Hi,
your question is certainly interesting to everyone who refurbishes or builds electrostatic panels. The tension itself is not a value of interest. What counts is the resonance frequency, that surely depends on several factors such as material and material thickness, mass, dimensions and the foil tension after the foil is glued to its frame and the weights are released.
The original Quad stretching jig uses steel weights which total to 45kg for the mid/high-panel and some 60 kg for the bass-panels. I have seen the original jig, which is now in Germany, and if I recollect correctly they have 30 weights. So, each weight is about 1,5kg for the mid/high and 2kg for the bass. If you apply these values you will be very close to the original way Quad is tensioning their panels. This is only valid if the tensioned area is in the vicinity of some 40 by 80 cm. The panel itself is apparently smaller. So if you intend to refurbish Quad panels or build new ones with the same dimensions this should put you very close to Quads specifications.
Coming to your question, Quad uses an indirect method to achieve, respectively to monitor the desired resonance. What they still do today, is to position the panel under test horizontal above a bass speaker ( in that case they use KEF B139). Again above the panel sits a microphone that measures the sound pressure coming through the panel. To be clear at this point the panel is not connected, it is the B139 that is working. Then they look for the maximum voltage at the microphone by running through the lower frequencies of the audio spectrum to find the resonance of the pre-finished panel. They apply this method prior to coating, assembly etc. By this they are able to decide weather the panels resonance is within the accepted tolerances. This also allows to match panels, if desired.
If you still would like to measure the tension, there are some different methods, most of them are not suitable for the hobbyist. What I did in the beginning was to lay down the one side of the panel that holds the mylar on a workbench and to measure the bending of the film by applying a constant weight (200g) in the very center of the film. To do so I used a micrometer. This for sure not a very scientific approach, but it works and the measured values correspond quite well to the resonance of the finished panel.
I hope this helps a bit and I would appreciate if someone could comment on the relation between the applied stretching weights and the achieved resonance – if there is a simple relation at all.
your question is certainly interesting to everyone who refurbishes or builds electrostatic panels. The tension itself is not a value of interest. What counts is the resonance frequency, that surely depends on several factors such as material and material thickness, mass, dimensions and the foil tension after the foil is glued to its frame and the weights are released.
The original Quad stretching jig uses steel weights which total to 45kg for the mid/high-panel and some 60 kg for the bass-panels. I have seen the original jig, which is now in Germany, and if I recollect correctly they have 30 weights. So, each weight is about 1,5kg for the mid/high and 2kg for the bass. If you apply these values you will be very close to the original way Quad is tensioning their panels. This is only valid if the tensioned area is in the vicinity of some 40 by 80 cm. The panel itself is apparently smaller. So if you intend to refurbish Quad panels or build new ones with the same dimensions this should put you very close to Quads specifications.
Coming to your question, Quad uses an indirect method to achieve, respectively to monitor the desired resonance. What they still do today, is to position the panel under test horizontal above a bass speaker ( in that case they use KEF B139). Again above the panel sits a microphone that measures the sound pressure coming through the panel. To be clear at this point the panel is not connected, it is the B139 that is working. Then they look for the maximum voltage at the microphone by running through the lower frequencies of the audio spectrum to find the resonance of the pre-finished panel. They apply this method prior to coating, assembly etc. By this they are able to decide weather the panels resonance is within the accepted tolerances. This also allows to match panels, if desired.
If you still would like to measure the tension, there are some different methods, most of them are not suitable for the hobbyist. What I did in the beginning was to lay down the one side of the panel that holds the mylar on a workbench and to measure the bending of the film by applying a constant weight (200g) in the very center of the film. To do so I used a micrometer. This for sure not a very scientific approach, but it works and the measured values correspond quite well to the resonance of the finished panel.
I hope this helps a bit and I would appreciate if someone could comment on the relation between the applied stretching weights and the achieved resonance – if there is a simple relation at all.
I have no experience with replacing the diaphragm, but I replaced the dust cover of my ESL-57s a while ago. I followed this method: I layed down the dust cover foil on a flat surface (kitchen desk 
). Then I applied glue on the wooden frame and pushed it down to the foil. After assembling together and sealed around the edge, I went through the surface of the dust cover with a hair dryer. I took care not to stretch too much, because I wanted to keep the tension (and the resonance) as low as possible. Theoretically, tension of the dust cover is necessary only to avoid rattle.
Is there any scientific explanation why is a relatively high tension used at the diaphragm? I guess even a relatively low tension can prevent the diaphragm from attracted by one of the stators.
