- go to chapter II
- go to
chapter III, coating
- go to
chapter IV, dissasembling of a 9710 driver.
scan of original Philips 9710/M8 brochure here. 370 KB.
Thanks to Darryl for not throwing out this
Philips brochure like I did many years ago.
9710 M/01 alnico version
all impedance of the two drivers: Nice! Both
having a Fs of around 44-45 Hz. The coating of
surround works fine and I hadn't expected to hit
the same Fs for the drivers. Remarkably there's
only a minor increase in impedance towards higher
frequencies. This is usually only seen from
symmetric drive (SD) speakers where the pole
piece is covered by thin copper. As can be seen
from the TS data sheet below, the induction at 10
kHz is low.
Since I wrote the above notes I
found this on a website from Holland:
Philips 9710 with the pole piece covered with copper.
So, indeed Philips used "symmetric drive" long
before Scan Speak.
No wonder the impedance curve is so flat.
Nothing new under the sun - only evolution.
TS-data. Note that calculated sensitivity is 91 dB.
Calculating a box for the 9710 isn't that
easy as math dictates an excessively large cabinet. More
than 100 litres closed for a Qtc of 0.7.
A 40 litres vented cabinet of 40 litres only provides a
F3 of 56 Hz. If we had had the Fs = 35 Hz and Qt = 0.4,
we would get an F3 of 43 Hz from a 40 litres cabinet.
We'll see if the drivers change over time. Usually Fs
decreases and so does the Qt. Philips recommends a 40
litres cabinet and this it's likely to be.
Frequency response at 1 metre distance
from microphone to front panel. Both drivers shown.
Disregard what's below 300 Hz due to gating. This looks
pretty good in terms on drivers' consistency. Same SPL
response from drivers. The peak at 3 kHz is certainly to
be heard when listening to music. On-axis it's severe,
but the strange thing is that it's an ear-friendly kind
of peak. It's excessively loud, but it doesn't shred your
ears the same way the 18W/8535 did in the 2.5 clone at 2
kHz. The FR is clearly consistent with the Philips
measurements seen below.
Left: Philips 9710 anechoic response.
Right: Philips' claimed response in living room
conditions. Now, this all of a sudden almost looks too
good to be true. Taking the speaker into your living room
all has suddenly changed for the better, and +/- 3 dB
from 70-12000 Hz isn't bad at all. If I extend my
measuring window to 320 milliseconds and add some heavy
smoothing, I can simulate living-room conditions and it
looks like this:
It appears that the
Philips claims are not all that wrong and what's
depicted above can take any shape depending on
actual placement in your living room. If I put
the speaker close to the wall, I would probably
get even more response in the bass and lower
midrange area. My workshop has a severe dip at
60-70 Hz, so this will account for lack of
response in this area. However, it doesn't
relieve the 9710 of having a severe peak at 3
Impulse response. Hmm...some peaks and in all directions.
Probably not that unusual from a twin-cone driver.
Step response. Two major peaks; I guess the second one is
from the whizzer cone.
Equalisation of 9710 - LspCAD:
A single coil and a series LCR network to
reduce the 3 kHz bump should make the 9710 perform
like seen above. And so it does as seen below.
This makes the upper treble week and I added two old 4
ohms Peerless cone tweeters connected in series and
driven from a 2nd order network consisting of 1.5 uF +
0.1 mH to ground and ended up with this FR:
Quite some dips and bumps, but it sounds
better than you would think.
The tweeters were simply placed on top of the cabinet,
thus the irregularities in upper treble.
This sounds good on vocals, the tweeter balance is fine
and this will be my first set-up in the test cabinets.
- if only I could find some more of these...
Nov-05: Finally got some of these Peerless
cone tweeters, the CT62H.
February 2006, Chapter II
The Peerless CT62H option
It's very much time to take another look
at the Philips 9710 fullrange drivers. One of the drivers
has been at the corner of my workshop in a 30 litre test
cab. Two of the Peerless cone tweeters were floating on
the top, spewing treble in all directions and making
enjoyable noise when doing other project. A simple 1st
order network was connected to the drivers.
A pair of 37 litres cabs were constructed and I very much
enjoyed making a pair of old-time cabs again - close to
how I did more than thirty years ago - except for the
driver rebating. I didn't have a router then - much too
expensive. And we didn't have MDF either. Chipboard was
the menu of the day.
Left: First time set-up, the 9710 with two
Peerless cone tweeters, CT62H. Right: preliminary
Cabinet, basic dimensions:
Cabinet made from 19 mm MDF.
