[Jim Griffin]

A few comments. 

First, AJ pointed out the AT-1 low frequency shortcoming (38 Hz 3 dB down point vs. 29 Hz claimed)  in the Home Theater report.  If you read the report, some of that difference likely is room placement and measurement issues experienced by the author during his testing.  In room measurements can be difficult to nail down.  But the speaker was short of specified capabilities though.

Second, Danny points out the problem with running the dual Jordan drivers full band in the
FocalStage floor standing speakers.  A few years ago I remember a British Hi Fi magazine review of an earlier Aurousal speaker which had two Jordan JX92S drivers run full range with the same arrangement.  The published plot showed a major dip in the treble area and I suspect the same from their latest implementation.  This FocalStage speaker really needs a proper crossover to enable these drivers to excel.

Third, it is great to see Martin King and JLM check-in on this thread.

Jim

PS for AJ.  I'll be in your area most of January per current plans. 

[planet10]

I. M. Fried Model H (the 'famous' coffin stereo sub, with 8 inch woofer, 

Are you sure it was as early as 1976? We were selling these i don't recall them being introduced that early.

http://www.t-linespeakers.org/classics/friedH/fried_H.html

was rated to 114 dB at 17 Hz

Probably a bit of an exaggeration.

dave

[JLM]

Are you sure it was as early as 1976? We were selling these i don't recall them being introduced that early.

http://www.t-linespeakers.org/classics/friedH/fried_H.html

Probably a bit of an exaggeration.

dave

I think it must have been 1976.  Heard them over the summer and took a year as a collage student to comprehend what I'd heard (real music versus rock monitors) and save for the Model B 2-way satellites that I had my senior year in college (1977-78).  And based on what I heard, including the organ music I played in that chapel, I believe Bud's numbers could at least be possible (they went way deeper/louder than my current 'babies').  Frankly Bud (Fried) didn't recommend them for home use (but try telling that to a enthusiastic young audio nut). 

I visited the old boy once in Philly.  Audio was his hobby.  His 'listening' room was 3 tall plaster walls, old plate glass on the 4th wall, plaster ceiling, wooden floor (if you could find it through all the speakers, magazines, etc.).  His ordinary turntable was set up on a card table, the garden variety receiver rested underneath on a pile a audio magazines on the floor.  A real 'speaker guy'.

[Kevin Warne]

'The real point is whether the stated sensitivity matches up to the stated extension given the TS parameters and box volume.  And of course, small drivers are inherently very limited as to the amount of low bass they can generate without simply falling apart.'

This is another myth. Sensitivity is nothing at all to do with bass extension, but efficiency. If you quote -3d, -6db values then these usually imply constant voltage. Of course if the design has wild imedance drops at low frequency then it may be drawing more power than the amplifier can handle and amplifier ability becomes a limit (or more accurately the power handling ability of the ampliifier). Therefore it is not a question of the drivers falling apart, but the amplifier, and I have never known a decently designed solid state amp to fall apart in this manner unless it is dropped from a considerable height 

The speakers will actually start distorting well before anything actually falls apart.  Its another myth that you need giant bass woofers to generate great sounding bass. Of course the maximum bass volume will increase but dynamic range nearly always suffers - at the lower end of the volume curve you simply don't have the delicacy and accuracy of a well designed full-range driver.

Contrary to common misconceptions transient response is also not helped at all by just scaling everything up. Nor does using 2 drivers instead of one necessarily increase the volume capability across the whole spectrum if the frequency bandwidth is split between the drivers. Using 2 or more full-rangers does, as effective power is doubled at all frequencies. I can go into more detail here if required.

[Guy 13]

Hi Kevin and all Audio Circle members.
I love when someone post something that makes sense
and is easy to understand.
Thanks.

Guy 13

[Kevin Warne]

Second, Danny points out the problem with running the dual Jordan drivers full band in the
FocalStage floor standing speakers.  A few years ago I remember a British Hi Fi magazine review of an earlier Aurousal speaker which had two Jordan JX92S drivers run full range with the same arrangement.  The published plot showed a major dip in the treble area and I suspect the same from their latest implementation.  This FocalStage speaker really needs a proper crossover to enable these drivers to excel.

Third, it is great to see Martin King and JLM check-in on this thread.

