JBL XPL90 Speaker Review

[GXAlan]

I had trouble finding what model the 3 way version would have been, thanks. These honestly look like very well engineered speakers, vastly superior to a lot of pricey models even today.

The PDF I included in my first post is the whole brochure.

 

[Dennis Murphy]

About a filter a 3kHz crossover point not working, is there an indication why in the spins e.g. directivity error around 3khz?

I'm also trying to understand the difference between the on axis response in the spin and the fundamental in the distortion response. I think there must either be something I'm not understanding about the measurements or the speaker behavior differences at these 2 levels, especially at that 3khz mark. Does the rise in distortion basically 'suck out' the energy at the fundamental so the fundamental response in the distortion test is much lower than in the spin test?

Anyway, it seems like there is all sorts of interesting/nasty stuff going on around that 3khz mark. But it's also interesting that the horzonital problems are at 3kHz and the vertical problems are a bit above 3kHz. Anyone have any ideas about that?

I'm also trying to figure out the frequency plot in the first distortion measurement. I've never seen anything like that in the NRC distortion tests or Amir's other reviews. That huge dip at 3 kHz looks all the world like a reverse null, where the tweeter has been deliberately wired with the wrong polarity to show the crossover point and the degree to which the tweeter and woofer are in phase at the crossover point with correct polarity. The steeper the dip, the better the phase matching with no reversal.

 

[617]

This is also my speaker (along with the 4319). The XPL-200 was JBL’s “audiophile” flagship at the time with both a Titanium tweeter and a titanium midrange. The XPL-90 was supposed to offer a similar sound at lower SPLs and Bass. For a bookshelf, it was surprisingly rated to peaks of 400W although it was nominally 100W continuous handling. Among many JBL fans, the XPL-200 is their favorite as it is more “forgiving” (less accurate) than the M2. Amir has listened to a lot of very high end speakers so the fact that he likes it, is saying a lot. Here is one discussion of the XPL-200.
https://www.audiophilenirvana.com/audiophile-equipment-reviews/jbl-xpl-200a-holy-grail/

The original retail was $650 each ($1300/pair.). This has the tweeter and mid bass from the XPL200. The titanium midrange was the signature feature for the series which this one is missing.

You can occasionally find beat up ones on EBay for $250. This one cost more since it was in better cosmetic condition.

The tweeter uses ferrofluid which MAY have dried out. This was also one of JBL’s first attempts to control diffraction by using the neoprene around the tweeter to deaden some of then effects of the enclosure. With the official stands the speakers were slightly tilted back.

For midrange accuracy, the XPL90 was Stereophile’s most impressive in its era:
https://www.stereophile.com/content/ja-loudspeaker-measurements-table-1-loudspeaker-reviews

Cabinetry on this line was considered to be exceptional, on par with the K2 and Everest line, and some even say that it was better than the S4700 and 4367.

Brochure attached.

Wow, great brochure. These appear to be very high quality drive units for the time period. I would have questions about the big metal dome mid (3 inch voice coil!,) but there's a reason they used it in 4 way speakers. If I saw a good deal on the big guys I would be very tempted.

These last speakers from JBL, infinity and a few others, before HT kind of messed up everything, strike me as a gold mine of good values. These bigger companies knew how to design fundamentally sound speakers with excellent driver engineering. After HT became dominant, the 'high end' started taking off, offering inferior value and poor engineering in many cases.

 

[amirm]

I'm also trying to figure out the frequency plot in the first distortion measurement. I've never seen anything like that in the NRC distortion tests or Amir's other reviews. That huge dip at 3 kHz looks all the world like a reverse null, where the tweeter has been deliberately wired with the wrong polarity to show the crossover point and the degree to which the tweeter and woofer are in phase at the crossover point with correct polarity. The steeper the dip, the better the phase matching with no reversal.

I have seen that a few times although this is one of the worst. It might be a reflection from the fixture at just the right angle and distance.

 

[GXAlan]

I'm also trying to figure out the frequency plot in the first distortion measurement. I've never seen anything like that in the NRC distortion tests or Amir's other reviews. That huge dip at 3 kHz looks all the world like a reverse null, where the tweeter has been deliberately wired with the wrong polarity to show the crossover point and the degree to which the tweeter and woofer are in phase at the crossover point with correct polarity. The steeper the dip, the better the phase matching with no reversal.

Here's the Crossover Diagram.

 

[andreasmaaan]

Regarding the distortion measurement, I think once again people are expecting a bit much from a moderately-priced 6" mid-woofer driven to 96dB/1m

Also, this blip in the %THD in the tweeter's range appears to be a result of some kind of interference in the measurement, and is almost certainly not a true reflection of the drive unit's distortion (see where I've circled the measurements in blue around 3kHz):

 

[617]

Here's the Crossover Diagram.

