Using 2 full range drivers per side

[maggiesnmacs]

Does anyone use 2 full range drivers on each side? It's pretty "common knowledge" that this will sound terrible and lead to horrible comb filtering, yet when taken to the extreme with line arrays it doesn't seem to be an issue and in room there are so many reflections that my belief is that comb filtering is a non-issue in most real applications. So I used 2 Dayton RS100-4's in a ported box on each side crossed at 90hz and running up to 20k. At first listen I felt I could hear comb-filtering and that it sounded weird somehow. So I played a sine wave and could definitely hear the comb filtering and assumed my experiment concluded. I let the speakers sit until I got to returning them to 1 RS100 per side. But being the lazy person that I am I left it for a while until I noticed, I didn't hear that weird comb filtering effect I thought I had heard before. I think it may have been placebo and now that I've had this setup for about 6 months I'm convinced comb filtering is a non issue in most real world applications. It makes sense to me that the acoustics of being in a room would cause all of the reflections to make the two main signals being out of phase an unnoticeable issue.

Now I have no proof to this other than my own observation, but I would like to do an abx test where I compare 1 Dayton to 2 at a low enough volume that the excursion is a non issue. Has anyone else tried just using 2 drivers like this? Also the reason that I did this is similar to why line arrays are done, to decrease distortion and also be able to lower the crossover

 

[Squeak]

That's my observation too. Comb filtering happens when there is interference happening between the two wavefronts. With two very similar drivers placed very close to each other, I don't think interference will be that much of an issue.
It's only when two very different drivers overlap like a tweeter and a mid that it becomes objectionable.
With fullrangers it is easily corrected with equalization.

 

[maggiesnmacs]

That's my observation too. Comb filtering happens when there is interference happening between the two wavefronts. With two very similar drivers placed very close to each other, I don't think interference will be that much of an issue.
It's only when two very different drivers overlap like a tweeter and a mid that it becomes objectionable.
With fullrangers it is easily corrected with equalization.

I actually disagree with this reasoning for why comb filtering isn't an issue. I believe having a tweeter and a mid would actually be even less bad since their responses aren't identical you have even fewer instances where the spl responses are even identical let alone canceling each other out. And equalization I don't believe is necessary. I believe that the idea that the wavefront canceling causes an audible null in frequency response is unimportant since 90% of the sound that is actually heard, especially at high frequencies is actually reflections of the walls which have completely different path lengths and amplitudes than the main wavefront.

 

[ewollowe]

I've been following this thread about a very nice line array, pros and cons including interesting comb filtering impressions.

http://www.diyaudio.com/forums/multi-way/288784-tc6wd-pe-buyout-line-array-lessons-learned.html

I'm using two 3" per side in a castle style arrangement (one firing up and one forward.) early days but sounds fine so far.

 

[zman01]

The P10 library has a few designs with 2 wide-band units on the front baffle (Twin CHR-Ken, Twin Mar-Ken, Wessex, Mercia).

You can run 2 FR drivers in 1.5 config and that should reduce concerns on comb filtering.

 

[Ray21]

You can run 2 FR drivers in 1.5 config and that should reduce concerns on comb filtering.

That's what I'd do!

 

[planet10]

Comb filtering looks far worse in a measure than it does to your ear/brain. The ear/brain expects comb filtering and mostly ignores it. If you are far enuff away it becomes mostly a non-issue.

Combing does affect the sonics, but many systems just do not have the resolution to show it, or what the multiple drivers bring is more to the taste of the user than what multiple drivers subtracts.

Speakers have huge compromises, even the best ones. Choose your poison.

And as Zia has pointed out a 2 (or more) driver system can be run as a 1.5 way. Either in series (with a cap bypassing one driver) to keep SPL the same as a single driver or in parallel (with a series inductor on one driver) to give ~6 dB increase in levels at low frequencies. In both cases you get 2x the cone area down low.

dave

 

[dynomike]

Agree with the above poster that a 1.5 crossover would be a better choice. Two drivers would not be enough to form a "line source" at any frequencies below the center-to-center wavelength so you basically would get "worst of both worlds". The one advantage is greater radiating area for low frequencies, which you would get with the 1.5. At this point you might also consider FAST

Don't let this theorizing stop you from trying it of course... I would wire it up such that I could audition both drivers together, or just one driver. Level match the single driver vs. both drivers so you have a fair comparison. This would require having separate cabinets for each driver, otherwise you'll get some unintended passive radiator effect happening.

I have some issues with the following post...

I actually disagree with this reasoning for why comb filtering isn't an issue. I believe having a tweeter and a mid would actually be even less bad since their responses aren't identical you have even fewer instances where the spl responses are even identical let alone canceling each other out. And equalization I don't believe is necessary. I believe that the idea that the wavefront canceling causes an audible null in frequency response is unimportant since 90% of the sound that is actually heard, especially at high frequencies is actually reflections of the walls which have completely different path lengths and amplitudes than the main wavefront.

Agree that comb filtering is usually inaudible in reverberent listening environments. In the upper mids / lower treble the on-axis response is smoothed out by reflections, which is why diffraction ripples usually aren't as audible as they are measurable. But, up top (like above 10kHz) the on-axis response totally dominates due to basically everything including paint absorbing ultra-high frequencies.