Jim

Hi Jim. Interesting observation and worth commenting upon. Yes the FocalStage does use the drivers in twin formation. This by itself does not necessarily cause any frequency irregularities (comb filtering) as the listener is typically on a central axis and they are equi-distant from the listener (or so close to equi-distant that such filtering is much less than any perceived room acoustic anomalies. The theoretical issue is with the teeter on and yes there is no crossover here. It's a design consideration but it was deemed that the advantages of having a switchable tweeter that compliments the dispersion profile of the Jordan and can be adjustable outweighs the advantages of having a crossover so potentially smoother response curve at the very top end. This is because:

1. The dispersion profile of the tweeter is very different to (and compliments) the Jordan so the slight comb filtering effect is actually not noticeable at distances greater than 1M and the usual listening distance for such speakers is excess fo 3M
2. Deliterious effects of any crossover on transient response and distortion as we know
3. Deliterious effect of completely splitting the frequency response on coherence
4. You lose any ability to adjust the dispersion profile with a full crossover

At the end of the day it's swings and roundabouts and there are differing opinions and preferences. We opted for the crossover free design as the improvement is clearly audible to us and more flexible at 3M listening distance - you get a number of advantages and no audible disadvantage at normal listening distance.
The problem with the standard frequency test is that it is done close-field so will show up comb-filering effects that are just not audible at normal listening distance (where they are far outweighed by room nodes).

One example where listening really is the all-important test and a standard frequency plot does not give you the full picture.

[Kevin Warne]

Something else I forgot to mention. You can't avoid comb filtering effects in any normal stereo setup anyway as you are listening to 2 independent high frequency sources that are separated by a relatively large distance. You will get very large HF dropouts as you move you head from left to right in the soundfield, even if you rotate your head slightly, and this is commonly not deemed an issue. If you measured the effect using a standard frequency plot it would look horrendous compared to any theoretical issue with doing away with the crossover. Of course it's rarely an issue in reality with real music in a real room, or we would have done away with stereo decades ago!

[Rick Craig]

A good test would be to take and build two systems with the same woofer, one ported and the other a MLTL. Measure them outdoors far away from boundaries and see how the bass response compares. Also measure the woofer nearfield to check for distortion using the same drive level. Any takers?

[Danny Richie]

Interesting observation and worth commenting upon. Yes the FocalStage does use the drivers in twin formation. This by itself does not necessarily cause any frequency irregularities (comb filtering) as the listener is typically on a central axis and they are equi-distant from the listener (or so close to equi-distant that such filtering is much less than any perceived room acoustic anomalies.

This is not true at all.

You don't just hear the on axis response. Your total in room response includes on axis and off axis information. When the off axis is irregular then you get an irregular room response. You'll have hot spots and dead spots that disrupt a sense of ambiance and imaging.

It's a design consideration but it was deemed that the advantages of having a switchable tweeter that compliments the dispersion profile of the Jordan and can be adjustable outweighs the advantages of having a crossover so potentially smoother response curve at the very top end. This is because:

1. The dispersion profile of the tweeter is very different to (and compliments) the Jordan so the slight comb filtering effect is actually not noticeable at distances greater than 1M and the usual listening distance for such speakers is excess fo 3M

This isn't true either. With multiple high frequency drivers the tweeter will not compliment the dispersion of the Jordan. It will disrupt the top end response with cancellation.

And why worry about the horizontal off axis response of the Jordan falling off and not worry about the vertical off axis response? The largest and almost always untreated part of the room is the ceiling. Room treatments and typical room furnishings will diffuse the side wall reflections to some degree. But the ceiling reflection will be a direct reflection of the vertical off axis response.

Let me give you some examples with some measurements I made three years ago. These are using 3" full range drivers mounted very close to one another. So the comb filtering effects are going to be near as bad as using 4" drivers with the acoustic centers further apart.

The Red and Green lines are the response of each driver shot independently. The Blue line is the two together. To get the Blue line to not show any deviation in the response was not easy. It had to be dead on between them. Moving the mic up or down even a half of an inch starts showing upper frequency range cancellation effects.



This is moving the mic up 4".As you can see the top octave has dropped out over 15db.



Here is the response 8" up, and already there are multiple cancellations.



And here is is 12" up. Now we have three cancellation points that effects the response down to 2kHz.



And again, this will be worse with 4" drivers and a longer acoustic center spacing.

And adding a tweeter will compound this effect with even more cancellation points.

The problem with the standard frequency test is that it is done close-field so will show up comb-filering effects that are just not audible at normal listening distance (where they are far outweighed by room nodes).

Also not true. They are easily audible. Especially the vertical off axis nulls.

