3-way questions

[angsuman]

Hello everyone, i'm ready to finalize my 3-way project before i order the parts. I have a few questions.

My original design calls for a dayton RS250 8in woofer, with a peerless Exclusive 830881 4" woofer, and a vifa xt19 tweeter. All drivers use 2nd order electrical slopes all equaling something close to a 4th order acoustic rolloff. My questions include:

1. The lowpass part of the bandpass filter on the peerless midrange uses a 3rd order electrical filter to reduce the level of peaking as i couldn't control it with just a 2nd order slope. I believe i've read somewhere that since this shifts the signal 270 degrees out of phase it actually will cause a delay when compared to the 180 degree delay as introduced by the other crossovers. Will this time delay have a significant effect on sound quality?

2. I have no measuring equipment except a multimeter but I will make some DIY electret microphones. I'm relying mostly on traces made in SPLtrace from published data than using Passive Crossover Designer (PCD) to design my crossover. How accurate is this program? What have been people's experiences as far as simulated to measured response goes.

3. After looking at Zaph's 2in RS52 dome midrange tests, the distortion numbers look amazing, very surprising. I'm thinking of redesigning the speaker to incorpate the 2in dome midrange as it seems to be a better performer than the peerless through all but the lower midrange. However since i will be running the woofer fullrange i'm curious if this will raise intermodulation distortions to higher levels as the woofer will now be playing higher around the 500-700 hz region versus the 250-300hz crossover with a peerless midrange and the dayton will also go as low as possible in a sealed box.

Sorry for the long post but this is going to be my first reference loudspeaker as i've built much simpler designs before.

Thank you,
Angsuman Roy

[infinia]

Hi
Sorry I can't answer your questions directly, as it would be too painful for me to think about. All I can say is, as your 2 crossover points start to converge into the midrange area it becomes more difficult in regards to crossover design, taking into account some of the things you are starting to contemplate like filter order, phasing, polarity and group delay. Any resulting mistakes made in the integration of the 3 drivers will be certainly easier to hear. I would guess if your chosen drivers response and phase are stable well beyond the crossover points then the crossover design will be the determining factor in the final response you are after.
My design process on 3 way designs, is firstly find a midrange driver with the widest BW possible. The midbass selected will hopefully be covering the whole critical range of the human voice and still be somewhat linear at the max SPL target. Secondly integrate the selected midbass with a suitable tweeter. I concentrate mainly on dispersion characteristics in the crossover region for a smooth transition from the midbass driver to the tweeter. The least problematic midrange, integration wise, has a BW (esp. including off axis) that overlaps the selected tweeters lower crossover point by some margin. The tweeters lowest crossover point and order of filter is usually driven by max SPL and IMD distortion tradeoffs. This is always the toughest crossover point to master. I have found once you get this crossover point right, or at least sounding good, then proceed to design the midbass and woofer transition.
My personal experience has been that it is very difficult to find a good midrange that can do all of this and still able to handle the lower registers of the male voice. I find myself constantly to be on the lookout for 6-7 inch cone drivers with smoother highend rolloffs. I concentrate at the relation to on axis to the 30 - 60 deg responses and use it for defining the upper mid bass crossoverpoint. Ideally I look for midbass drivers with electrical and acoustic rolloffs that are falling at the same frequencies. Thus can more easily be combined with an added electrical filter for the final desired crossover filter order. All this means I usually stay away from mid domes and cones with phase plugs. as always YMMV

[angsuman]

Ahh yes, that is one way to 3-way design, I was also thinking of the peerless 5.5inch nomex exclusive but found data on DIYmobile audio that showed the 4in have less distortion from 250hz up if i recall correctly. I was just shocked at the low distortion of the dayton RS52, and the off axis dispersion just adds icing to the cake.

Well if no one can answer my questions i'm really curious to see how accurate the passive crossover designer spreadsheet is as i'm relying on that until I get parts for a diy mic.

[tinitus]

Spreadsheet is probably very accurate but real life speakers are not - simple design tools rely on straight frequency and impedance
And remember that your mic have to be calibrated to be reliable
And I think you can forget about driver distortion figures...your crossover will distort much worse if it is not working 100% correct...which very few does...if any at all

[Paul W]

Crossover accuracy problems will likely not be with the PCD, but with the data feeding it. Crossover programs like PCD depend on accurate input from your measurement system, including baffle step etc....it's just a case of GIGO.

[angsuman]

Tinitus, what exactly are you referring to as crossover distortion? Is it just the fact that crossovers don't always work as theoretically made or are there significant distortions in the components themselves as well?

[tinitus]

Crossover distortion - I am not sure exactly what does it, I have only experienced that even very small adjustments can have a huge effect on how clean it sounds - could be phase and how our ears react to it, could be interaction with amp, coherence or whatever, cant say really

Oh yes, I believe that internal Q factor of various component of the filter plays a very important role - not very easy to mess with

I wish I was skilled enough to give it a more scientifically approach