[Carlman]

I have a Hypex 200 watt plate amp in my sub:
http://www.sfb.net/acatalog/HYPEX_Subwoofer_amplifier.html  (see the HS200 model)
This plate amp uses a 4th order crossover network.

And, I have Green Mountain Europa's.  Their crossover network is 1st order.

I cannot get them to blend seamlessly.  I am thinking about using a Marchand electronic crossover to split the frequencies and therefore reduce the integration issue.

I do not know enough to understand why it's not integrating.  However, I was told it had to do with timing or phase issues caused by having 1st order monitors and 4th order sub.

So, another option I was thinking is, how could I make the sub 1st order?  

Thanks if anyone can help.

-Carl

[Shamrock Audio]

Hi,

The reason you're having difficulty integrating your sub has nothing to do with the x-over in the Europa's. It's because of the phase rotation that takes place in all ported systems. Basically, as the output from the woofer "hands-off" to the output of the port, there is an acoustical phase rotation. This puts the sub and Europa's essentially out of phase with each other through those frequencies. The easy solution is to reverse the phase on the sub if your amp allows.

Hope this helps.

[Carlman]

oops.. guess I should've mentioned it's a sealed sub, front firing, and uses an Adire 10" driver.  I've tried inverting the phase.  In or out of phase makes it louder or not as loud.  Since it's not as loud in one of the 2 positions, it seems to integrate better.  However, if I put it in phase and turn down the volume, the effect is identical.  Overall, the bass is a little better in the louder sounding position.  (sorry, can't remember which is the 'better' position, 0 or 180)

It's not like it's way off, terrible sounding... just a little off to the point to where I want to tweak it.

One other thing I tried was using the Hypex internal crossover.  That sounded like I'd added poor quality interconnects to the equation and reduced extension and realism in the highs.  (and soundstage seemed less deep)  The bass integrated slightly better but the cost of the other issues that setup introduced was not acceptable.

[mgalusha]

Carl,

You could try inserting a first order filter in front of the sub amp and bypassing it's crossover if possible or at least set it as high as possible. I suspect part of the integration issue is how fast the sub crossover is rolling off or it could very well be a phase problem. A 4th order filter creates a 360 degree lag in phase, so even though the waveforms are in the same direction they are 1 cycle behind. 'Course there are folks who say that doesn't matter but I'm not one of them.

If you can bypass the internal XO and put a 1st order XO in line with the amp input, it may be easier to blend. It's certainly not expensive to try. A passive 1st order filter need only be a cap and resistor.

mike g

[Carlman]

Mike,
When you said this:
 "so even though the waveforms are in the same direction they are 1 cycle behind"  a light went on.

This is exactly what I think is happening and what I'm hearing.  I believe this is what Roy was explaining as well.  The bass sounds fast but the pace is slow and the details aren't meshing well.

I need to build a simple first order crossover, then.  I don't know anything about how to piece the right parts together though.  Looks like I'll be heading to the bookstore.

I would think a 100Hz filter would be sufficient.  The search is on.

Thank you!!!

edit: PS, I found this to be a nifty tool:
http://www.loudspeakers101.com/CrosFrm1.htm

[doug s.]

are you running the europa's thru the sub amps' x-over?  if you are, stop & run them full-range.  these types of subwoofer amp x-overs typically muck things up.  

re: the issue of the sub being 1 cycle behind, this is true, but typically not audible at typical (<100hz) subwoofer x-over frequencies.  my vmps' are crossed this way w/a marchand x-over, & have worked seamlessly w/a wide variety of monitors.  of course, you could put your subwoofers in the near-field, & this 1-cycle lag becomes a non-issue.  even if you can't run your set-up this way full-time, experimenting w/nearfield placement will tell you if this is a problem.

using a 100hz 1st order x-over on yer subs will mean you will get lotsa output from the subs up into the lower midrange.  not the best solution, imo...

if you are running the europa's full-range, & just use the sub w/its internal x-over, then try a line-level x-over from aci, to insert between the preamp & europa's amp:

http://www.audioc.com/accessories/Hipass.htm

hth,

doug s.

