# speakers in serie and parallel question

#### yeroen

I have a technical question, some advise is welcome
I have a set of 4 x 8 inch speakers with a 4 ohm impedance and 2 8 ohms speakers (same model and brand, only different impedances) . my amps don;t like 2 ohm impedances, So i have to fiddle with the setup to keep the impedance at a value around 4-6 ohms. But I also want to have an equal outpunt of all speakers.

I was thinking of this setup:

|---- + 4 ohms - ------|----- + 4 ohms - ------|
|--------------- + 8 ohms - ---------------------|

+ amp out -

this way my amp sees 4 ohms as total impedance.
but my guess is that the 8 ohms speaker will get all power and output 100%, and the 2 x 4 ohm speaker in series will get 50% of the power, so are outputting half of the power each. (so basically I expect the output to be the same as 2 x 8 inch speakers in stead of 3 speakers).
Is that correct?
If so, that is not the desired situation and I'll have to look further for a solution

The setup is for a PA system, the amp is a 2 x 1000w LEM procon amp. The system should be able to produce high volumes

Last edited:

#### AllenB

Paid Member
The 4 ohm drivers will each use a quarter the power of one on it's own.

In any case, their sensitivity is a factor. The two in series will have the same sensitivity as one on it's own, except for using half the total power to get there.

1 user

#### yeroen

they all have the same sensitivity (95db/1watt/1m)

I expect the following output at 1W output on the amp:

| ---- 0,25W -> 89db ------ 0,25W -> 89db ------|
| ------------------0,5 W -> 92db ------------------|

whereas in full parallel the 4 ohms speaker would get more of the amp output
but the impedance would be way too low

Last edited:

#### AllenB

Paid Member
Ok, so the same 95dB for each combination...
1x 4 ohm,
2x 4 ohm in series,
1x 8 ohm.

#### yeroen

I am building a column speaker, with 3 mid-bass units and 1 compression driver. kinda like the RCF nx24a.
3 units is difficult I see now. even with 3 units of the same impedance. 2 is easy, 4 is easy, but three.....
So adding the extra unit does not give me any gain in volume, only gain is box size and weight ?

Last edited:

#### AllenB

Paid Member
2x 4ohm in series, in parallel with 8 ohm.. gives you 101dB

#### AllenB

Paid Member
(Wait, I see you added to your last post specifying by power. I have to look at that)

#### yeroen

Ehmmm, but the two 4 ohm speakers (A and B) will get less power than the 8 ohm speaker (C). So probably dont output their specified volume in that configuration right? Would that be audible?

In this config the amp will see about 4 ohms at the output:

the 101 dB you mention would be a fine output

#### AllenB

Paid Member
Let's state that the amp is applying 2.83V

The 8 ohm driver gives 95dB

The 4 ohm drivers each get half a Watt, putting the total for the two at 98dB

The total would be almost 103dB

#### yeroen

"The 4 ohm drivers each get half a Watt, putting the total for the two at 98dB"
isn't it so that a double in power (or an extra speaker) gives +3 db, so v.v. a reduction of power in 50% deminishes the output by 3 db -> so 95-3= 92db output at half a watt times two = 96 dB?
I'm a bit puzzeled...

addition: the idea behind the colums speaker is steering the sound in the vertical plane and keeping the dispersion in the horizontal plae as it is. I am afraid that because of the not equal output of the 3 speakers in a row, the sound is steered (vertically) to the upper part if the 8 ohm woofer is placed as lowest speaker in the row, (or lower part if the 8 ohm speaker is placed as top speaker)

#### AllenB

Paid Member
"The 4 ohm drivers each get half a Watt, putting the total for the two at 98dB"
isn't it so that a double in power (or an extra speaker) gives +3 db, so v.v. a reduction of power in 50% deminishes the output by 3 db -> so 95-3= 92db output at half a watt times two = 96 dB?
I'm a bit puzzeled...
Each has half a Watt, 95-3=92. There are two, so power wise 92+3=95, and working together wise 95+3=98

addition: the idea behind the colums speaker is steering the sound in the vertical plane and keeping the dispersion in the horizontal plae as it is. I am afraid that because of the not equal output of the 3 speakers in a row, the sound is steered (vertically) to the upper part if the 8 ohm woofer is placed as lowest speaker in the row, (or lower part if the 8 ohm speaker is placed as top speaker)
This sounds like a job for some drawings and calculations..
..or, a simulation. Do you have one prepared? I can run one if not..

