[Rick Craig]

I had someone send me a HiVi RT8-II planar ribbon for evaluation of use in a line array. Here's the first measurement, taken on a 22" long by 9" wide baffle:



Looks decent for an on-axis curve; however, considering the size of the diaphragm I had expected more extension on the low end. With a -4db point of 2K relative to the 5-10K area a low crossover point will have some definite issues.


Next is the horizontal off-axis at 30 degrees...



Here the wide element really starts to beam due to its' 2" width. Can't beat the laws of physics. 


And here's the on-axis versus off-axis at 45 degrees...




Next I decided to take a look at the impedance from 1K on up...



Nothing unusual here - the expected flat impedance from a planar driver.


To get an idea of where the fundamental resonance (Fs) is I moved on to running 2nd and 3rd harmonic distortion plots. With planars and ribbons the Fs isn't usually very visible on an impedance curve - unlike what you see for most dome tweeters. The HD plots help determine where the crossover point should be.



The strong peak of the 2nd harmonic @1.5K gives a good indication of the Fs with 2% distortion. Everything else above 1.5K looks acceptable with a moderate (cavity?) resonance at 9.5K.


Next I also measured the Fountek CD2.0M and overlaid the response curves for both tweeters at the same distance and input level.



The top curve is the Fountek and both are pretty close in sensitivity. The main differences are at 10-20K and below 3K. The upper octave peak of the Fountek will be not much of a problem to handle in the crossover; in addition, some of the inherent comb filtering of an array with smooth this out. The response below 3K becomes more problematic for the HiVi, especially considering that it has a wide faceplate with increased C-T-C spacing from the woofers in an array.

This is where the Fountek's horn loading of the diaphragm becomes an advantage because it extends the low frequency response. If you use a passive crossover with the HiVi all you can do is shelve down the response above say 2K to achieve a lower crossover point (but not too low because of the 1.5K resonance). The result is that you lose sensitivity (about 3db) but now you've created another problem - the off-axis response above 2K is reduced and we already know from the horizontal plots that there's a lack of energy in the upper octaves. You're also going to have to double the power input (+3db) to achieve the same output level that was available before tailoring the response. This increases the distortion of the driver but since there will be several of them in a line we're not as concerned about that because the thermal power handling will still be very high.

If the planar owner decides to use the HiVi, it's a good thing that he has an active system with a DSP crossover driving it 

[JoshK]

I agree those HiVi's don't look too promising. 

[FredT300B]

I wouldn't assume measurement results that indicate a response or distortion issue will result in an audible problem in a real world listening situation. 

John K's measurements of the Fountek ribbons and his comments about these measurements indicate their sound would be unimpressive: "...Of the three true ribbons, this has the best performance above 4kHz. It has the worst performance at 2kHz, rendered somewhat irrelevant because they are all terrible at 2Khz..." This suggests the Founteks would sound distorted in the critical range below 4K, but the ones in my line arrays and my new Jim Holtz WMTMW "Statements" sound pretty sweet to me. (Acutally, the Fountek in the Statements crosses at 3.7khz, so it isn't used in the problem frequency area).

Likewise, if planars test so badly why do they sound so good? The "tweeter" on my Maggie 1.6's was at least two inches wide, but I was never bothered, or even aware of, a lack of treble dispersion when listening slightly off axis. Ditto for the Neo 8's in the Alpha LS, LS-6, and LS-9, and the HiVi planar in the $22,000/pr Coincident Technology Total Rererence loudspeaker. It seems to me the issue of high frequency dispersion is more a theoretical concern than a practical one. You can measure it, and under certain circumstances you can hear it, but in real listening rooms, with proper speaker placement and toe-in, it doesn't seem to be problem.

[doug s.]

i'd investigate the rt2 series drivers.  these have much better horizontal dispersion.  wertical dispersion is lousy, but not an issue w/line source.  i have heard them in the swans speakers (hi-vi's own company), & i have a pair of coincident victory's which use a (supposedly custom iteration) rt2c-a driver, & its treble response is phenomenal.  but, it's crossed over at ~4.5khz...

doug s.

[BrunoB]

Next I also measured the Fountek CD2.0M and overlaid the response curves for both tweeters at the same distance and input level.



The top curve is the Fountek and both are pretty close in sensitivity. The main differences are at 10-20K and below 3K. The upper octave peak of the Fountek will be not much of a problem to handle in the crossover; in addition, some of the inherent comb filtering of an array with smooth this out. The response below 3K becomes more problematic for the HiVi, especially considering that it has a wide faceplate with increased C-T-C spacing from the woofers in an array.

The Fountek CD2.0M frequency response is typical of  ribbon tweeters having  a  strong resonance at 1.5KHz. This resonance moves up the FR below 2KHz. Not so with the HiVi.  I am not saying that the HiVi does not have a resonance, but is not visible on the frequency response (no bump). It would be interesting to see the waterfall plot or the impulse plot of both driver.

Bruno

[Rick Craig]

Next I also measured the Fountek CD2.0M and overlaid the response curves for both tweeters at the same distance and input level.



The top curve is the Fountek and both are pretty close in sensitivity. The main differences are at 10-20K and below 3K. The upper octave peak of the Fountek will be not much of a problem to handle in the crossover; in addition, some of the inherent comb filtering of an array with smooth this out. The response below 3K becomes more problematic for the HiVi, especially considering that it has a wide faceplate with increased C-T-C spacing from the woofers in an array.

The Fountek CD2.0M frequency response is typical of  ribbon tweeters having  a  strong resonance at 1.5KHz. This resonance moves up the FR below 2KHz. Not so with the HiVi.  I am not saying that the HiVi does not have a resonance, but is not visible on the frequency response (no bump). It would be interesting to see the waterfall plot or the impulse plot of both driver.

Bruno

The low end really depends on how the driver is loaded on the front (horn coupling) as well as the rear chamber damping and fundamental resonance of the radiating element. There's a tendency for designers to cross ribbons and planars too low because they often overlook distortion plots and implement them like you would a typical dome tweeter.