[JohnR]

You should perhaps try reducing the frequency of the sub lowpass filter so that it is "under" the Europas natural rolloff. I don't think a first order highpass will help your issue, the way you are describing it.

[Carlman]

Doug... I'm running the Europa's full range and using the sub's xover/amp all in one plate amp... the Hypex 200 to be exact.

I definitely don't want a high pass filter.  I want the opposite and I would bypass the xover in the Hypex with a simple, 1st order low pass type xover or filter.  

I will try moving it into the room a bit more.  I have it slightly behind the main speakers now.  That's just where it ended up last time I went through this.

-C

[doug s.]

Doug... I'm running the Europa's full range and using the sub's xover/amp all in one plate amp... the Hypex 200 to be exact.

I definitely don't want a high pass filter.  I want the opposite and I would bypass the xover in the Hypex with a simple, 1st order low pass type xover or filter.  

I will try moving it into the room a bit more.  I have it slightly behind the main speakers now.  That's just where it ended up last time I went through this.

-C

you certainly do not want a speaker-level hi-pass filter.  but, i'd strongly recommend trying the aci hi-pass filter.  as a line-level filter that goes between amp & preamp, it will be virtually transparent.  much more transparent than a filter between amp & speaker.  prolly the next best thing to a fully active x-over, and for $30/pair, ya cannot really lose.  if ya ask, aci may even offer a demo.  i plan on trying these for a small system i am setting up, w/my proac reference 8 signatures - extremely revealing speakers that i certainly do not want to muck up...  

also, i would recommend what johnr sez - try lowering the subs' x-over point, if ya haven't awreddy done so...  

and, imo re: what yure contemplating regarding another low-pass filter on yer sub amp, what is awreddy there is likely as good as yure gonna get, unless ya yust go w/a fully active x-over like the marchand.  in this manner, ya would set the subs' x-over point at its max, w/the marchand at the *real* (much lower) x-over point.  and, the europa's would also be set at the marchand's x-over point.  whatever worries ya have about losing transparency by crossing over the europas (active, *or* w/the aci line-level filters) should be more than outweighed by the fact that relieving the europa's woofer of having to strain on low-frequency signals, should still result in a net improved sound from them.

one other thing - for the best two-channel audio subwoofin', ya really need two subs.  best case w/one sub - directly centered between the monitors, & x-over point set low - will still result in a small reduction of soundstage info, vs no sub, imo.  and, this is opposed to an increased soundstage w/a pair of stereo subs.

good luck,

doug s.

[Brian Bunge]

Carl,

You don't exactly have 1st order monitors.  Not completely anyway.  I can't remember if they are sealed or ported, but a sealed enclosure has a natural 12dB/octave rolloff.  This is effectively a second order high pass filter.  If it's a ported enclosure it will have a natural 24dB/octave rolloff.  This is effectively a fourth order high pass filter.  If they are ported try stuffing the port and playing with the sub's XO to see if you can get a better blend.

EDIT:  I also wanted to point out that these ported enclosure rolloffs are somewhat generalizations.  Different ported alignments have slightly different rolloffs.

Also, I really like Doug's idea of stereo subs.  But I might be a bit biased!

[Carlman]

They are front ported, that's a good point.  So far I've found playing with the position a bit more to really help...  I'll try plugging the port to see what happens.

Thanks,
Carl

[Brian Bunge]

Carl,

Keep in mind that stuffing the ports will affect the low frequency extension of the speakers.  Without knowing the specs of the woofer I can't approximate what the response will look like, but there's no harm in trying it out to see how it works for you.

I also meant to mention that Mike Dzurko at ACI very often recommends stuffing the ports in their speakers when mating them to a sub.

[Mike B.]