[Danny Richie]

Something else I forgot to mention. You can't avoid comb filtering effects in any normal stereo setup anyway as you are listening to 2 independent high frequency sources that are separated by a relatively large distance. You will get very large HF dropouts as you move you head from left to right in the soundfield, even if you rotate your head slightly, and this is commonly not deemed an issue. If you measured the effect using a standard frequency plot it would look horrendous compared to any theoretical issue with doing away with the crossover. Of course it's rarely an issue in reality with real music in a real room, or we would have done away with stereo decades ago!

It doesn't work that way. Stereo signals are two different signals.

You are causing cancellation effects at the source.

[Danny Richie]

Oh yeah. None of these are true either.

2. Deliterious effects of any crossover on transient response and distortion as we know
3. Deliterious effect of completely splitting the frequency response on coherence
4. You lose any ability to adjust the dispersion profile with a full crossover

A crossover has little effect on transient response or distortion. You can cause distortion of the signal from the smearing effects of a slow discharging capacitor, but that's another story.

Splitting the signal to different drivers does not necessarily cause any loss in coherence either. It can if not designed properly, but not typically.

And you don't loose any ability to adjust the dispersion using a crossover in the design. Quite the opposite. You can control the off axis response using a crossover and running the drivers without a crossover will cause off axis cancellation.

And you run the tweeter with no crossover? Really?

[planet10]

And again, this will be worse with 4" drivers and a longer acoustic center spacing.

The Jordans should be classified as 5 1/4" drivers.

And you run the tweeter with no crossover? Really?

In post #1 Kevin says he is using a high pass on the the tweeter at 9k,so there is an XO, just not on the FRs.

dave

[planet10]

From the industry thread linked in my post just above:

The speakers have 4.5 db of baffle-step compensation built in to cater from some rear-loss at low frequencies.

To the question of efficiency: We have 2 drivers rated at 84 dB which gives 87 dB efficiency (90 dB sensitivity given parallel connection) less 4.5 dB BSC yields 82.5 dB efficiency.

And one would assume some sort of filter (ie XO) to achieve the BSC?

dave

[Danny Richie]

The Jordans should be classified as 5 1/4" drivers.

That would mean that the acoustic centers are even FURTHER apart and the effects will be much worse than what I posted above.

In post #1 Kevin says he is using a high pass on the the tweeter at 9k,so there is an XO, just not on the FRs.

I knew he at least had to have a cap on it but this statement made it sound like he was not.

The theoretical issue is with the teeter on and yes there is no crossover here.

[Jim Griffin]

Danny has already covered most of the comments by Kevin on the use of two full range drivers full band such.

My experience with the Jordan JX92S driver goes back more than 10 years with first a full range mini-monitor (ported 7-9 liter box), next a Jordan with a ribbon mini-monitor, and finally a Jordan with a ribbon MLTL version.  Baffle step compensation was necessary with the Jordan for any of these designs so some bit of circuitry was included to balance the frequency response for most room environments.  The ribbon driver was the Aurum Cantus G2si.  Akin to most full range drivers, the Jordan JX92S beams (reduced dispersion in both vertical and horizontal plane) starting at 3000 Hz.  A crossover to the ribbon was at 3000 Hz.  The results were exceptional as the horizontal plane performance was wide and airy across the entire frequency band.  The small ribbon had a very minor impact on the vertical plane coverage.  My prototype MLTL box has a switch so I can change between the Jordan driver run full range or insert the crossover with the ribbon.  I've yet to find a listener who did not prefer the crossover settling vs. running the Jordan full range.   A full range driver gives up significant (and audible) dispersion when they are run full range vs. a proper tweeter. 

If I were improving your FocalStage implementation, my advice would be to implement a 2.5 way crossover.  The 0.5 part would be to roll off the bottom woofer to achieve baffle step compensation and then have a crossover to the tweeter in the 3000 Hz area.  You would have a much more balanced frequency response and proper dispersion within the room.    You get the full impact in the bass region of the two drivers and you have a substantial improvement with in the treble range.   

[Danny Richie]

To the question of efficiency: We have 2 drivers rated at 84 dB which gives 87 dB efficiency (90 dB sensitivity given parallel connection) less 4.5 dB BSC yields 82.5 dB efficiency.

And one would assume some sort of filter (ie XO) to achieve the BSC?

I think that would make it 85.5db on the sensitivity scale.

And some BSC would suggest the use of a filter. Or maybe he is just seeing the comb filtering effects eating up the response above 1 or 2kHz.   

