As promised, here is the result of sine wave testing
of the ER Audio ESL-IIIB to examine the limits for applied bass signals.

The conditions of the test were as follows:-

Amplifier, 300W SS, thd 0.005% at 40vrms output,

Oscillator, wien bridge, thd 0.05% between 10Hz and 10kHz.

ESL speaker set up with 27 ohms plus 5mH in series network
connected across the input Step Up Tranny primary input coils,
one end grounded to 0V from amp, live end with 150uF series cap,
and with 1.5 ohms series R.

The 150uF and 32mH of shunt input inductance and 27 ohms form a damped
second order input filter to obstruct very low F from the speakers.

The following voltages were able to be applied with
less than 2% or audible distortions.

An oscilliscope was connected to a microphone signal
which confirmed the gross distortions which occured above the applied
voltages
listed below 220Hz, and also confirmed the clean undistorted signals
above 220Hz.

220Hz and above, 20Vrms OK.
140Hz, 4Vrms, NOT OK,
100Hz, 3Vrms,
85Hz, 1.7vrms,
70Hz, 3Vrms,
50Hz, 2Vrms,
44Hz, 0.7Vrms,
35Hz and below, less than 1Vrms.

The distortions heard coming from the speakers
were either sharp increases in distortions or what sounded like buzzes,
or resonance provoked artifacts.
waveforms from the mic confirmed what i was hearing.

The voltage figures would be a lot worse if there was no LF reducing
input filter.

The performance at bass F compared to my dynamic speakers is truly
100% dismal and unnacceptable, and these speakers just don't like bass
signals at all.

EHT was set at a mild -2.7kV level.

The poor sound heard during last saturday's testing with music is due
therefore
to these panels just being completely unable to handle any bass signals
adequately
below 200Hz.

There is a chance the membranes are being pulled off centre but without
being
pulled right over against a stator so that might explain the buzzes.

But to the best of my ability, the panels have been constructed as per
the dimensions
set by the materials in the kits and its not possible that the distances
either side of the membrane
to stators is grossly different when the membrane is uncharged with EHT
supply turned off.

It is very hard to tell if any parts or whole of the membrane is being
pulled off centre just by gazing at the darn things.

So now an effort will be made to configure the panels for a second order
cut off
at 250Hz, and my bass bins connected to replace the bass produced by the
ESL,
and they should perform well, but with no change to low sensitivity.
so a suitable drive to the bass speaker will have to be devised
if I can find a suitable LF transformer, or use bi-amping
with appropriate filters.


Patrick Turner.

The tests performed below in my last post for
distortion thresholds for ESL at bass frequencies were performed
after the speakers EHT had been turned on for 1/2 an hour, and I
concluded the speakers were
giving very poor bass performance.

But when re-tested after 2 hours, the thresholds for
audible distortions raised.
Results obtained after greater time to allow EHT charge up
are as follows:-

220Hz and above, 20Vrms OK, even 28Vrms OK.
140Hz, 28Vrms, OK,
100Hz, 28Vrms,
85Hz, 27Vrms,
70Hz, 14Vrms,
50Hz, 10Vrms,
44Hz, 1.0Vrms,
40Hz, 1.5vrms,
35Hz and below, less than 1Vrms.

I will leave the speakers turnd on all night and see what happens.

The above show a steep cut off for the the allowable bass signal
amplitude below 80Hz.

I will see if this improves further with EHT supply left turned on.

The input to the step up tranny is subject to the second order filter
which gives a -3dB pole in
transformer voltages at 50Hz, and so F lower than 50Hz will not tend to
saturate the tranny core.

The SUT has 55 turns on the P winding and 5,000 on the sec, on a core
with 40mm tongue x 50 stack.
With 28Vrms applied, to cause a B max of 1.5 tesla where core saturation
has begun,
the frequency where this occurs is at (28 x 22.6 x 10,000) / ( 40 x 50 x
55 x 1.5 ) = 38 Hz,
which means that indeed pink noise or any other music test signal should
be
filtered to prevent high level bass giving problems, hence my use of the
150uF ahead of the input tranny.

