The aim of this
study is to modify the Audax
TW034XO-P47N tweeter, making it suitable for
points of crossover in the 1-2 kHz range. Using
the TW034 with the usual point of crossover in
the 2-3 kHz range does not to my experience
justify the trouble of modification.
Pay notice to the possible
need for liniarising impedance when taking point
of crossover this low.
For the modified dome here 1.2 mH + 26.7 uF
(22+4.7) + 8R2 will make a flat impedance.
Whether this is needed or not will depend on the
actual point of crossover and not least the slope
of the high-pass filter.
Using TW034 from 1.5 kHz with an e.g. first order
filter may soon fry your voice coil, so be
Audax TW034XO-P47N comes
with an solid center pole piece. A dome of foamed
rubber is glued to the pole piece providing the
necessary damping of the cavity behind the dome.
Below the voice coil gap is a fairly large
un-damped cavity with only little (negative)
impact on performance.
My initial thought
was that overall performance could be improved by
venting the pole piece and adding a damped rear
chamber as is common practice. This turned out
not to be the case. Venting the center pole piece
is essential, but that's about it. No extra
chamber was needed although everything wasn't
Drilling 4 x 4 mm holes to the cavity behind the
voice coil proved beneficial and in particular
when this cavity was coupled to a small highly
Left: TW034 dome. Right: The damping material on pole
piece feels like foamed rubber.
The finished tweeters. Some black paint might do
well...Next pair will be nicer.
mm hole is drilled in center pole piece and chamfered 45
deg to a depth of 3 mm.
4 x 4 mm holes are made at back plate at a diameter of 38
- next pair -
Magnetic gap is covered with alu tape to reduce amount of
debris getting stuck here. Completely avoiding this is
This alu tape didn't perform well. It took an hour/each
to clean the magnet system.
Next time I'll seal the magnet gap with a ring of
Superfix and wait for the next day. Superfix can always
be removed again.
Preventing iron chips from drilling the rear holes to
ventilate magnet gap is not possible, so some work can be
I usually start out with a 4 mm drill and proceed with 6,
7, 8, 9, 10, 11 and finally 12 mm. Last thing is
chamfering the hole as seen to the right.
Allow the magnet system to cool in between. If iron parts
get too hot, it may loosen the glue holding it all
Cleaning the magnet gap is a drag! To the right the
cleaned magnet system.
Left: Rear view with the 4 x 4 mm holes to the voice coil
gap. Make a circle of Ø 38 mm for guide.
Right: A slightly chamfered 3 mm felt ring is placed on
center pole piece and MDM3 is used to dampen center pole
Routing for the rear panel.
Left: Felt rings in place. Right: The finished magnet
system. Seal routing center hole with a drop of glue.
Left: TW034 as-is (red) vs. modded magnet system. Right:
Performance of two modded tweeters.
Left: CSD (40 dB scaling) of un-modded tweeter. Right:
CSD (40 dB scaling) of modded tweeter.
Left: Impedance of tweeter as-is (yellow) and two modified
domes (red and blue).
Note no bump at 2.2 kHz and Fs has been lowered from 900
Hz to 700 Hz.
Left: TW034-modded step response. Right: Same ETC plot.
Overall a well-damped system.