Stereo Sub Integration

[Chuck Elliot]

As I posted up in my other thread, I just received my F15HP subs. While I waiting for the casters, I’m working on my integration plans for adding the subs to my DIY speakers.

I have measured my DIY with REW. The F3 seems to be about 60Hz @ 12 dB. I always knew that these guys were going to need subs. I could have gotten more bass out of them, but the cabinets would have been huge.

Equipment:

REW software
UMIK-1 microphone
Behringer UCA222 (2 In 2 Out USB->Unbalanced.)
XDA-2 -> XSP-1->(XPA1Ls & F15HPs)

Plan:

1. Re-measure F3 of DIY both left and right individually with REW. XSP-1 set full range.
2. Connect UCA222 to XSP-1. Switch XSP-1 to high-pass. XPA-1Ls off. Dial in F3 to same as DIY speakers with REW.
3. Remove UCA222 and turn XPA-1L on. Re-measure. Slope should now be 24 dB with same F3.
4. Setup F15HPs to PEQ= Off and crossover frequency at something below the F3 from above.
5. Level match the F15HP using REW. (Note: No low-pass bass management from XSP-1)
6. Move crossover point up until bump begins to appear at F3.

I’ve never done anything like this before and really would appreciate comments. Remember this is a 2 channel system and there’s no DSP in the XSP-1.

[Boomzilla]

Hi Chuck -

If the F3 is 60 Hz, then be sure to remember to sum both the acoustic roll-off of the speakers AND the electronic roll-off in the crossover to get a true slope. You (of all people) know this for sure, but just a reminder. The conventional wisdom is to use an electronic crossover (only) at a crossover frequency an octave above the F3 of the speakers to avoid that acoustic roll off. Although you might try a crossover at 120 Hz, I doubt that it would sound too good.

Since your (awesome) satellite speakers are acoustic suspension, you might just use the acoustic roll off of the main speakers alone and then cut the subs in at 60 Hz. Phasing might need to be adjusted, but I'm sure you've got the tools for that too.

The plan you propose seems sound to me (no pun intended...). The only fly in the ointment will be that there are no "calibration frequencies" on the XSP-1's bass management. It's all "by guess, by God." When I had my XSP-1 (and my USP-1), my impression was that the high and low pass filters weren't exactly matched. In other words, the low pass frequency at 1-o-clock didn't necessarily match the high pass frequency at 1-o-clock.

Since you have the REW, you shouldn't be troubled by this. I've found that I get an unnoticeable match in most cases "by ear" without having to actually measure at all, but being the technician that you are, I'm sure you'll be happier with the measurement method.

ENJOY and please post your impressions once up & running.

Thanks - Boomzilla

[Chuck Elliot]

Nov 23, 2013 13:16:07 GMT -5 Boomzilla said:

If the F3 is 60 Hz, then be sure to remember to sum both the acoustic roll-off of the speakers AND the electronic roll-off in the crossover to get a true slope. You (of all people) know this for sure, but just a reminder. The conventional wisdom is to use an electronic crossover (only) at a crossover frequency an octave above the F3 of the speakers to avoid that acoustic roll off. Although you might try a crossover at 120 Hz, I doubt that it would sound too good.

Since your (awesome) satellite speakers are acoustic suspension, you might just use the acoustic roll off of the main speakers alone and then cut the subs in at 60 Hz. Phasing might need to be adjusted, but I'm sure you've got the tools for that too.

The whole intent is to match the natural roll-off of the mains to get a 24dB slope. I'm using the effect for a purpose. A side benefit is that the XPA-1Ls will have some strain removed and stay in Class A longer. Since the speakers are stacked, phasing "should" be less of a problem.

Nov 23, 2013 13:16:07 GMT -5 Boomzilla said:

The plan you propose seems sound to me (no pun intended...). The only fly in the ointment will be that there are no "calibration frequencies" on the XSP-1's bass management. It's all "by guess, by God." When I had my XSP-1 (and my USP-1), my impression was that the high and low pass filters weren't exactly matched. In other words, the low pass frequency at 1-o-clock didn't necessarily match the high pass frequency at 1-o-clock.

I don't have to guess, the purpose of the Behringer step is to measure where the XSP-1 is.

[Boomzilla]

So with the even-order slope, the subs will be 180-degrees out of phase with your mains to match at the crossover frequency?

[Chuck Elliot]

Nov 23, 2013 13:36:17 GMT -5 Boomzilla said:

So with the even-order slope, the subs will be 180-degrees out of phase with your mains to match at the crossover frequency?

I wish I knew the type of crossover the XSP-1 had.

Not all even order crossovers have a phase reversal. Linkwitz-Riley crossovers sum flat at the crossover point with no phase difference.

