Super tweeter for FullRange Driver

I have a pair of Audio Nirvana 15in Alinco Classics that I just got done getting into cabinets. After some listening impressions, I've realized I cant live with the high frequency response these things put out by themselves. They only go to about 15k hz. and it's made music lose its sparkle and sense of space and spacial cues. I was thinking of buying a pair of Fostek T90As and pointing them at the ceiling for a more even dispersion and better integration with the Audio Nirvana driver. My main question is what  capacitor volume would you suggest? Where in the frequency response would you crossover to the tweeter? Would wiring the 8ohm tweeter and the 8ohm Audio Nirvana in parallel make me need to change my amp(Audio Note Kit 1 300b Set) from 8ohm to 4ohms? Any other brands of tweeters recommended for a fullrange driver that is around 96bd efficent? Sorry for all the questions, incredibly new to trying to implement crossovers with full rangers. If this sounds like a bad idea. Please speak up. 
Your plan makes a lot of sense to me, and imo the T90A is a good choice. My first commercial effort used a fullrange driver augmented by a rear-firing T90A.

In a case like this, where you don’t have the benefit of measurements, you find the capacitor value for the supertweeter by trial and error. There is simply not enough information available to reliably predict what the correct value should be. I’d recommend you not go over 3 microfarads, and I’ll go out on a limb and toss out 1.5 microfarads as a suggested starting point.

Capacitor values add when you wire them in parallel. If you want to end up with a boutique capacitor, find the correct value using inexpensive film capacitors before you spend the big bucks on a boutique cap. Capacitor quality cannot make up for not using the right value.

I don’t think you’ll end up with an impedance curve that will cause your amp any distress. There is very little energy in the top octave so an impedance dip up there is unlikely to be an issue, and the capacitor in series with the tweeter will keep the net impedance from dropping very low anyway.

Thanks for the response Duke. Do you suggest I throw in a resistor or two to help attenuate the volume, or some kind of Lpad? I'm guessing the T90As will be significantly louder than the Audio Nirvanas due to there higher sensitivity. 
ak, I would try without first but if they run too hot then yes but you want to put it in front of the capacitor or you will change your cross over point.
Oh, I would adjust the L pad to suit then measure it's resistance. I would get a matching 10 watt resister and wire that in permanently.
Those are nice full range drivers and you have a good amount invested in them. If it were me I would consider getting something like the Dayton Omnimic system to test your design.  Danny at GR Research did a video on adding a super tweeter to a full range and was of the view that there are often phase issues that cause unpredictable response unless you test and compensate. 
My Bache Audio Tribeca’s use the Fostex FT96 EX-2 tweeter with a modified Tang Band wide band driver. I believe it’s crossed over around 8K. It’s a great sounding tweeter. You may get good results aiming it at the ceiling. Another option would be to aim it straight ahead into the room. Assuming that the Audio Nirvana drivers are positioned with some toe-in.
I had the Aperion Audio super tweeters on top of my 8" AN speakers and they did enough for me. They've got a decent amount of adjustment which is handy. 
I strongly 2nd @cedargrove's suggestion to watch the GR Research Tech Talk Tuesday video on You Tube, wherein Danny Richie explains and demonstrates all that is involved in adding a tweeter to a single-driver design. You'll be very glad you did!
I use Fostex T90 forward-firing with Altec 604E.
This combination works great.
I use first order crossover and Werner Jagusch auto-transformer to decrease output of super-tweeters. I use 1uF AN Copper capacitors. I’m not sure they are the best for this application but they work good.

Back of ceiling firing tweeters sound horrible!!! It completely destroys focus and sound-stage.
I don’t understand why some people use back of ceiling pointing of tweeters.
Probably people are too lazy to make a phase alignment of tweeters. When alignment is done properly, super-tweeters sound together with the main speakers.
Alexberger wrote: " Back of ceiling firing tweeters sound horrible!!! It completely destroys focus and sound-stage."

I disagree. Quite the contrary in fact when done right, and I can go into detail if you’d like.

The reason for keeping the supertweeter out of the first-arrival sound is so that it doesn’t screw up the phase of that first-arrival sound coming from your fullrange driver.

By increasing the amount of high-frequency energy in the reverberant field while leaving the first-arrival sound intact, you are fixing the tonal balance without degrading anything else, as long as the reflection path length is long enough. And a ceiling bounce usually is long enough at high frequencies.

