I think the loudspeakers were similar to the currently available SVEN 230 computer speakers (available in Russia). Here's the photo of internals, it's a "barebone" opamp-based amplifier (I suppose) with a built-in transformer. I haven't disassembled speakers I had in 2011, but considering the cost ($10 per pair), it's the only option. Maybe I'm mistaken about power rating, it could have been 10W per pair, but that's not so important I think.

Each one of the speakers had its own power plug to electrical outlet? or  just one for all of  the speakers ?
Please correct me  if I'm wrong.

I'm talking about 3 speaker-pairs. Each pair has 1 stereo input which I converted to mono by twisting the L/R input leads, so they were receiving the same in-phase signal, obviously at a lower voltage. Then all 3 twisted input stereo pairs were twisted together with the output channel of the soundcard (I can't remember exactly, but I've probably only used 1 output lead as to not cause any interference on the soundcard); ground leads were twisted together, of course. Very loose and lame way to do it I know, but as a result all speakers received the same in-phase signal, from the soundcard which produced a mono sinewave signal output.


Each speaker pair was connected to the electric tee independently, 3 connections overall. Note that the fuse on the tee was blown not due to connection awkwardness, but only after I hit a certain frequency on computer's sinewave generator plugin.


I'll try to do it all again soon, I've just ordered the required parts.


I do not think soundcard chip model is relevant, and I do not know which one it was. Probably, some common Realtek soundcard chip.

So I understand that  e.g right channel from the sound card was not connected at all .
and only left was used ?

Yes, most probably. That's what common sense tells me at the moment. There was no need to twist output leads together as only a single source signal was needed. That was the whole idea in fact - hexagonal geometry and a single signal source. I now think that 6 speakers in a hexagon were a bit overengineered, but I guess otherwise I would not get big enough energy surge to notice anything. Now I do not think distances (108cm) or the number of speaker pairs are important overall, but to myself it was the only way as I'm not an EE.

Question  about the tee:
The tee you are talking about- was it the electrical extender  allowing from one wall electrical outlet to  have  3 individual  outlets like  T splitter  connected into the 220V outlet that had its own 15A fuse?
Wesley

Yes, it was a simple electrical extender with 15A thermal fuse, the simplest one that costs about 5$ at the moment.


The apartment has 360V input, but as I was doing the experiment in the same room I think both the computer and the speakers were connected to the same 220V phase, just different outlets.

1.
Exactly in the middle of the distance between the speakers the two pressure waves collides
squeezing air - when the both membranes from the 2 speakers are at OUT  position exactly in the same time.
If we place on the geometrically perfect circle equally spread 6 speakers than all of the 6 pressure waves from the 6 speakers collides
squeezing air - when all membranes of the speakers are at OUT position exactly in the same time.

Sorry, but you are missing the frequency component from your reasoning. They can be in-phase or out-of-phase at a given point in space *depending* on the frequency of the soundwave. There's no much sense to consider the central point as soundwaves mostly do not interact with each other. It's important what the phase is at the point where the diaphragm is located.

Was it possible that the speakers from the same pair were placed  exactly at opposite sides facing each other?

This is an interesting question, but I do not remember exactly. This may or not be important relative to the amplifying cascade. Anyway, for any experimenter there are only two options available, not too much work to find out. As I've replied earlier, I'll retest, and hopefully I'll get a blown fuse again-I've ordered several for replacement.

Speed of electrical signal in the wire is: 10 ⁸ m/s
speed of the sound  in air is 331 m / s
So if the electrical signal is delivered to  our 6  speakers  than all of them are in phase at any time and frequency  is not important 

Well, it's not important how electrical signal is converted to pressure, it may be out-of-phase depending on how speakers were wired. What's important is which phase of the *opposing* soundwave you get on the diaphragms.


I much prefer Genelecs myself. But this is a different kind of application, it does not require full-frequency-range precision. One needs to reproduce just a single sinewave, and for that kind of application any junk loudspeakers can be used.

it may be out-of-phase depending on how speakers were wired.

We are assuming that speakers are connected exactly the same. red to red ,black to black. As you didn't change polarity in your original experiment right?

Well, it's not important how electrical signal is converted to pressure,

We are  collecting information now and not yet  formulating any conclusions
Wesley

Dear Alex my time is very well paid in American dollars, and I'm dedicating it  to you for free.
We are not going yet to  the nonsense that makes sense for you.
We may, but not now.