You do point out an important step though that I missed. REW also measures phase, so I will need to adjust the phase on the F15HPs to match what the combination of the XSP-1 + the natural rolloff is.

Thanks Boom

[Boomzilla]

You're probably right, Chuck - My experience "back in the day" was that anything with odd order slopes (6, 18, or 36 dB per octave) had to be wired in phase for the -3 dB points to sum at the crossover point. All even order slopes (12, 24, or 48 dB per octave) had to be wired 180 degrees out of phase for the crossover points to sum. As you point out, though, my "knowledge" may be somewhat obsolete...

I've currently got my Magnepan 1.6 speakers running full range. They die at about 45 Hz in my room, so I've got the sub dialed in at that frequency and 180 degrees out of phase with the source. It sounds very smooth to me.

[Chuck Elliot]

Linkwitz-Riley crossovers sum at the -6dB half power point. My older Klipsch all have various driver phase reversals, but they're all 6 dB crossovers.

Love to hear current Magnepan speakers sometime. Last time for me was the '80s. I bet yours sound great.

[Boomzilla]

The Maggies DO sound great. That large wave launch and box-less sound can be electrifying. I can understand why the critics love them.

That said, the DefTech SM65s are flatter and wider range. I may keep both...

[frenchyfranky]

Nov 23, 2013 13:55:03 GMT -5 Boomzilla said:

You're probably right, Chuck - My experience "back in the day" was that anything with odd order slopes (6, 18, or 36 dB per octave) had to be wired in phase for the -3 dB points to sum at the crossover point. All even order slopes (12, 24, or 48 dB per octave) had to be wired 180 degrees out of phase for the crossover points to sum. As you point out, though, my "knowledge" may be somewhat obsolete...

I've currently got my Magnepan 1.6 speakers running full range. They die at about 45 Hz in my room, so I've got the sub dialed in at that frequency and 180 degrees out of phase with the source. It sounds very smooth to me.

I have a pair of ERT-8.3 starting roll off around 45 Hz and I let the XSP-1 at full range to benefit the chain ERC-2 XSP-1 XPA-1s to stay in differential reference mode but the sub crossover on the XSP is set to min (50 Hz) and the sub is set to 180 deg out of phase and everythings blended perfectly.

[garym]

Chuck,

I'm assuming you're feeding left and right main channels separately to the two subs (not a summed signal to both). Correct?

If so, where to set the crossover should take into account the comparative THD of the mains and the subs. With most speakers distortion begins to rise as frequency drops, typically beginning around 100 Hz. So the question becomes: which is cleaner at the chosen xover freq, the mains, or the subs? Since you're using two subs you don't have the directionality problem that you'd have with a single sub crossed too high. So you have more flexibility in choosing the optimum xover.

[Chuck Elliot]

Nov 23, 2013 21:31:30 GMT -5 garym said:

Chuck,

I'm assuming you're feeding left and right main channels separately to the two subs (not a summed signal to both). Correct?

If so, where to set the crossover should take into account the comparative THD of the mains and the subs. With most speakers distortion begins to rise as frequency drops, typically beginning around 100 Hz. So the question becomes: which is cleaner at the chosen xover freq, the mains, or the subs? Since you're using two subs you don't have the directionality problem that you'd have with a single sub crossed too high. So you have more flexibility in choosing the optimum xover.

Yes, full stereo subs, not summed.

This is going to be one big science experiment with lots of different things to play with.

The number one thing is to get more proficient with REW. I have already used it to match the level of the main drivers. My DIY speakers use autoformers to adjust the level between the 2 woofers and the horn. The horn is extremely efficient at 108 dB and I got a good match at -14dB for the woofers. It would have been a far more tedious task by ear or with just a SPL meter.

I'm going to dial in by ear on the first pass pretty much as frenchyfranky explained to get to something somewhat usable and continue from there.

I plan on detailing the entire adventure in this thread.

As I tell my wife, the perfect present(Christmas or otherwise) for me has batteries or plugs into the wall and has lots of lights and switches! It doesn't matter what it is I'll love it!

[aboroth00]

I wouldn't worry about the THD of the mains and subs where the crossover would occur. You should be setting it much higher than the F3 anyhow. The whole point of using the subwoofer is because it's more efficient in the bass region than the mains with the added perks they you don't have to tax the speakers or amp to produce those low freq. You will be dealing with room modes depending on the dimensions of your room in that freq region so that might play into where you set your XO point. You can see it on REW so at least you'll know what you're dealing with.

I would sum the signals as it would help more dealing with evening the response in room. The sub should be set low enough to not be discernible in stereo.