The sensitivity has not changed for the whole band,
and has remained the same as measured in other posts 29May07,
ie, it takes 4V fed into the ERA ESLs to produce the same SPL as 1V fed
into
my dynamics.

Patrick Turner.






Patrick Turner wrote:

As promised, here is the result of sine wave testing
of the ER Audio ESL-IIIB to examine the limits for applied bass signals.

The conditions of the test were as follows:-

Amplifier, 300W SS, thd 0.005% at 40vrms output,

Oscillator, wien bridge, thd 0.05% between 10Hz and 10kHz.

ESL speaker set up with 27 ohms plus 5mH in series network
connected across the input Step Up Tranny primary input coils,
one end grounded to 0V from amp, live end with 150uF series cap,
and with 1.5 ohms series R.

The 150uF and 32mH of shunt input inductance and 27 ohms form a damped
second order input filter to obstruct very low F from the speakers.

The following voltages were able to be applied with
less than 2% or audible distortions.

An oscilliscope was connected to a microphone signal
which confirmed the gross distortions which occured above the applied
voltages
listed below 220Hz, and also confirmed the clean undistorted signals
above 220Hz.

220Hz and above, 20Vrms OK.
140Hz, 4Vrms, NOT OK,
100Hz, 3Vrms,
85Hz, 1.7vrms,
70Hz, 3Vrms,
50Hz, 2Vrms,
44Hz, 0.7Vrms,
35Hz and below, less than 1Vrms.

The distortions heard coming from the speakers
were either sharp increases in distortions or what sounded like buzzes,
or resonance provoked artifacts.
waveforms from the mic confirmed what i was hearing.

The voltage figures would be a lot worse if there was no LF reducing
input filter.

The performance at bass F compared to my dynamic speakers is truly
100% dismal and unnacceptable, and these speakers just don't like bass
signals at all.

EHT was set at a mild -2.7kV level.

The poor sound heard during last saturday's testing with music is due
therefore
to these panels just being completely unable to handle any bass signals
adequately
below 200Hz.

There is a chance the membranes are being pulled off centre but without
being
pulled right over against a stator so that might explain the buzzes.

But to the best of my ability, the panels have been constructed as per
the dimensions
set by the materials in the kits and its not possible that the distances
either side of the membrane
to stators is grossly different when the membrane is uncharged with EHT
supply turned off.

It is very hard to tell if any parts or whole of the membrane is being
pulled off centre just by gazing at the darn things.

So now an effort will be made to configure the panels for a second order
cut off
at 250Hz, and my bass bins connected to replace the bass produced by the
ESL,
and they should perform well, but with no change to low sensitivity.
so a suitable drive to the bass speaker will have to be devised
if I can find a suitable LF transformer, or use bi-amping
with appropriate filters.

Patrick Turner.

Reading Patrick Turner in
. au:

There is a chance the membranes are being pulled off centre but without
being
pulled right over against a stator so that might explain the buzzes.

Aha!

There was talk earlier of the sensibility, even necessity, of coating the
membrane on both sides. IIRC you dismissed the idea, at least to having any
effect, and further wrote of practical difficulties in this particular case.

Yet it occurs to me thinking of it that it must make a difference.

In the double coated case each side of the membrane each side sees an ideal
equal distance to each stator, with a dielectric of air.

In your single side coated case, one side sees its distance to stator with air
dielectric, and the other side sees that distance + the membrane thckness,
with a mixed dielectric of air and the membrane. The mylar will increase the
capacitance, and the distance increase decrease it; by less, I imagine ...

Then there's the slightly unequal by mylar thickness electrostatic forces.

Did the stiction always occur just on one side? Which? Coated or uncoated?