). Then I applied glue on the wooden frame and pushed it down to the foil. After assembling together and sealed around the edge, I went through the surface of the dust cover with a hair dryer. I took care not to stretch too much, because I wanted to keep the tension (and the resonance) as low as possible. Theoretically, tension of the dust cover is necessary only to avoid rattle.Is there any scientific explanation why is a relatively high tension used at the diaphragm? I guess even a relatively low tension can prevent the diaphragm from attracted by one of the stators.
StanJ,
Concerning your remark:
<<I hope this helps a bit and I would appreciate if someone could comment on the relation between the applied stretching weights and the achieved resonance – if there is a simple relation at all.>>
This helps me at least ! This information is of great interest.
I am currently trying to make the link between the weights applied and the resonance frequency(s).
I already have a computing sheet but I need some information or confirmation:
* the tensioned area is always 80*40 whatever the panel ? (bass or treble)
* this 80*40cm area is flat during the process ? I mean the attachement between weights and PET sheet is in the horizontal plan. Or saying it differently the weights are evenly distributed on the full (80+40)*2 cm ?
* what is the density of the PET sheets ? I assume around 900kg/m3. Same for bass and treble ? Same question for the coating itself but I assume this part is neglectable
* do you have some estimation of the expected resonance frequencies ? I heard "supersonic" for treble meaning above 20kHz but this is achievable for horizontal direction and not vertical.
* finally I will need the effective dimension of the moving parts (=the coated part) but this will be easy to find
Thank you
Concerning your remark:
<<I hope this helps a bit and I would appreciate if someone could comment on the relation between the applied stretching weights and the achieved resonance – if there is a simple relation at all.>>
This helps me at least ! This information is of great interest.
I am currently trying to make the link between the weights applied and the resonance frequency(s).
I already have a computing sheet but I need some information or confirmation:
* the tensioned area is always 80*40 whatever the panel ? (bass or treble)
* this 80*40cm area is flat during the process ? I mean the attachement between weights and PET sheet is in the horizontal plan. Or saying it differently the weights are evenly distributed on the full (80+40)*2 cm ?
* what is the density of the PET sheets ? I assume around 900kg/m3. Same for bass and treble ? Same question for the coating itself but I assume this part is neglectable
* do you have some estimation of the expected resonance frequencies ? I heard "supersonic" for treble meaning above 20kHz but this is achievable for horizontal direction and not vertical.
* finally I will need the effective dimension of the moving parts (=the coated part) but this will be easy to find
Thank you
Salut,
it is an excellent idea to develop an algorithm that describes the relation between the applied stretching force and the corresponding resonance frequency. As far as know this has never been done before and published.
I am happy to answer your questions as follows:
1) Concerning the Quad stretching jig: 40 by 80 cm² is correct for both panels.
2) The sheets are flat while tensioned and glued to the frames
3) The specific density is the same for both panels. The value is –as far as I remember - 1,4 g/cm³ (check Mylar by Dupont). However Quad uses or at least used 6 ym for the mid/treble unit and 12 ym for the bass unit.
For the coating you may apply an extra 5-10 %.
4) The resonance of the Quad 57 bass panel is between 50 to 70 Hz.
Resonance supersonic? : You really mean that?: ...Impossible!
5) Dimensions: google!
Hope this helps a bit – and good luck,
StanJ
PS: I propose you open up a new “non-Quad-related” thread that focuses only on your subject. I am sure, if you do so, you will get plenty of help and appreciation.
it is an excellent idea to develop an algorithm that describes the relation between the applied stretching force and the corresponding resonance frequency. As far as know this has never been done before and published.
I am happy to answer your questions as follows:
1) Concerning the Quad stretching jig: 40 by 80 cm² is correct for both panels.
2) The sheets are flat while tensioned and glued to the frames
3) The specific density is the same for both panels. The value is –as far as I remember - 1,4 g/cm³ (check Mylar by Dupont). However Quad uses or at least used 6 ym for the mid/treble unit and 12 ym for the bass unit.
For the coating you may apply an extra 5-10 %.
4) The resonance of the Quad 57 bass panel is between 50 to 70 Hz.
Resonance supersonic? : You really mean that?: ...Impossible!
5) Dimensions: google!