300 mm (W) x 550 mm (H) x 319 mm (D), net volume ~ 37
Vent = 73 mm (ID) x 110 mm length.
seen the frequency response of the 9710 driver
alone with no crossover attached.
I've recently had a mail exchange with a fellow
running two 9710s in series on an open baffle and
I'm still wondering how he can live with this.
The open baffle may sound very different from
what I hear from my boxed 9710s, but it won't
remove the peaky treble. The frequency response
of the 9710 not only looks awful on paper, it
sounds awful too - to my ears. 10 dB peak at 3.5
kHz! Exactly where the ear is most sensitive! I
honestly don't get it. As said previously, it's
not an ear shredding kind of peak, but this is by
any standard low-fi. I'll call it amplitude
distortion as it surely changes the way
instruments and voices sound.
I did survive this in the late Sixties having the
speakers at the corners and on the floor of my
room, where on-axis listening was impossible
unless I was lying on the floor. And my B&O
amp had tone controls... I don't recall if I used
to lower the treble. However, the 9710 didn't
last long and in came Peerless 3-25 kits, fitted
with 12" bass drivers, 5" midranges and
cone tweeters. Much better. Well, my hifi story
some other time. Had I been able to make a
crossover as seen below, my hifi career might
have looked very much different. But we didn't
have personal computers, CLIO and LspCAD back
response of individual drivers. The Peerless
tweeters are connected in parallel and the FR is
measured right between the tweeters at 1 metre
distance. Placing a pair of tweeter on a
horizontal axis has its problems as we will see
of drivers. Note the remarkably flat impedance of
the two Peerless tweeters in parallel. Blue and
yellow is phase. Vent tuning is 43-44 Hz.
the final system impedance (crossover V4) from
LspCAD. 6.3 ohms minimum where it hurts and a
minimum of 4.5 ohms at 3 kHz. This is a dream for
every low wattage valve amp.
version 4. Producing a linear response from the
9710 proved much more difficult than anticipated
and quite some tweaking had to be tried before
rendering what is seen above.
Every large 8" driver - and the 9710 is
a large driver - will start to sound honky when
it has to handle the upper midrange/lower treble.
BUT, honkyness is not only a matter of
dispersion, but also on level and not having an
elvated level at 750-1500 Hz is of paramount
importance. My V2 crossover ran for some time and
I thought there wasn't much to do about it, but
bypassing the first 1 mH coil with a 47 ohms
resistor and bypassing the second coil with (2R7
+ 1.5 uF) helped smooth this troublesome area and
surprisingly the forwardness and slight honkyness
was much reduced. Maybe we really can have an
8" running everything up 2.5 kHz... Well,
it's never going to be like what a 4" can
do, but possibly something we can live with.
The tweeters took a little equalisation too,
hence the 0.1 mH paralleled by 2R7 at the entry.
Phase tracking may not be the best I seen, but I
value amplitude higher than phase tracking here.
The 9710 is lively cousin, and if you
don't treat it right it'll shred your ears.
did the modelling fit with reality? Very good
indeed as depicted above. Now the overall
response is basically flat and vocals have
improved compared to previous crossover versions.
Red = FR, blue = minimum phase.
A guy visiting me brought along an MingDa MC34-A
integrated amp (EL34 pp) and we ran the 9710s
from this amp and don't say hifi can't be
simplistic: An amplifier like this and a decent
pair of 92 dB/2.8V speakers. Try Chris Minh Doky:
"Listen Up", the intro track, and
you'll jump in the seat when the drummer hits the
snare drum hard. And try an old Garrard 401
turntable + the Denon 103 cartridge and you have
a system that we could have listened to
some 40 years ago. Amazing.
System sensitivity is 92 dB/2.8V based on average
dispersion. Very good indeed.
Red is at 1 metre distance at a height between
9710 and tweeters. The other graphs at +/- 10 and
20 cm from this position corresponding to +/- 30
and 60 cm at 3 metres listening distance.
dispersion. Very bad indeed.
No surprise, placing two tweeter on a horizontal
axis will produce this picture. Red = on axis,
yellow = 10°, purple = 20° and green = 30°.
But it's not as bad as you may think. This is
what a single, tiny microphone capsule picks up
at a specific position, not similar to what our
two ears perceive at different locations. But
obviously it's not the perfect solution .
Fountek NeoCD3 option
Fountek NeoCD3 ribbon tweeter
Let me say that the 9710s are great fun!
It may not be the end of your search for the
sound, but they may provide some very enjoyable hours in
front of your "stereo".