Regarding whether or not to use a resistor or an L-pad, again that’s "play it by ear". DO NOT put an L-pad or any parallel resistor (or inductor) legs AHEAD of the crossover capacitor. Any resistor (or inductor) leg in parallel with any driver’s coil NEEDS the crossover capacitor in between it and the amplifier to filter out the rest of the signal, otherwise your resistor (or inductor) leg is seeing a full-power, full-spectrum signal from the amplifier.

If your crossover (more precisely, high-pass filter) is first order - i.e. just a capacitor - and if you’re just using a series resistor to adjust the level, then it does not matter whether you place that resistor before the capacitor or after the capacitor. Either way, the resistance that the capacitor "sees" is the supertweeter + the resistor.

@akwilson501 === (They only go to about 15k hz.) --
 How you know you driver go to 15k hz ? and what is level -10 db or more?    The sparks is 9-10Khz. 15Khz is not listenable and just add more air. To add super tweeter and make pretty flat response ( not to bright) is very complicated, You have to use you ears and also measuring
You might want to check your hearing first.  I bought a nice pair of VIFA super-tweeters and didn't notice any difference.

Ran the test cd that came with them and found that I no longer hear anything over 12 kHz.  :~(

Fortunately I was able to re-sell them for what I paid.

Disappointing.  I was looking for a little more shimmer on cymbals.

How many hours have you logged on your AN Classic Alnico's? I have owned mines over 9 months and they will improve dramatically over a 3 month period and will get even better with more use.

These AN have more clarity and transparency compared to a previous speaker I owned with a frequency response of 20 Hz - 40 Khz and I haven't noticed any sound degradation, they have the ability to sound live and natural...

the Fostex is a good choice. I’ve used the Heil AMT 1D tweeters with great results as well
After some listening impressions, I’ve realized I cant live with the high frequency response these things put out by themselves. They only go to about 15k hz. and it’s made music lose its sparkle and sense of space and spacial cues.

If you want the very best. Get a pair of these Plasma tweeters that I’ve got, no mass. Highs just dance in front of you like firefly’s. They’ll sit on top of your Audio Nirvana Classics and look the part too.

Adjustable level, and can go down to 5khz and reach up to a staggering 100khz!!!!!.
He's dreaming on the price, I'd ask around $2.5k for mine.
Cheers George

This discussion remain me how is different the taste,  i dont want to tell the famous brands on the market now which use one full range driver
with double cones  , main cone and whizzer, which play a tweeter-super tweeter role, and say the sound is beatifically, Conclusion-   some like piece a paper as supertweeter , some need plazma . sorry for my english 
My speakers Altec 604E. These speakers have high frequency extension just up to 12KHz.
I did experiments with back of ceiling firing tweeters. I used Morel MDT30 tweeters with 1st order filter. The effect of  back of ceiling firing was like a saliva of singers where scattered around the room.
Then I used  Morel MDT30  in forward  firing way. I adjusted the phase by moving tweeters back and forward. Even 1mm makes a difference because at 10KHz the wave length is 340mm.
After that, I used Dick Sequerra Pyramid T9 ribbon super twitters.
A couple years ago, I started use a better parts in crossover and Duelund hook up wires. As result, I can use speakers without supertweeters at all with still good tonal balance.
Now I started using T90a. T90a suted much better to Altecs 604E. They give a huge positive effect. I use auto-transformer (-7.5dB tap). This solution is most SET amp friendly. The extra load of addition supertweeters loaded thought the auto-transformer is very mild.

Duke, I think the L pad or resistor goes in series in front of the capacitor and driver. No?
Ribbon tweeters are mini line sources. They do not send energy up or down. In other words their vertical dispersion stinks. Their horizontal dispersion depends on the width of the driver. The narrower it is the greater is the horizontal dispersion. I would think these characteristics would make a ribbon tweeter excellent for an ambient driver. In thinking about it I would aim it at the opposite side wall just in front of the listening position. This would send the sound off the side wall into the rear wall then on to the opposite side wall before getting to the listener, the longest path. 
George, you would be the guy to buy Hill Plasmatronics. Mesmerized by the purple light. Turn on that gas!
Those tweeters are banned from import to the US. Ozone factories. The Hills avoided this by using Helium to create the plasma. Thus my comment about the gas. The Hills were awful by the way. I helped set a demo pair up. Must have been 1978 or so. The plasma tweeter worked fine if you like point source speakers but the rest of the speaker was terrible not to mention having huge helium tanks and being butt ugly. They did not last long.
In my case adding T90a tweeters didn't change tonal balance too much
but made sound much more refined.
Not just added air, soundstage and high frequency extension (as was expected), 
but also made better separation, more realistic tone of instruments, better bass definition and texture, better soundstage layering.
Thank a lot to @zipost who convinced me to try Fostex super-tweeters.

mijostyn wrote: "Duke, I think the L pad or resistor goes in series in front of the capacitor and driver. No?"