How long have you had the Adire woofers? I have a pair of
dpl 12's and it took them about three months to break in. At least that is what I am guessing was going on. For the first months they were to bassy and about a half measure behind the upper range drivers. I guess you could call them slow?? Now they have just disappeared. Perfect integration. I have a similar crossover situation. Just a thought.

[Roy Johnson]

Carlman-
Because, like many speakers, the Europa is a two-way, ported design, its low-frequency response has certain phase shift characteristics which must be matched/complemented by an electronic crossover for the "best" blend.

One achieves the best blend by applying a 12dB/octave high-pass filter to the Europas and a 24dB/octave low-pass filter to the subwoofer. The crossover point is ~70Hz.

Why this specific?
Because the Europa woofer, although ported, still behaves like a sealed-box woofer, with the port's output "coming along for the ride" if you will. This is true of other ported woofers too.

So, as in a sealed-box design, its woofer already has a 12dB/octave mechanical (acoustic) rolloff beginning at ~70Hz. Therefore, the output from this woofer also has the phase-shift of the equivalent 12dB/oct. electrical filter = 90 degrees at that 70Hz.

This means a 70Hz tone is time-delayed w.r.t. the midrange tones by 1/4 (90/360) of 1/70 seconds (the period of 70Hz), or 1/280 second- about 3.5 milliseconds (msec or ms). Sorry, back to the topic...

When you add 12dB/octave, electrically, at 70Hz, you now have 180 degrees of delay (1/140 second or 7ms).

When you put 24dB/octave on the subwoofer at 70Hz, a "time-advance" is generated within that filter.
(a poor way to say what is going on, as this is all relative to tones where the time-delay is zero)
Thus, that subwoofer filter has 180 degrees of "advance" at 70Hz.

Put the two together: (-180 + 180)= 360 degrees total difference

The result is that the sub output and Europa 70Hz output are "in phase", in terms of looking at all the peaks and valleys of the sine waves. On transients, they are not, which we perceive as a "mumble" in that range, and a lack of depth of image (depth = time delays).

But we at least hear that the integration is not too bad- that we rarely fiddle with the crossover controls after the subwoofer is broken in (which takes some hard pounding, perhaps at least 50 Hours; Mike B. makes a good point).

Stuffing the Europa port would likely keep the crossover blending a little smoother, since we would hear less port resonance. However, the Europa's port resonance is quite well-damped, so don't expect too much difference, unlike the designs using a small woofer in a large cabinet (much  higher Q = longer-lasting resonance).

This 12/24 crossover technique originated in pro-sound and was "adopted" by the THX company as the THX Crossover. Only the actual crossover frequency is varied, with the sub's polarity kept "normal".

One would need to put a mono subwoofer between the Europas (not necessarily dead on center for decent blending), and approximately "in the plane" of the Europas (depends on the sub). Tune for blending by playing string bass recordings (Christian McBride's "Gettin' To It" or Patricia Barber CDs) that pass through that 70Hz range. You may find the sub needs to be brought forward by 6-10 inches, to help make up for its own time delay, which reaches a maximum at say 30Hz (delay = 1/4th of 1/30 sec down there) for a low free-air resonance sealed-box subwoofer (few and far between).

Drawbacks:

an electronic crossover on the way to the Europa, along with another set of patchcords.
the subwoofer may have its own built-in crossover, of say 12dB/octave, so that new outboard electronic crossover would actually be ordered with a 12dB/octave filter (instead of 24) to add to that sub's 12dB/octave crossover.[/list:u]
Possible immediate remedies:
Set your subwoofer's crossover point to ~47Hz, leave the Europas as "full-range", keep that sub's polarity "normal", and try placing it at the same tape-measure distance to the Europas' woofers, or closer.

For

this crossover point, listen to kick drum, which consists of 45-55Hz tones, for tightness (attack).

