For this "detour" design I have to hand a pair of potted Coral 5M-12, 8 ohm, 40W, 800hz-14kHz, plus a pair of Vifa D19TD-05-08, 8 ohm.
Why is it a fair analysis? I agree with the tweeter being not much of an issue with a high crossover point but not sure I agree with most of the rest.This seems to be a fair analysis; the last words just put... a last word onto the project 😱
There are a number of standard range woofers that are more suitable for this application than the expensive Scan-Speak which requires more volume than 40 litres to get tight bass. Several have been given in the thread.
The CSS has too limited a power handling/sensitivity to keep up with 2 x 8" woofers. Purpose designed midrange drivers usually have a higher resonance frequency compared to woofers and midwoofers. It is not a problem for a midrange crossing at 400-500Hz in the way it would be for a midwoofer.
If you want ideas about drivers and passive crossovers there are a number of DIY centre channel speakers using this configuration or close.
Brisso some basic investigation on your Wharfedale drivers is in order here:
What we got here? 4 or 8 ohm drivers? Wired in series or parallel? They look like those familiar old well behaved Elac units that work reflex or closed box.
The reason the Gale had such terrible low impedance was the 4" mid was being asked to go so loud and low. But if you wire the bass in series, it'll still be quite efficient if I know those Elac units. I expect 50-60L will be OK.
Really don't fancy that Coral cone mid, and flimsy D19 tweeter will need a highish 3.5kHz crossover. All you're doing really is building a closed three way with doubled bass units. A simple enough conversion.
ScanSpeak-3W-Discovery
What we got here? 4 or 8 ohm drivers? Wired in series or parallel? They look like those familiar old well behaved Elac units that work reflex or closed box.
The reason the Gale had such terrible low impedance was the 4" mid was being asked to go so loud and low. But if you wire the bass in series, it'll still be quite efficient if I know those Elac units. I expect 50-60L will be OK.
Really don't fancy that Coral cone mid, and flimsy D19 tweeter will need a highish 3.5kHz crossover. All you're doing really is building a closed three way with doubled bass units. A simple enough conversion.
ScanSpeak-3W-Discovery
Attachments
Steve, is third order the key to these D19s? IIRC I didn't like them below 5k but they were OK otherwise.
http://www.parts-express.com/pedocs/specs/264-500-vifa-specifications-44783.pdf
Recommended second order or higher.
@SYstem7
Thanks for the XO diagram Steve.
I did trace the 308's XO and I'll post it tonight when I get home.
Doug
Thanks for the XO diagram Steve.
I did trace the 308's XO and I'll post it tonight when I get home.
Doug
I think it was Troels Gravesen who pointed out that 19mm domes really don't like working much below 4.5kHz. For all that, the mylar Celestion HF2000 wasn't much different. On flat baffle in a regular three way the tweeter is always third order to align phase. The Gale is just lifting the standard Celestion HF2000 tweeter filter.
FWIW, those Wharfedale tweeters are plastic by the look of them. We had their designer, well from something similar from something like the Wharfedale Delta 30 here a little while back, and he thought they were some of his best work. Apparently this little vintage speaker was an absolute bestseller. So had something right. 😎
FWIW, those Wharfedale tweeters are plastic by the look of them. We had their designer, well from something similar from something like the Wharfedale Delta 30 here a little while back, and he thought they were some of his best work. Apparently this little vintage speaker was an absolute bestseller. So had something right. 😎
Attachments
Because he throws some numbers and it's ok 😱
On my side, a 8" woofer is already too big. Doubled, I can't imagine
If I were to make such a box, given the basic instructions, I would
maybe mantain the same volume, probably making a heavy contoured baffle with double 4", like the Mivoc WAL 16 -as I seem to find only that brand that makes little 16 Ω speakers to be paralleled easily...similar to TB- but ok, power ratings are rather low for a "force" project.
In the same fashion, little 2" FRs can be adopted as midrange.
