Series vs parallel (again!)

[Elipse]

I know this one has been rumbling around for ages but the last posts I can find are some years ago now and although I think series seems to be the generally accepted fave..................

Opinions please. 2 x "identical" woofers for a single centre channel, 8 ohms nominal each and a sensitivitly of 85.5dB (2.83V/1m). In series, therefore, 16 ohm, 4 ohm in parallel. The separate mono amp to drive these quotes (RMS) 190W into 8 ohms, 300W into 4 Ohm and also 300W into 2 Ohm. So it would appear it should be able to "happily" drive these units in parallel.

On paper, correct me if I'm wrong, (be gentle, novice here!), I would prefer them in parallel due to their relatively low sensitivity compared to the single midrange to be used which has a quoted sensitivity quoted of 91dB. So, in parallel the combined woofer output would be relatively greater, therefore requiring less "correction" in the XO due to the sensitivity difference?? The less "correction" required in the XO the better IMHO.

I would be grateful for and comments and any other factors to consider. Thanks.

 

[chris661]

You'll be somewhere near. Due to the Fletcher-Munson curve, you might need to pad the mid down a little, but this will be dependant on personal tastes and listening volume, so don't worry about it too much. Just be prepared to tweak if it's not quite there.

When drivers wired in parallel are given the same voltage as before, you get a 6dB increase in level. The current for a set voltage doubles as there's half the impedance, and the radiating area doubles, each of which gives +3dB.

You'd end up with around 91dB at 2.83 volts through the lows and mids, so, as I said above, you'll be somewhere near.

Chris

 

[Elipse]

Thanks very much. Sounds good, hope so anyway!

 

[Cal Weldon]

What woofers are they?

 

[Elipse]

Hi

Midwoofers really, space being somewhat limited. 2 x Scanspeak 15W/8531K00 5.25" in a nominal 15.6L enclosure.

 

[Cal Weldon]

Yes, looking at the FR curves of your woofs, you will likely need to pad the mids a bit.

 

[Spatz]

Another problem when going serial is the slighty different impedance response of the woofers around their resonance frequency. Voltage drop will be different, thus will be the SPL of the woofers.

For example, one woofer has its resonance frequency at 50 Hz, the other one at 45 Hz (10% tolerance, that's realistic). So one woofer will get much more power at 45 Hz, as his impedance is much lower at this point than the impedance of the other woofer, and the other way round.

I would always connect woofers in parallel to avoid this problem. A pack of two parallel woofers can then be connected with another pair of parallel woofers as the tolerances should be less this way. Midranges and tweeters that are not used at their resonance frequency can be connected in serial, if thats necessary.

Bye,

Spatz

 

[Elipse]

Many thanks to you all.

I feel much more confident progressing this project now.

 

[Pan]

You'll be somewhere near. Due to the Fletcher-Munson curve, you might need to pad the mid down a little, but this will be dependant on personal tastes and listening volume, so don't worry about it too much. Just be prepared to tweak if it's not quite there.

When drivers wired in parallel are given the same voltage as before, you get a 6dB increase in level. The current for a set voltage doubles as there's half the impedance, and the radiating area doubles, each of which gives +3dB.

You'd end up with around 91dB at 2.83 volts through the lows and mids, so, as I said above, you'll be somewhere near.

Chris

Chris, a speakers response should not be shaped with equal loudness curves in mind.

Speakers are and should be more or less flat. Sound shaping should be left to the artist, audio engineers (recording/mix/master) and architects.. ;-)


/Peter

 

[chris661]

You have a point, though I did mention that it would be down to personal taste/preference.

That said, equal loudness curves are also volume dependant (which is why the old loudness button is still going strong), so this will also be a factor to consider when setting the levels.

Chris

 

[sreten]

Hi,

The issue and loudspeaker design has nothing to do with F&M curves.

Dependent on the bass/ mid c/o frequency it has everything to with
the ~ 6dB baffle step, the two drivers could be wired as 0.5 way
for a passive crossover to give full baffle step, or just in parallel
with the usual overlarge inductor for baffle step.

Due to baffle step midrange senstivity is 4dB to 6dB lower than
the nominal bass sentistivity into half space, the reason is that
that bass is into full space lowering senstivity by 6dB.

rgds, sreten.

 

[Legis]

Sorry for resurrecting an old thread, just wanted to share my experience.

I have usually tried to maximize the parallel connections in my line arrays. Then yesterday I connected one of the midwoofer arrays fully in series. Ten 8ohm woofers in series, so the nominal impedance was 80 ohms. I liked the sound! Even with somewhat current limited amplifier I could generate huge SPLs with everything the amp could swing. The way that the quality of the bass stayed the same regardless of the volume, was unheard of. The control and the attack, which were very good, were nearly identical regardless of the volume.

I was speculating that maybe the back emf behaves slightly differently with series connection. I think normally the generated back emf goes first to the output of the amplifier and from there it can go back to the driver and damp it. The smaller the output impedance that the amplifier has, the less the back emf is dissipated at the amp, and the more it can damp the driver (the basics of the "damping factor").

With series connection the back emf travels through the voice coils of the speakers to the amp, when it travels there for the first time. Most of it might be dissipated at the woofers before it even reaches the amplifier. Maybe the generated back emf damps the series connected drivers more efficiently... If so, wouldn't it also mean that the amplifier's negative feedback loop does not get affected by the back emf so much, because the emf is almost dissipated before it reaches the amp? Just speculating.