A good test is to "skip" rapidly thru the warble test-tone tracks of the Stereophile Test CD #1, not letting each segment last long enough to strongly excite the room. This will tell you clearly how flat the amplitude response is, better than a sound-level meter in my experience.

Also consider that a sub is not necessary. After our new Audio Magic Illusion interconnects were broken in (400 hours), we've never heard such a decent low-end response from our Europas. The rest of that particular system is an Edge M-10 power amp driven directly by a Birdland DAC w/CEC transport, so they are the only interconnects in that system. The digital cable is an Illuminati and the speaker wires are 4-year old Audio Magic Sorcerers.

Hope this helps! These 12/24dB crossover slopes are quite useful on most speaker/subwoofer blends. No other combo of slopes brings the phase shift difference to 360 degrees for dynamic loudspeakers. Not even a first-order crossover to the sub, as the Europas are not a "first-order" rolloff (no dynamic speaker has that).

Carlman, thanks for your kind comments on the Europas, and for your listening party (wish I could have been there).

Best regards,
Roy Johnson
Green Mountain Audio

[doug s.]

Drawbacks:

an electronic crossover on the way to the Europa, along with another set of patchcords.
the subwoofer may have its own built-in crossover, of say 12dB/octave, so that new outboard electronic crossover would actually be ordered with a 12dB/octave filter (instead of 24) to add to that sub's 12dB/octave crossover.

my experience has been that whatever minute loss of transparency may that be added w/the use of a hi-quality active x-over is more than made up for by reducing the stress of the monitors' woofer.  i have never experienced a loss of transparency w/the varied (and some highly detailed) speakers i have had in my rig, when run thru my marchand xm-9 x-over...

re: the added x-over to a powered subs' x-over: if the powered subs' x-over is set at its highest point - say 140hz typical - then its effects would be minimal, imo, if the outboard active x-over is at your recommended 70hz setting...

regards,

doug s.

[Carlman]

Wow!  Thanks for taking the time to go into so much detail, Roy.  I couldn't remember everything you'd said and I wasn't about to ask you to write it all down for me... but you did anyway!  Thank you very much.

This just goes to show what a great company GMA is.  I'd like to think new speaker cables will give me the bass I want but I feel certain they will not.  

I'd rather just wait and see what GMA comes out with in the next few months and just get rid of the sub.  Buying very expensive wires just isn't the right answer to me.  If I were in the market, I'd appreciate the recommendation but I've already spent several hundred on my wires and that's the last set I'm buying for many years.

In any case, thanks for recommendations, facts, and tips.  I've tried some but not all.  Plus, I think something may be wrong with my sub or amp.  We'll see.

-C

[Brian Bunge]

Carl,

One thing's for sure, there's nothing wrong with the enclosure!:)

[Roy Johnson]

doug s.
You're right about the benefits of a good crossover- the Marchand is one we recommend.
Thanks Carlman- glad you enjoyed.
Cool stuff you got there, Mr. Rutledge!

The 140Hz subwoofer crossover example doug s. mentions- here are some things to ponder:

Any crossover circuit, filter, has phase shifts (time delays).
Filter-design graphs show the phase-shifts (delays) gradually go away as you move away from the crossover point, and are mostly gone by ~1/3rd the crossover point for a low-pass filter, and ~3x the crossover frequency for a high-pass filter.
So, a 140Hz low pass filter has a phase error that gradually goes away below 50Hz (1/3rd of 140Hz).

If you want to look up the phase shift degrees for a particular filter, look at "electrical filter design" texts- for I think 3rd yr EE's, in a college library. They are also nicely laid out in some B&O papers reprinted in the AES Anthology from Old Colony.

Any woofer is a mechanical filter, so it too has phase shifts, virtually the same as an electrical filter.
If the sealed 12" has a 30Hz tuning, it has its max time delay down at that 30Hz, an amount of time delay that gradually lessens as one goes up the scale, mostly gone by 90-100Hz (which is ~3x 30Hz).