( 20-30 W max but Hey! A Celestion plastic frame 3" neo can reach 60 W ! )
I haven't experience with double speakers so I really can't tell
You are the people I trust
Peerless and Vifa have apparently been bought by Tympani, who also owns Scan speak, if I remember correctly). But my experience with Peerless drivers is that they compete very well with the best Seas, Scan Speaks and others. They just don't look as fancy. I'm talking about the 8 inch Nomex cone (treated paper) model 830869. $80/each here in the U.S. Also the Peerless TG9FD10-08 3.5 inch glass fiber cone driver, $22 each here in the U.S. (at Madisound). These drivers are exceptional from 400HZ - 15kHZ, which means you can keep crossover points well away from that area where the ear is so sensitive (600HZ-5kHZ).
As for tweeters, I wouldn't rule out a Peerless ring radiator. Seas makes some seriously great 1 inch domes for about $50 per, that I would consider. Since I would position the crossover points at 500HZ and 7kHZ (long story short), I would also consider using a Fountek 1.5 inch ribbon tweeter (rated one of the best ribbons by Zaph, who tested many exhaustively - Google Zaphaudio). With that high crossover point, I would also look for any 1/2 inch or 3/4 inch dome tweeters, which have wider dispersion and often seem to sound more real to me.
Whatever tweeter you choose, make sure you like the frequency response graph. When there are fancy words like special metals, ceramic, or "diamond" whatever, those are often high priced exotic experimental and didn't do as well in the zaph tests that I saw. Personally I'd go with the Fountek 1.5 inch ribbon, and accept the limited off axis dispersion, part because ribbons do cymbals so darn nicely, and part because there would be a less abrupt change in dispersion at the 7kHZ crossover with the tweeter being slightly larger than the original Celestion tweeter (which I think was 3/4 inch?).
The Peerless midrange drivers I recommended above are a little difficult to mount since the flange is minimal, and you want to insure reasonable rear air flow around the ceramic magnet (The Peerless neodymium magnet drivers have crappy FR graphs). And they don't come with a rear chamber. Personally I think they are worth the trouble. Roger Russel formerly of McIntosh uses those drivers in his $19000 line array system.
I made a sub-enclosure for them out of an ABS (plastic) pipe (that plumbers use) bought at a Hardware store, that has a 4 inch inner diameter and 1/4 inch walls. I ran this pipe all the way to the back of the enclosure, which helped brace the enclosure and provided the back end termination of the pipe.
I put several discs of fairly dense carpet in the bottom, then several discs of foam rubber, then the "acousta-stuff" variation of fiberglass insulation (not as dusty - a plastic rather than glass) on that, leaving about half of the pipe near the driver empty. I glued all of that in with silicone rubber glue. I also glued regular thin felt to the sides up near the driver. It measures VERY flat acoustically from 400HZ on up to at least 10kHZ. Above that I don't necessarily trust my calibrated mic (Behringer).
As for the crossover, I'm one of those guys who prefers active crossovers, for their steep slopes (4 pole LinkwitzRiley), easy differential level adjustment, general accuracy (doesn't care about the impedance curve of the drivers), and the less reactive loading of the poweramps.
I'd bring the driver connections out on the rear panel to banana jacks, and if I've just got to have a passive crossover, I'd build it on a big piece of masonite or plexiglass such that it could plug onto the banana jacks on the back of the cabinet (along with some screws and spacers for better rigidity).
Passive crossovers done right are actually fairly difficult. You have to accurately measure the SPL AND impedance at the frequencies you want to crossover at. Published specs may not be good enough. Everthing is more sensitive (and interactive) when you go for more than a one pole cutoff rate.
If instead you go active on the crossover, once you've got a chassis and power supply for the crossover circuit, I always stick in another circuit that pumps up the low bass such that it's acoustically flat down to 30HZ. I LOVE low bass, and all music regardless of the frequency delivered is in an envelops that contain spectra down very low in frequency. So having good extension of bass will improve the sense of being there more than just giving you those low notes on the bass instrument. A speaker that rolls off at 50HZ can be pretty good, 40HZ can be very good, but 30 HZ can be awesum. Every percussive device will sound more real. When you have the closed box "infinite baffle" it's a shame to not take advantage of active EQ pumping up that low end. You can't use that technique with bass reflex nearly as well since the damping of the woofer is flimsy below the tuned frequency. Too likely to bottom out the driver.