Hope this helps a bit – and good luck,
StanJ
PS: I propose you open up a new “non-Quad-related” thread that focuses only on your subject. I am sure, if you do so, you will get plenty of help and appreciation.
oshifis said:
Actually the tension has to be sufficient to keep the diaphragm from being attracted to the stators... then there is the relationship between the excursion at resonance vs. power input vs. overall frequency response. In the case of the Quad we just got a very good run down of where that optimal tension point is for the Quads... too high a tension and obviously the resonant point is too high as well... a balancing act between the factors...
_-_-bear
First results of computation: first of all, let's make the difference
between intrinsic resonance(s) of the diaphragm and resonance
when stimulated with a source and considering coupling with air.
Using your figures, the stretched unbonded panel (80*40cm) will
have a main mode of resonance at 207Hz for treble and 169Hz for
bass. Meaning that hitting the diaphragm once it is set on the
stretching jig will give this sound. A nice way to check that it is
correctly stretched. This is by the way how it is done here:
http://loudspeaker-repair-service.reromanus.net/refurbish_quad.htm
Very interesting site wrt stretching and annealing !
These figures come from the classical computing of the resonance
modes of a square drum.
Once bonded, the diaphragm will give a different sound if
stimulated. Because the moving part is smaller. That was the
purpose of my question: the treble and bass panels are in fact
divided in three moving parts. But the exact dimension is not widely
available thus I cannot tell what will be the sound of the diaphragm
once bonded on the stator.
Next step will be to compute resonance frequency when stimulated
by an electric source and radiating in the air: these are not the
same frequencies as the ones given above because now the
"radiation resistance" will be implied.
Hoping I am clear !
between intrinsic resonance(s) of the diaphragm and resonance
when stimulated with a source and considering coupling with air.
Using your figures, the stretched unbonded panel (80*40cm) will
have a main mode of resonance at 207Hz for treble and 169Hz for
bass. Meaning that hitting the diaphragm once it is set on the
stretching jig will give this sound. A nice way to check that it is
correctly stretched. This is by the way how it is done here:
http://loudspeaker-repair-service.reromanus.net/refurbish_quad.htm
Very interesting site wrt stretching and annealing !
These figures come from the classical computing of the resonance
modes of a square drum.
Once bonded, the diaphragm will give a different sound if
stimulated. Because the moving part is smaller. That was the
purpose of my question: the treble and bass panels are in fact
divided in three moving parts. But the exact dimension is not widely
available thus I cannot tell what will be the sound of the diaphragm
once bonded on the stator.
Next step will be to compute resonance frequency when stimulated
by an electric source and radiating in the air: these are not the
same frequencies as the ones given above because now the
"radiation resistance" will be implied.
Hoping I am clear !
Hi,
not only this will happen,
...if the diaphragm ist polarizied and between the stators,
it will attempt to attract the two stators,...
...it acts as if it has a negativ compliance...
...and will move the resonce frequency down...
(~The Art of Sound Reproducktion from John Watkinson)
Hope it helps
Regards,
Frank
not only this will happen,
...if the diaphragm ist polarizied and between the stators,
it will attempt to attract the two stators,...
...it acts as if it has a negativ compliance...
...and will move the resonce frequency down...
(~The Art of Sound Reproducktion from John Watkinson)
Hope it helps
Regards,
Frank
If you want to read:
http://books.google.de/books?id=01u...X&oi=book_result&ct=result&resnum=2#PPA182,M1
http://books.google.de/books?id=01u...X&oi=book_result&ct=result&resnum=2#PPA182,M1
I tried to add the effect of radiation through the air.
This brings an additional mass (which lowers resonance) but an additional spring effect (which highers resonance).
Note that all input figures are supposed to be verified except the size of the moving part (10*70cm below).
Any comment welcome !
Bass panel:
P0 101325
T0 293,15
R 8,32
Gamma 1,4
Weight applied with jig 60 kg 60 kg
Linear force 245,25 N/m
Stretched diaphragm Width 40 cm 0,4 m
Height 80 cm 0,8 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 5,3376 g 0,0053376 kg
k (spring) 6051,301198
Resonance frequency 169,4616788 Hz
Equivalent radiating surface 0,129691115 m2
Additional mass due to air 15,81035605 g 0,015810356 kg
Additional k due to air 99121,6412
Resonance frequency with radiation effect 354,9261059 Hz
Moving part Width 10 cm 0,1 m
Height 70 cm 0,7 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 1,1676 g 0,0011676 kg
k (spring) 17289,432
Resonance frequency 612,439864 Hz
Equivalent radiating surface 0,028369931 m2
Additional mass due to air 1,617572455 g 0,001617572 kg
Additional k due to air 4743,125409
Resonance frequency with radiation effect 447,6375819 Hz
Celerity 121,2569066 m/s
This brings an additional mass (which lowers resonance) but an additional spring effect (which highers resonance).
