Speakers in Parallel or Series... does it matter?

[dbishopbliss]

I am looking to buy two speakers to put in a cabinet for an 8 ohm load. Is there a difference in sound or how the amp/tranformers react between running two 4 ohm speakers in series vs two 8 ohm speakers in parallel?

[sluckey]

I am looking to buy two speakers to put in a cabinet for an 8 ohm load. Is there a difference in sound or how the amp/tranformers react between running two 4 ohm speakers in series vs two 8 ohm speakers in parallel?

Two 4Ω in series would match your 8Ω load spec but two 8Ω in parallel will not match your 8Ω load spec.

[shooter]

I got a 16Ω as payment for fixing an amp, bought another to make an 8Ω 2X12 cab

found 16Ω speakers cheaper than their 4 or 8 counterparts 

[jjasilli]

Impedance matching is an important factor.  Usually a 2:1 mismatch either way is considered OK.  No guarantees.


Other factors:


*  Failsafe.  If speakers are in series and one blows, the whole cab is "dead"; so the amp has no load.  This could harm the amp.  If speakers are in parallel and one blows, the amp still has a load.


*  SPL.  With identical speakers in series matched to the OT secondary: they equally divide-up the drive voltage from the amp.  The sonic output of ea is cut in half for a drop of 3dB ea, totaling a 6dB drop; but there are two of them, so the SPL increases by 3dB for a net SPL loss of 3dB!


             With identical speakers in parallel matched to the OT secondary, there is a 3dB increase in SPL.

[dbishopbliss]

Sounds like two 16 ohm in parallel (that is what I meant to type above) is the way to go.

[jjasilli]

Agreed, so long as you have that as a happy option.  E.g., once I got a good deal on 2X Celestion speakers and the only option was 2X 4 Ohm in series for an 8 Ohm load.

[arno]

Hi, maybe a strange first post, but I can’t resist the urge to respond here.

... The sonic output of ea is cut in half for a drop of 3dB ea, totaling a 6dB drop; but there are two of them, so the SPL increases by 3dB for a net SPL loss of 3dB!

Where did that other 3dB go? You'd get -3dB when you put each speaker in a separate room.

If the OP expects the same total impedance from 2 identical speakers wired either in series or parallel, the total power will be the same considering all other factors are equal.

one 8ohm speaker:
P = U^2 / 8

two speakers wired in series (4ohm + 4ohm):
each speaker gets half the voltage but the same current must flow through both speakers
P = (U/2)^2 / 4 for each speaker =>
P = 2(U/2)^2 / 4 for 2 speakers

two speakers wired in parallel (16ohm || 16ohm)
each speaker gets the same voltage but current is divided over two speakers
P = U^2 / 8 / 2 for each speaker =>
P = U^2 / 8 for 2 speakers

Most speakers have a different sensitivity spec and frequency response for different impedance versions from the same speaker model so you will get a different sound depending on which model you choose.

[shooter]

here's a quicky on SPL, speaker guys like the silly math 

[jjasilli]

one 8ohm speaker:P = U^2 / 8

I don't know what this symbolism means.


If the OP expects the same total impedance from 2 identical speakers wired either in series or parallel, the total power will be the same considering all other factors are equal.
This is a compound statement which is confusing:  if 2 speakers are identical they will not present the same impedance if wired in series vs. parallel, nor can that properly be compared to just one of those speakers.  It would be helpful to "unpack" the compound statement.

You'd get -3dB when you put each speaker in a separate room.
Not applicable.

Series:  At any specific frequency, a speaker's impedance presents as a fixed resistance to that frequency.  Let's use 1000 Hz; and the amp wants to see an 8 Ohm load.  At 1000 Hz each nominal 4 Ohm speaker presents as a 4 Ohm resistor, for a total 8 Ohm load.  Ea speaker (4 Ohm resistor) in series drops 1/2 half the drive voltage.  So, if the amp is outputting, say, 12VAC  @ 1000 Hz, each speaker in series gets only 6VAC.  However, 1X 8 Ohm speaker would be getting all 12VAC.

[arno]

I presume the OP meant to ask the following question, slightly refrased 

"Should I buy two 4 ohm speakers and wire them in series or should I get two 16 ohm speakers and wire them parallel, either way I will have an 8 ohm impedance connected to the amp. What's the difference in sound?"

I think it's a legitimate question and maybe some general properties may be associated with higher vs lower impedance coils. Maybe not, I know next to nothing about speakers.
Your statement confused me but maybe it wasn't even replied to the exact same question, as the OP made an error initially.
I did not intend to directly assault your post, just want to avoid confusion

I don't know what this symbolism means.

this reads as Power = Voltage squared over resistance (8ohm)
the "^" symbol is often used to initiate an exponent.

