I have a few n00b questions for the knowledgeable folk here.
1.Should a diaphragm's height to width ratio follow the golden ratio?
2.Is there a benefit to making many small sections or is this not good?
3.When breaking it up into sections with the spacer tape, should these sections follow the golden ratio too?
4.I have heard that a narrow panel or a wide panel can negate the negative effects of directivity. What would be narrow and what would be wide?
5.What difference is there running the tape across or down to make the sections?
6.Is there a point to making a relatively narrow panel if I plan to use wings or should I just go for a wide panel in that case?
1.Should a diaphragm's height to width ratio follow the golden ratio?
2.Is there a benefit to making many small sections or is this not good?
3.When breaking it up into sections with the spacer tape, should these sections follow the golden ratio too?
4.I have heard that a narrow panel or a wide panel can negate the negative effects of directivity. What would be narrow and what would be wide?
5.What difference is there running the tape across or down to make the sections?
6.Is there a point to making a relatively narrow panel if I plan to use wings or should I just go for a wide panel in that case?
I don't think the golden ratio has any particular benefit in terms of the sound, but it would be great marketing BS. You could wax poetic about ancient Greek philosophers and do a fine job of flim-flamming customers (or dinner guests) and keep everyone entertained for hours! In fact, I think the golden ratio was invented for the purpose of high-end audio marketing, though the inventors/discoverers didn't realize it at the time. It's a lot like the laser back in the 70s when it was referred to as "a solution in search of a problem". It would be a refreshing change from all the quantum mechanics crap that's out there...
The speaker would be at its best when playing music composed using scales based on the golden ratio.
You could make very small sections but I think that would tend to restrict the diaphragm too much and you'd end up with reduced low frequency output. I think you really want to make the sections as large as possible without the diaphragm sticking to the stators.
You can use wings to broaden frequency response of small/narrow panels, but wings won't increase the maximum SPL. Building a larger driver the same size as the small driver plus wings will increase the SPL and provide better low frequency output.
I_F
The speaker would be at its best when playing music composed using scales based on the golden ratio.
You could make very small sections but I think that would tend to restrict the diaphragm too much and you'd end up with reduced low frequency output. I think you really want to make the sections as large as possible without the diaphragm sticking to the stators.
You can use wings to broaden frequency response of small/narrow panels, but wings won't increase the maximum SPL. Building a larger driver the same size as the small driver plus wings will increase the SPL and provide better low frequency output.
I_F
Hi,
1) the golden rule for ESLs is 8:1 (H:W)
2) as Radio Eriwan tells us:"...it depends!"
Basically the answer is no....but it may differ, depending on the building circumstances and the used materials
3) no, the golden number for this is 70:1 to 100:1 of the d/s, which means that the Diaphragm should have supporting points that are positioned, depending on the diaphragm/stator-distance.
4) no, directivity is a parameter that is directly affected by the panels (or speakers) dimensions. There exist some tricks with electrostats to improve the behaviour, but those rely on driving not the complete membrane with the full signal range, but to filter the signal (certain similarity to a classical crossover) in a way, that virtually the membrane area increases with sinking frequency.
5) since the area where the tape is is not contributing to sound, but increases the losses, You like to keep this area as small as possible. With flat panels vertical tapes most often need less area than horzontal ones (following the distance-rule from No.3). With curved panels You need horizontal tapes to centre the diaphragm in the middle between the stators. What they don´t do: They don´t work as a mechanic crossover, as it is stated in Final´s papers (that´s complete BS)
6) go for BIG. Size matters!!
jauu
Calvin
1) the golden rule for ESLs is 8:1 (H:W)
2) as Radio Eriwan tells us:"...it depends!"
Basically the answer is no....but it may differ, depending on the building circumstances and the used materials
3) no, the golden number for this is 70:1 to 100:1 of the d/s, which means that the Diaphragm should have supporting points that are positioned, depending on the diaphragm/stator-distance.
4) no, directivity is a parameter that is directly affected by the panels (or speakers) dimensions. There exist some tricks with electrostats to improve the behaviour, but those rely on driving not the complete membrane with the full signal range, but to filter the signal (certain similarity to a classical crossover) in a way, that virtually the membrane area increases with sinking frequency.
5) since the area where the tape is is not contributing to sound, but increases the losses, You like to keep this area as small as possible. With flat panels vertical tapes most often need less area than horzontal ones (following the distance-rule from No.3). With curved panels You need horizontal tapes to centre the diaphragm in the middle between the stators. What they don´t do: They don´t work as a mechanic crossover, as it is stated in Final´s papers (that´s complete BS)
6) go for BIG. Size matters!!
jauu
Calvin
I_Forgot said:
Thanks for your comments, I_F
If this is so, I could have a 1830mm (6'0") high, 230mm wide panel with a diaphragm to stator spacing of 3mm. Since the width of the entire panel is 77 times the d/s spacing, it would make no sense to use a vertical spacer, but maybe horizontal ones at 300mm distances or so?Calvin said:
Does 4.6' ^ 2 sound OK for full range or maybe 200Hz if it comes to that?6) go for BIG. Size matters!!