I dreamed of maybe one day building an ESL pair some while back, and trawled
around the DIY circuits and sites of the time. Now I seem to have accumulated
large ex elegant office tall wide wooden doors for large smooth work surfaces,
and a couple of roll ends of what looks like possibly useful diaphragm film,
and an old Tanner table saw for frames (and dynamic speaker boxes), so perhaps
I will sometime still do so; everything in me (and perhaps being a Libran
influences this wants to see the film coated on both sides, for symmetry.

How much these potential distance effects may matter, one should at least, if
unable to test, fully consider, do you not think? Theorise, discuss, etc?

Then again, it occurs to me - in a sense, with two coatings, effective, one
each side, not only do we get symmetry, but maybe double the charge capacity?

Plus the very sandwich itself has two equal sign charges either side of a
dieletric, the whole central. I intuit that this is far better than one-sided.

I am sure others here may have more understandings and competence in these
fields ahem than I evidence in these rambling thoughts ... let's hope so!

Regards,

Ross Matheson
Auckland, NZ.

On May 30, 10:11 am, RdM wrote:

Yet it occurs to me thinking of it that it must make a difference.

In your single side coated case, one side sees its distance to stator with air
dielectric, and the other side sees that distance + the membrane thckness,
with a mixed dielectric of air and the membrane. The mylar will increase the
capacitance, and the distance increase decrease it; by less, I imagine ...

Then there's the slightly unequal by mylar thickness electrostatic forces.
Theorise, discuss, etc?

Then again, it occurs to me - in a sense, with two coatings, effective, one
each side, not only do we get symmetry, but maybe double the charge capacity?

Plus the very sandwich itself has two equal sign charges either side of a
dieletric, the whole central. I intuit that this is far better than one-sided.

All of the above (in my opinion) pales against the most basic reason I
can see for coating both sides of the mylar. Allow me to generalize
from some specific cases:

a) All good quality plywood is an odd number of layers _usually_ in
symmetrical pairs about a slightly thicker central layer. By
symmetrical, I mean in grain direction, thickness and general
composition.

b) High-quality melamine-laminated furniture - Formica - (almost a
contradiction-in-terms) has a "paper backing" on the blind-side to
prevent warping of the substrate.

c) Bi-metalic coils warp based on the different coefficients-of-
expansion of each metal...

It would seem from these general examples of materials-behavior that
unless the coating and the mylar are absolutely and exactly equally as
responsive to heat, humidity, static, skin-effects and other
environmental phenomena that the combined material would have some
unpredictable response as a two-layer material that could be obviated
by making it a three-layer material. Even if the conductive admixture
were not included in the 'outside' coating.

I am speculating that the increased effect of the charge on two vs.
one layer would make the system more responsive if the additional
attraction did not cause contact.

It would seem that the conditions in Australia range from severely dry
to near rain-forest, and in NZ from temperate to near-rain-forest. I
can see these factors being at issue as well. Would a double coating
allow the mylar diaphram to be positioned further from the stator to
get equal/better results? Would this be useful under more-than-normal
humidity?

Interesting point. Thanks for bringing it up. My brother keeps a pair
of Maggies (MG1) that I snagged for him some years ago... they were
far too large for my use then and now, but he keeps them under near-
ideal conditions by way of humidity and temperature. They are lovely
speakers which he treasures.

Maybe some day.

Peter Wieck
Wyncote, PA

RdM wrote:

Reading Patrick Turner in
. au:

There is a chance the membranes are being pulled off centre but without
being
pulled right over against a stator so that might explain the buzzes.

Aha!

There was talk earlier of the sensibility, even necessity, of coating the
membrane on both sides. IIRC you dismissed the idea, at least to having any
effect, and further wrote of practical difficulties in this particular case.

Indeed there was a discussion, and after having talked to a couple more
guys, one
of whom has been making ESL for 20 years as a hobby, what I said was OK,
ie, there
doesn't need to be a coating to both sides of the membrane.

