[konut]

Hi Danny,
      Greatly  enjoy your posts both here and on AV123. Seeing as how your using the NEO 3 in your line arrays, I was wondering if you had considered offering this tweeter as an add-on super tweeter ala the Rocket ERT? The ERT does not have the sensetivity to match my Visaton B200, but need something to add the last octave to the B200. If you could impliment a design to match the features of the  ERT, I'd think you'd have a real winner on your hands for those looking for a 'helper' tweeter. Your thoughts?

[Brad]

Konut,

I had that same idea - but I was actually thinking about new speakers for my car.
After some close measurement, there was just barely room to stuff a B200 and a Neo3 into the doors.

The sensitivity of the ribbon (and speed) looks like a good match for the B200.

This should work for home use too   

[Danny Richie]

Adding a tweeter like that actually doesn't work as you'd think.

What happens is that it overlaps onto the same frequency range that the main tweeter is covering and causes cancellation. You can't get around it. Depending on distance and wavelength it will couple in some areas causing as much as a 6db peak then it will cancel in other areas causing as much as a 15db hole in the response. Then the top end looks more like a roller coster ride than any resemblance of an accurate response curve.

[konut]

Even with using a capacitor between 1.5uf to 0.68uf, and the fall off of the B200 as you noted in another thread?

[Danny Richie]

That would cross them at about 10kHz. But then you still have energy from the woofer at 20kHz. If it is in phase at 10kHz then at 20kHz it might be out of phase and at 5 kHz it might be out of phase. So now you have a 20db suck out at 20kHz and a 15db suck out at 5kHz.

Moving the mic (or your ear) up or down a few inches and it can all flip flop. Dips become peaks and peaks become dips.

To cross at 10kHz you need to use about 120db per octave slopes on each driver.

If I get time in a few days I'll measure on of our M-130's. It plays to about 10kHz. Than I'll add a tweeter with a small cap and you can see what it will do.

Keep this in mind too. As good as that B200 sounds, from 2kHz and up most any soft dome tweeter will clobber it in performance. If you are going to have to add one any way, and with that driver you will, then cross it off no higher than about 3kHz. I'd try it as low as 2kHz because the off axis response of the 8" woofer will already be falling off by then.

[Danny Richie]

Okay here are some measured responses to illustrate what I said yesterday.

I took an A/V-1 speaker and used it to show what happens when you use a high crossover point with a low order crossover.

I used a baffle step compensation circuit on the woofer just like the one proposed for the B200.

I used a single resistor and a 1.0 uF cap on the tweeter.



The red line is the woofer. The Green line is the tweeter, and the Blue line is the two together. Note that from about 6,500Hz to 13kHz they are in phase and show gain.

Notice that from about 14kHz to 18kHz they are out of phase and causing cancellation. Also from 6kHz and down they are out of phase and causing cancellation.

Now look at a vertical off axis response to see the destructive interaction between the drivers as they become out of phase and cancel each other out all over the place as I make slight changes in microphone height.

The Red is on axis. I was a little lower than tweeter axis for this one. I then moved the mic up about 4" each time and the measurements are shown going from Orange, to Yellow, and then to Green.



Looks pretty bad huh? Want to take a guess at what the room response would look like? Pretty bad as well.

Note the tweeter only measured the same way.



Then have a look at an A/V-1 speaker with its standard series crossover and measured the same way.



Next I started all over but used a second order crossover on the tweeter.



Note that there are still areas that are out of phase slightly above and below the crossover point.

Now look what happens to the vertical off axis measured the same way as before moving the mic up 4" at a time and going from Red, to Orange, to Yellow, and then Green.



Pretty bad huh?

[konut]

WOW! Danny, really appreciate the effort involved in making those measurements! It does serve to illustrate the care which must be taken to insure a linear presentation. And while the curves yielded do look 'bad', its only in relation to the excellent response of your standard A/V-1. On axis, its less than a + or - 5db variation. I'm afraid, though, you missed the essence of my intent. A more illustrative exercise would have been to add the NEO 3, with a cap, to the existing A/V-1. This would have been a more accurate representation of what I'm proposing. Don't get me wrong, I fully appreciate your expertise in the design and implementation of speaker systems, and would welcome any suggestions you might have regarding those of us trying to augment the bandwidth of the B200. I thought that, as your using the NEO 3 in one of your products, you could give us more insight to how this might be done effectively, without destroying the properties of the B200 that so many of us admire and appreciate.  aa

[Danny Richie]

And while the curves yielded do look 'bad', its only in relation to the excellent response of your standard A/V-1. On axis, its less than a + or - 5db variation.

Actually a standard A/V-1 is typically within +/-1.5db, and sometimes even better.

http://www.gr-research.com/kits/av1freq.htm

A more illustrative exercise would have been to add the NEO 3, with a cap, to the existing A/V-1.

Yea, but I didn't have a box cut for a Neo 3. I only had a standard A/V-1 test box laying around and it already had an M-130 and a GR-T2 tweeter in it.

I fully appreciate your expertise in the design and implementation of speaker systems, and would welcome any suggestions you might have regarding those of us trying to augment the bandwidth of the B200. I thought that, as your using the NEO 3 in one of your products, you could give us more insight to how this might be done effectively, without destroying the properties of the B200 that so many of us admire and appreciate.

I'd do this.

1) Run a pair of the B200's in an open baffle. Run them in series with each other but by-pass the lower one with about a 62 to 75uF poly cap. This will allow the two of them to share the low frequency ranges equally. This will double power handling and cut distortion in half. This also allows only the top driver covering the upper frequency ranges alone.

I have done this with small full range drivers and it works out great.

2) Cross the upper one to a Neo 3 pdr at about 2kHz or so.

[konut]

HMMMMM......A lot to think about here. What most enjoy about the B200 is its simplicity, and lack of the need for a crossover. The appeal of a simple cap and add on tweeter to get that last octave was too much of a quick and easy solution to a more complicated problem, which only goes to show 'there ain't no free lunch'!    With the thought of trying to preserve as much of the midrange 'purity' as possible, and after looking at the response curve of the B200, would it be possible to use an asymmetrical 1st order crossover in the following manner. Start rolling off the the B200 at say, 1000hz, and rolling in the NEO3 at say 4000hz. With the rising response of the B200 this might not result in the hole that would be expected with a more linear driver. Which also begs the question, how low can the NEO3 be crossed over with a 1st order Xover without running into problems?

[Danny Richie]

If you are going to do it then you might as well do it right. If you are going to put a network on it then you might as well design one that will give you an accurate response.

If you want to keep things simple on the B200 and straighten it out a little then try the LCR network to flatten out the response.