Wessel Dirksen wrote:

Patrick Turner wrote:
Wessel Dirksen wrote:

"Patrick Turner" wrote in message
...


Wessel Dirksen wrote:

Hi West,

The whole Infinity RS Series, particularly the RS-III's and II's
were
notorius for being hard to drive. The reason for this is that
the EMIT
ribbon tweeters of the time (which had to be used there because
they were
Infinity's pride and joy back then) were quite capacitive as all
ribbons
tend to be.

. . . down low.

Well why?

I have no idea why actually. Older ribbons and also to an extent
electrostatics tend to have this characteristic below the pass band. It
certainly makes sense from an inheirant 1st order HP filter standpoint,
but I just see it. I prefer good old electromagnetic drivers myself.

See the ribbon speakers made in Oz at

http://www.ambiencespeakers.com.au/

These all have bass speakers attached, hence
their Z character becomes like a capacitor at some low F
The interaction of speaker to the box causes the Z to change
to largle reactive.
If a speaker Z rises and is reactive it does not matter if the sensitivity
is maintained,
since the amp will always cope.
But where a speaker's reactance, L or C reduces the Z to a low value,
then the amp may have trouble maintaining the output voltage.

Electrostats are large capacitors, but with some series R and
parallel R across them.




If you have only an 8 ohm outlet, and the speakers dip to
2.5 ohms or some low Z, then a separate 8 ohm to 3 ohm
tranny will change the Z that the amp sees.

The tranny can be used just for the midrange/tweeters if that is the
low Z.
Depends on exactly what impedance you have to drive.
But low Z trannies can be wound to overcome the mismatch if there is
one.
Such trannies are not available off the shelf; its a special item,
but its all thick wire, and easy to wind.

Well I was reluctantly thinking you meant this Patrick. I can't comment
from experience but a tranny in the loudspeaker after amplification
just feels bad to me but that could be superstition on my part.

All ESL have step up transformers, Martin Logan, Quad, etc.
These are regarded by many as extremely good speakers.
The extra distortions added by the transformers don't seem to wreck the
music.

Now
that I am just beginning to understand tube amplification I can
understand why you mentioned this as an option. You may be right, but
it just seems weird to me. 100V tranny systems never sound good but of
course aren't meant to sound good.

Most 100v systems are used in PA work.

A purpose wound 8 ohm to 4 ohm tranny will have a
turn ratio of 1.41 : 1.0, and if wound to
to saturate at full powwer at less than 14 Hz and wound with
a many interleavings, the bw can be 400 kHz.
if the core is GOSS, the thd will be quite negligible.



But if you have decent impedance matching, a tube amp will cope.
It just needs to have lots of tubes to get the power if the
sensitivity low 80s,

if one likes loud levels.
I also like to impedance equalise all speakers.
The high impedance peaks of the woofers isn't a problem though.

This is a good point. Just a matter of curiosity. This happens more
often that non-compensated capacitive loads occur in the x-over region
of production loudspeakers.

Midranges appear as a capacitor load at LF.
Its not a worry.

ESL are mainly capacitive for their whole range.


It seems that "good" amps tend to sound
better with these types of loads.

A well made amp isn't upset by capacitor loads.

So how much is a negative phase
impedance load less of a crucial factor when the Z magnitude is higher?
(as it often is)

If Z goes high, the current is low, and the amp copes fine.
A 10 uF cap at 100 Hz is a 159 ohm reactive load.
Conecting it across the output terminals with a 100 Hz signal
cannot do any damage, and a low current flows
even at clipping.
But at 10 kHz, the same 10 uF becomes 1.59 ohms, and
most amps will poop themselves if asked to
maintain the mid band power levels.
50 watts into 8 ohms is 20 vrms, and I = 2.5 amps rms.
4 vrms vrms into 1.59 ohms = 2.5 amps rms.
20 vrms would mean I = 12.5 amps, and
this current flows in the output devices will cause gross
overheating if maintained for some time.

But levels of HF in music decline rapidly above 5 kHz, so
strapping a 10 uF across the output of a 50 watt
amp playing music at an average 1 watt level
won't usually have much effect on the sound if the amp has
a low Rout.
If you don't think so, try it with an old amp.
Just don't do it for longer than 10 seconds to see that not much happens.

Patrick Turner.