This man has panels with his own coating which are much more sensitive
and drivavble than the ERA
panels I am struggling with, and he's able to apply 6kV without stiction
of the membrane,
and the dspacer distance is the same at 2.5mm.

Yet it occurs to me thinking of it that it must make a difference.


In the double coated case each side of the membrane each side sees an ideal
equal distance to each stator, with a dielectric of air.

The thickness of the membrane = 0.0035mm, which is 1/685 of the spacer
distance.
On one side of membrane, the frames have a copper foil strip about 0.2mm
thick
thus unbalancing spacer distance far more than the thickness of the
membrane.

There is more occuring that one would expect.
Y


In your single side coated case, one side sees its distance to stator with air
dielectric, and the other side sees that distance + the membrane thckness,
with a mixed dielectric of air and the membrane. The mylar will increase the
capacitance, and the distance increase decrease it; by less, I imagine ...

In Quad ESL57, the stator is a 2mm thick plate of PVC, painted with
highly conductive paint
to make the actual stator.
The PVC has dielectric properties very different to air, and may assist
the operation.


Then there's the slightly unequal by mylar thickness electrostatic forces.

The mylar thickness is entirely irrelevant; slight irregularities in
physical tolerances elsewhere
swamp any effect the mylar thickness will have.
Other reasons cause the membrane to suck over to one of the stators.

Did the stiction always occur just on one side? Which? Coated or uncoated?

The uncoated side always tends to stick. at over 3,000V EHT, one can
push the membrane over against
a stator with a coton bud and watch the membrane tending to stick, then
releasing because there is just enough tension.
At 3,500V, it just swings over on its own. First time I set up a panel,
both bass membranes just went over against a stator, and arcs appeared
all over to the stators,
through the membrane, and through the lousy powder coating on the
stators.
When dismantled later, little black balls of **** were found where arcs
had repeatedly occurred.

I doubled the tension specified in the kit instructions but all that's
done is slightly
raise threshold for the EHT level needed before a membrane heads over
against a stator.

I dreamed of maybe one day building an ESL pair some while back, and trawled
around the DIY circuits and sites of the time. Now I seem to have accumulated
large ex elegant office tall wide wooden doors for large smooth work surfaces,
and a couple of roll ends of what looks like possibly useful diaphragm film,
and an old Tanner table saw for frames (and dynamic speaker boxes), so perhaps
I will sometime still do so; everything in me (and perhaps being a Libran
influences this wants to see the film coated on both sides, for symmetry.

The double sided coating may work better because of weird electrostaic
phenomona,
but not because of the thickness of the material.

Coating both sides means you have to be able to apply the EHT from a
surrounding strip of copper
around each frame, and so perhaps the spacers which will be only 2.5mm
thick for bass
have to be made into a frame or ladder lattice that will support the
tensioned membrane
when its glued on, and then coatings can be applied.
But you still have to coat the "underside' of the membrane where is
glues on to the spacer support,
and provide some means of getting EHT to flow in. This is the hard bit.
ERA have simplified the construction to allow untrained ppl to build the
panels.
But it appears to me that every second person who buys a kit either
gives up, is bitterly dissapointed,
or who abandons construction because a friend has so many problems.

In short, ER Audio ESL kits are for masochists.

I cannot endorse their product at all.

I'd have prefered the kit had gone together without hitches, and in the
20 hours they say is doable,
and be able to recommend ppl to buying kits and paying me to put them
together.
But no, this isn't possible, the kits do give problems, and take far too
long.

Rob Mackinlay of ERA is quite aware of the bad press he's had here.
He's made ZERO attempt to turn up here to answer criticisms.
His mate in the UK, Colin Topps, the ERA sales rep did reply in private
emails to me
when I first began to post publicly about the problems I had. He tried
to insist
I couldn't have any problems, and eventually told me to sod off.

Hi Colin! I know your'e reading this! nice weather in UK?
Just because you told me to sod off and never ever to email
you again didn't mean I would shut the **** up mate!

