Successful implimentations of ribbon tweeters?

A couple of rooms I spent a fair bit of time in at Bristol last week were the Rountree and Ypsillon. Both, I thought, did something very pleasant with soprano voice. With the obvious common factor being the ribbon I wondered whether it was that that I was liking. Seems rather unlikely I'll ever buy either so has anyone else done nice things with ribbons?

 

I've just posted this elsewhere:

"OK, so I've bought Neat MFS with ribbon super-tweeters and Dali Ikon 1's with combination ribbon tweeters. I like the treble on each of these. The Dali's, as SSM has said, are particularly good on voice.

My ATC's provide scale and timbre on mid and bass that the small speakers can't match. Air-moving effects that cannot be obtained from smaller drivers.

I'm always searching for a speaker that plays accoustic instruments and voice well, including piano. Mixing ribbons and cones seems to be a matching issue in the mids, though I know full range ribbons / electrostatics plus subs are an accepted combination.

Who has had full-range ribbons or electrostatics, and would you share your thoughts please?"

 

My experience of speakers that use a ribbon tweeter is that (to me) they often lack integration in the sense that the ribbon has a very different sonic 'charachter' to the unit it crosses over to. To be specific, the ribbon often sounds much quicker and more dynamic than the moving coil unit it crosses to and (to me) the sonic charachter of the different driver types is all too evident. I noticed this with the Rountree speakers and with the small Monitor Audio speakers (amongst others). I understand that one of the problems with using a ribbon unit is that most ribbons will not handle much power and so need to be crossed over at a relatively high frequency; often a frequency that is a little too high for the mid/bass driver to reach too in a linear manner - going to a three way design brings its own problems in terms of system complexity.

 

Magnepans have quite a good ribbon tweeter. The best one so far IMO is the Raven tweeter. They also make a mid/tweeter combo thats a killer. Expensive but worth a look. Nola uses them in their ref. model.


rollo

 

I like the approach of Dali with the Helicons. Their treble unit is a dome and a ribbon which have been designed to work together, they are not seperate tweeters but part of the same tweeter. It seems to work very well.

The standmounts are the ones I really rate.. these..

http://www.prokare.com.tr/image1/dali_helicon300_125.jpg

 

http://estore.websitepros.com/1736754/Detail.bok?no=562

That looks interesting, a new Raven driver.

 

Sounds even better. The Mid/treble unit is their ultimate effort in design .


rollo

 

cant help boasting...
My friend has built me a speaker...with Boehlender Graebener tweeters.
http://members.chello.cz/alexander/
wonderful, particularly for classical music.... airy, realistic, and also has bass despite the reasonable size.

 

Are there a lot crap manufacturers out there or something? Its hardly the test to integrate a ribbon and a cone. Pathetic if you ask me.

Problem with most affordable speakers is they use absolutely dog shit ribbons. And a relatively low distortion mid crossing to tall order harmonic distortion generator(however 'airy' that might be) usually would sound like two completely different transducers working at the same time.

 

Shin,

Going back to when you were specifying the AT drivers for your new speakers I remember that you were looking for a mid/woofer with very good acceleration and a light cone to match well to the speed of the RAAL ribbon.

Something I wondered at the time but never asked was, how does the acceleration and weight of the cone effect the sound? Surely the 'speed' of the driver is directly related to the frequency extension. A driver moving 2000 time a second will reproduce 2KHz and nothing more, whether it is a light cone or a heavy one.

Am I missing something about how to match cones to ribbons? Is it more to do with the decay pattern of a light cone, or the spectral make-up of the distortion of a light cone matching better or something?

RAAL ribbons are the best true ribbons I have heard. Most of the commercial speakers out there do seem to use crap ribbons.

 

And yet people seem to love them.
Does this mean the distortion produced by many ribbons is subjectively quite 'nice' to listen to?

 

Exactly 2Khz is 2Khz. Its more interesting to consider what happens when producing thousands of frequencies at once from a single driver. What does matter is impulse rise time and this has little to do with the moving mass and suspension parts which might seem the obvious factor for how something like acceleration factor is determined. Instead a large part is voice coil inductance. Drivers with high inductance suffer from hysteresis or memory effect and tend to have rise times that differ depending upon where the voice coil is within the gap. Such a driver covering bass and mid frequencies will be using some excursion and trying to reproduce midrange. The end result is easy to imagine.

Subs are notoriously high inductance drivers but because of their limited bandwidth, the wavelength of the frequencies produced and our relative insensitivity to bass, we don't notice. Shift the frequencies higher and the range wider and drivers with high inductance just don't cut it. High inductance alsol naturally creates a low pass behaviour on the top end of the response.

Audiotechnology use SD or Symmetric drive in their motors which is a theme on a copper sleeve through pole piece. Nicked from the AT website, this results in:

A greater usable frequency range.
Lower intermodulation distortion.
Linear rise time no matter where the voice coil is situated in the gap.
A more consistent impedance curve.
Much lower inductance of the voice coil.
Rise time is increased up to a factor 10

Many of the high end drivers use copper in the motors and in the right places all the above will be influenced for the better. Seemingly most effective, but dependant on the motor design, is the copper sleeve through the pole piece.

CSD rather than rise time might be worth looking at. TBH I don't see much point in obsessing too much here because good drivers of today, unless breaking up or whatever, are doing well enough. Just look for a clean nature at the point where your crossing the ribbon and that's not really a big task since 2.5-3Khz doesn't have a whole lot of energy. I'm more worried about getting the power response right with a ribbon than matching the CSD.

