Parallel vs. Series Speaker Connection

[Cogito]

If the amp is marginally underpowered, it will drive the speakers without clipping when operated within limits, but the dynamics will be rounded off. A more efficient speakers ( ie: two 8 ohms in parallel vs a single 4 ohm) will only play the sounds little louder it will still have "lost" dynamics.

[Roscoe Primrose]

Assuming the 4ohm speaker has the same power sensitivity as the individual 8ohm drivers, the amp will have to produce only half as much power to generate the same acoustic output. The amp will have 3dB more headroom, how is that going to lose dynamics?!?!

Roscoe

[SoundMods]

Actually, since the power sensitivity went up by 3dB, it ought to be EASIER for the amp to drive the 4ohm load created by two 8ohm speakers in parallel, assuming the OPT has a 4ohm tap. You're only looking at 20V across all three scenarios if you don't change output taps or use an amp (SS probably) with a vanishingly low output impedance. Remember, if you use the 16ohm tap to drive the series connection, and the 4ohm tap to drive the parallel connection, you're only getting half the voltage at the amp output for the same power delivered to the speakers.

BTW, it would be a lot simpler to follow if you just used V^2/R for power

Roscoe[/quote]

To translate what Rosco said -- it's not rocket science and making a mathematical mountain out of a mole-hill gets confusing for no reason.

[Stuart Polansky]

Like any good internet thread, this one has gone way off of the rails and taken on a life of its own.

Walter had the best answer, which was "gee,just try it and see".

As to the power formula and Ohm's law, thanks for the tutorial, but I'm fairly familiar with them.

As to comparing voltage input to the two variations, and as stated, in my case that means 8 ohms or 32 ohms, the notion of constant voltage input to either is wrong. If they were connected to an OTL or SS amp, sure, but (AS STATED) I'm connecting them to different taps of a transformer coupled amp: 4 or 6 ohms, respectively. This because the design I'm using actually requires OPTs with different turns ratios to achieve about 8K PP. I'll probably end up around 10k, so power will suffer a bit.

If the gain is set to provide a set power level, the voltage applied to the 32 ohm speaker will be twice that applied to the 8 ohm speaker, since effective TR is doubled, right? So power level comparisons using a set voltage level don't work.

Tom P told me in a phone conversation, that there was no acoustic power efficiency gain when series connecting speakers. (Unless I misunderstood him), I just accepted that as fact. He's a sharp guy, has messed with more speakers than Roscoe has bullets.Okay, maybe not that many. In either case, of course, the drivers are each delivering half the sound pressure, so distortion should be lower.

All that crap said, again, I don't care about the efficiency. I am not using low powered or flea powered amps. I know that the speakers in question will BLAST Hell's Bells as loud as I can stand with a mere 16 watt amp. The new application will have more power than that.

MY QUESTION is more about what happens sonically when two voice coils and all of their imperfections (parasitics?) are connected in series? In parallel, they are not interacting with each other as much, although, I suppose a non-global feedback amp with a substantial "non-zero" output impedance might allow the drivers to affect each other electrically. But in series, does anything, positive or negative in quality, happen sonically that would have one prefer a parallel connection? Anyone have any actual experience with a comparison between the two?

The other half of that question was regarding the output transformers. Using the 16 Ohm tap halves the TR vs the 4 Ohm tap. I keep reading about the wonders of lower turns ratios and improved HF performance as a result. Perhaps that advantage is only seen with turns ratios which are a magnitude lower, not just half. When connecting to the 4 Ohm tap, the feedback connection isn't directly coupled to the speaker, it's typically left hanging off of the 16 ohm tap. Half of the output winding is generating an unterminated/unloaded signal. Seems like a less-than-ideal situation.

I haven't experimented in this realm yet, was just looking for someone else's practical experience.

Not a guide to the basic calculations, for goodness sake.

But thank you.

[dberning]

Stuart,

By all means, using the entire winding of the transformer provides better coupling and thus is more desirable in theory. Now on a comparison in listing, all bets are off.

I don't see any fundamental problem with the two drivers in series.

David

[Stuart Polansky]

Stuart,

By all means, using the entire winding of the transformer provides better coupling and thus is more desirable in theory. Now on a comparison in listing, all bets are off.

I don't see any fundamental problem with the two drivers in series.

David

Thanks David!

[Cogito]

Assuming the 4ohm speaker has the same power sensitivity as the individual 8ohm drivers, the amp will have to produce only half as much power to generate the same acoustic output. The amp will have 3dB more headroom, how is that going to lose dynamics?!?!

Roscoe

The point I was trying to make is, once the signal is screwed up due to AMP's inability to produce sufficient power on demand, having higher efficient speakers wont help.

[Cogito]

Yes, the thread sidetracked a little from your original post. I was trying to clarify what was being said.

Tom P told me in a phone conversation, that there was no acoustic power efficiency gain when series connecting speakers.

Thats exactly what I said in post #3 in this thread.

On the acoustic side, a pair of parallel speakers will result in +6dB gain and no gain in serial speakers.

[tomp]

In addition to the sensitivity issue there is another factor to consider if you are going to use the drivers around resonance. As the driver goes from just below resonance to just above resonance, not only does the impedance change but it goes from leading to lagging. I'm enclosing a plot of a typical impedance curve. A few things are noticeable.

First at resonance the impedance increases dramatically which greatly reduces the ability of the amplifier to control cone motion because of the limited ability of the amp to provide corrective power to the driver. Second, notice that the driver goes from being more capacitive to more inductive. Not every driver of the same make and model has an identical Fs. If they are in parallel it is of no concern. If they are in series, one driver affects the motion of the other. Each driver sees the other driver as a generator with different characteristics. I have connected the same models of driver in series and using a stroboscope have noticed that at resonance the motion of the drivers can get out of phase with one leading the other. That did not happen in a parallel configuration because in almost all cases the amplifier can swamp out the effects of the drivers because the amp is connected to both as separate devices. In series the amp output is fighting with the out of phase energy of the other driver. With a series configuration since the cones are out of phase, the acoustical output is reduced and you will get a response dip at resonance.

At the Midwest Audiofest in 2016 where I was a judge, after the competition was over one of the competitors came up to me and complained about a response dip in his arrangement of multiple 15" drivers connected in a series arrangement. He couldn't understand it because the drivers did not have that dip when measured individually. When I asked him if the dip was at resonance he replied yes, how did you know? I explained this condition to him and it then made sense.

I have since avoided using series connected drivers if they are going to be used around Fs. I do use them in series when they will be in the area above resonance, for example with my dipole mid bass array where I have three series connected pairs of drivers then wired in parallel. The Fs of the drivers is 25 Hz but I use them from 100 to 300 Hz and have no problems.

Attachments

[Stuart Polansky]

Tom,

Was waiting for you to reply on this. I remember you telling me there was an issue, but forgot the details. Thanks.

My drivers have an FS of 37.7Hz. Plan is to use them over 100-125 Hz, depending...., so hopefully no issues.

Stuart