Now back to the discussion.....

How much these potential distance effects may matter, one should at least, if
unable to test, fully consider, do you not think? Theorise, discuss, etc?

All I know is that Quad can make far better performing reliable panels
than ER Audio.

Unless any maker of ESL kits displays his complete understanding of the
product on his website, has all the test data and has done all the
tests, and admits the problems,
then don't buy from them.

I think that coating to one side only leads to an imbalance of charge to
one side, hence stiction,
and it has nothing to do with even distances.


Then again, it occurs to me - in a sense, with two coatings, effective, one
each side, not only do we get symmetry, but maybe double the charge capacity?

Would not the charge on one side repel the charge in the other?

The coating needs to be very high resistance, and various coatings and
mylar
work in different ways together so a -ve charge to one side may be best,
Quad have both sides charged,
and +ve.

I can only really learn from those who have built ESL for awhile.


Plus the very sandwich itself has two equal sign charges either side of a
dieletric, the whole central. I intuit that this is far better than one-sided.

Sure, but as I say, there are ppl who have had good results from single
coated membranes.



I am sure others here may have more understandings and competence in these
fields ahem than I evidence in these rambling thoughts ... let's hope so!

No, there are very few here who have had any experience with building or
repairs to ESL.

People don't like open public discussions because if they bull****, they
get their arses fried.
I try to be reasonable about it. But nothing good gets done unless
all the problems are admitted, faced, then overcome, and it always
involves discussions, and when I try to do something good,
I'll discuss matters with myself, and sometimes I have to admit I was
fool for
doing something that doesn't work.

People not used the testing of their ideas don't like this forum.

You cannot get way with BS here for long.

I enjoy the input from ppl whose experience has taught them to
understand what they are doing,
and who can tell me what works, and what doesn't, and why.

In a second post 2 hours after than the one you have replied to
I mentioned the amount of bass voltage which could be applied rose after
leaving the panels
alone to charge up for longer.
Yesterday I re-tested the response and bass voltages which could be
applied after leaving
the panels turned on overnight but found they'd become unable to take
more than a few volts
at 70Hz, and when a large voltage was applied, sounded like someting
became un-stuck.
I lowered the EHT from 2,700V to 2,000V,
and they seemed fine except for a huge 12dB resonance peak at 41 Hz,
which seemed to mar most music when turned up.
So I concluded that the charge in the membrane is unstable, and stiction
or membrane pulling is
unpredictable. I am putting in silicone "button" to maybe stop the
resonance where the membrane will flap
badly, and maybe glue felt strips 70mm wide down the centres of each
bass panel
on the outside of the rear stators. Perhaps the two measures wil damp
the resonances.

With EHT = -2,000V, sensitivity is down to 77dB/W/M, perfectly lousy, so
that an even greater drive voltage
is required.

I am investigating alternative coatings and techniques for the membrane
and stators.

I am back at square one with these speakers.



Patrick Turner.




Regards,

Ross Matheson
Auckland, NZ.

Patrick Turner wrote:


I doubled the tension specified in the kit instructions but all that's
done is slightly
raise threshold for the EHT level needed before a membrane heads over
against a stator.

How does ERA suggest you tension the diaphragm?
Is there a figure for the diaphragm's resonant frequency?
I would vote for charging both sides.

By the way, did you know that 'west' is a Jute sockpuppet?

--
Eiron.

May contain traces of irony.

Eiron wrote:

Patrick Turner wrote:

I doubled the tension specified in the kit instructions but all that's
done is slightly
raise threshold for the EHT level needed before a membrane heads over
against a stator.

How does ERA suggest you tension the diaphragm?