This is a compound statement which is confusing:  if 2 speakers are identical they will not present the same impedance if wired in series vs. parallel, nor can that properly be compared to just one of those speakers.  It would be helpful to "unpack" the compound statement.

I'm sorry if I didn't make myself clear. I meant two identical speakers, as in two identical 4 ohm speakers in series vs two identical 16 ohm speakers paralleled (or a single 8 ohm speaker)

... So, if the amp is outputting, say, 12VAC  @ 1000 Hz, each speaker in series gets only 6VAC.  However, 1X 8 Ohm speaker would be getting all 12VAC.

In both cases the total power seen by all speakers is 18W. 6^2/4 + 6^2/4 = 12^2/8 = 18W
Two speakers providing half the power each (series or parallel) will add up and provide the same power as a single speaker (or any number of speakers, in series or parallel) as long as the total load to the amp is the same. Maybe you're saying the same thing?

...The sonic output of ea is cut in half for a drop of 3dB ea, totaling a 6dB drop; but there are two of them, so the SPL increases by 3dB for a net SPL loss of 3dB!

This is what's confusing me. how exactly did you get at -6dB? if the total power is divided by two speakers, each will get 1/2 power. when acoustically combined, they will add up and give the same power.

(In fact two speakers will usually sound louder than one, given the same output power as a result of the greater surface area and wider spread but this is off-topic)

[jjasilli]

EDIT:  Yes, I think you are right.  If one 8 Ohm speaker puts out 100 SPL; then 2X 4 Ohm speakers in series put out 50 SPL ea X 2 = 100 SPL.

Speaker "sensitivity", which produces SPL, is governed solely by volts, not watts.  (Watts are important for a different reason - to make sure the amp has enough power to meet the speaker load demand.)


For a given drive voltage, 2X 4 Ohm speakers in series each see 1/2 the drive voltage compared to 1X 8 Ohm speaker.  1/2 voltage = -6dB.  Two identical speakers in phase & in close proximity = +3dB. 


-6dB + 3dB = -3dB.

[jjasilli]

bUMP.

[arno]

Speaker "sensitivity", which produces SPL, is governed solely by volts, not watts.  (Watts are important for a different reason - to make sure the amp has enough power to meet the speaker load demand.)

If speaker output were a function of just voltage, not power, doubling the voltage would only yield a 3dB increase in SPL.

the table on this page shows the relationship between power and spl
http://www.puiaudio.com/resources-white-papers-speaker-power.aspx

to increase spl by 3dB, you need twice the power or 1,414 * voltage

Besides, the amp doesn't care how many speakers you connect given the same total impedance.
If we're assuming equal sensitivity for all speakers and perfect linearity.
Where are the other watts going?
Only half the power would be converted to sound, but since our speaker sensitivity has been fixed, the other half must be lost.

unless when black matter is involved, a physics theorie is generally flawed if power gets lost along the way.

Maybe this is where the confusion is coming from:

2,83v in 8 ohms = 1 watt
you get 25% power for 50% voltage
1,415v in 8 ohms = 1/4 watt
but in this case the speaker impedance is also 50% to get to the total of 8 ohm (4+4=8)
1,415v in 4 ohms = 0,5 watt per speaker = 1 watt total

[pdf64]

Speaker "sensitivity", which produces SPL, is governed solely by volts, not watts...

Can you provide a citation for that?

My understanding is that mutual coupling (the thing that produces the 3dB gain when the number of closely aligned drivers is doubled) is unrelated to series or parallel connection. It just needs each driver to be identical and to receive the same signal.

I prefer series, as my perception is that one speaker doesn't damp the response of the other.

[jjasilli]

Will research, but meanwhile: amplitude = height of the curve.  For electricity, amplitude = volts.  Think of the height of a sine wave.  The trace of the curve is volts on the y-axis (time or frequency on the x-axis).  The Area Under the Curve = the Power required to raise the curve to the desired height in Volts. 


If you run an amp into an 8 Ohm speaker load, and feed it 2.83VAC signal; that's 1 W.  Let's say 1X 8 Ohm speaker thereby generates an SPL of 100 @ 1 meter.


2X 4 Ohm speakers in series also = an 8 Ohm load.  So the same 1W is dissipated.


But, 2X 4Ohm speakers in series are at 1/2 voltage,  I.e., electrically, 2X 4Ohm speakers in series are the equivalent of 1X 8 Ohm speaker at 1/2 voltage. 1/2 voltage = -6dB.

Acoustically, 2X 4 Ohm speakers are 2 speakers = +3dB.

