Super High-End 3-way floor-stander

[ScottG]

Originally Posted by gornir

Hi ScottG,

Thanks for your comments.

..Zaphs profile is to build nice speakers from cheap drivers and I think that’s ok, but everything more expensive gets low value for him, despite the fact that you often actually get something more from the driver that costs more. It would really interesting to hear anyone’s impression of this woofer used in a diy build.

/Goran

Your welcome!

Actually I don't really look at his comments or his ratings. Looking at the 8" driver roundup *data* both the Visaton and the Usher are lower in distortion (below 100 Hz). Of course both have lower excursion and a higher fs.

I couldn't tell you about the lower freq. distortion performance of the Lambda 10, could be better, could be worse. I would say however that you have more excursion with higher compliance, and greater sd - which should result in better performance. As for cabinet size, the 3.5 cubic feet is a recommendation but not an absolute by any means.. basically it's all down to how low in freq. you want to go vs. power and maximum spl.

[perceptionchanges]

Maybe you could take a look at the Newform ribbons. They sound terrific and with their length and ablity to be adjusted up and down on the mounting pole make them very versatile. The 1000hz + crossover point allows you more options with your other drivers also...

LINESOURCE REFERENCE - Ribbon Array Ultra High End Audio Speakers for Home Theater and HiFi Music Systems

[Helmuth]

BG have also line array's that play the hole midrange usable as dipole or monopole. They will do a lot of things better then a cone driven speaker I guess, and they will also have their drawbacks I don't know because I never used them.

RD40 300-20000
RD48 250-20000



Stig Eric likes them a lot I read of his reply's.

[gornir]

Originally Posted by perceptionchanges

Maybe you could take a look at the Newform ribbons. They sound terrific and with their length and ablity to be adjusted up and down on the mounting pole make them very versatile. The 1000hz + crossover point allows you more options with your other drivers also...

I’m familiar with the Newform tweeters. I almost bought a pair for about 15 years ago. As I recall they were quite expensive and I never went through with the order.

I’m extremely found of the ScanSpeak midrange sound so I think I stick with it. The question for me right now is which woofer and tweeter to combine with the midrange. The smaller ribbon R8 could be interesting, but it's quite expensive and has a low sensitivity.

Regards

/Goran

[ultrakaz]

Originally Posted by gornir

Hi everyone,

I have begun sketching up a new 3-way floor-stander design. My wife is tired of how many different loudspeakers have in the living room Ok, I said let me build a super high-end loudspeaker that I can live with for a long time, then I promise you that I will only have one set of loudspeaker in the living room.

She said…….. Do it!

I realize that this project will take some considerable time to finish, especially since I most likely will have to design and build the enclosures by myself and they must match the drivers in build quality and looks.

There are so many different loudspeaker drivers and manufactures that can be considered as high-end, but I have so far chosen two brands of loudspeakers, ScanSpeak for mid and woofer and RAAL for the tweeter section.

Why these two brands? Ok, I just love the build quality and looks of the ScanSpeak drivers. I also have some experience with ScanSpeak drivers and I like the performance of them. Why the RAAL ribbon tweeter? What I so far have managed to research, it seems like the RAAL ribbon is the “mother” of all ribbons. I have used Fountek ribbons in several designs and I must say I favor the ribbon sound when used properly over the dome tweeter sound. I know this is a hot topic, dome vs. ribbon, but I have to follow my perception that I subjectively like ribbons more.

Since I’m unfortunately isn’t a rich man I can’t buy and test every driver I consider for this design and here is where I really need your help!

Please share any thoughts, experience and measurements you have on the listed drivers below. I cannot do these design choices without your help. Especially ScanSpeak is notorious of not measuring as published (TS-parameters).

My contribution will be a thread of my build where I will share my thoughts, measurements and the cross-over design etc.


Ok, here are my 3 different design choices so far:

Alternative 1:
Tweeter: 1x RAAL 140-15D
http://www.raalribbon.com/download/raal_140-15d.pdf

Mid: 1x ScanSpeak 15M/4531K00
http://www.scan-speak.dk/datasheet/pdf/15m-4531k00.pdf

Woofer: 2x ScanSpeak 22W/8851T00
http://www.scan-speak.dk/datasheet/pdf/22w-8851t00.pdf

Estimated cross-over frequencies: 300-500Hz & 2000-3000Hz

Estimated driver unit price/pair: 2540€

Alternative 2:
Tweeter: 1x RAAL 70-10
http://www.raalribbon.com/download/raal_70-10.pdf

Mid: 1x ScanSpeak 12M/4631G00
http://www.scan-speak.dk/datasheet/pdf/12m-4631g00.pdf

Woofer: 2x ScanSpeak 18W/8531G00
http://www.scan-speak.dk/datasheet/pdf/18w-8531g00.pdf

Estimated cross-over frequencies: 500-700Hz & 3000-4000Hz

Estimated driver unit price/pair: 1870€

Alternative 3:
Tweeter: 1x RAAL 70-10
http://www.raalribbon.com/download/raal_70-10.pdf