If you go active on the crossover, I'd only bother at the 500HZ Xover, and go passive one pole at 7kHZ. I did that in one of my systems and it sounds great. The high crossover point allows the one pole rate. You could also move that crossover point higher, but then the off axis dispersion change becomes more abrupt.
I Hope this helps.
As for tweeters, I wouldn't rule out a Peerless ring radiator. Seas makes some seriously great 1 inch domes for about $50 per, that I would consider. Since I would position the crossover points at 500HZ and 7kHZ (long story short), I would also consider using a Fountek 1.5 inch ribbon tweeter (rated one of the best ribbons by Zaph, who tested many exhaustively - Google Zaphaudio). With that high crossover point, I would also look for any 1/2 inch or 3/4 inch dome tweeters, which have wider dispersion and often seem to sound more real to me.
Whatever tweeter you choose, make sure you like the frequency response graph. When there are fancy words like special metals, ceramic, or "diamond" whatever, those are often high priced exotic experimental and didn't do as well in the zaph tests that I saw. Personally I'd go with the Fountek 1.5 inch ribbon, and accept the limited off axis dispersion, part because ribbons do cymbals so darn nicely, and part because there would be a less abrupt change in dispersion at the 7kHZ crossover with the tweeter being slightly larger than the original Celestion tweeter (which I think was 3/4 inch?).
The Peerless midrange drivers I recommended above are a little difficult to mount since the flange is minimal, and you want to insure reasonable rear air flow around the ceramic magnet (The Peerless neodymium magnet drivers have crappy FR graphs). And they don't come with a rear chamber. Personally I think they are worth the trouble. Roger Russel formerly of McIntosh uses those drivers in his $19000 line array system.
I made a sub-enclosure for them out of an ABS (plastic) pipe (that plumbers use) bought at a Hardware store, that has a 4 inch inner diameter and 1/4 inch walls. I ran this pipe all the way to the back of the enclosure, which helped brace the enclosure and provided the back end termination of the pipe.
I put several discs of fairly dense carpet in the bottom, then several discs of foam rubber, then the "acousta-stuff" variation of fiberglass insulation (not as dusty - a plastic rather than glass) on that, leaving about half of the pipe near the driver empty. I glued all of that in with silicone rubber glue. I also glued regular thin felt to the sides up near the driver. It measures VERY flat acoustically from 400HZ on up to at least 10kHZ. Above that I don't necessarily trust my calibrated mic (Behringer).
As for the crossover, I'm one of those guys who prefers active crossovers, for their steep slopes (4 pole LinkwitzRiley), easy differential level adjustment, general accuracy (doesn't care about the impedance curve of the drivers), and the less reactive loading of the poweramps.
I'd bring the driver connections out on the rear panel to banana jacks, and if I've just got to have a passive crossover, I'd build it on a big piece of masonite or plexiglass such that it could plug onto the banana jacks on the back of the cabinet (along with some screws and spacers for better rigidity).
Passive crossovers done right are actually fairly difficult. You have to accurately measure the SPL AND impedance at the frequencies you want to crossover at. Published specs may not be good enough. Everthing is more sensitive (and interactive) when you go for more than a one pole cutoff rate.
If instead you go active on the crossover, once you've got a chassis and power supply for the crossover circuit, I always stick in another circuit that pumps up the low bass such that it's acoustically flat down to 30HZ. I LOVE low bass, and all music regardless of the frequency delivered is in an envelops that contain spectra down very low in frequency. So having good extension of bass will improve the sense of being there more than just giving you those low notes on the bass instrument. A speaker that rolls off at 50HZ can be pretty good, 40HZ can be very good, but 30 HZ can be awesum. Every percussive device will sound more real. When you have the closed box "infinite baffle" it's a shame to not take advantage of active EQ pumping up that low end. You can't use that technique with bass reflex nearly as well since the damping of the woofer is flimsy below the tuned frequency. Too likely to bottom out the driver.