If the input power levels are low then you can probably get away with placing the L-pad in front of the capacitor, but you LOSE the L-pad’s feature of not shifting the crossover point. This is because the impedance of the load that the capacitor "sees" is no longer constant as you turn the knob on the L-pad.  At high power levels, you run the risk of cooking the L-pad because it's seeing a full-range signal from the amp. 

The L-pad is designed to go just in front of the tweeter, where it maintains an 8-ohm load regardless of how much attenuation is cranked in so the crossover point doesn’t change. Also in this location, it is protected by the crossover capacitor.


Those tweeters are banned from import to the US. Ozone factories.

Really!!! someone better tell these guys they’re breaking the law, as these Lansche speakers also use Plasma tweeters, even though they are horn loaded and directional to get the db levels, the Magnat MP-02 are better, no horn and have a 360’ radiation pattern.

Nelson Pass on plasma tweeters
As Nelson Pass said of his massless “ion-cloud loudspeaker” described in the sidebar “[the speaker] gave new meaning to the word ‘transparency.’” And so it was with the Lansche Audio No.7. This loudspeaker’s midrange and treble reproduction was absolutely sensational, and different from that of conventional loudspeakers, whether cone, electrostatic, or ribbon. The Lansche simply disappears as a sound source, not just spatially (which it, along with many other great loudspeakers, does), but also mechanically. By that I mean the physicality of the loudspeaker’s operation—the mechanism by which is creates sound— disappears, replaced by the physicality of the instrument it is reproducing. It’s the kind of sound that produces a “fool-you” realism of timbre, as well as “fool-you” palpability and immediacy. There was an ethereal character to the sound, as though the music existed independently of any electro-mechanical contrivance—“conjured out of thin air” to use Jonathan Valin’s wonderful description.

Absolute Sound on them
The treble had the wonderful quality of being highly resolved without sounding bright. The No.7’s reproduction of cymbals, and of brushes on snares, was revelatory. The corona-plasma tweeter beautifully resolved the transient detail of the stick hitting the cymbal, the shimmer that changes character slightly as it decays, and then revealed the finest inner detail at the end of the decay. This lack of smearing of fine transient detail was spectacular and alone worth the price of admission. When combined with the top- end openness and transparency, this treble resolution produced a stunningly lifelike feeling of actually being in the same room as the instrument.

This is why I said my Magnat's doing highs, are like fireflys dancing in 3d space in front of you.

Cheers George
I’m using enigmacoustics sopraninios permanently charged electrostatic supertweeters to goo effect, if they are in your price range .  
I appreciate the feedback from everybody. Why I would love to try both the magnats and the enigmacoustics, I'm going to work with some fostex t90a's at first to see if I like what they add. If so, I will upgrade at somepoint, so thanks for the advice on those. I'm trying to cross my tweeter over around 15khz, wouldn't a .33uf capacitor be ideal, in a 1st order filter? Considering the 106 db on the fostex, and the 96db for the fullrange.
ENIGM Acoustics Sopranino electrostatic supertweeter

Never seen these before. Wow! killer load for an amp, like most esl’s in the highs.

"Sopranino’s magnitude increases with decreasing frequency. However, with the 8kHz filter setting, there is a very high inductive phase angle between 7 and 12kHz, a region where the impedance magnitude drops to less than 2 ohms. This makes the Sopranino very difficult to drive at this setting"
Cheers George
  • Werner Jagusch autotransformers:

Makes super tweeter load very mild for amplifier.
For example if  you use Fostex T90a and you need align it with speaker sensitivity 97dB.
You connect T90a to -9dB tap. The super tweeter load that see an amplifier is 64 Ohm.
You must put HP filter before autotransformer. You can use potentiometer in parallel to  autotransformer input for adjust cut off frequency.
akwilson501 wrote:

"I’m trying to cross my tweeter over around 15khz, wouldn’t a .33uf capacitor be ideal, in a 1st order filter? Considering the 106 db on the fostex, and the 96db for the fullrange...

"3uf and 1.5uf seem as little high doesn’t it."