Another question, I have spoken with a film supplier who doesn't feel that his product responds well to heat treatment. I was hoping that I could make up for my inexperience by hitting it with a hair dryer once I apply the diaphragm to one stator. Which film should I be searching for?
Hi,
I can only recommend that You don´t start Your first project with a fullrange ESL. Instead I suggest You concentrate on panels working from ~300Hz upwards.
Its relatively easy to build panels that supply just the mid-highs, but it is nearly impossible to build a really good fullranger even for longtime profs!
When You follow this advice, than 3mm d/s is far too much. 1.5mm is a better number to start with. And when Your good at it You might reduce this value to even lower numbers (1.0mm) in future.
The reason is, that efficiency/output should be the prime design goal. So You should choose the d/s as small as possible to get highest output. You won´t need more than 0.5mm for a freq-range of 300Hz and above, but You need some reserve against the ´bowing´ of the membrane when the polarization is On and to cope with acoustic coupling. So 1.5mm(max) is the right starting point. 10" (250mm) width is ok for a crossover-freq of ~300Hz, optionally You might choose a bigger width. With a wider panel the acoustic cancellation begins at lower frequencies. Hence with a wider panel You need less equing, which is good with regard to efficiency and dynamics. With a 40cm wide panel You already don´t need wings aside the panel any more, but just a thin but strong stabilizing frame.
Height doesn´t need to follow the 8:1 rule. The rule applies to fullrangers. Following this rule the membrane works on a real acoustic impedance in its complete frequency range (the ideal state of affairs). Crossing the panel over 1 or 2 octaves above its resonance freq, the 8:1 rule doesn´t apply any more and You can reduce the diaphragm area (height) considerably.
With a panel height of 40" or more You´ll have enough membrane area (my own panels feature 250x1250mm). Heights of something between 50"and 60" will allow for good listening whilst sitting as well as standing. A standard format for metal sheets is 1250x2500mm, so You might be able to order 8 equal sheets of metal of 250x1250mm size, without much costs for cutting or labour.
There are just very few materials that work reliably as diaphragm materials. DuPont´s Mylar type S and C and Mitsubishi/Hoechst´s Hostaphan RE have proven to work. The differences between Mylar S and C is only, that type C is manufactured to tighter tolerances than type S. All 3 possibilites shrink well when using a hot air gun. A hair dryer won´t work, because You need to reach ~150°C.
The thickness can be chosen between ~3µm and 6µm, depending on the amount of mechanical tension You need (e.g. curved panels like MartinLogan style need extremely high tension, while planar panels get away with lower tension). A film greater thickness, i.e. 12µm can already be heard! Less resolution and tilted down highs are the result. Keep in mind that the coating of the membrane introduces some additional mass, which can be considerably higher than the membrane´s mass itself. You have to use a thin diaphragm AND a lowmass coating to reach the best.
jauu
Calvin
I can only recommend that You don´t start Your first project with a fullrange ESL. Instead I suggest You concentrate on panels working from ~300Hz upwards.
Its relatively easy to build panels that supply just the mid-highs, but it is nearly impossible to build a really good fullranger even for longtime profs!
When You follow this advice, than 3mm d/s is far too much. 1.5mm is a better number to start with. And when Your good at it You might reduce this value to even lower numbers (1.0mm) in future.
The reason is, that efficiency/output should be the prime design goal. So You should choose the d/s as small as possible to get highest output. You won´t need more than 0.5mm for a freq-range of 300Hz and above, but You need some reserve against the ´bowing´ of the membrane when the polarization is On and to cope with acoustic coupling. So 1.5mm(max) is the right starting point. 10" (250mm) width is ok for a crossover-freq of ~300Hz, optionally You might choose a bigger width. With a wider panel the acoustic cancellation begins at lower frequencies. Hence with a wider panel You need less equing, which is good with regard to efficiency and dynamics. With a 40cm wide panel You already don´t need wings aside the panel any more, but just a thin but strong stabilizing frame.
Height doesn´t need to follow the 8:1 rule. The rule applies to fullrangers. Following this rule the membrane works on a real acoustic impedance in its complete frequency range (the ideal state of affairs). Crossing the panel over 1 or 2 octaves above its resonance freq, the 8:1 rule doesn´t apply any more and You can reduce the diaphragm area (height) considerably.
