Another "Jer-Pondering" - Speakers in series

[jbeckva]

One thing I used to use (and somewhat still do) to evaluate an amp, receiver, whatever "business end" of the system, is whether or not it would put the "B" pair of speakers in series or keep it in parallel with the mains. To me, when they kept them in parallel then that is a mark of confidence in the design to where it's built to handle low impedance loads (put two of the same impedance in parallel, half the impedance).

But also.. I was thinking about when the configuration is "series" instead. Aside from the obvious - the output's just about halved for each speaker - I always had some suspicion that the overall quality of sound decreases.

So I'm thinking (pondering) about this now, and hmmm... Most of us know about the "impedance curve" right? That's where the actual instantaneous impedance of the speaker will vary depending on frequency. Some speakers out there - dramatically. They call those a "tough load" to drive.

So say you had two sets.. one that's fairly constant, and then you have one of these "wild stallions" that will jump to 20 ohms at 100hz, then all the way down to 2 at 14Khz (or something like that... )...

So given the voltage output at the amp, designed to be "true" to the source and fairly constant, wouldn't the speaker that has large variances in impedance affect the other more "constant" speaker?

Take the theoretical example above... 20 ohms at 100 hz.. So - let's say the other speaker was a lot flatter - 5 ohms 100 hz but on average 4 ohms nominal. Given a 10 volt input, 25 ohms total (we're in series here) would yield 400mA of current. So with that current going through, you would see 8 volts present across the speaker with 20 ohms, and only 2 volts across the 5 ohm speaker. Power wise - that's 800mW for the 5 ohm speaker, but a whopping 3.2W for the 20 ohm speaker.

So... with all the "pondering" above... wouldn't putting speakers in series with each other cause actual sound degradation??

HMMMMM...

 

[laatsch55]

I always thought so, plus, in series wouldn't they be out of phase???

 

[Pure_Brew]

I haven't looked this over for awhile but I believe the two speakers will sum to 1 specific load on an amplifier, as the two speakers make up one complete circuit as far as that is concerned. Objectionable sound quality will likely come from subtle, and not so subtle, differences in the starting and stopping points of the individual diaphragms and their overall response patterns. The result is an erratic comb filtering effect where the response will look like what you get in a poorly behaved room, likely worse.

As far as what each speaker "gets" is a bit beyond me, except for things that are generally relative like efficiency and frequency response individually, which I believe would not change. If there is a change, I believe it would be due to the ability for the amplifier in question to drive a more difficult load, which itself would have to be defined by the system and phase angles presented to it.

 

[Pure_Brew]

I always thought so, plus, in series wouldn't they be out of phase???

The whole system would be in phase but 2 different speakers playing at the same time would have some in and out of phase response anomalies.

 

[Gepetto]

You should not do this if you care about sound quality. As an example of what you are talking about, consider that you put two slinky spring toys in series with a ping pong ball in the intersection of the two slinkys. One end of the series slinky pair is tied down to the floor. The other end of the series slinky is in your hand and you have stretched the slinkys to establish a rough equalibrium between the two springs.

Your goal is to try and control the ping pong ball in the center node between the 2 slinkys such that the ping pong ball is always half way between your hand and the floor while you are moving your hand up and down. You will find this task virtually impossible to achieve. The ping pong ball is analogous to the center node between the 2 series connected speakers.

Compare that to the 2 slinkys being in parallel. That combination is easy to control.

With 2 speakers in series, you basically have killed the high amplifier damping factor.

 

[Pure_Brew]

I like your analogy. The entire frequency response spectrum would then be infinite slinkys. If they were in parallel, that would be the same issue. How would a parallel connection effect damping factor then?

 

[jbeckva]

Exactly what I was thinking would happen... the slinky analogy is dead on.

I think tho (PB) that in parallel it would be separating the two slinky's ... one in one hand, the other in another. The voltage the amp puts out would still be constant across both, to the limit of the amp's current capabilities. Dunno how it would affect damping factor tho... hmmmm... good question...

 

[Gepetto]

Technically if you parallel 2 speakers with equal impedance then the damping factor is reduced by a factor of 2. For example, an amplifier with a damping factor of 1000 rated into 8 ohms will have a damping factor of 500 driving 4 ohms (or 2-8 ohm speakers in parallel).

Practically the damping factor of most SS amps with large global negative feedback is so high that the reduction from 1000 to 500 is not even noticeable (or audible).

 

[Pure_Brew]

but if we go to 16ohms then....?