100 Ohm loudspeaker

[graaf]

Hi to all!

let me quote from "Tube CAD Journal":

The big problem tube OTL amplifier face is having to deliver the high current needed to drive the low impedance loudspeaker. One watt into 8 ohms requires a peak current delivery of 500 mA, as (0.5² x 8) / 2 = 1. Which means that a total of four EL509s is needed to meat the 250 mA idle current demand of the push-pull amplifier. If the load impedance were 285 ohms, then the same 500 mA peak current swing would yield 35 Class-A watts. What the world needs is a great 100 ohm loudspeaker!

so the world needs a great 100 ohm loudspeaker

or even better a 300 ohm loudspeaker

how can it be done?
what about a line array of multiple drivers wired in series?
a line array of 24 8-ohm drivers wired in series is 192 ohm
is it feasible? any obstacles, difficulties?

is anything fundamentally wrong in wiring woofers in series?

dr Griffin (of " www.audiodiycentral.com/resource/pdf/nflawp.pdf " fame) recommends wiring woofers in line arrays in series/parallel connections where for instance four woofers are wired in series giving a 32 ohm total impedance
theory and measurements also show that there is nothing wrong in series wiring of woofers because "the speaker never knows that there is another speaker wired in series with it": http://www.monstercable.com/mpc/stab...ng_Woofers.pdf

so why not?
any ideas?

best,
graaf

[18thell]

In older design, there was an amp. using 8X 6080 in parallel, and requires a 16 ohm speaker, the circuit employed OTL design. This told us if more tubes in parallel, the output impedance could be lower, that's why we use output transformer for impedance matching. Usually the primary impedance for tube is from 2K to !0K ohms, depends on which tube to be used. The sound quality depends on the output tranny quality. A pair of high quality output tranny is very expensive, but we still have to use it.

If the in / out impedance doesn't match ideal, lost of power or low efficiency or distortion would occur. Careful design the matching and select good quality tranny is a must for best audio reproduction.

[Sch3mat1c]

625 x 6V6 in parallel would work quite excellently.

The classic EL36 circuit (Philips IIRC) used two in SRPP OTL into a custom 600 ohm speaker. Thanks to the EL36's high perveance, this is quite acceptable.

Tim

[mashaffer]

So it sounds like an amp for an infinite baffle subwoofer would be the ideal application of OTL. Two manifolds with four 16ohm speakers in each.

But seriously I suppose a high impedance speaker with normal coil system would have trouble with high Frequencies due to coil inductance but some sort of electrostatic or piezo setup should work.

mike

[Sch3mat1c]

Just as much inductance as any other, as a matter of fact. It scales with voltage and current.

Besides, inductance has little place in a loudspeaker because velocity is proportional to EMF.

Tim

[Dave Cigna]

Originally posted by 18thell
In older design, there was an amp. using 8X 6080 in parallel, and requires a 16 ohm speaker, the circuit employed OTL design. This told us if more tubes in parallel, the output impedance could be lower, that's why we use output transformer for impedance matching.

I think the point that Broskie was making is that there's more to it than output impedance; you need the capability to deliver current as well.

One section of 6080 used as a cathode follower has an output impedance of about 150 ohms, so 16 of them in parallel would be roughly 10 ohms. You might get 50mA of drive current from each one swinging 40 Vrms, so 800mA total.

We can use Ohm's Law to figure out how much current we would need to swing 40 Vrms across 16 ohms:

i = v/R = 40/16 = 2.5 amps!

That's three times what all those parallel 6080 cathode followers can deliver despite the fact that the output impedance would seem to be low enough.

We can work the problem from the other direction. Suppose we expect our 8x6080 amp to deliver 20W into a 16 ohm load.

Power P=i^2 x R, or i = sqrt(P/R) = sqrt(20/16) = 1.1 amp. Not gonna do it. The best we might do (based on my assumptions) is

P = i^2 x R = 0.8^2 x 16 = 10 watts.

If that seems disappointing it's all because each 6080 just can't deliver as much current as the low output impedance seems to suggest.

-- Dave

[Nikolas Ojala]

Originally posted by graaf

or even better a 300 ohm loudspeaker

how can it be done?

Just like voice coils are made: Plenty of carefully wound thin copper wire.

I would choose that approach, because if you connect many woofers in series, you get many separate acoustic radiators. It is not a problem in very low frequencies, but if you want a full range speaker or even bass/midrange, you will get a comb filter effect.

[Tube_Dude]

Originally posted by Sch3mat1c
The classic EL36 circuit (Philips IIRC) used two in SRPP OTL into a custom 600 ohm speaker.

Hi Tim

Actually Philips at the time used 800 Ohms and 400 Ohms.

The Philips speakers with 800 Ohm voice coil can be recognized by an A in the type number, e.g., 9710A, and 400 Ohm speakers have a B .

http://people.cs.uu.nl/gerard/RadioC.../SerBalOut.htm

Regards

[graaf]

hm
I see

and what about a line array of multiple drivers wired in series?
a line array of 24 8-ohm drivers wired in series is 192 ohm
is it feasible? any obstacles, difficulties?

best,
graaf

[AndrewT]

don't electrocute yourself on the speaker terminals.