L 101 crossover point and polarity

[davidpou]

Hi everyone,
I have been torturing my L 101 since I have them (around 6 years) and recently my new torture instrument is a mini DSP 2x4 HD. So the passive crossover has been removed (a condo and a pad is protecting the compressions) and the speakers are run in bi amp with two crown D75 in mono for the woofers and a MC275 mkII for the tweeters.
I am playing with the crossover part. I know JBL crossed the drivers at 1500 Hz but there is plenty of space to move this point as both drivers once equalized, overlap from 400 Hz up to 4000 Hz ! (Amazing drivers to me…)
My question bears on polarity: I was convinced that both drivers were of negative polarity. (I checked my self the woofer with a 9V battery). But if I cross them in LR 24dB and put both drivers in phase I get a dip while if I inverse the tweeter polarity I get the nice flat result you see on the picture….
I precise that I respected JBL principle and double checked the wiring: everything is black on black and red on red and the polarity is reversed (for both drivers) at the speaker terminal….
By the ear I have difficulty to tell the difference honestly but since I inverted the tweeters it seemed more harmonious and homogeneous…..
How come ?
What crossover point would you choose ?
7488774886

[grumpy]

Both amp types preserve or invert phase similarly?

[davidpou]

Well I would think an amp cannot invert phase... can it ?
They are on different plugs though and that might be the cause....if the plugs are phase inverted ... I will check this...

[johnlcnm]

A L.R. filter must have the drivers phase aligned in space to have a flat response on axis. a member here, 4313b used an alignment method that seems to work well. He set the two drivers electrically out of phase. Varies the time delay to the woofer for a max null at the crossover. Set the phase back to in phase. The response should be flat on axis at the crossover. This is a common problem with horn systems. The horn driver is not in the same plane as the bass driver.

Regards,

John

[grumpy]

Well I would think an amp cannot invert phase... can it ?

Certainly it can. Some do, some do not. Between using different manufacturer amps
and bridging, it's definitely worth checking.

[hsosdrum]

Certainly it can. Some do, some do not. Between using different manufacturer amps
and bridging, it's definitely worth checking.

He may be quibbling the difference between *phase* and *polarity*.

[grumpy]

Good point...perhaps I've sine waves on the brain.

Signal inversion (positive voltage in, negative voltage out) with no delay
added is the same as polarity swapping at the driver terminals... hence my question
about the amps. Sounds like it was sorted out sufficiently/empirically :)

What's being done in the DSP re crossover slopes (24LR filters should sum flat in magnitude at crossover)
or driver phase-center + filter delay compensations could affect things as well, but I'm assuming those
are not the issues being questioned.

[Earl K]

You know, the 1/2 wave length of ( 1000hz ) is right around 6.75". ( ie; 1124/1000x12/2 )

It's not hard to imagine that the all important "acoustic center" (http://orbit.dtu.dk/fedora/objects/orbit:20068/datastreams/file_4439883/content)of the horn driver (combo) is behind that of the woofer ( by about, just that distance ).

That half-wave differential ( in acoustic centers ) will cause a null around the crossover point (with an LR24 slope ) .

http://www.lansingheritage.org/images/jbl/specs/home-comp/175dlh/page2.jpg

David, your understanding of how crossovers really work ( & how they interact with acoustic devices ), is under-developed ( & only partial ).

You need to overlay ( account for ) the acoustic realities onto the (predicable) electrical slopes, to arrive at what will really occur ( at crossover ).

:)

[Earl K]

A L.R. filter must have the drivers phase aligned in space to have a flat response on axis. a member here, 4313b used an alignment method that seems to work well. He set the two drivers electrically out of phase. Varies the time delay to the woofer for a max null at the crossover. Set the phase back to in phase. The response should be flat on axis at the crossover. This is a common problem with horn systems. The horn driver is not in the same plane as the bass driver.

Regards,

John

David, delay the woofer just like John has suggested .

Then your L.R.24 slopes should combine properly ( with everything being held in positive polarity ).

FWIW, I haven't used a pro amp that inverted the output signal ( in decades ) .

That's one good reason to use pro amps I guess .

:)

[boputnam]

First of all, David - some nice measurements.


A L.R. filter must have the drivers phase aligned in space to have a flat response on axis. a member here, 4313b used an alignment method that seems to work well. He set the two drivers electrically out of phase. Varies the time delay to the woofer for a max null at the crossover. Set the phase back to in phase. The response should be flat on axis at the crossover. This is a common problem with horn systems. The horn driver is not in the same plane as the bass driver.
JohnThis is a pretty important post - what it does, without saying, is - it discriminates between electric and acoustic phase. The LR filter needs things in-phase electronically to produce a flat response at crossover. Fine. But what you are measuring is the acoustic phase - this has to do with the effects of (delays caused by) the LR crossover filters, and the acoustic centers (separation) of the drivers.

There's one slight caveat to John's post: There is always acoustic time delay from the LF - the magnitude of which varies by the knee of the LPF (i.e., the amount of delay is frequency dependent). 4313B didn't "vary" the time delay "to" the LF (that couldn't be done this way), he just inverted the LF (put it 180° electrically out of phase) which in that instance smoothed the acoustic response.



My question bears on polarity: I was convinced that both drivers were of negative polarity. (I checked my self the woofer with a 9V battery). But if I cross them in LR 24dB and put both drivers in phase I get a dip while if I inverse the tweeter polarity I get the nice flat result you see on the picture….
Unless a purist on what JBL had done, I wouldn't get too consumed by mimicking their settings. As you note, the frequency response of those drivers gives great latitude in setting a crossover point. Bear in mind the vocal range, and strive to keep the point out of the most audible portion of that range. You may find you like an acoustic dip at crossover - that is a cancellation which may clean-up the acoustic response in your setting for your taste. Or, maybe you don't.

FWIW, I toggle to inverted subs at every show I mix - during system check, sound check and live during the show. Physics give us a good starting point but the venue behavior is dynamically affected by temperature, humidity and body count, and if I can clean-up the interaction between the subs and mains, I'm all over it...:bouncy: