Chained speakers signal connected in series

I didn't confirm your hook-up but it will work fine. The next question is your amps. Cannot they take the lower impedance of a parallel hookup? If they can that is better and more power. 

As long as the cable is LARGE enough and a good quality cable it should be the same. BUT you're doubling the ohms and 1/2 the output of the amp. In turn low distortion and much higher dampening. Probably sounds BETTER in series for sure.. Second it will surly help the amps run cooler.

Just something to think about.. 

You can connect two speakers in series and it won’t harm the amplifier.
BUT because of two sets of crossover connected in series the sound it produce will be totally different from just one speaker connect to the amplifier! Both woofer in each speaker will get much lower crossover point and the tweeter will get a band-pass filter instead of high pass.

Unless the two loudspeakers are identical, there will be some modification of the frequency response of BOTH speakers. It does not matter which is first and which is second in the series.

The difference may or may not be small, it depends on the specifics. Try it and see what you think.

I disagree with imhififan’s opinion that the crossover points will change. In my experience, they do not. But each speaker’s frequency response will be modified by the other speaker’s impedance curve being in its signal path. If the two impedance curves are identical, then so is each speaker’s sensitivity to the other’s impedance curve, and the net frequency response of each is unchanged. But if their impedance curves are different, then there will be changes in the frequency response of both speakers.

(For geeks only: It is tempting to think of each speaker as a resistor in series with the other, thus ruining the amplifier damping factor for both and effectively doubling the woofers’ Qes, but this is also incorrect, because each speaker also has its own motor strength.  Assuming generally similar speakers the DC resistance is approximately doubled but so is the total motor strength, therefore woofer Qes is not degraded. This is a different situation from adding an equivalent series resistance, in which case only the DC resistance would be doubled [with total motor strength remaining the same], in which case the effective driver Qes would also be doubled, as predicted by the calculated degradation in effective amplifier damping factor.)

Duke 
speaker designer

Both woofer in each speaker will get much lower crossover point and the tweeter will get a band-pass filter instead of high pass.

<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

No it does NOT.. The impedance does not change at the driver, Only what the amp sees at the XO terminal, not the driver. The ONLY way to change the resistance is to CHANGE the resistance.. We haven’t

If that were the case then running in parallel would produce the opposite effect AGAIN. Twice the current and double the crossover point? Double the current draw but it doesn’t change AT the driver. The driver resistance hasn’t changed a bit..

<<<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>

I did this 15 (?) years ago. READ ON!!

Let us say that both channels were being fed with a pure 1kHz sine wave and that both halves of the amplifier were identical in every respect.

If you connect your stereo speakers conventionally, then for each positive part of the sine wave, the amplifier drives each speaker cone forwards. For each negative part, the amplifier drives the speaker cone in the opposite direction. Each speaker then produces a 1kHz tone.

The positive speaker terminal is producing a voltage that represents your sine wave. The negative (or Common/Ground) terminal is the reference for this voltage. (across two points right)

Now, connect a single speaker across the two positive outputs and ignore the grounds. If you have identical voltages with respect to ground coming out of your terminals, then the net voltage (and hence speaker cone displacement) is zero and so you will hear nothing.

Only very very expensive amplifiers have this degree of perfect matching so it is likely you may just barely discern some sound.

Now, connect a conventional stereo source to the amplifier and what you now get is a voltage on each positive terminal that is slightly different from the other. In this case, you will hear something. Often, vocals are fairly central so the same signal is present on both channels and will cancel out. Instruments that are left or right of center (i.e. more on one channel or the other) will be discernible to a greater or lesser extent depending on the amount of stereo separation.

A Little trick to try.

Connect a pair of speakers to your amplifier but instead of the two negative wires returned to their terminals, connect them to each other. Effectively you have just connected two speakers in antiphase series across your left and right channels.

Play your tunes and enjoy "separated stereo with little or no center sound stage"!

This is not as stupid as it sounds. The A speakers, (Left and Right) are wired conventionally. The B speakers are wired as I have just described and are placed further apart than the normal speakers but slightly behind the listening position.

By running the amp with just the A speakers selected, you get conventional stereo.

By running the amp with A + B selected, I get a wider stereo effect with no additional electronic processing involved.

By connecting a suitable value resistor or variable resistor between the speaker commons and the amplifier negatives (they are usually common so you only need one connection) and allowing some of the signal to go the conventional route and some to go the "wrong way round" you can vary the WIDTH of your stereo separation.

You can go one better than this.