[Danny Richie]

If I were improving your FocalStage implementation, my advice would be to implement a 2.5 way crossover.  The 0.5 part would be to roll off the bottom woofer to achieve baffle step compensation and then have a crossover to the tweeter in the 3000 Hz area.  You would have a much more balanced frequency response and proper dispersion within the room.    You get the full impact in the bass region of the two drivers and you have a substantial improvement with in the treble range.

I second that.

Or try this on the two woofers.


____________________ +


                                          -
                                         /
                             ____ +
                            /
                         cap
_____________/______ -


That puts the two of them in series, but you by-pass the lower one with a cap value between 47uF and 72uF. You might settle for a 56uF or a 62uF value.

This lets the two of them share the lower ranges (so power handling is good) while only one driver plays the upper ranges. This does away with the comb filtering effects, and it keeps the two woofers in phase.

I still recommend a crossover point no higher than 3kHz to the tweeter. Without question, it will sound better and have MUCH better off axis response.

[Danny Richie]

I am going to offer you some more help with these Kevin.

Jim and I will tell you without question, and from experience, that a crossover between the tweeter and upper woofer will sound better.

However, if you are dead set on using no crossover or letting the Jordan's play full range then try this....

Mount the tweeter on top of the box (facing up) or on the back of the box. This will still give you an added since of air, spacial cues, and improved imaging without adversely effecting the on axis response with comb filtering effects. You get the benefits of the added tweeter without the adverse effects. You might even play with any in line resisters to let the level come up a little.

[Kevin Warne]

I am going to offer you some more help with these Kevin.

Jim and I will tell you without question, and from experience, that a crossover between the tweeter and upper woofer will sound better.

However, if you are dead set on using no crossover or letting the Jordan's play full range then try this....

Mount the tweeter on top of the box (facing up) or on the back of the box. This will still give you an added since of air, spacial cues, and improved imaging without adversely effecting the on axis response with comb filtering effects. You get the benefits of the added tweeter without the adverse effects. You might even play with any in line resisters to let the level come up a little.

Nice to hear but I have considered the options and don't require any help thanks very much. Yes I agree the Jordan sounds better with a tweeter but disagree that a crossover improves the sound at normal listening distance. The Focalstage does not imploy a full crossover, only a high pass filter, which is effectively half a crossover so it is a 1/1.5 way design. You can test the effect of the tweeter by switching it in/out as many times as you want and it sounds better and smoother 'in' with no comb filtering noticeable I can assure you. I am not a great fan of layers of electronics for issues that don't exist in practice and am very happy with the design, as are very many others.

And yes stereo will display comb filtering effect with any HF mono signal according to your definition, such as a flute playing centre-stage for example. It;s just not a issue unless you are a pedantic addict of frequency plots that tell you little to nothing about sound realism. I understand you are trying to offer advice as as how you would do it and much appreciated but I have already done years of research and listening (over 30 years actually) so I know a bit about what sounds decent.  I am alsoa musician/composer and have an extremely good ear for what sounds right. I have rained myself play by ear alone.

If I was interested in a 3-way multi-way I would not use 2 very capable full-range drivers for a start. As for adjustable dispersion I mean the ability of he listener to adjust the dispersion. (to some degree as that is one of the effects of increasing the tweeter level) The Jordan will start to beam above around 3K and the tweeter half x-over point (capacitor) is a great deal higher than this. Room reflection anomalies are not any issue at all as with a semi-line source ceiling and floor reflections are much reduced as you probably know.

Thanks for telling me how a 3-way speaker works though - I never would have guessed!

[Danny Richie]

Kevin, it may sound just fine to you and you may be real happy with it. However, as we are trying to tell you, the design will cause cancellation effects that will cause the response to be far from accurate.

For instance here are some other measurements that I took about the same time as the ones I posted previously.

This uses one of the 3" full range drivers with a ribbon tweeter on top of the cabinet. It had a 1uF cap on it. Finding a position on top of the box that did not cancel out the on axis response was not easy. It took careful movements of about 1/4" or so at a time to find the right distance that did not cause comb filtering effects.

Here is the woofer and tweeter response independently of each other.



Now this is as good as I could make the response by playing with distance and mic height. Note there is some cancellation going on even in the lower ranges because of the first order roll off.



Now check out what happens when I move the mic up a few inches each time going from Red to Orange to Yellow.



We are talking 15db swings easy from a few inches of movement.

This is not the same thing as room related cancellation effects of two speakers. Two accurate speakers will produce two independent room responses that are a reflection of the accuracy on and off axis of each speaker.

You are creating cancellation effects at the source.

Again, if you are dead set on this type of design then try facing the tweeter up or to the rear of the speaker and you can avoid these problems.