My own design forced a shift from 3 to 3.5 way, rolling off the lower midrange in the MTM and letting the other carry on up to meet the tweeter. This change was one of those big pieces that really just sounds right in comparison to the 3-way. A look at the offaxis vertical measurements show a much smoother transition with 3.5way. On axis measurements are only a small part of the picture.

RAAL seem to be doing things well. The distortion performance is unusually linear in its relationship with rising SPL. It starts relative high (in comparison to the best dome tweeter) at low level ie. 85dB/1m but crank it upto 113dB/1m and its risen much less than a dome would have:



Stick a crossover on that at 2.5Khz and you'd be looking at less than 2% distortion across the operating bandwidth of the RAAL and that at 113dB/1m.

The RAAL's or ribbons in general suffer from virtually no thermal compression. The ribbon element is fully exposed to air and has a huge radiating area compared to a normal dome tweeters voice coil enclosed in a gap.

I don't think you'll see the RAAL used in affordable speakers though.

 

I have Dali Ikon 6 speakers and I'm yet to notice one single instance where they have sounded wrong. The treble is very very sweet indeed.

 

Depends on the person and the situation I suppose! Personally I find most ribbons initially impressive but if you listen to a wide variety of music over a long period they can be fatiguing and make everything sound kind of 'the same'. Especially with heavy music like metal or emo. I think they work well for acoustic stuff or anything quite gentle, but its a party trick that wears quickly for me. I think a large part of why people like them is 'cos they look cool!

I noticed at the show that the speakers with ribbons like Monitor Audio, Rountree, and Dali all seem to have a slightly boosted treble. Its almost like they are saying 'hey, we have ribbons!'.

 

Shin, what were you going to say there? I'm not sure I quite understand the relation of acceleration factor to driver sound. The impulse rise time will be directly related to the upper frequency limit of the driver. So what is this separate measurement of 'acceleration'? Sorry if this is OT for the thread!

I hear what you are saying about dispersion matching. This is something people ignore all too often in speaker design. Its very, very important IMO.

 

Not sure what happened to the opening bit of that post, I swapped a couple of the sentences around so it must have gotten lost/deleted. I'll restate what was lost and go into a bit more detail at the same time.

The term acceleration factor(AF) has nothing to do with the speed of a driver that's operating within its usable bandwidth, by that I mean where phase and amplitude characteristics are optimal, and this is with good reason. If you think about the speed of sound and the cone oscillation required to produce a particular frequency, these are FIXED values. If they changed to bring about this "faster sound" you'd have a driver that bends the laws of physics. This is the reason why I want to politely slap people who say ten light and nimble 8" drivers produce faster sub bass than one big heavy 24".

Instead, acceleration factor is indicating the upper frequency limit a driver can produce whilst maintaining those same optimum amplitude/phase characteristics.
As an example, producing 20Khz vs 20hz requires a much higher rate of oscillation. A driver with high acceleration factor likely has the qualities to potentially produce that extended range. Note that I say likely as there's other factors which limit a drivers upper limit.

Just to confuse things a little further, acceleration factor also takes into account driver sensitivity. Eg. A high sensitivity driver has higher AF than a lower sensitivity model that's otherwise identically spec'd.

To investigate for yourself, see this:

Acceleration Factor = BL * Sd / Mms

Try applying that to some drivers and see how the AF product relates to the drivers upper frequency limit and sensitivity. Most obvious example is a 15" sub vs. 1" tweeter.

Yes I've found it to be revelatory to consider the sound radiation pattern and the how smooth the overall power response is. Designing for just on axis performance might sometimes get you a good sound design but consider that even from a listening distance of around 8 foot, 50% of what your hearing in an average room isn't direct sound. If you don't get the power response(regardless of narrow or wide directivity) right then your engaging in a serious design flaw and its hardly surprising that this is very audible.

I've just bought a pair of the new DXT waveguide Seas tweeters:





Off axis performance is exceptional(0, 15, 30, 45 and 60degrees shown. BTW graph taken from Zaph's website):



Add a filter to flatten the waveguide loading hump and you have something that's flat and smooth from 0 to 30 degrees offaxis all the way upto 15Khz and still good even at 60 degrees.

I'm looking forward to building a compact surround speaker with this wide and even dispersion design coupled with the Seas Excel W12CY003 Nextel driver.

 

Well, I can't say that I am familiar with the Ikon 6 so it is possible that the character of the ribbon on this speaker is better integrated than most. On the other hand it is possible that this lack of integration (in terms of presentation rather than frequency response) is rather subjective and therefore listener dependant; perhaps I would find this an 'issue' with the Dali's whilst you may not - it is quite likely that there are qualities that my speaks posses that you would have an issue with whilst I may barely notice them.

Perhaps I should add that I was not criticising the subjective sweetness of the treble in my post.

 

Thanks for the explanation, makes sense now!

Those DXT tweeters do look nice. The waveguide looks a bit... lumpy, though. Probably the cause of that little bump above 15KHz. I wonder if the same could be achieved by adding a similar size waveguide to the 22TAFG.

I hope you post your experience building with them, I will follow it with interest!

 

Its all intentional. See the DXT website:

http://www.dxt.dk/pages/webside.asp?articleGuid=35007&menuGuid=15292

 

it's interesting you've mentioned ATCs. as a previous owner of active 20s i tend to believe that ATC should be charged with genocide against audio humanity for using that crap and cheap vifa and seas tweeters instead of something more decent that would match a quality of their bass/mid-bass/mid drivers.

there was a certain point when i wanted to buy a short RAAL ribbon, build a new baffle for the 20s and see what i would get out of it but i couldn't bother eventually. even the crossing frequency was similar as a recommended one for the RAAL (2.8k i think) and sensitivities were quite similar so it would probably produce quite a significant improvement but now i'm happy i sold them before making them probably unsalable.