Are those....bypass capacitors??

 

[wwenze]

Are those....bypass capacitors??

U mean like the .01 parallel with a 12uF and similar?

Probably an electrolytic with a not electrolytic (wanted to say film but considering the age it might be paper for all we know)

 

[xarkkon]

Interesting, another speaker with poorer objective performance but better subjective review!

One might look at these and denounce the subjective reviews, but I think these sort of reviews helps one better understand what the numbers mean and how they actually translate to user experience. Thanks as always!

 

[beefkabob]

Take all the subjective and the panthers with a grain of salt.

 

[Dennis Murphy]

I have seen that a few times although this is one of the worst. It might be a reflection from the fixture at just the right angle and distance.

That probably explains it, unless the panther jumped in and switched the tweeter polarity while you weren't looking. Still, it's kind of strange that it just happens to be at the crossover frequency.

 

[amirm]

That probably explains it, unless the panther jumped in and switched the tweeter polarity while you weren't looking. Still, it's kind of strange that it just happens to be at the crossover frequency.

That is odd indeed. Klippel doesn't have an RTA so it is time consuming to try things but I may experiment a bit to get to the bottom of it.

 

[pierre]

 

[ctrl]

I'm also trying to figure out the frequency plot in the first distortion measurement. I've never seen anything like that in the NRC distortion tests or Amir's other reviews. That huge dip at 3 kHz looks all the world like a reverse null, where the tweeter has been deliberately wired with the wrong polarity to show the crossover point and the degree to which the tweeter and woofer are in phase at the crossover point with correct polarity. The steeper the dip, the better the phase matching with no reversal.

I would agree with that. Could it be that during a restoration operation the polarity of one of the two chassis was reversed?

In the Spinorama measurement the resolution seems not fine enough to show the steep frequency response dip at 3kHz.

The steep frequency response dip at 3kHz is also the reason for the percentage increase in distortion. If the drivers were connected correctly, the spot would be completely inconspicuous.

Whether the woofer was connected with reversed polarity can be easily checked with a 1.5V battery. In case of contact the woofer should move outwards.

 

[andreasmaaan]

In the Spinorama measurement the resolution seems not fine enough to show the steep frequency response dip at 3kHz.

The spinorama measurement and the polar response measurements should be plenty fine enough to show a dip like that if it were there...

The FR of the HD measurement does look suspiciously similar to the 60-degree vertically off-axis measurement, however.

EDIT: scrap that last point - there's no top-end roll-off.

 

[edechamps]

I have added the JBL XPL 90 to Loudspeaker Explorer where it can be compared to other speakers.

Good consistency within the listening window, except when listening above speaker height:

 

[ctrl]

The spinorama measurement and the polar response measurements should be plenty fine enough to show a dip like that if it were there...

You're probably right.

My first suspicion was a reflection from the bevelled edge under the tweeter, but since the frequency response dip around 3kHz is so nicely balanced and exactly on the cossover frequency...

Another possibility would be that the filter slopes of the crossover reflect the desired filter (BU3 or LR4), but the phase shift between tweeter and bass driver is not compensated due to the chassis depth offset - although the woofer has already been moved to the front.

This would then lead to the sound pressure holes above and below the axis at the crossover frequency - as shown in the vertical sonogram.
This would also explain the frequency response dip at vertical +10°.

But then Amir would have had to perform the THD measurement at e.g. ver+10° instead of 0° to get the result shown.

 

[andreasmaaan]

But then Amir would have had to perform the THD measurement at e.g. ver+10° instead of 0° to get the result shown.

I suspect this will prove to be the explanation.

 

[Bruce Morgen]

Adding series resistance, even without a voltage divider, tends to shift the whole response down fairly evenly. The driver integration is going to differ either way. Certainly you can get away with a broadband 3db attenuation.

It honestly bothers me how few audiophile speakers allowed for some simple mid and treble tuning. Totally required for room compensation and can be accomplished with minimal cost. Don't even need l pads or anything, just a few switches.

I've used simple series resistance alone to tame the high end of cheaper speakers with good results -- I just figured that given the fairly sophisticated JBL crossover circuit I'd better keep the "R" component of the high-pass LCR arithmetic in the same ballpark as the factory DCR spec for the tweeter alone, which is what the voltage divider approach does.

 

[Juhazi]

The reason why this vertical off-axis null shows at different angle may come from NFS interpolation to different distance in different tables/pots. At 3kHz wavelength is so short that even few degrees make a difference. And obviously acoustic slopes aren't perfectly symmetric and timing is not perfect either. Again, this is a problem with the NFS - so many variables and so many ways to show response.

There is also wiggle at 1200Hz and red spl peak at 180 deg - leaking sound form the port making interference with on-axis sound.