#### yeroen

Would be very much appreceated
What do you need to simulate things?

I was planning on experimenting with active EQ and bi-amping, so no need to design compicated passive filters
The speakers are 18Sound MB400 (mb400)
The compression driver is going to be a faital pro HF108r with faitalpro-sth100 horn
crossing freq somewhere around 1.5kHz
Box size around 80cm x 26cm x 30cm

there are 4 12" subs to take care of the lower part from 100Hz down.
Plan is to create a top end sounding speaker for my band

#### AllenB

Paid Member
How far apart will the 8" drivers be?

Where will you cross them (upper)?

#### DonVK

@yeroen There is a simple way to ball park this, however a simulator will give a more accurate answer because the driver has inductance as well. The amplifier output is voltage, but the driver sensitivity is specified as power 95dB/W (2.8V*2.8V/8R=1W). Lets assume 4R speaker and 8R speaker are the same sensitivity and we'll look at the "change" in sensitivity w.r.t a single driver.

- 2 drivers in series have 2x the resistance (-3dB electrical gain) however you also have 2x cone area Sd (+3dB acoustic gain). Net is -3dB + 3dB =0db change to sensitivity. This is from 10*log(2)=+3dB, and 10*log(0.5)=-3dB

- 2 drivers in parallel have 1/2 the resistance (+3dB electrical, more current) however you also have 2x cone area (+3dB acoustic). Net gain is +3dB +3db = 6dB change in sensitivity.

Your example, the 2 series drivers have net 0dB sensitivity change. Then a 3rd driver (8R) is added in parallel. So electrically +3dB, but you only have 50% more Sd so 10*log(1.5)=+1.7dB. The net change is = +3dB + 1.7dB = +4.7dB relative to the sensitivity of a single driver, with net resistance 4R load the amplifier. This is approximate because it excludes effect from speaker placement and all T/S parameters.

 - I just saw @AllenB is doing a sim, which will be better

#### AllenB

Paid Member
Assuming 225mm spacing with the louder driver on the top..

500Hz

Paid Member

Last edited:

#### yeroen

drivers will be placed close together, like this:

my plan is to align the top horn a bit tilted down, because these speakers will be stacked on top op the subs, or on a pole, so the horn and upper woofer is always at least above crowd heads.
So a vertical directivity with a 'lobe' down would fit as in your sim (thanks for that!)

But if I see you sim, I see dips arising around 400Hz around 60 degr. off axis
Do I read it correctly that vertical dispersion is falling from 400Hz and further off axis?
I was planning on crossing the compression driver at around 1.5 khz

And what about horizontal dispersion? Is there a beaming effect in this placement/coluns design?

#### AllenB

Paid Member
Do I read it correctly that vertical dispersion is falling from 400Hz and further off axis?
Yes, and side lobes appear in the usual way. With real drivers that are not point sources, it may be less distinct.

just the usual for 8" drivers, narrowing towards 2kHz.

my plan is to align the top horn a bit tilted down, because these speakers will be stacked on top
Sometimes pointing it up is better for wavefront integration. There is some uncertainty casually discussing it this way and many keep it straight in this situation.

#### yeroen

At home I build a d'appollito system that sounds really nice. Would i be an idea to use that kind of setup here too?
So W(8) -W(4) - T -W(4)
or W(4) -W(4) - T -W(8)

And may I expect some benefits for low throw performance compared to a point source speaker?

#### AllenB

Paid Member
It's not easy to guess.. but you'll be crunching a different set of compromises. In this case the woofers will be far apart for the crossover frequency covered by a waveguide that is smaller..

1 user