by 15mm. That's half a 30mm wavelength (11300hz). Half of that is pretty close to 6000hz, and twice that is out of the hearing range. So there may be nulls around there. I could go on about this, and diagrams help a lot. Do a google search of acoustic center offset if you need to know more.
This is my theory, take it or leave it. But I'm not a XO minimalist. To me, as soon as you add a XO, you've done the damage. Especially a high XO like 6khz. I'd add components to make it do it's job right, rather than more wrong than without, just to stay purist. I try and eliminate components as I can, but if I need 4 tweeter components to make it cross well, better than making it cross poorly. However, as you add components, xo design complexity steps up and design software and measurement is required. YMMV.
Do you have the tweeter already? If so, place some tones through it at 1/2 watt or so. Every tweeter is a little different, but you should be able to hear what I'm talking about. I recently did this with an SB29RDCN-004. Fs is at 550hz on that tweeter and at a mere 1/4watt or less, it went zingggg. Oh it sounded gross. I've got that thing so the transfer function is -30db at Fs. I felt that was necessary as a design decision.
The AC offset is a little more difficult to explain. The sound comes from the back of the cone on the EL70 and the back of the dome on the tweeter. If both are flush mounted on the baffle, the sound is coming from about 15 or so millimeters difference between the two (15mm is a guess, you'd have to measure to know). Well as the sound waves leave the baffle, they'll be out of sync with each other by 15mm. That's half a 30mm wavelength (11300hz). Half of that is pretty close to 6000hz, and twice that is out of the hearing range. So there may be nulls around there. I could go on about this, and diagrams help a lot. Do a google search of acoustic center offset if you need to know more.
This is my theory, take it or leave it. But I'm not a XO minimalist. To me, as soon as you add a XO, you've done the damage. Especially a high XO like 6khz. I'd add components to make it do it's job right, rather than more wrong than without, just to stay purist. I try and eliminate components as I can, but if I need 4 tweeter components to make it cross well, better than making it cross poorly. However, as you add components, xo design complexity steps up and design software and measurement is required. YMMV.
It isn't really a "dual woofer" system. Unfortunately the online tool that I used to design the crossover called them "woofers" and I failed to edit that out of the drawing.
This would more accurately be called "dual full range with tweeter helper".
This is also not a "new build". I already have the dual EL70s in microTowers ("Castle style, with one driver on the front and one driver on the top). That is why this tweeter in particular interests me, because I can just drill a hole in the upper corner of the front baffle and shove the tweeter in it. There's not much room for any other type of tweeter, unless I mount it lower on the tower (which I do not want to do).
Thank you for your perspective on this. I will do some more reading and research before making any decisions or purchases (I have not bought the tweeters yet). In the meantime, would a proper second-order crossed at 12000 be a good choice, do you think?
Ive never done it, but a popular way to add tweeter air to a full range setup, is to not filter the woofer and just cap the tweeter. You could try that with a very small cap (1uF). But like I say, I've never done it, so who knows. That would protect the tweeter better though, and the electrical phase difference may fix the AC offset. Otherwise, yes id go 2nd order, even still down around 6khz. That extra component helps a lot.
Time alignment would be great / necessary if this were a traditional two-way but with the double widebanders physically at 90 degrees, I think it could be a non-issue unless the widebanders are reconfigured for 1.5 duty.
If you wanted to try the simplest thing that could possibly work, you might consider:
* a much higher first-order crossover (or as tuxedocivic says, just a cap but aim for a higher high pass)
* a contour filter to rebalance the widebander response
* computer-based digital EQ (it's free) just to nail down what sounds better
I would avoid cutting into the cabs though, because you might have to try lots and lots of tweeters to find the one you can live with. Just one person's random opinion though!
If you wanted to try the simplest thing that could possibly work, you might consider:
* a much higher first-order crossover (or as tuxedocivic says, just a cap but aim for a higher high pass)
* a contour filter to rebalance the widebander response
* computer-based digital EQ (it's free) just to nail down what sounds better
I would avoid cutting into the cabs though, because you might have to try lots and lots of tweeters to find the one you can live with. Just one person's random opinion though!
Last edited:
How would having 2 widebanders at 90 degrees from each other make it a non-issue?
Hi tuxedocivic,
How can we align the widebanders to each other in such an arrangement (and then time-align the tweeter to boot)? One driver is literally on top of the other, so it's not a question of a simple one-dimensional offset.
But let's assume the first widebander is time-aligned to the tweeter, which as you suggest, is a good idea. So now how would you time-align those to the second upward-firing widebander which is running fullrange? The null we sought to avoid with the first widebander will appear thanks to the second widebander, unless the second widebander is a .5.
But I could be wrong! Wouldn't be the first time...
How can we align the widebanders to each other in such an arrangement (and then time-align the tweeter to boot)? One driver is literally on top of the other, so it's not a question of a simple one-dimensional offset.
But let's assume the first widebander is time-aligned to the tweeter, which as you suggest, is a good idea. So now how would you time-align those to the second upward-firing widebander which is running fullrange? The null we sought to avoid with the first widebander will appear thanks to the second widebander, unless the second widebander is a .5.
But I could be wrong! Wouldn't be the first time...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.