If you go active on the crossover, I'd only bother at the 500HZ Xover, and go passive one pole at 7kHZ. I did that in one of my systems and it sounds great. The high crossover point allows the one pole rate. You could also move that crossover point higher, but then the off axis dispersion change becomes more abrupt.
I Hope this helps.
I suspect very few, possibly no, small "full range" drivers are well suited for use as a midrange in a 3 way with a reasonable amount of woofer surface area. They tend to have a combination of low sensitivity and low power handling which means they cannot produce the maximum SPL required to keep up with the rest of the speaker. This can be addressed by using multiple drivers but it seems simpler to use a driver designed to be a midrange for the midrange. Not that there many purpose designed midrange drivers around these days with most new drivers being midwoofers which often need doubling up to work over the midrange.
Having recently used it as a midwoofer I heartily recommend the FaitalPro 5FE120 as a mid here.
It's a 5" as the name suggests, 88dB/1W, 5mm Xmax and the response is flat out to 6k, takes 80Wrms, it is dirt cheap at £24ea and it sounds very, very good indeed.
It will easily keep up with a couple of 8" woofers, I'm thinking of pairing it with a 12" in a 3way.
It's a 5" as the name suggests, 88dB/1W, 5mm Xmax and the response is flat out to 6k, takes 80Wrms, it is dirt cheap at £24ea and it sounds very, very good indeed.
It will easily keep up with a couple of 8" woofers, I'm thinking of pairing it with a 12" in a 3way.
Why you would want to link mechanically the passive components to the box once you've resolved to attach the binding posts directly to the speakers ?
Wisely, I would leave the components out of the vibrational path.
Distributing them along the line from the amplifier is better.
Cable discussions would disappear as a collateral effect 🙂😛
Scanspeak parted from Typhany some time ago. Their parent company is now KYET although Scanspeak continues to develop and operate independently. Here is the news announcement:
Scan-Speak
Unfortunately, the freight to import these is a killer. AUD$155 each, landed.
That is a scary mark up!
I'm paying £20ea (Aus$40) incl 20%VAT and p&p, for that or even double they are a complete no brainer.
Have you tried their distributors in Malaysia, Indonesia, Macau or Hong Kong?
Btw I used them with SEAS H0737 which is a nice little tweeter and it didn't need any padding either which is why I used it. The speakers were only meant to be a cheap&cheerful stereo for my youngest daughter when she starts Uni.
Kinda backfired because apart from bass extension they came out substantially better than my own Tannoy DC-based speakers.
I have to build two more pairs now: One for myself and one for a mate who heard them.
I've never found cabinet vibration of crossover components to be a significant distortion mechanism. That's a theoretical idea, much like the speaker wire issue. True, but insignificant. With speaker wires, just about any wire is fine, (I buy 16AWG AC line cord at the hardware store for I think $0.50/foot unless the wire will be over 20ft., then I might bother to go with 14AWG - not a big deal), but the connectors can be a weak link, especially over time. The one time speaker wire could contribute in a bad way is if the power amp phase margin is weak, and the capacitance of a long speaker cable could cause spurious oscillations. With opamps, we put a 100-200 ohm R in series with the output just for that reason. With a speaker, it's arguably wise to put a 0.2 - 0.5 ohm R in series with the output for this reason. You throw away a small amount of power, and reduce the damping factor slightly (it's already limited by the DCR of the drivers).
I want to add that I read and have become convinced that the way the ear-brain mechanism works, our sensitivity to stereo imaging resolution changes when we get above about 5kHZ. For best integration of stereo imaging, you usually want to have the drivers aligned vertically, but if the crossover point is up at 7kHZ or above, it's much less of an issue.