With a panel height of 40" or more You´ll have enough membrane area (my own panels feature 250x1250mm). Heights of something between 50"and 60" will allow for good listening whilst sitting as well as standing. A standard format for metal sheets is 1250x2500mm, so You might be able to order 8 equal sheets of metal of 250x1250mm size, without much costs for cutting or labour.
There are just very few materials that work reliably as diaphragm materials. DuPont´s Mylar type S and C and Mitsubishi/Hoechst´s Hostaphan RE have proven to work. The differences between Mylar S and C is only, that type C is manufactured to tighter tolerances than type S. All 3 possibilites shrink well when using a hot air gun. A hair dryer won´t work, because You need to reach ~150°C.
The thickness can be chosen between ~3µm and 6µm, depending on the amount of mechanical tension You need (e.g. curved panels like MartinLogan style need extremely high tension, while planar panels get away with lower tension). A film greater thickness, i.e. 12µm can already be heard! Less resolution and tilted down highs are the result. Keep in mind that the coating of the membrane introduces some additional mass, which can be considerably higher than the membrane´s mass itself. You have to use a thin diaphragm AND a lowmass coating to reach the best.
jauu
Calvin
Hi,
with my panels it becomes damn loud
Rather forget a simple first order crossover!
The reason is, that the panels resonance will likely be of app. +10dB of magnitude. To cope with that You have to use notch filter. Even when using a single series cap as 1st order filter You´ll end up with a higher order because of the notch. And this is just the electrical response. Since the panel itself drops steeply below the resonance You´ll probabely end up with 4th to 6th order acoustical response. Since different filter responses are a major factor why most hybrids don´t sound so well, You´d have to design a high order lowpass for the bass. Additionally You have to equalize nearly each and every panel around 1kHz. All of this costs on output (ML kills up to 10dB with their passive filters).
I recommend active filtering. First it sounds better, second its easier to handle and implement and third it saves the output.
jauu
Calvin
with my panels it becomes damn loud
Rather forget a simple first order crossover!
The reason is, that the panels resonance will likely be of app. +10dB of magnitude. To cope with that You have to use notch filter. Even when using a single series cap as 1st order filter You´ll end up with a higher order because of the notch. And this is just the electrical response. Since the panel itself drops steeply below the resonance You´ll probabely end up with 4th to 6th order acoustical response. Since different filter responses are a major factor why most hybrids don´t sound so well, You´d have to design a high order lowpass for the bass. Additionally You have to equalize nearly each and every panel around 1kHz. All of this costs on output (ML kills up to 10dB with their passive filters).
I recommend active filtering. First it sounds better, second its easier to handle and implement and third it saves the output.
jauu
Calvin
I will be filtering actively. I was thinking of using only partial EQ above about 600Hz so that the response is naturally down 6dB at 300Hz. Then I could just hit the resonance hard out of band, thus having a first or second order acoustic response around 300Hz for an octave or so.
If I understand correctly, I would need perhaps 10dB more EQ if I wanted to fully EQ and actively cross over as an after thought, so I may now have some sensitivity to play with. Hopefully this means that I won't need to build a full ranger to achieve good output above 300Hz using this method.
If I understand correctly, I would need perhaps 10dB more EQ if I wanted to fully EQ and actively cross over as an after thought, so I may now have some sensitivity to play with. Hopefully this means that I won't need to build a full ranger to achieve good output above 300Hz using this method.
Hi,
here´s some measurements:
red: ML Prodigy, black: My panel of ML Sequel-size, just the panels, without frames or baffles.
You can easily see the quite serious resonances at 160Hz/250Hz.
These have to be notched. Even though he Prodigy panel is of larger size it features a higher Fs-value because the thicker membrane can withstand higher tension values.
It follows a range from ~300Hz to 1kHz where phase cancellation rules and a audible suckout takes place (keep in mind that the curves are nearfield. The suckout is much stronger at the listening position). You have to equalize this range.
The range from ~1kHz to 4kHz is a little bit too loud in most cases. This adds ´speed´ and ´livelyness´ to the sound, but sound may become itchy at elevated levels, especially when playing high pitched female voices. The lower SPL above 7kHz with the Prodigy panel is solely due to the thicker membrane (12µm vers 3.5µm).
ML crosses the Prodigy over at 250Hz thereby using the panels´s resonance for eqing purposes. You can do this easily with a 2nd order HP with high Q and crossover freqency raised to around 400Hz, thereby equalizing the suckout region and giving a wonderful linear response down to 250Hz and highest output. The drawback is, that it only looks good on paper. Not only does the HP feature a higher Q, but the panel exhibits a long long long decay on its Fs too. And this is clearly audible. Below 250Hz SPL drops with ~6th order! Couple a typical heavy bass to it (with lots of boom), cross over with just 2nd order and voila, we have the typical ML/Hybrid- sound that so many people take as typical for hybrids and that lead to the bad image of this type of ESL.