These graphs, say for Butterworth filters, plot "degrees of phase shift at a certain frequency". A graph might show 90 degrees shift at the crossover point, so the time delay at that frequency is 1/4th of the period of that crossover frequency.

Where the graph shows the phase shift as just a few degrees, say 6, then the time delay at that frequency = (6/360) x (period of that frequency) = period/60 = 1/(60xf).

That is how to get time delay out of phase graphs- you normally cannot just look up the time delays.
Plotting time-delay vs frequency tells us a lot:
on semi-log paper, the frequency runs from L-R along the bottom as usual, and the time delay starts from 0 and goes on "up" (indicating more delay). Time Delay is "+"!!
So, on the same paper:

Plot the sub's delay at each frequency from knowing its box LF tuning (port output not considered).
Plot the sub's filter delay.
Plot the delay from the upper-range speaker's woofer tuning.
Plot the delay from the upper-range speaker's woofer/mid crossover.
Plot the delay that represents the tape measure

difference from the sub to your ear vs. the main speaker's woofer/mid to your ear. It is a constant delay at every frequency- a straight, horizontal line at t = (distance/13500inches) seconds. Move the sub closer/farther, and thus drop this horizontal line down or raise it, respectively.
Plot the delay for the built-in sub equalizer.
Plot the delay for the possibly variable!!! low-filter many subs have below 20-25Hz.
Plot the delay (and flip the polarity) of the sub's port output (which is modeled as two filters- one high, one low, with possibly different slopes, but both "crossing over" at the port's tuning frequency.[/list:u]
You can see how two curves from your filter and woofer tuning might be mostly parallel, or diverging, or converging.

When they are parallel, that means there is even more, but smoothly changing time delay- changing at each frequency. Just add up the time delays from those two curves at each frequency if the two devices are in series- like the LF EQ feeding the sub.
If the two devices are in parallel (like the marchand operating on both the sub and the main speaker), then add up the differences. If they are a constant difference, great, but a constant difference of "zero" is even better.

When they diverge, the phase shift (time delay) could suddenly start to increase at a certain frequency, or the total delay could wind up being just a constant difference between them at many frequencies (a good thing- as maybe one could just pick up the sub and move it).

When they converge, the time delay might suddenly increase/decrease, but just in a narrow tone range.

Knowing the time delays at different frequencies, you can predict certain performance from your system:

Image smearing, because time delay means motion in the performer from front to rear.
Image collapse from front to rear- if the echo arrives too late (too early), you can't hear it if the tweeter is making an entirely new (different) sound.
Transient distortion- taking a moment to get going, and then

always the same time to stop, at a certain note. Not really "ringing", but sounds a little like it.
Timbre distortion- the timing between an instrument's fundamental tone and its harmonics, which we see on the `scope as "the" wave. Change the phase of the harmonics vs the fundamental and vs. each other, you change the timbre, the sound of the instrument, and its attack, and decay, and thus the phrasing, so making the musical intent harder to hear.
It can become harder on your ears, literally, to try and hear through that, especially when playing louder, complex, poorly recorded music (which explains the popularity of soft, simple, audiophile recordings).[/list:u]

Sorry for the length- I can't figure out how to say this any shorter. I wish the magazines had written more on this, as we are only looking up numbers on existing filter graphs in electrical engineering texts, converting their degrees into time, then adding up the delays to see what we've got at each frequency. So we know how to proceed.

Hope this helps! The 12/24 crossover is consistently the "best" solution, but in Doug's case, some of that 24dB/octave for the sub might come from the Marchand, and the rest from the sub's own crossover, and would be modified from near-perfect blending by only two things: the LF tuning of the sub (the lower the better), and the distance differential between sub and main speaker to the ear (should be zero difference). The 12dB/octave part of the filter pretty much takes care of the upper speaker, no matter what its woofer's tuning- as the crossover point set to the main speaker's woofer's tuning, so the phase shift from that upper system is predictable.


Best,
Roy