These speakers are discriminating. They will leave some
of your CDs and vinyls on the shelf, but from not too
poor recordings they can provide a wide soundstage, a
vastly enhanced transparency compared to running the
drivers fullrange and an overall balanced sound that make
the saxophone of Paul Desmond sound even better than
ever. They provide a 92 dB/2.8V sensitivity and an easy
load on your amplifier. Low-wattage valve amps will love
the 6-7 ohms impedance.
I have a range of 9710 drivers, three with the common
whizzer cones and two without. The whizzer cone only has
some impact above 8 kHz and removing it doesn't change
the huge bump at 4-8 kHz, much to my surprise. The three
whizzer cone drivers leave one for experimentation and
the 9710 is easy to take apart. Three screws holds the
basket in place and by removing these you can lift off
the basket with cone and voice coil.
Check images here and preliminary
Crossover for 9710 and
Fountek NeoCD3 ribbon tweeter.
This is a pretty straight forward 4th order
filter and if you think 4th order filters
destroys transient attack, try this!
And please don't confuse severe peaks in
frequency response with good transient
performance. This is very often the case.
response of system (red) and individual response from
dispersion from 1 metre distance.
Green = between treble and bass.
Red = + 20 cm, blue = +10 cm, yellow = -10 cm, purple =
Removing whizzer cone and coating of 9710
Removing the whizzer cone
has the effect seen above. Red = un-modified and
blue = whizzer cone removed and rubber
"phase plug" inserted. The whizzer cone
is mostly active above 6-7 kHz and the response
is reasonably linear up to 16 kHz. Not bad for an
8" driver. It also means the bump above
2-2.5 kHz is smoother and the response declines
nicely above 7 kHz.
Coating of the 9710 cone
is a serious matter. What type of coating should
be used? If it doesn't work there's no way back.
I don't have a whole range of coating materials -
and not a whole range of 9710 drivers as well -
so to cut a long story short, I chose to use the
DAMAR coating. The 9710 cone is rather soft and
the DAMAR will most likely make the cone more
rigid. This may have several consequences: One is
that the cone will operate more like a piston and
the other is that rigid cones tend to break up -
somewhere. Hopefully not in the area where it's
supposed to operate. Thirdly the coating material
will add weight to the cone and reduce
Actually the DAMAR coating worked rather well on
this cone, where I in other cases have seen
serious cone break-ups around 2.5-4.5 kHz, just
where the point of crossover is supposed to be.
Seen below is the response from the 9710 before
(red) and 7 days after one DAMAR application
Red = un-coated 9710
without whizzer cone. The frequency response is
here merged with the nearfield response at 350
What can be seen is an approx. 1-1.5 dB loss in
sensitivity due to the coating but also that the
steep rise in response above 2 kHz has moved some
4-500 Hz up. Good.
Green = un-treated cone.
Red is 5 min. after application of DAMAR. Some
serious loss in sensitivity, but most of the
"DAMAR" is solvent that will evaporate
Green is after 30 minutes. The response after 7
days is seen above.
Without doing further
modifications to the crossover, here's the
response of the 9710 + NeoCD3 before and after
coating of the 9710.
Red = after coating. Blue = without coating +
1. removing the whizzer cone removes the response
above 7 kHz, but leaves the area where I want to
use the driver mostly un-affected.
2. Applying DAMAR coating reduces the sensitivity
by some 1-1½ dB and moves the bump above 2 kHz
up some 4-500 Hz.
So what about the sound?
The sound changes very much.
Getting rid of the whizzer cone immediately makes
a more clean sound. The 9710 used as a
"common" midbass driver really do not
need a whizzer cone but I have a feeling that
some will miss the more colourful sound of the
Adding a single thin layer of DAMAR coating makes
yet another powerful impact on sound. Now this
set-up really starts sounding like hifi - so take
great care! - but it doesn't relieve the 9710 of
having a sound much dictated by the very deep and
The DAMAR coating greatly enhances the level of
transparency and it is obvious that the very soft
paper cones smear a lot of detail.
The 9710 still has "something" that the
Monacor SP200X is short of. The sense of speed
and sparkle. Monacor.
The transient attack of the 9710 is very hard to
find elsewhere, but it comes at a price.
It has been great fun reviewing
the 9710s here. I've enjoyed it a lot getting my
old Sixties' drivers back and finding out what it
was I was listening to back then.
The 9710s are not the end of story and I'm sure
we can find better drivers today built from the
same concept of a very lightweight 8" paper
cone and maybe I should take a look at those Supravox drivers
and find out what they can do.
- return to top.