By all means go with whatever works! You may well be right.

My thinking was this: The T90A’s contribution will all be going into the reverberant field, and its radiation pattern is fairly narrow so that on-axis 106 dB doesn’t apply. What does apply is the power response - that is, the sum of the driver’s output across all angles. So when you eyeball the factory curves, the off-axis curves are more representative of what the T90A’s contribution will be:

Mentally average the 30 degree and 60 degree curves and imo THAT is a reasonable approximation of what you’re starting out with.

In my specific case it is 15 degree. I measured it.
You want supertweeter will add more high frequencies over 10KHz. If the angle is more than 30 degree it seems to be not good.

I’ve experimented for a long time with implementing a super tweeter into a single driver design and after about a year of back and forth, I decided that it never sounded completely correct. I used many values and brands of resistors, capacitors and even tried a L pad. I was working with the Fostex T90a.  I came to the conclusion that to me I always felt I could hear a gap where the highs of my full range single driver cut off and where the T90a came in. Also the very highs of a super tweeter can sound bright. I had been using high quality resistors and I still experience this. I felt that the T90a never blended in as well as a good 2 way design.  So if you ask me, if I had a single driver speaker and needed more energy in the top, I would be looking at high resolution cables, tubes or other tweaks that would help.  Also, I strongly felt that instead of incorporating a super tweeter into my speakers, I should had went with a really good regular tweeter. Pretty sure it would had blended in better. BTW, I still have my single driver speakers and now I enjoy them very much. I use No super tweeters, just the single driver alone. I just used better cables and different tube amp. I also own these speakers, Rogers LS3/5a, Martin Logan Request, and about 10 tube amps, mostly 300b, and EL34 and 1 KT66 amp.

Scroll down to the bottom of the page.
Dave is as good as it gets. Plus he can do any custom tapers you might need.

Cheers,             Crazy Bill
P.S. I am a devoted single driver user. The Decware HDT. Never felt any need for a super tweeter.

I run a pair of The Horns (from The Horn Shoppe) loaded with Fostex Fe126e drivers. I always loved the directness and transparency but eventually decided to add a little to the top in the form of T90A (and a little to the bottom with a powered sub). 

I built a setup very similar to this (and much cheaper) and really enjoyed being able to dial in the volume on the T90As.
Hi, I have a pair of Lowther 'TP1' back-horn-loaded loud speakers
with Lowther PM5A drivers. The speakers are around 105db / 1 watt.
(a 1 watt SET valve amp drives them to party sound levels with ease!)
I use a pair of Fostex T90a which came with L pad attenuators.
I found I didn't need the attenuators as the main speakers and the super tweeters are of similar sensitivity. I experimented with different values and ended up with 0.33 mic capacitors to 'roll-in' the super tweeters at around 8 k Hz.
I have the super tweeters wired into the same terminals as the speaker cables and have them facing toward the listener, same as the main speakers. 
Assuming that you are correct and your current 15 is 96db, as long as that 96db is based on 2.83v @ 1 meter, you will need around a 10db pad on the Fostex.  The Fostex impedance curve shows that from 10k to 15k,  this tweeter runs nearly a flat 8 ohm,  so the tweeter pad is fairly simple.... For an Lpad,  you'll need a series resistor of 5.47 and a parallel behind that of 3.7.  Then for 6db crossover slope,  I would recommend several caps on hand between 1.3mfd and 2.0mfd and experiment with which blends best.  
I hope this helps, 
Hey everyone, thanks for the help, the community has came together wonderfully for me as usual. I am just getting around to adding the Fostek supertweeters to the Audio Nirvana 15in. I ended up going with a .33uf capacitor with an lpad to attenuate the volume of the tweeter. It pairs up wonderfully with the driver and does exactly what I wanted it to. Really opened up the sound stage and added extra detail to everything. Music is much smoother and bass seems more realistic as well. Everything just flows better now. Couldnt recommend more. I will be saving up for some better tweeters at somepoint, for now I'm satisfied. 
Hi @akwilson501 ,

What kind of capacitor do you use?
The capacitor quality and jumper wire quality is very import too.
I prefer paper in oil capacitors for smother sound.

I don't think this tweeters have enough sensitivity (93dB) for Omega Super 3 high output monitors (97.5dB).
I use Fostex T90A and it is incredibly good. Very detailed, natural and smooth. 

I bought Fostex t90a supertweeters for my Altec 604E 1.5 years ago.