Note that all input figures are supposed to be verified except the size of the moving part (10*70cm below).
Any comment welcome !
Bass panel:
P0 101325
T0 293,15
R 8,32
Gamma 1,4
Weight applied with jig 60 kg 60 kg
Linear force 245,25 N/m
Stretched diaphragm Width 40 cm 0,4 m
Height 80 cm 0,8 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 5,3376 g 0,0053376 kg
k (spring) 6051,301198
Resonance frequency 169,4616788 Hz
Equivalent radiating surface 0,129691115 m2
Additional mass due to air 15,81035605 g 0,015810356 kg
Additional k due to air 99121,6412
Resonance frequency with radiation effect 354,9261059 Hz
Moving part Width 10 cm 0,1 m
Height 70 cm 0,7 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 1,1676 g 0,0011676 kg
k (spring) 17289,432
Resonance frequency 612,439864 Hz
Equivalent radiating surface 0,028369931 m2
Additional mass due to air 1,617572455 g 0,001617572 kg
Additional k due to air 4743,125409
Resonance frequency with radiation effect 447,6375819 Hz
Celerity 121,2569066 m/s
StanJ said:
I found the source of what I said, regarding "supersonic resonance": http://www.quadesl.org/Fixit_/TreblePanel/treblepanel.html
I do not really understand their point.
I am progressing of my side. First of all I removed the "additional k due to the air" because we are not using closed box. Provided enough room size, this is not for us.
Then I re-computed the "equivalent piston area" which is now the total area because
we are working in the mass control zone.
Latest results. They are very sensitive to the distance stator-diaphragm. Therefore I must check this value.
Bass
P0 101325 Pa
T0 293,15 K
R 8,32
Gamma 1,4 -
Eps0 8,85419E-12 F/m
Ratio piston/non-piston 0,405284735 -
Weight applied with jig 60 kg 60 kg
Linear force 245,25 N/m
Unbonded diaphragm Width 40 cm 0,4 m
Height 80 cm 0,8 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 5,3376 g 0,0053376 kg
k (spring) 6051,301198 Shape is in "sine*sine"
Resonant frequency in vacuum 169,4616788 Hz
Equivalent "piston" surface 0,129691115 m2 Shape is in "sine*sine"
Added mass due to radiation in the air 15,81035605 g 0,015810356 kg
k due to radiation (compress the air) 0 Open: not in a closed box
Resonant frequency (radiation effect included) 85,13542671 Hz
Bonded diaphragm Width 10 cm 0,1 m
Height 70 cm 0,7 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 1,1676 g 0,0011676 kg
k (spring) 17289,432
Resonant frequency in vacuum 612,439864 Hz
Equivalent "piston" surface = total surface 0,07 m2 Behaving like a piston
Added mass due to radiation in the air 6,269362388 g 0,006269362 kg
k due to radiation (compress the air) 0 Open: not in a closed box
Voltage 6000 V
Spacing stator-diaphragm 3,175 mm 0,003175 m
k due to electrostatic force -1394,273707 Behaving like a piston
Resonant frequency (radiation effect included) 232,6777029 Hz
StanJ,
Last question from me. I still have some doubts on few points of
my computing sheet but whatever the model I take for each
parameter, I find that the mentioned weigths are quite high to target
low resonance (below 100Hz).
Knowing that the jig comes from England, is there any chance that
the "45kg" and "60kg" are in fact 45 and 60 pounds ?
My intention is certainly not to modify the facts so that my computing sheet finaly works. But I prefer to check.
Thank you.
FL
Last question from me. I still have some doubts on few points of
my computing sheet but whatever the model I take for each
parameter, I find that the mentioned weigths are quite high to target
low resonance (below 100Hz).
Knowing that the jig comes from England, is there any chance that
the "45kg" and "60kg" are in fact 45 and 60 pounds ?
My intention is certainly not to modify the facts so that my computing sheet finaly works. But I prefer to check.
Thank you.
FL
Calvin said:
Not till recently. But as I derived the air mass from a model of a perfect
pulsating sphere (http://www.brouchier.com/livre/node33.html), it
seems indeed wise to double this value. Because the diaphram pushes
on one side while it pulls on the other side. Which is not considered
in the model.
