[JeffB]

Thanks Rudolf.

1) If the problem is diffraction at the edges, I have trouble seeing why baffle width matters.  It seems like the wider the baffle the further the edges are away and thus less likely to interfere with the on-axis waveform.  This is hard for me to visualize.

2) Do you think a properly designed waveguide on just a tweeter might allow for a wider baffle on the tweeter.  Thus still providing a smooth response, but extending the low end a bit more?

3) Does rounding of the edges help, and if so what do you recommend for baffle thickness and rounding?

[JeffB]

Could edge rounding be used as any kind of a substitute for changing the baffle width?
Or is any improvement related to rounding small compared to changing the width?

[pedroskova]

You have two things at work in regards to dipole baffle width: diffraction and the dipole peak.  With diffraction, you can chose a very narrow baffle (pushing the effected frequencies up to where they are less objectionable), or a very wide baffle, where, again, the effected frequencies (this time on the low end) are not located where our hearing is most acute.  On the other hand, you have the "dipole peak" to contend with.  The dipole peak is that frequency at which the backwave wraps around the baffle and adds constuctively to the front firing wave.  This creates the peaks...or mounds...in the example sims that Rudolph has been posting.

All of this becomes a compromise depending on the dipole driver chosen, and how low you need it to play - the narrower the baffle, the earlier it begins to roll off...but possibly...with a better diffraction profile, and a better/worse dipole peak (depending on the driver and its intended bandpass).  Rounding edges lowers diffraction (the bigger the radius the better), but does nothing to prevent dipole peaks or dipole roll off.

I was left choosing between what I felt was the best passive solution (coaxial) and active...and I chose active... because, if the baffle is too narrow for my target f3, no problem, EQ it with an eye on using drivers that can take the needed EQ. Dipole peak? Try to put it in a crossover region, and if that doesn't work, EQ it.

...plus, my basic philosophy is that the more things I need to turn on at the start...the better. 

[panomaniac]

Well, I just found graphs of the P Audio drivers, they look a little scary to me in the upper frequencies.

Be not afraid, good sir!   They are actually excellent drivers.   Worry more about the low end of the horn and get the crossover right.

Funny about the dipole peak.  I read so much about it, see it in the sims, but don't often hear it or measure it  - much.
Could the room be filling in or killing a lot of the peak?

[Rudolf]

1) If the problem is diffraction at the edges, I have trouble seeing why baffle width matters.  It seems like the wider the baffle the further the edges are away and thus less likely to interfere with the on-axis waveform.  This is hard for me to visualize.

pedroskova already explained your issues better than I could do. So this is just some extension:
What you hear of baffle edge diffraction at the listening position is generated by the delay between the direct wave radiated from the cone and the diffracted wave from the baffle edge. The delay is a function of the baffle width. Linkwitz has a detailed description of how it works on his site:http://www.linkwitzlab.com/faq.htm#Q8

2) Do you think a properly designed waveguide on just a tweeter might allow for a wider baffle on the tweeter.  Thus still providing a smooth response, but extending the low end a bit more?

The function of a properly designed waveguide is a controlled attenuation of off-axis radiation. If done right, a tweeter wave guide does not need a baffle as extension. That´s exactly what I do. My tweeters have a small waveguide built in:

 

3) Does rounding of the edges help, and if so what do you recommend for baffle thickness and rounding?

Rounding of the edges helps, if the edge radius is comparable to the wavelength in question. So rounding of baffle edges besides a woofer is pure cosmetics. At 2 kHz the radius needs to be at least 1 cm - 2 cm would be better. Personally I don´t round edges any longer - I try to keep baffle width small enough to avoid diffraction issues.

Funny about the dipole peak.  I read so much about it, see it in the sims, but don't often hear it or measure it  - much.
Could the room be filling in or killing a lot of the peak?

Look at my last sim. The dipole "peak" is < 2 dB and stretched over almost 2 octaves. You can see it in the sim, but for your ear it may be just a little "forwardness" -nothing more. And as pedroskova said: "Try to put it in a crossover region". At 500 Hz that surely is the best solution.

[panomaniac]

LOL.  Yes -  always look at sims and think "Holy-Cow! Look at that bump!"  But it's really only 2 or 3 dB at the most.  That's easily lost in room.    And I build asymertical wings, so that should be even less bumpy than the sims.

[Telstar]

Could you also run a simulation for 35cm? The 12" driver is centered at 62,5cm not at 80cm btw.

This looks VERY nice to me!!

If you can also set the driver specs in this software, those are Qt=0.6 and Fs=50hz with no crossover.

This simulation is with an ideal driver (linear response from 0 to infinity). I can do simulations for specific drivers if I have the specific parameter set for Boxsim (which includes more than the TSP). It will be easier for you to estimate  the Q induced drop off than for me to build even a minimised parameter set. Took me about three hours to get the "ideal driver" model running correctly.

The drivers are kind of proprietary and those are the only avaliable specs.

I played with EDGE (mic at 90cm and 2.5m which is my listening point) and with a single midrange the above looks indeed the best:


Using two drivers changes things completely. I had to make the baffle larger and positions the drivers not parallel.
Why do i want to use two drivers? Because i heard a pair of speakers built this way and they sounded wonderfully. It enriches the midrange.



This looks better above 1000hz to about 5khz. If you were to NOT use a crossover in front of these drivers, which of the two setups would you use?

[Rudolf]

Using two drivers changes things completely. I had to make the baffle larger and positions the drivers not parallel.
Why do i want to use two drivers? Because i heard a pair of speakers built this way and they sounded wonderfully. It enriches the midrange.