-6dB = 3dB = -3dB.
***

Contrast with parallel operation.  2X 16 Ohm speakers in parallel also = an 8 Ohm load.  So the same 1W is dissipated. But ea 16 Ohm speaker in parallel gets a full compliment of 2.83V.  Electrically, 2X 16 Ohm speakers in parallel, are the equivalent of 1X 8 Ohm speaker at full voltage.  Electrically we have a -0-dB situation.

Acoustically, 2X 16 Ohm speakers are 2 speakers = +3dB. (compared to 1X 8 Ohm speaker)

0dB + 3dB = +3dB


***
EDIT:  https://www.jpro.co.nz/news-article/wattage-vs-spl-what-should-you-look-at-when-selecting-a-loudspeaker-978 :  :The relationship between voltage input and SPL output is called voltage sensitivity. "

[arno]

2X 4 Ohm speakers in series also = an 8 Ohm load.  So the same 1W is dissipated.

Exactly. But if your series speakers are only converting half the power to sound waves, where is the other half going?

But, 2X 4Ohm speakers in series are at 1/2 voltage,  I.e., electrically, 2X 4Ohm speakers in series are the equivalent of 1X 8 Ohm speaker at 1/2 voltage. 1/2 voltage = -6dB.

Not one 8 ohm speaker at 1/2v, two 4ohm speakers, 1/2v each
The same voltage, same load just divided by two speakers.

2 * 1/2 voltage squared divided by 1/2 impedance = voltage squared / impedance

lets say 9 volt output:

2 * 4,5v^2 / 4r = 9v^2 / 8r
2 * 5,0625w = 10,125w

Contrast with parallel operation.  2X 16 Ohm speakers in parallel also = an 8 Ohm load.  So the same 1W is dissipated. But ea 16 Ohm speaker in parallel gets a full compliment of 2.83V.  Electrically, 2X 16 Ohm speakers in parallel, are the equivalent of 1X 8 Ohm speaker at full voltage.

Did you consider the fact that parallel speakers need to get along well and divide the current?

P = I^2 * R
9v over 8 ohm = 1,125A so 0,5625A in 16ohm each
P = 5,0625w each or 10,125w combined

Do you have a specific idée fixe or maybe experienced something that contradicts this?
You reluctantly seem to be defending this +/-3dB idea, with different arguments every time.
I can only repeat the same formula, if it's wrong I'd be happy to admit.

[sluckey]

Hi, maybe a strange first post, but I can’t resist the urge to respond here.

I agree. You seem to enjoy being argumentative rather than trying to contribute to the forum. Not a warm friendly introduction at all.

[arno]

Hi, maybe a strange first post, but I can’t resist the urge to respond here.

I agree. You seem to enjoy being argumentative rather than trying to contribute to the forum. Not a warm friendly introduction at all.

I understand, I figured a technical discussion is a contribution but I guess we all made our point.
Let's move on and a warm hello from my side nonetheless.

[jjasilli]

I'm not making different arguments.  I'm re-stating the same concepts in different ways.


Your error is in confusing speaker dissipation in Watts with amp output power in Watts:  2.83V X 1/2 = 1.415V;  1.415V² / 4 Ohms  = .5 Watts. That's 50% of 1W, so ea. 4 Ohm speaker should be down only 3dB (not 6dB).  This is wrong, because amp output power in Watts has not changed.  The amp is still putting out the same 1W into 8 Ohms.

What has changed is the voltage seen by ea 4 Ohm speaker.



If we feed 2.83V into 2X 4 Ohm speakers in series, ea. sees 1.415V = 1/2 voltage; Again 1/2 voltage = -6dB. So ea speaker is down 6dB from 100 dB = 94 dB.  But there are 2 speakers.  So in the math world of logarithms: 94 + 94 = 97 dB.  That's down 3dB from 100 dB.

[pdf64]

...amplitude = height of the curve.  For electricity, amplitude = volts.  Think of the height of a sine wave.  The trace of the curve is volts on the y-axis (time or frequency on the x-axis).  The Area Under the Curve = the Power required to raise the curve to the desired height in Volts...

The y axis of a sine wave trace can volts, current, power, some point on the cross-section of a rotating shaft, whatever (time has to be the x axis here for the concept of a sine wave to make sense). see https://en.wikipedia.org/wiki/Sine_wave

...The Area Under the Curve = the Power required to raise the curve to the desired height in Volts...

That's an interesting way of looking at things but I don't think it's correct; consider that voltage can exist without current flow (hence no power used).

...2X 16 Ohm speakers in parallel also = an 8 Ohm load.  So the same 1W is dissipated. But ea 16 Ohm speaker in parallel gets a full compliment of 2.83V.  Electrically, 2X 16 Ohm speakers in parallel, are the equivalent of 1X 8 Ohm speaker at full voltage.  Electrically we have a -0-dB situation.

