So What's More Important..., a Smooth Frequency Response or Time Alignment?

[toddalin]

I have noticed that when using the RTA in the seating position with the Ethyl Mermans, I get a smoother frequency response at the bottom end of the frequency range (big dip is mostly removed) when the cabinet with the 10" sits well behind the plane of the 18". The middle and upper portions of the range are just a bit smoother when the cabinet is brought forward.

I attribute this to the fact that when the 10" sits more forward, it is more subject to floor bounce that causes destructive interference, whereas moving the 10" cabinet way back removes this bounce because now the 18" cabinet serves as the "floor" and is much, much closer to the 10" driver.

This is not something that I can instantanously A/B. Thoughts?

[badman]

Time alignment between the 10" and 18" shouldn't be a significant issue, what's the XO point? The bigger concern is the reflection from the top of the 18" enclosure, that's what's causing the roughness higher in frequency. For the floor bounce you can either move the crossover point up, so that you're using the floor-loaded 18", or even easier, use a throw pillow at the reflection point for critical listening. Moving the XO is a little more iffy, I'm betting the notch is at 500 or so if I'm picturing the system right.

[Ian Mackenzie]

I have noticed that when using the RTA in the seating position with the Ethyl Mermans, I get a smoother frequency response at the bottom end of the frequency range (big dip is mostly removed) when the cabinet with the 10" sits well behind the plane of the 18". The middle and upper portions of the range are just a bit smoother when the cabinet is brought forward.

I attribute this to the fact that when the 10" sits more forward, it is more subject to floor bounce that causes destructive interference, whereas moving the 10" cabinet way back removes this bounce because now the 18" cabinet serves as the "floor" and is much, much closer to the 10" driver.

This is not something that I can instantanously A/B. Thoughts?


When you say the bottom end what frequency is that?
What is the crossover frequency and slope of the 18 inch and the 10 inch driver?
I pic would help as l don’t recall the baffle.

An educated guess might be phase cancellation of the 18 and10 inch drivers at a particular frequency. Moving the 10 inch driver back is like adding phase compensation. Low pass filters on woofers tend to exhibit some group delay in designs with lower crossover points. Even if the frequency is attenuated by the crossover there can still be interference.

As l said if the facts are mentioned such effects are more easily understood.

[toddalin]

When you say the bottom end what frequency is that?
What is the crossover frequency and slope of the 18 inch and the 10 inch driver?
I pic would help as l don’t recall the baffle.

An educated guess might be phase cancellation of the 18 and10 inch drivers at a particular frequency. Moving the 10 inch driver back is like adding phase compensation. Low pass filters on woofers tend to exhibit some group delay in designs with lower crossover points. Even if the frequency is attenuated by the crossover there can still be interference.

As l said if the facts are mentioned such effects are more easily understood.

Crossover is at 225 Hz. I will have to run the RTA again to determine the dip frequency. But know that connecting the 10" in phase creates a broad dip in one area, whereas out of phase largely fills in that dip and creates another at a lower frequency, that is much tighter.

http://www.audioheritage.org/photopost/data//500/medium/DSC_00251.JPG

[Ian Mackenzie]

I had a look at Robs simulation.

Behind the scenes your 2351J has an impedance in the enclosure.
I don’t know if enclosure is seal, open back or vented.
But that impedance and the 6 dB bandpass will become reactive at resonance.
The resonance will cause modulation of the voltage drive to the 2251J causing response variations.

In simple terms your passive crossover won’t work properly unless it sees a relatively flat impedance from the 2251J.

The way around this is to measure the impedance. If you shunt the 2251J with an inductor and a series resister it will compensate for this impedance peak.
It will be ballpark 3.5mH and 8 ohms.

The above is likely a contributing factor to the effect your seeing.

[Ian Mackenzie]

With your theory on the floor bounce look at the direct listening distance and then the distance if a sound wave bounces off the floor midway along the direct listening distance. That second longer wave will cancel at a particular frequency.

I would sort out the impedance of the 2251J first.

[toddalin]

The 2251J cabinets are ported on the rear panels.

Current crossover configuration.
http://www.audioheritage.org/photopost/data//500/medium/ScreenHunter_216_Nov_12_12_13.jpg

Rob's simulation is not quite accurate because it does not consider the effects of the inductance of the 2241 on the 2251.