As for the mid/tweeter crossover causing less issues, well, no. Best case scenario would be to have perfect amplitude and phase matched response between both drivers throughout the crossover region, so that the nulls will be aimed (vertically) off-axis. Deviation from identical driver behaviour will make matters worse in the crossover region, not better.

 

[planet10]

As for the mid/tweeter crossover causing less issues, well, no. Best case scenario would be to have perfect amplitude and phase matched response between both drivers throughout the crossover region

And the C-C distance between the drivers being crossed less than a ¼ wavelength of the XO frequency. And either a 1st order XO or one of the more exotic digital XOs that can maintain phase coherence with a higher order XO.

dave

 

[midrange]

Bandor made a design with 2 50mm units per box, run full range. I have come across a current design that uses this layout with these drive units. I think from memory they were meant to have a crossover at 300 Hz.

 

[maggiesnmacs]

The 1.5 way design would work except eventually I want to add a tweeter right between the two drivers crossed at 6khz or so. I see your point that above 10k the problems would get worse since the on-axis response dominates and everything absorbs those reflections for the most part, but I haven't found that to be audible personally. A tweeter would alleviate that anyways, but I would still be crossing above the center to center rule of thumb. My goal though is to have the virtual center of the full range get be the same as the tweeter as I believe this is important, especially at high frequencies to your perception of the sound coming from an individual source. And I believe the 1.5 design adds unnecessary complexity when in my experience just driving the full ranges in series doesn't lead to noticeable comb filtering all the way up to 20k. If not audible up to that frequency I highly doubt it makes almost any difference at 10k and below.

 

[picowallspeaker]

If not audible up to that frequency I highly doubt it makes almost any difference at 10k and below.

It'll depend on the driver, the geometry of the cone, the profile of the surround suspension, the dustcap. When doubled, who knows !?
Then, phase difference is MOST notable the more you go down in frequency; with vocals is mostly horrid as you notice it on the...uhm, ahhr ehhhr- vocals.
Going lower to the very bass notes you'll hear real "peaks & dips". But this happens when a woofer and a midrange driver gets overlayed and not properly crossed, or when the exit of a TL, say, is out of phase at some frequencies with the woofer in the nearby.
It's true that above 10 kHz phase changes every 3,4 cm for a 10 kHz tone, or, during the travel from speaker to ears it has done many rotations

 

[dynomike]

It's true that above 10 kHz phase changes every 3,4 cm for a 10 kHz tone, or, during the travel from speaker to ears it has done many rotations

Huh? short wavelength =/= phase rotation. Could you explain?

maggies, now that we know you are looking to do a MTM design, it's a different question... but 2.5 way MTM designs do exist commercially. Any multiway system will create lobing issues, so the design is veering away from "full range" methinks.

I do appreciate the idea of wanting "point source sound", but the easiest way to do that is to use a single driver. Having no experience with MTM designs I'll bow out now... good luck!

 

[planet10]

The 1.5 way design would work except eventually I want to add a tweeter right between the two drivers crossed at 6khz or so.

Pushing the FRs apart more (or using bigger drivers) will exacerbate combing issues. For a dual FR system i was thinking 3 or 4" drivers. The top end of the 3" drivers we use will perform as well (or better) as many good tweeters.

If you want to add a tweeter going 2.5 way is an even better idea. AFAIC with an MTM keeping c-c within ¼ wavelength is even more important. At 6k that is a distance of about 6/10 inches.

dave

 

[wesayso]

Huh? short wavelength =/= phase rotation. Could you explain?

maggies, now that we know you are looking to do a MTM design, it's a different question... but 2.5 way MTM designs do exist commercially. Any multiway system will create lobing issues, so the design is veering away from "full range" methinks.

I do appreciate the idea of wanting "point source sound", but the easiest way to do that is to use a single driver. Having no experience with MTM designs I'll bow out now... good luck!

In a way a Unity or Synergy is MTM with small distances between drivers more easily to get it within 1/4 wavelength. A smart/cute way to do the MTM in my opinion. It has that point source potential.

 

[weltersys]

If you want to add a tweeter going 2.5 way is an even better idea. AFAIC with an MTM keeping c-c within ¼ wavelength is even more important. At 6k that is a distance of about 6/10 inches.

Dave,

The wavelength of 6000 Hz is 2.26 inches. ¼ wavelength of 6kHz is only .565 of an inch.

Cheers,
Art

 

[planet10]

I rounded to the nearest 10th… it does illustrate the impossibility of the task.

dave

 

[weltersys]

I rounded to the nearest 10th… it does illustrate the impossibility of the task.

dave

I get it now, when you wrote "6/10 inches", you meant 6/10ths of an inch, equal to 3/5ths of an inch...

 

[planet10]

Yes -- probably from my heavy use of metric.

dave

 

[chrisb]

Re two full-range drivers per side - I've had good results with a pair of unfiltered units in the "Castle" micro-tower configuration; actually a total of three builds of those - CSS EL70, CHN70 and a two-way with Dave R designed XO for MTM using CHBW70 and ERT26? tweeter.

I can discern no obvious / distracting artifacts of the drivers mounted approx 166mm / 6.5" apart at an 80dg angle.