[Dick Pierce[_2_]]

Peter Larsen wrote:
Hi guys,

see http://muyiovatki.dk/yorkville/

comments appreciated, is the 4 Ohm box broken?

Looking at one of many assertions made on the web page:

"the 4 Ohm box is less efficient, a LOT less efficient
than the 8 Ohm box. It should be 3 dB louder, not some
6 dB less loud, but that is what it is."

Having actually designed a number of drivers in 8-ohm,
4-ohm (and, for that matter, 13 ohm for one client),
this assertion makes some assumptions that simply do not
hold. For example, it assumes that the elctro-mechanical
transduction factor remains constant, it assumes mass
remains constant. Neither of these are true in designing
the same "equivalent" driver in different impedances, and
both can substantially affect voltage sensitivity.

For example, suppose you have an 8 ohm woofer and you want
to make it 4 ohms. Well, the simplest way to do it is to
simply have half the amount of wire on the voice coil.
While it will reduce the resistance by a factor of 2,
it also reduces the Bl product by factor of 2 as well.

Let's take a practical example. Suppose we start with a
17 cm woofer with a 3.8 cm, 4-layer voice coil, with a
winding length of 1.9 cm, a gap length of 0.8 cm, a
moving mass of 20 g and a compliance of 1 mm/d. We end up
with a woofer with a DC resistance of 6.8 ohm, s Bl
product of 10.2 N/A, and a reference efficiency of 0.43%.

Now, all we do is drop this to a two-layer coil. The DC
resistance drops, not surprisingly, to about 3.4 ohms,
and the BL product drops to 5.1 N/A. At the same time, we
lose about 5 grams off the voice coil, dropping the total
moving mass to 15 g. The new 4-ohm driver now has a
reference efficiency of of 0.35%, alone a drop of about
2 dB.

Now, in fact, that's not the way most designers change
efficiency: the do it by playing with a combination of
the wire gauge, number of layers, winding length and
more.

Often, to accomodate a different winding, the magnetic gap
of a 4-ohm driver has to be a bit wider. That increases the
losses in the magnet, reducing its efficiency and reducing
the flux desnity (B) in the gap.

Another issue is that the two most difficult parameters
for a manufacturer to control is sompliance and Bl product.
The latter is irrelevant here because compliance has no
appreciable effect on efficiency in the passband. But the
efficiency goes as the square of the Bl product. While the
l (length of wire in the principle portion of the magnetic
field) can be tightly controlled on a routine basis, keeping
driver-to-driver variations of the field density B is not
in the slightest bit easy. +-10% variation in high-quality
woofers is common, and that alone can account for a +-1dB
range in efficiency.

I'd be quite surprised if these two system had even
approximately the same efficiency: if they did, I'd
bet the the 8 ohm one was the broken one.

Another comment: the frequency response curves seen on the
site are , well, pretty awful, and it's mostly because of
poor measurement technique, not because of intrinsically
bad response. The curves show tremendous amount of interference
artifacts and more, and it's difficult to derive any
meaningful results from this data.

This is not a flat out criticism of whoever made the measurements
per se, but atests more to the fact that measuring speakers
properly is DAMNED hard to do. Measuring them well repeatedly
across several samples is even harder.

One and/or the other pair may well be broken. The difference
in efficiency itself is no indication of some problem. But
I'd be hard pressed to declare which was broken, if any, from
these measurements.

--
+--------------------------------+
+ Dick Pierce |
+ Professional Audio Development |
+--------------------------------+