Use TWO amplifiers (a matching pair). Connect the source to both amplifiers using a Y splitter lead on each channel. Drive two pairs of speakers from each amp so that you have 8 speakers in total. Position four of the speakers around the lower half of your listening area. Position the other four higher up in the room so that you have the same left-right and front-rear orientation (but two layers if you like), separated by a meter or more. Ideally half a meter above and half a meter below your ear position. Do not connect any speaker grounds to the amplifiers - wire all four pairs as "antiphase series".

Now, instead of using a variable resistor on each back channel, use a 2 axis potentiometer (joystick) with the center taps of each axis connected to ground and the ends of the tracks (two per pot) connected to the junction of your speaker negatives.

You can now use the joystick to completely mess with your aural experience. Purple Haze comes to mind... "Movova, and Let Jimi take OVA"

TINKER TIME FELLA... Back to the lab.... Come on dog...

Regards

I disagree with imhififan’s opinion that the crossover points will change.

I believe the Tannoy SGM 10B has parallel crossover, if connect two speakers in series, the LF section will be #1 speaker low pass inductor --> #1 speaker woofer --> #2 speaker low pass inductor --> #2 speaker woofer.
Correct me if I'm wrong, doesn't each woofer will see two inductor and a woofer voice coil in another speaker in series?

Interesting - I spent a little time to sketch the schematic. Granted the amplifier output will see 2 inductors in series, the crossover frequency of the driver is based on the L and C values. That ratio does not change whether the speakers are connected in series or parallel. My vote is that it does not change the crossover frequency. As far as the series or parallel connection of the speakers to the amplifier, the series connection does make it typically easier for the amp to drive the load. But this is only true up to about 70% of the maximum voltage capability of the amplifier. A parallel connection of the speakers will yield a 3dB increase in SPL, if your amplifier can handle the load resistance and deliver enough current. I preferred to connect them in parallel. That being said, I frown on multiple drivers in series OR parallel, especially in the midrange and tweeter frequencies. Whether they are point source of line source drivers, the sound will lose detail and become fractured and unclear. Happy listening!

Good summary.

A 6db first order XO in series, the SECOND speaker drivers will get what is left over after being filtered by the inductor. IF the XO is built in parallel
NOT series.  Parallel the speakers would NOT change the Crossover points. In series it would.. 

To keep the XOs points all the SAME. Paralleled XOs and Paralleled Speaker connections. 

In series WILL change the XO point. I'm not sure how much but a little grade school math will get you close. It would be LOWER.

I'm not sure you could hear anything BUT a bass signal out of the second speaker box in series. Parallel and use a stout class d, that will work.. :-)

That wiring set up is exactly how you wire a MB and Sub in series by just adding an additional resistor between the two. The first would be at 280 and down, THEN between drivers ad a second resistor to drop to the 80-120 db first order. It's part of a Quasi 2nd order paralleled passive XO, I actually use.. Sure helps out a tube amp IF it has to do bass duty, TOO.
Shares the load more evenly. LOL It's HEAVY all the time..:-)

It's a great design for planars too.. Fast, transparent and VERY revealing.. What  you're feeding the drivers is exactly what your gonna hear.. SO a high floor noise, bad source or funky sounding (BRIGHT) SS STUFF is still gonna sound BAD.. High E speakers 90+. One ribbon I use is 103 db and 100 watts RMS. 10 watts your ears are bleeding..
1-40khz. I can point that thing and break glass. I'm pretty sure..

Regards

oldhvymec - You are funny, thanks. I'm going to actually do some "grade school" math on a 2nd order low-pass filter. I will connect them in series first, then in parallel and the numbers should speak the truth, unless my math is bad. I will share the numbers once I'm done! Your analogy of a 103dB speaker also got a chuckle out of me. 103dB @ 1watt,1meter distance is not that loud, really. Break glass you're flat out wrong, even 1000watts, that only 133dB. Jet engines at a distance of more than 100 feet are 140dB, that doesn't break glass. 123dB @ 100watts is loud, but rock concert levels and threshold of pain are about 130dB. While I don't listen to music at those levels now, I wish I NEVER did... 

I did some testing this morning by connecting and disconnecting the speakers from the amps. The speakers connected in series (two speakers per channel chained) sound different from what they sound when they are connected separately (single speaker per channel). Interesting, that while both speakers in series produce a duller sound, one of the speakers degrades the sound more than another. Unless it is OK to connect them in parallel I would have to loose one pair from the system as it appears a single speaker per channel sounds way better. 