Mid: 1x ScanSpeak 12MU/4731T00
http://www.scan-speak.dk/datasheet/pdf/12mu-4731t00.pdf

Woofer: 2x ScanSpeak 18WU/8741T00
http://www.scan-speak.dk/datasheet/pdf/18wu-8741t00.pdf

Or

Woofer: 2x ScanSpeak 18WU/8747T00
http://www.scan-speak.dk/datasheet/pdf/18wu-8747t00.pdf

Estimated cross-over frequencies: 400-600Hz & 3000-4000Hz

Estimated driver unit price/pair: 2250€ or 2200€

Comments:

I will first do an active version (DEQX) of the loudspeaker and later on a passive cross-over version. At this stage I have no idea of how the loudspeaker shape and design will look like, but considering the amount of drivers and size it will be a fairly large enclosure (60-80litres internal).

I want the loudspeaker to be as efficient as possible and my estimate, depending on driver choices is that it will be around 88db (+- 2db) 2.83v/1m.

My estimate of cross-over frequencies is just an estimate based on driver size and estimated distortion numbers. When the drivers are placed in an enclosure and measured properly I will know more accurately.

My estimate of the price is based on the driver units only without shipping or other fees. The prices are collected from Europe Audio - Home page (ScanSpeak) and www.loudspeakershop.eu (RAAL).

For now, I’m most interested in alternative 1, but I would greatly appreciate any information/experience you guys have on the above drivers.

Help me choose the right design.

Thank you!

Regards

/Goran

Congratulations on clearing the biggest hurdle (WAF) to your future high end build. No doubt, given your experience, methodology, and SS drivers you are going to build something really awesome. To the less experienced it would be a gamble and we would be better off building a proven design.

All the Alternatives would make great speakers, but I like "Alternative #1", because the crossover frequencies are outside the critical "telephone" passband. Have not heard the 15M, but I've heard the 15W and it is about the best mid/woof you can buy. Along with the RAAL you are at the end of the road.

If you can give a listen to some Wilson Audio or Dunlavy speakers. You may not like them overall, but they do some things really well, especially the larger ones.

I might add, if it is possible, you might try to have a separate upper bass driver dedicated to around 75Hz to 250Hz. Some may even argue that this is actually where most of the visceral energy or life of the music is. Unfortunately, this is really expensive (almost prohibitive) to do passively, but there is an effortless impact to the sound that conventional 3-ways lack. Of course, a properly designed 3-way might be all that you will ever need and is miles ahead of any 2-way.

[gornir]

Originally Posted by ultrakaz

Congratulations on clearing the biggest hurdle (WAF) to your future high end build. No doubt, given your experience, methodology, and SS drivers you are going to build something really awesome. To the less experienced it would be a gamble and we would be better off building a proven design.

All the Alternatives would make great speakers, but I like "Alternative #1", because the crossover frequencies are outside the critical "telephone" passband. Have not heard the 15M, but I've heard the 15W and it is about the best mid/woof you can buy. Along with the RAAL you are at the end of the road.

If you can give a listen to some Wilson Audio or Dunlavy speakers. You may not like them overall, but they do some things really well, especially the larger ones.

I might add, if it is possible, you might try to have a separate upper bass driver dedicated to around 75Hz to 250Hz. Some may even argue that this is actually where most of the visceral energy or life of the music is. Unfortunately, this is really expensive (almost prohibitive) to do passively, but there is an effortless impact to the sound that conventional 3-ways lack. Of course, a properly designed 3-way might be all that you will ever need and is miles ahead of any 2-way.

Hi ultrakaz,

Thank you for your kind words

I have decided to scrap my alternative 2 and 3. I will stick to my alternative 1 with the 8” Scans. However, this design sketching has grown to 3 different 8” configurations with 3 different tweeter and 2 midrange options.

I was a bit curious on the Illuminator series, but I couldn’t find enough information about them to make it worthwhile to proceed on that track.

Soon I will post my new design thoughts and some box sims around the SS 8”.

Regards

/Goran

[seebert]

Thanks for your comments. That’s a cool and futuristic design, but I think it would be a challenge to build.

From wood yes, but what about another material (mix) and a pour in solution. Just a thought. Good luck.

Sorry - posted out of sequence.

[Lynn Olson]

You might consider Skaaning. Speakers are not designed by corporations, but by skilled engineers, and I consider Skaaning to be one of the most talented engineers in the field - he's the one that designed some of the best-known Scan-Speak drivers.

True, they're not cheap, but that's not surprising when you consider these are not Chinese-made drivers that mimic the look of Scandinavian drivers, but the real thing, made in Europe. I've heard speakers designed around the Skaaning drivers, and I weren't so deeply into the high-efficiency thing, those are the ones I'd choose.

RAAL makes superb ribbons, but keep in mind that even the large ribbon is at its best well above 2 kHz - 3 to 4 kHz is a better choice for the crossover point. So if your heart is set on ribbon treble, do yourself a favor and select a midrange that needs as little equalization as possible through 5 kHz, or above.