Telstar, you heard two 12" drivers mounted one upon the other in OB without a crossover and they sounded wonderful to you? Our tastes in listening MUST be very different. 

This looks better above 1000hz to about 5khz. If you were to NOT use a crossover in front of these drivers, which of the two setups would you use?

Sorry, but I can´t answer that question. Are you really intending to use a 12" driver up to 5 kHz? Or are we talking about a coax driver?
Under no circumstances would I let a 12" operate much above 1 kHz. It will beam like crazy, and while the on-axis response may be ok, your ears will tell you that the reverberant field lacks treble heavily. At least that´s my experience even with much smaller widerange drivers.
Possibly people with some PA experience know better than me how to deal with treble from such large drivers.

[Telstar]

Telstar, you heard two 12" drivers mounted one upon the other in OB without a crossover and they sounded wonderful to you? Our tastes in listening MUST be very different. 

Of course they weren't the only drivers used. Likewise if I do a smaller baffle with only one driver, I'll also add a woofer and a tweeter.
And there was a slight waveguide which, looking at the edge graphs must fix the HF response.

I wonder if the non parallel vertical arrangement would still have a positive impact.
I sent you an email because I do not want to disclose more details publicly.

[panomaniac]

The Beta 12LTA is a 12" that works just great up to 6K, then you have to get out, or it gets ugly.  Ask JBSpeakerman, he knows this driver well.

Don't think I'd use two of them together tho....

Have heard the Beta 8 used in pairs for midrange up to 6K.  Worked very well, yes indeed. (Gary Pimm's system).

[Rudolf]

The Beta 12LTA is a 12" that works just great up to 6K, then you have to get out, or it gets ugly. 
Have heard the Beta 8 used in pairs for midrange up to 6K.  Worked very well, yes indeed.

I´ve got no experience with drivers of that size going up to 6 kHz. Of course whizzer cones help to widen the radiation pattern again. But you still need to listen on axis, don´t you? And above 6 kHz a horn or WG driver will be mandatory to keep radiation narrow?

[Telstar]

The Beta 12LTA is a 12" that works just great up to 6K, then you have to get out, or it gets ugly.  Ask JBSpeakerman, he knows this driver well.

Don't think I'd use two of them together tho....

Why not?

[panomaniac]

Hmmmm.....  good questions.   

I sure don't remember the Beta LTA beaming when used with the tweeter @ 6K.  It is actaully an amazing midrange driver - not that you could tell by looking at it.  Its sound is very dynamic and yet delicate and detailed at the same time.  Almost electrostatic sounding, but more dynamic.  Quite a surprise.

Why not use 2? - uh..... comb filtering?  I dunno, just doesn't seem like a good idea.  Although perhaps it could work well if one was rolled off lower than the other.  The nice Beta 8s I heard were done that way.

Darn, now you've got me thinking.......

[ttan98]

Hmmmm.....  good questions.   

I sure don't remember the Beta LTA beaming when used with the tweeter @ 6K.  It is actaully an amazing midrange driver - not that you could tell by looking at it.  Its sound is very dynamic and yet delicate and detailed at the same time.  Almost electrostatic sounding, but more dynamic.  Quite a surprise.

Why not use 2? - uh..... comb filtering?  I dunno, just doesn't seem like a good idea.  Although perhaps it could work well if one was rolled off lower than the other.  The nice Beta 8s I heard were done that way.

Darn, now you've got me thinking.......

Hi,

I am looking for a  good mid-range at a reasonable price, there are few Beta 8" models can you tell me the actual model Gary Pimm is using?

Thanks.

[panomaniac]

This one.  It has a light purple cone, if that helps.
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=290-404

I think Gary uses lots of rope caulk (window putty) on the baskets to tame them.  No other treaments that I know of.

[ttan98]

This one.  It has a light purple cone, if that helps.
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=290-404

I think Gary uses lots of rope caulk (window putty) on the baskets to tame them.  No other treaments that I know of.

Why to reduce vibration caused by the basket? If not please explain.

[panomaniac]

Correct!  It's a pressed frame basket.  Not a bad one, but the putty kills a lot of vibration.  It can also help kill reflections coming from the inside of the basket out thru the cone.  Felt works best for that.

The tricks like putty and felt are the frosting on the cake.  Without them you still have cake, but it's not as pretty - nor as tastey.  But all the frosting in the world won't change a bad cake into a good one - so start with a good one.

[ttan98]

The Beta 12LTA is a 12" that works just great up to 6K, then you have to get out, or it gets ugly.  Ask JBSpeakerman, he knows this driver well.

Don't think I'd use two of them together tho....

Have heard the Beta 8 used in pairs for midrange up to 6K.  Worked very well, yes indeed. (Gary Pimm's system).

I had a private communications with G. Pimm, he informed me that the Beta 8A started to roll off quickly starting @ 4Khz(from his FR curve) and he used the natural roll off(no cone break up) to x-over to the tweeter. Very neat.

The combo gave him about 95dB spl. Very neat as well.

I may be give configuration a go.

[panomaniac]

IIRC, Gary put a small inductor on the Beta 8 to flatten the rising response.  Not really a crossover - just a gentle EQ. Hi pass is active - I think.   But you should confirm that with him.

He uses one of those big Hi-Vi planar tweeters up top.  To my ears it had no "tweeter sound" at all - which is high praise coming from me.

[ttan98]

IIRC, Gary put a small inductor on the Beta 8 to flatten the rising response.  Not really a crossover - just a gentle EQ. Hi pass is active - I think.   But you should confirm that with him.

He uses one of those big Hi-Vi planar tweeters up top.  To my ears it had no "tweeter sound" at all - which is high praise coming from me.

I got the important info from him. Thanks panomaniac.