Since most people use these tweeters with a 1st order filter, so I used with a 1st order filter. Moreover, it was technically simple. I had quality Audio Note Cu 1uf copacitors and autotransformers that make it easy to adjust the volume of super tweeters.

The effect of the supertweeters was very positive. More detailed, refined sound, more air and 3D soundstage.

But over time, I began to come to the conclusion, that the high frequencies are still a little lacking, and if I increase the volume of the super tweeters, the upper mids become too bright.

Then I remembered what my friend Sergey ( an experience DIY speaker builder) said that for supertweeters it is advisable to use filters of the 3rd or 4th order. (For example, the Harbath Hl5 has a supertweeter connected by the 3rd order.)

I thought that such a filter would allow me to raise the high friquencies without climbing to the upper mids.

I contacted Sergey for advices and began to build a filter. I used 2 pairs of capacitors that I had: 1uf Audio Note Cu and MultiCap RTX.

The coils were calculated by the online calculator Bessel filter of the 4th order. I bought cheap coils at First, I set the filter to 12K Hertz, but then I got a small dip in the frequency response in the region of 10-12K. So I decided to rebuilt the filter to 10K hertz.

The 4th order filter greatly weakens the output of super-tweeters, so I had to abandon the autotransformer and connect directly through the Dell 0.5 Ohm resistor, as Sergey advised.

It is too early to draw final conclusions. The parts have not passed break in and I do not listen to it for so long enough. In addition, Sergey advised to shunt the 1uf capacitors with small polyester capacitors.

But the impressions are as follows:

The upper middle is less bright, but cleaner, smoother, more detailed. Vocals, saxophone sound clear, without distortion typical for a 1st order filter (when the tweeter and supertweeter are played in parallel in this frequency range). The chorus sounds smoother and cleaner without mud or squeals.

The focus of the instruments is sharper, the soundstage is more transparent.

There are many more upper high frequencies. A lot more details of cymbals, brushes and so on. Flutes, violins have more overtones. More air and better presence effect.

cabinets. After some listening impressions, I’ve realized I cant live with the high frequency response these things put out by themselves. They only go to about 15k hz. and it’s made music lose its sparkle and sense of space and spacial cues.

Yes correct,
A high tech WBer 8 has highs to 12khz+, but its also nice to add the 12k-15khz sheen, glimmer, sparkles.
I don’t like super tweets, planar/nor ribbon tweets.
I plan to add a
DavidLouis clone of a famous Beryllium tweet, has a incredible neo dymium magnet motor.
Under $200,
Running a fantaistic 3.5 inch paper tweeter w a  2.2 Mundorf cap, Cuts off like 9khz,,,
so there is a overlap in fq's at 9k-12k, or what ever fq the DavidLOuis VX8 rolls off, I have no idea. 
I/m willing to fork out $200 to try the Be tweeter. and will post a note on its performance. 92db sens, same as the paper cone tweeter.

Hi @mozartfan ,

I you use 1st order, you will hear your supertweeter in parallel with your speaker 3 octave down crossover frequency.
It can cause modulation distortions in upper midrange where our hearing is very sensitive.
3rd and 4th order work much better for supertweeter. 
I did 1st order and 4th order with the same speakers and supertweeters. When I moved to 4th order I got much more smooth, detailed, and clean sound in the upper midrange. Much less brightness, harshness, graine. It also afforded me to increase level of upper high frequencies (over 12KHz).

My Vandersteen speakers have an additional rear firing tweeter that is adjustable/off.   Using the tweeter, it sounds bigger, but the pinpoint imaging is gone. Off for me is preferred.
String - Me 2, flat to something like 30 khz, 100% out of phase distortion that some find attractive in an overly dead room, or a standup cocktail party… sweet spot there….
My primary speakers for years in my 2.1 channel system were a pair of Voight Towers with Fostex FE-206 full range drivers. I added a Fostex FT17H Horn super tweeter to the speaker just to see if I could expand the soundstage a little more. I had already added a small Martin Logan sub that worked really well with the FE206s. I followed the Fostex recommendation and only added a 1.0uf capacitor ahead of the FT17H. Much to my surprise, adding the tweeter did make a difference. The best way I can describe the change is the soundstage ceiling (I see it as a sound field bubble - I know that sounds weird) was higher or more expansive in the room. That sound field bubble was larger and made the previous field seem cramped by comparison. 

I no longer use the speakers. My son and his roommates are enjoying them at university now. They aren't focused on quality of the sound, but they love how loud they can play!