Acoustically, 2X 16 Ohm speakers are 2 speakers = +3dB. (compared to 1X 8 Ohm speaker)...

Consider that 16 ohm speakers need 4Vac signal for their nominal 1 watt, in order to generate 100dB. With 2.83V they will only be getting 0.5 watt each; see below for the power ratio dB change of that.

... https://www.jpro.co.nz/news-article/wattage-vs-spl-what-should-you-look-at-when-selecting-a-loudspeaker-978 :  :The relationship between voltage input and SPL output is called voltage sensitivity. "...

Due to the significant variation of speaker impedance against frequency, it's not feasible (or useful) to power the speaker with 1 watt over the frequency test bandwidth. Hence the test signal amplitude is set to the voltage required to dissipate 1 W across a resistor of the same resistance as the speaker's nominal impedance, using a very low output impedance amplifier feeding the speaker. Then that voltage is used for the frequency sweep of the speaker. So yes, technically it is voltage sensitivity, because at some frequencies the speaker power will be nowhere near 1 watt.

...If we feed 2.83V into 2X 4 Ohm speakers in series, ea. sees 1.415V = 1/2 voltage; Again 1/2 voltage = -6dB...

As we're dealing with power, half power is -3dB.
See https://en.wikipedia.org/wiki/Decibel and https://en.wikipedia.org/wiki/Field,_power,_and_root-power_quantities

[jjasilli]

Bare voltage is a red herring.  The point in contention is the resulting SPL of the speaker configurations. A speaker cannot be driven without Power.  Power in Watts = Volts X Current.  Or, W = V sq. / R.


The 2 positions can be reconciled as to the net dB of 2X 4 Ohm in series vs. 1X 8 Ohm.  The devil is in the detail of how literal the meaning of 1W @ 1M is.  This is the nominal standard at which SPL is measured.  However in reality 2.83V is the "presumptive" standard, see:  http://www.psbspeakers.com/articles/Guide-to-Speaker-Specifications



"Voltage Sensitivity
Because today’s solid state amplifiers do a good job across the board of maintaining a voltage output of 2.83 volts, many companies consider this as their standard of measurement. Here again, 2.83 volts are inputted and measured at 1 meter. [Note: 2.83 volts into an 8 ohm load is equal to 1 watt. Ohm’s Law: Power (watts) = Voltage (V) x Current (I) or Power = V sq. /R (impedance in Ohms)]


Because a speakers’ efficiency in transforming (transducer) power into sound is greatly determined by the impedance of a speaker, (see more on impedance below) 2.83 volts becomes greater about 1.5 watts at 6 ohms (as many PSB Speakers are rated.) and 2 watts at 4 ohms — a 3dB increase. "


But, if instead we take the 1W @ 1M literally, then for a 4 Ohm speaker 1W = V sq. / 4 Ohms; V = only 2V not 2.83V!  In that case, the 4 Ohm speaker is more sensitive than the 8 Ohm speaker, because it puts out 100 SPL with a mere 2V input; while the 8 Ohm speaker requires 2.83V to output 100 SPL.  (EDIT: i.e., the 4 Ohm speaker is more electrically sensitive.)


This violates the Given Condition of our discussion that the speakers be identical except for their impedance rating.  If they have different sensitivities, then they are not otherwise identical.  This entirely accounts for the differences in our respective calculations of the resulting SPL.

[jjasilli]

Supplementing my Reply immediately above:  yesterday I emailed Jensen on this topic (who differs from PBS Speakers in their SPL measurement):

Email to Jensen
 Would you be kind enough to clarify the SPL ratings for the Jensen P-10R 4, 8 &; 16 Ohms. You state the the SPL for 1W @ 1M. 1W into 8 Ohms is 2.83 signal drive voltage. 1W into 4 Ohms is 2V. Are you using different voltages for ea impedance rating to literally achieve 1W? Or, are you using 2.83V for each impedance rating? Thank-you for your attention to this request!
 /John


Jensen didn't know! So they contacted their factory in Italy, then just gave this response:

John, 
We got lucky and the guys were wrapping up the day in Italy and got us an answer already.  We are using different voltages for each impedance rating.

We use the nominal impedance to calculate the reference voltage: 2.83Vrms for the 8Ohms, 2Vrms for the 4Ohms, 4Vrms for the 16Ohms.   


Thought this might be interesting info.

[amp_troll]

The best is parallel.  I could go into a detailed discourse why, but basically the amp can "SEE" both loads and
"KNOW" what they are.  If you want a detailed answer I will give do so.