I would avoid any mid that has resonances in the 3~5 kHz range - sure, these can be equalized, but after equalization, there are still noticeable colorations, particularly in comparison to a ribbon tweeter. When drivers break up, strange things happen to dispersion (thanks to multiple emission surfaces on the cone that go out of phase with each other), and equalization cannot correct for dispersion anomalies.

It's easy to be hypnotized by the power of digital equalization, but it doesn't magically make drivers flat, despite the measurements (at one point in space). Move the microphone around, or check the distortion spectra, and you'll find the resonances are still causing trouble - it's just been shuffled around to another domain that isn't as easy to measure. Try and choose a midrange that is a good complement to the RAAL - flat response, with very fast decay characteristics, and not requiring a notch filter in the passband or above the passband. Very few mids meet that criteria.

As you may have noticed, many commercial high-end systems with ribbon tweeters have poor subjective integration between the ribbon driver and bass/mid driver. Following these guidelines will save you a lot of expensive mistakes.

[pieter t]

Originally Posted by Lynn Olson

............. do yourself a favor and select a midrange that needs as little equalization as possible through 5 kHz, or above.

Try and choose a midrange that is a good complement to the RAAL - flat response, with very fast decay characteristics, and not requiring a notch filter in the passband or above the passband. Very few mids meet that criteria.

Completely agreed. That's why I suggested the Yamaha beryllium midrange dome, which works flawlessly from 500Hz - 6kHz (2nd order filtered) in the NS1000. Here we have the "presence" range handled by a single driver. I have no experience with Skaanings drivers but other drivers of the cone variety meeting the criteria will be very hard, maybe impossible to find.

[gornir]

Hi Lynn,

Thank you for your insightful comments.

Originally Posted by Lynn Olson

You might consider Skaaning. Speakers are not designed by corporations, but by skilled engineers, and I consider Skaaning to be one of the most talented engineers in the field - he's the one that designed some of the best-known Scan-Speak drivers.

Yes, I agree the Audio Technology drivers are quite amazing and the fact that you more or less can have them custom built for your application is quite unique for us DIY:ers. Actually ScanSpeak and AT drivers where almost the only two options I had for this project besides the Accuton drivers. I gave up the Accuton:s because of their hard cones and for the cone break-up reasons you mention below. I’m not the greatest fan of hard cone drivers, woofers and tweeter domes, even though there are examples of nice ones.

Originally Posted by Lynn Olson

True, they're not cheap

Yes, unfortunately they are a bit more expensive than the Scans.

Originally Posted by Lynn Olson

I would avoid any mid that has resonances in the 3~5 kHz range - sure, these can be equalized, but after equalization, there are still noticeable colorations, particularly in comparison to a ribbon tweeter. When drivers break up, strange things happen to dispersion (thanks to multiple emission surfaces on the cone that go out of phase with each other), and equalization cannot correct for dispersion anomalies.

Yes, I fully agree and that’s why I have chosen the ScanSpeak 15W/15M as the midrange for this project. Besides that it sounds absolutely wonderful it measures like a dream without any visible cone break-ups and it has one of the smoothest roll-offs I’ve measured. Picture 1 & 2 shows my measurements of the ScanSpeak 15W/4531G00 on an 8.5” baffle. I think these Scans could be one of the few drivers that has a chance to marry well with the RAAL:s.

Picture 1: No smoothing , Tweeter-axis, Blue=0deg, Red=15deg, Green=30deg.
Picture 2: No smoothing , Tweeter-axis, Blue=0deg, Red=30deg, Green=45deg.

Originally Posted by Lynn Olson

RAAL makes superb ribbons, but keep in mind that even the large ribbon is at its best well above 2 kHz - 3 to 4 kHz is a better choice for the crossover point. So if your heart is set on ribbon treble, do yourself a favor and select a midrange that needs as little equalization as possible through 5 kHz, or above.

I’ve used Fountek ribbons in several designs without any problem, but they do need steep filters e.g. LR4 acoustical slopes and not lower than 3-3.5kHz without distortion becomes an issue. My research so far on the bigger RAAL indicates that it could be crossed somewhere between 2.5-3.5kHz without any distortion issues. I’ve seen that ShinOBIWAN on this forum has used the RAALS extensively. I must contact him and see if his willing to share some frequency and distortion measurements.

Originally Posted by Lynn Olson

It's easy to be hypnotized by the power of digital equalization, but it doesn't magically make drivers flat, despite the measurements (at one point in space). Move the microphone around, or check the distortion spectra, and you'll find the resonances are still causing trouble - it's just been shuffled around to another domain that isn't as easy to measure.

Yes, I agree there are many things to consider even when going active in the digital domain e.g. cone break-up, baffle step, baffle diffraction and off-axis dispersion etc. I have done a small study in this forum SU551-RS28F – a 2-way DEQX system loudspeaker measurement study on this topic, using my DEQX.

Originally Posted by Lynn Olson

As you may have noticed, many commercial high-end systems with ribbon tweeters have poor subjective integration between the ribbon driver and bass/mid driver. Following these guidelines will save you a lot of expensive mistakes.

Yes I agree

Regards

/Goran