Yes, OP that is basically what I said. One speaker connected to each amplifiers output will sound better than two, whether it's in series or parallel. If you connect them in parallel, if your amp can handle it, it will be about 3dB louder but NOT as clear, just like in series.
For those other Agoners that said the crossover frequency will change when connected in series, I have done the calculations and the crossover frequency DOES NOT change. I will post diagrams later, but if you want to do your own calculations, be my guest. Can we agree that Rt equals 16ohms in series? Can we agree that Lt = L1 + L2 in series? Can we agree that 1/Ct = 1/C1 + 1/C2 in series? If it's yes to the above, then simply plug those numbers into the equations. f = R/2πLQ and f = Q/2πRC. Both will come out the same with a very small error depending how many decimal points you take it out too. Happy listening to all.

     Personally: I don’t find the idea that an audio signal, after passing through a system* when fed into a subsequent and identical system, in series, might just sound a little different, the LEAST bit surprising.

     They’re rated at 6 Ohms, so: can only be run in parallel (a 3 Ohm load, but varies, fed a music signal), with an amp that can handle a 2 Ohm load, with stability (check your amp’s specs).

     *Referring, specifically, to the TANNOY SGM 10B Super Gold Monitor Speaker; comprised of two open-air-core inductors, two iron-core inductors, a bunch of wire wound resistors, a couple voice-coils (resistance, inductance and back EMF) and an electrolytic.

      Regarding those voice coils, impedances and such:    http://education.lenardaudio.com/en/05_speakers_3.html

      The chart shown: for an 8 Ohm (nominal) speaker

                 Note the lower freq impedance dip.        

       btw, for clarity: What I listed above is all found in ONE, SGM 10B crossover network and the system’s two drivers.

rodman - That's all good info, but when I saw "geek" letter for ohms instead of Greek letter, it made me skeptical😁 The clarity thing I mentioned is from my testing, I had to prove it to myself. The theory about multiple drivers over the same frequency range is pretty well known. This is for high frequencies mostly, but also applies to the midrange in my listening tests.
esputnix - To make your listening tests more meaningful and convincing, use an SPL meter to make sure the listen levels are both the same. I would also use the same program material, speaker placement and distance from the speakers to your listing position. All things being the same, something louder will appear to "sound" better.

@fiesta-    I had posted, regarding how humorous that was, then thought about my own failings at being perfect (ALL the time) and deleted that line.
(part of my typical typing/posting/editing method in here, as outlined elsewhere)

      I'd also noticed, "symbolised", in the same sentence, which, for all I know, may indicate: he's from Britain (or something, maybe), which still wouldn't account for a Geek omega.

       Far as, "skeptical": I've known everything in that portion of the article, for decades and am confident of it's accuracy.  😉

        btw: When I said, "...might just sound a little different...": I was being facetious.

         I can't imagine that sounding worth a crap, for the reasons mentioned (especially: back EMF, feeding the next system)!   But then: I've never connected Tannoys in series, so: what do I know?

                                         Happy listening!

 I will share the numbers once I'm done! Your analogy of a 103dB speaker also got a chuckle out of me. 103dB @ 1watt,1meter distance is not that loud, really.

<<<<<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>

At 1000 watts with 12 ribbons and planars drivers, I'm no where NEAR my system. Not to mention the other 20 MB and sub drivers, driven by 2 12K class ds. It would cave your chest in.. MB are 93% E 8" drivers

Add my center channel of 16 more neo 8 planars. We are close to BLASTING the meat off the bone at 3 meters. I would say 130 db would be conservative. I have guinea pigged a few smart A$$ kids.. It's left them with a nose bleed or two.. LOL

I'm sensitive I guess. :-)

I actually have ear plugs in right now!! Unless I'm sleeping or listening to music/TV they are in my ears.. At 66 I still spend a lot of time with noisy hand held tools.. My clunky fingers have a tough time after I use them for a while.. THUS the ear plugs stay in.. Not out.. 45 year now. 

I listen most of the time after 23-2400 hours. 2-6 watts with 6V6s and maybe 20 watts for the bass and 10-15 watts sub.. I can hear it EASY in a 25 x 40 x 16 wood shop.

BUT the issue is in series, imhififan made a good point. IN SERIES the second speaker is getting only what is NOT filtered from the first XO.

Sound different? Sound BAD is more like it, real bad.

Parallel will work with a stout class D no problem. I'd even say work well.