So far we have   probable configuration of :
6 speakers in 3 pairs where  3   power cords were connected to the same electrical outlet 220V 15A fuse protected
Each pair had its own  amplifier .
If we  draw the circle  we would spread them equally on it but each pair of speakers is  facing each other in the straight line
 drown through the center  of the circle.
Computer with the  sound card was connected to  separate outlet presumably from the same electrical phase.

_________________________________________________

Impedance match and load  specification:
In electronics, impedance matching is the practice of designing or adjusting the input impedance or output impedance of an electrical device for a desired value.
https://en.wikipedia.org/wiki/Impedance_matching

load by itself is specified in its general form as:
the quantity that can be carried at one time and also measured quantity of a commodity fixed for each type of carrier.
So  in the electrical outlet we have 220V potential and ZERO current.
The load will be measured by current drown from the outlet  and expressed by European formula P=U x I
 where P is power in W   
U is 220V
and I (current in Amps) is  unknown yet.

So Load  needs enough  power  drown from the electrical outlet .
That has nothing to do with gain or amplification  or number of speakers or feedback
Load is a Load.
But if we don't have purely resistive Load we may expect that  Load  doesn't have fix value.
We know that  the bigger the load the  higher is the current drown from the electrical outlet 220V

So we may project that outlet as a big Russian guy having enough of power to carry on the load needs-
-like  if it was not  some resistive load but a pretty  Ukrainian teenager girl.
She wants this  she wants that, and  the guy happy or not must deliver.
 
Impedance vs changing value of the Load:
That pretty  Ukrainian girl requires  a maintenance  resisting  in many ways if not happy.
In  electronics  the Load can be  purely resistive at 10% but   in many ways if not happy
manifesting additional  capacitive or inductive reactivity at 90%
Your own picture  here: Re: Acoustical feedback energy system shows the transformer
from power supply mounted  in  one of the  every two speakers.
This transformer has   coils- means inductive reactance and  capacitive reactance  of the winding:
https://en.wikipedia.org/wiki/Gyrator%E2%80%93capacitor_model
and  you have 3 of these  transformers in your 3 pairs  ( total 6 speakers .)


At resonance, both capacitive and inductive reactance will be equal to each other.
The inductor and capacitor will also be conducting more current at the resonant frequency.
Current flowing across both components is 180° out of phase, which results in a mutually canceling current.
exploring-the-resonant-frequency-of-an-rlc-circuit

So we have  something that looks like short circuit blowing the fuse .
 In other words you didn't have any gain, any magic energy produced,  you  just  sort of shorted the two terminals of that 220V outlet.
And by that  you are the double loser. You can't use your speakers at this   particular frequency and you don't have power in the outlet.
what-is-a-srf-this-is-used-in-ac-power-distribution-system-with-isolation-

The SRF is the frequency where the capacitive and inductive elements in the system are of equal magnitude and opposite sign.
This leads either to very low  impedances 
The effect is that the system does not work as intended. Keeping SRF well away from the operating frequency is a key design requirement.

To better understand  what I;m talking about  take look at the  DC and AC Relay. or  DC and AC electric door lock - the strike
A DC relay coil has resistance that limits the dc current.
An AC relay coil relies on its impedance for governing the current.
So it has few winds ~10% and ~90% is impedance

If you than decide to connect  the AC relay to DC voltage you will damage it as we speak.
Sometime it will blow in fumes.
because for DC that  relay or that electric lock doesn't  have any impedance - it Doesn't exist.

Additional  factor to know:
SWR of impedance match.
standing wave ratio (SWR) is a measure of impedance matching of loads to the characteristic impedance of a transmission line or waveguide
https://en.wikipedia.org/wiki/Standing_wave_ratio

If we don't want to explain the phenomena with use of   Counter-electromotive force (counter EMF, CEMF, back EMF), - the electromotive force (EMF) manifesting
as a voltage that opposes the change in current which induced it. CEMF is the EMF caused by electromagnetic induction.
we may  use -
The short explanation:
If There is no impedance patch to the load  than part of that energy   is reflected back and impacts the  source of energy   like Transmitter  of RF or the  outlet.
_____________________________________________________________


So you were not producing any magic energy but in  resonance  your power transformer  primary winding  had  more current than your outlet   fuse was able to withhold.
cheap Chinese design - that's all.

Note:
The electronic load  in tests : tests energy sources or energy conversion blocks with static or dynamic loading.

Wesley

weslie, please answer a question: will such loudspeaker (as I've used) short-circuit if the movement of the diaphragm be hardly constrained, or the voice-coil glued to the magnet?