Then I reconsidered the computing of the "k" factor when the
loudspeaker is stimulated: from a sine*sine shape we go to a
piston-like shape and thus the spring effect is different.
Finally, if considering pounds instead of kg, it gives results close to expected value:
Weight applied with jig 60 Pounds 27,21552 kg
Linear force 111,243438 N/m
208,3333333 kg/cm2
% of breakdown limit 10,96491228 %
Unbonded diaphragm Width 40 cm 0,4 m
Height 80 cm 0,8 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 5,3376 g 0,0053376 kg
k (spring) 2744,821813 "Shape is in ""sine*sine"""
Resonant frequency in vacuum 114,1311511 Hz
"Equivalent ""piston"" surface" 0,129691115 m2 "Shape is in ""sine*sine"""
Added mass due to radiation in the air 31,6207121 g 0,031620712 kg Model of a pulsating sphere, times 2 (pression on one side, depression on the other)
k due to radiation (compress the air of the box/room) 0 Open: not in a closed box
Resonant frequency (radiation effect included) 43,37316811 Hz
Bonded diaphragm Width 10 cm 0,1 m
Height 70 cm 0,7 m
Thickness 12 um 0,000012 m
Density 1390 kg/m3 1390 kg/m3
Mass 1,1676 g 0,0011676 kg
TBC k (spring) 4068,331447 Delta(Force)=Force.Delta(Surface)/Surface.z
Resonant frequency in vacuum 297,085361 Hz
"Equivalent ""piston"" surface = total surface" 0,07 m2 Behaving like a piston
Added mass due to radiation in the air 12,53872478 g 0,012538725 kg Model of a pulsating sphere, times 2 (pression on one side, depression on the other)
k due to radiation (compress the air of the box/room) 0 Open: not in a closed box
Voltage 6000 V
Spacing stator-diaphragm 3,175 mm 0,003175 m
k due to electrostatic force -1394,273707 Behaving like a piston
Resonant frequency (radiation effect included) 70,29833076 Hz
Frequency response
By the way, the frequency response of my ESL57: see below.
If I apply again "60kg" in the compute sheet (and no more 60 pounds), I find this resonance frequency: 118,3170111Hz (iso 70Hz).
Finaly...not so bad compared to my measurements. But knowing that the diaphragm is supposed to get stiffer when aging, I still wonder which value to take.
By the way, the frequency response of my ESL57: see below.
If I apply again "60kg" in the compute sheet (and no more 60 pounds), I find this resonance frequency: 118,3170111Hz (iso 70Hz).
Finaly...not so bad compared to my measurements. But knowing that the diaphragm is supposed to get stiffer when aging, I still wonder which value to take.
An externally hosted image should be here but it was not working when we last tested it.
Hi StanJ,
Did you already refurbish a panel for an ESL 57 ?
I just bought an additional pair of ESL57 for spare parts and I will try soon to repair my first panel. I hoped I could find 2 valid panels for treble among the 4 but the efficiencies of each panel are very different.
I will stick to your figures i.e. 45kg of weight for treble, 60kg for bass, 40x80cm,...
You mention a method here above to measure resonance. Un-coated means that the 40x80cm are freely moving ? I understood so.
Thank you.
Yes I did repair two sets of 57.
The discribed measurement of the resonance frequency is valid for the panel prior to coating. That is to say, that the panel is not driven itself, however it is stimulatetd by a B139. This measurement is applied to the panel not to the film in the streching jig. At this point the film is already glued to the frame and the weights are released. The resonance of the 40x80 cm² is not measured at all.
The weights are correct.
The discribed measurement of the resonance frequency is valid for the panel prior to coating. That is to say, that the panel is not driven itself, however it is stimulatetd by a B139. This measurement is applied to the panel not to the film in the streching jig. At this point the film is already glued to the frame and the weights are released. The resonance of the 40x80 cm² is not measured at all.
The weights are correct.
Thanks a lot for reply. Thus I did not understand well. My confusion comes
from the "prior to coating".
I call "coating" the slightly conductive compound which needs to be applied on diaphragms: Calaton, Elvamide,...
Because for me coating is applied to one side before glueing (after glueing
this side is no more accessible) then it is glued then coating can be applied
on the other side, for those who apply on both sides.
In that sense, "prior to coating" might be related to the step just after
glueing the diaphragm, and just before applying coating on its outside
face, thus before closing the panel with the second halves.
Regarding the measurement method, I understand the principles.
Thanks for confirming the weights: it would have been hard to guess from
the few pictures available.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.