I like mids 300 hz and UP AT least 6 ohms 10-16 is even better. Valves don't like that low of numbers. My Macs will run at 2 ohms.
Sound BAD at best.. 8 ohms or > they just twinkle.

Regards

I just read your post "DOES NOT change." fiesta75.

Sure it does IN series. The signal path has gone through EVERY component in the first speaker box IF it is a typical series XO. IF it is a parallel OX design (like a Quasi second order) and BI AMPABLE it would STILL pass through all the XO filters BACKWARDS (?) in the bass or the mid/highs XOs after it went through every component in the first speaker box..  + - + -, is every component including the speaker terminals and VCs.

It’s IN SERIES, not Series/parallel or Parallel/series BUT one AFTER the OTHER.

How you like that for Techno talk.. I use to have to teach smart apprentices. Oh yea, they were smart alright. .Just like I use to be.. :-)

How’s my math by the way, you notice, no numbers applied for the unsighted.. I’m close after 50+ years of welding.. I strike an arc I can see just fine. TIG, MIG, TWIG, TWIT, TWILL, TWASEN’T.. LOL

Regards

I will post some sketches tomorrow. If you connect the second speaker just across the woofer of the first speaker, the second speaker will see both inductors. I'll show you what I mean, but if you care to just plug some numbers into the "grade school" math equations, you will be able to see without me having to post anything... Elementary school?🤣

@fiesta75, 

Can we agree that Rt equals 16ohms in series?

Yes

Can we agree that Lt = L1 + L2 in series?

No, IMO, the first woofer will see the voice coil of second woofer as an inductor in series, same for the second woofer, it will see the voice coil of first woofer as an inductor.
I think the total inductance should be Lt = L1 + L2 + Lvc1 + Lvc2

Can we agree that 1/Ct = 1/C1 + 1/C2 in series?

Yes

it's yes to the above, then simply plug those numbers into the equations. f = R/2πLQ and f = Q/2πRC. 

May be a crossover design simulator software can give us a better answer.

Even if L1 and Lvc1 are present, L2 and Lvc2 are also present. If the frequency of the crossover is say 2kHz., it will remain at 2kHz., because L1+Lvc1 equals L2+Lvc2. Is that not correct? L1 = L2 and Lvc1 = Lvc2 for each speaker. Both inductances are present whether they are used to calculate the crossover frequency or not. Correct?
Perhaps a crossover design simulator would give us a better answer, I am willing to listen to your presentation. That could save me the hassle of posting sketches. I think we are both capable since we have been working to solve the problem with likklegerry's crossovers. I will be waiting.

 While you're listening through your pencils: you're going to simply ignore everything that a voice will induce to a signal, when actually presented with music?

                                        HILARIOUS (in a tragicomedic way)!

     Best broaden your horizons, as to what goes on (and can be measured/calculated) when certain elements (ie: caps and voice coils) are actually under the influence of *changing frequencies (ie: an actual audio signal) as opposed to DC (ie: a multimeter).

     *Ever looked at the impedance plot of a woofer, actually being used in a circuit?

       (https://audiojudgement.com/speaker-impedance-curve-explained/)    

           (https://www.apogeeweb.net/tools/reactance-calculator.html)

       Perhaps: even plug some of those facts, into your calculations?

  (http://oer2go.org/mods/en-boundless/www.boundless.com/definition/reactance/index.html)                
 
                       NAH, why bother with Physics, right? 🙄

          ps: I only bother posting such things, knowing they will be ignored by most, BUT: hoping someone out there remains interested in facts/learning.

                                             Happy LISTENING!

@fiesta75,
I think it was 7 or 8 years ago, my friend use Hornresp (http://www.hornresp.net/) did a simulation on his speakers and found that connect two speakers in series the low pass filter cutoff frequency get lowered, our conclusion was the first woofer will see the voice coil of second woofer as an inductor in series, same for the second woofer, it will see the voice coil of first woofer as an inductor in series, and caused crossover point lower than a single woofer alone.

If Lvc1 does not equal Lvc2 in a dynamic setting, then it maybe possible to perceive a change in the crossover frequency, maybe even measure it. I’m saying in a "grade school" math equation, the data does not support this. I personally don’t ever connect speakers in a series configuration, for MORE than one reason. It just would be nice to see some actual data from some physical testing or calculations that will support the crossover frequency change. Does anyone want me to post sketches or does everyone understand the circuit? I don’t want to spend more time on this subject unless someone has proof to show me. Enjoy the music!

I think audiokinesis nailed it here! I should have mentioned this before, thank you for sharing your knowledge.