[DanTheMan]

What can I say?  190Hz crossover, RS100 flush mounted, all modeled using the Edge by Tolvan for smoothest baffle contribution and drivers leveled using test tones and a RatShack analog sound meter which is very accurate in the crossover region.  Behringer CX3400 crossover in 2-way mode.  I only own a jigsaw, so that was the only power tool used in the build.  All edges with a large round over by chisel, file and sandpaper as is the cutout for the RS100.  The screws you see in the front will be painted black.  It's quite an impressive sound to say the least, and I am very pleased though it doesn't get to blistering SPLs comfortably.  All for under $200.  Measurements to come... well in a few weeks anyways. 











Dan

[DanTheMan]

Well here's the first graph I tried to make in room, mic at seated position no smoothing.  I think it looks OK, but I have too much midbass and not enough of the deep stuff.  It also looks like maybe some cavity resonance in the treble from the 3/4" board.  Any suggestions?


Dan

[DanTheMan]

Mostly what I want to know is are there any problem areas other than the woofer being 6-10dbs too loud?  I fix that in the second graph at the bottom of this post.  The woofer level is now set as low as I can make it.  I can still turn up the FRer if desired.

What would you expect the response to look like?  The filter is 24db/octave and I think it's LR.

What's up with the approx 70Hz dip that turns into a 140Hz dip in the next graph?  The mic position is about half the distance of the original at very close to the same height.

Does it look like there's a resonance in the treble?  Seems the response has a regular and repeating dip/hump cycle in stereo, but not as severe when just measuring one speaker.  Am I getting some sort of comb filtering or does this graph look pretty good and I'm just worried about nothing?

_______________________________________ _________________________

Here's my next graph I just took after playing with the crossover: one speaker only at about 4ft away with the mic pointed roughly between the woofer and tweeter.  The treble looks a little smoother, but it still has dips and peaks but they seem tighter together and also theres now a boost at 15k.

Do you see any response issues I should be concerned with?



Thanks again,

Dan

[Rudolf]

Smoothing the graphs to 1/6 oct would make things better visible.

Regarding the loss of treble in the first graph:
At the listening position you measure the full power response of the dipole. The FR driver does not radiate high frequencies to the back. Those are missing in the power response, but in measurements at a close distance you mainly measure the frontal radiation. Keep in mind, that those measurements don't distinguish between direct sound and later reflections. But your ear/brain does. So let your ear decide what treble level is correct.

At the listening position the graph will change with every inch you move the microphone in the room. I prefer to integrate over different positions by moving the microphone about a foot in every direction while measuring. There will be wind noise degrading the low frequency measurement, but above 200 Hz you will get a meaningful graph.

[DanTheMan]

Thanks Rudolf!  It sounds great in the room.  I can't say I want to change a thing, I just want to make sure.  With it EQed flat there is definitely less bass than what I've become used to, but it sounds very natural.  Instrument harmonics seem to be at the right volume.  It seems like too simple of a project to turn out so well.  I can't wait to get the deep bass all set up.

Dan

[Nyal Mellor]

If you're looking to analyze the frequency response of those nice OBs you should do a nearfield measurement i.e. maybe 12" from the speaker. You'll also need to change the gate time in Room EQ wizard to like 3ms or so so that you're not including any reflections.

Unfortunately that's not going to give you good resolution low frequency measurements, for that you'll need to go outside!

At the moment you're measuring the room + the speaker. That nice dip at 150hz is speaker boundary interference and that peak at 50hz looks like a room mode

[DanTheMan]

Thanks Nyal!  Yea, I went outside and did some measurements nearfield.  All those lumps smooth over nicely.  Heck, even nearfield indoors they smooth out.  The closer you get though, the higher that 15k peak reaches.  The whole response looks much better outdoors even though I haven't figured out how to change the gate.   All the effort to take things outside and I never tried to change the gate for a better measurement.  I'm either lazy or stupid.   

If I had it to do over again I would buy the cheaper 2-way behringer crossover and put the savings toward plate amps for subs.

Anyway, these are the first speakers I've had that don't have a lot of large, audible peak and dip cycles, let alone any.  Listening to a log sweep sounds so even it's amazing.  Only the frequency extremes get lost.  That's somewhat d/t my hearing and the fact that my average listening level is very low by theater and concert requirements.  Maybe when I'm old I'll still have it. aa

Dan

[DanTheMan]

Nyal, here's a graph of the treble response from seating position with the gating changed:


Dan

[DanTheMan]

I made a jig for measuring polars and did some graphs with REW of this build.  Smoothing applied for a clearer picture.  Notice the difference between L and R.  It just shows how much baffle width has to do with dispersion.
On axis left speaker:

On axis right speaker:

11.25 degrees off axis left speaker:

11.25 degrees off axis right speaker:

22.50 degrees off axis, left speaker:

22.50 degrees off axis, right speaker:

33.75 degrees off axis, left speaker:

33.75 degrees off axis, right channel:

45 degrees off axis, left speaker:

45 degrees off axis, right speaker:


Dan

[ttan98]

DantheMan,

If you mount a waveguide on the full ranger you will get better directivity, ie the dip at 45deg will not be so pronounced.
Normally to don't have the privilege because there is no off-the-shelf waveguide suitable for your full ranger, hence you may have to tip-toe in your speakers towards the centre for at least 30deg.

[DanTheMan]

The thing is, across my couch the response is pretty flat--a bit hot in the middle, but nearly flat across the board.  I was actually thinking about reversing the left and right.  Right now the inside edge has a lot broader of a response.  If I flip'em and toe that in further, I should be just about conforming to the Blumlein method--well at least the way it was explained to me.

Thanks for the insight!

Dan