The method they specify clearly is to cut up a lot of adhesive gaffer
tape 50mm wide,
and place them ready on the side of the bench. They supply a roll of it.
They also supply a small spring balance with a metal plate 50mm x 75mm
that hangs on the spring balance,
so that the tape pieces are pressed onto the mylar, and metal plate,
then tension is pulled up to a reading 0.8 Kg, and the tape is pressed
to the bench,
and the tension remains, and plate pulled carefully away from the tape.
10Kg could be tensioned up this way, but there is little need to tension
more than 2.5Kg/60mm
as that is the tension in the 3.5uM membrane in a Quad ESL63.
The distance betwen tapes is no more than 10mm, so there is 0.8Kg per
60mm of side length applied.
The tension is done by applying first across the middle of the maylar
sheet each way in a cross,
then at each corner, then outwards from the middle again on each side
one by one alternately
until you have worked outwards from where you start.
Because the mylar sheet is about 75mm larger than the panel size the
tension within the area of the
panel is fairly even if there are slight uneven tensions at the edges.

After 5 goes, i could tension up a membrane 1.5M x 0.55M in 20 minutes.
It becomes easy with practice, but many ppl would have bothers with
uneven tension.
The first one ripped apart into the panel area, so I just started again.

If you don't repeat some things, you'll never get any good doing them.


Is there a figure for the diaphragm's resonant frequency?

The first couple of membrane attempts gave about 32 Hz with 0.8Kg /
60mm,
but on the last one I used nearly 2Kg / 60mm, and Fo went up to 41Hz.


I would vote for charging both sides.

So would I but the construction sequence and details of the build up of
layers does not permit
coating of both sides of the membrane.



By the way, did you know that 'west' is a Jute sockpuppet?

West has had very cordial friendly private emails with me and I see
little evidence he is a sock
puppet of anyone else except his good self.

Are ye not a sock puppet for Eiron?

Just who is exactly themself while online in groups?

I try to be the same wherever I was online or offline; and no matter
whether I an dining at Her Majesty's Buckingham Palace,
or with a or with my visiting friend Nicole K at a local cafe.
Thinking, telling, and doing everything with the same game plan
where you appear consistent no matter what the circumstances is called
one-ness.

Its all I can manage, and two or three of me would be exceedingly
tiresome,
and one is bad enough :-)

West does ask a lot of simple questions, and I answer them as best i can
without being what I consider unpolite, or tactless, but nevertheless
firm, fair, bordering on terse.
Finding time to avoid stepping on peoples' toes while tip towing around
their questions
is difficult. He's never been grossly malicious, or foul mouthed, so
I can concentrate on playing the ballgame, and not have to deal with the
man.


Patrick Turner.



--
Eiron.

May contain traces of irony.

Peter Wieck said:


Interesting point. Thanks for bringing it up. My brother keeps a pair
of Maggies (MG1) that I snagged for him some years ago... they were
far too large for my use then and now, but he keeps them under near-
ideal conditions by way of humidity and temperature. They are lovely
speakers which he treasures.


Be very careful with storage, indeed. Moisture is killing.
When the aluminium wires oxidize (and they *will* ), they're on their
way out.
The lucky thing is, one can rewire them relatively easy:

http://www.integracoustics.com/MUG/M...aks/index.html

--

- Maggies are an addiction for life. -

On Jun 4, 11:41 am, Sander deWaal wrote:
Peter Wieck said:

Interesting point. Thanks for bringing it up. My brother keeps a pair
of Maggies (MG1) that I snagged for him some years ago... they were
far too large for my use then and now, but he keeps them under near-
ideal conditions by way of humidity and temperature. They are lovely
speakers which he treasures.

Be very careful with storage, indeed. Moisture is killing.
When the aluminium wires oxidize (and they *will* ), they're on their
way out.
The lucky thing is, one can rewire them relatively easy:

http://www.integracoustics.com/MUG/M...aks/index.html

--

- Maggies are an addiction for life. -

Thank you for this. I have archived it against the eventuality. I will
be the one doing the work. My brother is the writer, I am the one with
the good hands.

Peter Wieck
Wyncote, PA