The more complex but imo better sounding solution is to cross over at a higher freq. Than You need a mirrored frequency response with Your crossover network. A HP plus Notch plus some equing around 1-4kHz. In passive technology the notch alone may already cost You 10dB in output. In active technology You won´t loose output because of the possible gain of the filter network. Besides ESLs honor active drive by better sound and I can only recommend to drive panels actively when You are looking for best performance.
jauu
Calvin
here´s some measurements:
An externally hosted image should be here but it was not working when we last tested it.
red: ML Prodigy, black: My panel of ML Sequel-size, just the panels, without frames or baffles.
You can easily see the quite serious resonances at 160Hz/250Hz.
These have to be notched. Even though he Prodigy panel is of larger size it features a higher Fs-value because the thicker membrane can withstand higher tension values.
It follows a range from ~300Hz to 1kHz where phase cancellation rules and a audible suckout takes place (keep in mind that the curves are nearfield. The suckout is much stronger at the listening position). You have to equalize this range.
The range from ~1kHz to 4kHz is a little bit too loud in most cases. This adds ´speed´ and ´livelyness´ to the sound, but sound may become itchy at elevated levels, especially when playing high pitched female voices. The lower SPL above 7kHz with the Prodigy panel is solely due to the thicker membrane (12µm vers 3.5µm).
ML crosses the Prodigy over at 250Hz thereby using the panels´s resonance for eqing purposes. You can do this easily with a 2nd order HP with high Q and crossover freqency raised to around 400Hz, thereby equalizing the suckout region and giving a wonderful linear response down to 250Hz and highest output. The drawback is, that it only looks good on paper. Not only does the HP feature a higher Q, but the panel exhibits a long long long decay on its Fs too. And this is clearly audible. Below 250Hz SPL drops with ~6th order! Couple a typical heavy bass to it (with lots of boom), cross over with just 2nd order and voila, we have the typical ML/Hybrid- sound that so many people take as typical for hybrids and that lead to the bad image of this type of ESL.
The more complex but imo better sounding solution is to cross over at a higher freq. Than You need a mirrored frequency response with Your crossover network. A HP plus Notch plus some equing around 1-4kHz. In passive technology the notch alone may already cost You 10dB in output. In active technology You won´t loose output because of the possible gain of the filter network. Besides ESLs honor active drive by better sound and I can only recommend to drive panels actively when You are looking for best performance.
jauu
Calvin
Hi Calvin,
How much dB do you loose because of phase cancellation in the range of 300 Hz to 1 kHz?
Looking to several measurements of different esls, the cancellation dip is centered around 100 to 300 Hz. Above 300 Hz it is fairly lineair most times.
Maybe the 'wings' on for example the old sequel will have some benefits here.
MartinJan
How much dB do you loose because of phase cancellation in the range of 300 Hz to 1 kHz?
Looking to several measurements of different esls, the cancellation dip is centered around 100 to 300 Hz. Above 300 Hz it is fairly lineair most times.
Maybe the 'wings' on for example the old sequel will have some benefits here.
MartinJan
Hi
My plan was to avoid notch filters and equalizing as much as possible.
The notch filter probably can be avoided by silicone dots on the membrane. I've used this technique with several esls (it is also used by audiostatic) and this kills the resonance very effectively. However, in case of curved esls, the dots cannot be (partially) hidden by the horizontal bars as the dots must be placed in the middle of the membrane. So you will see them. Not a big problem for me.
Hopefully the equalizer can be avoided by the wings, but we can only know after measurements.
martinjan
My plan was to avoid notch filters and equalizing as much as possible.
The notch filter probably can be avoided by silicone dots on the membrane. I've used this technique with several esls (it is also used by audiostatic) and this kills the resonance very effectively. However, in case of curved esls, the dots cannot be (partially) hidden by the horizontal bars as the dots must be placed in the middle of the membrane. So you will see them. Not a big problem for me.
Hopefully the equalizer can be avoided by the wings, but we can only know after measurements.
martinjan
Hi,
You can add wings to lower the frequency where Phase-Cancellation starts, but You´d need unpractical wide wings. With a 10" wide panel P-C effects the range below ~800Hz, a 5" wide Panel will be effected below ~600Hz. So You won´t come around using some kind of equing.
Using less mechanical tension lowers the resonance freq, but doesn´t help with P-C. P-C is a phenomenon of panel size/baffle size only.
Since it is not difficult to build a panel the size You wish, be it larger or smaller, it is always beneficial to build as large as possible. Wings are just dead area (acoustical wise). Having active membrane area instead is far better. When You build ´transparent designs´ a larger panel looks better than a solid wing too.
What the dots do:
Putting dots between stator and membrane raises stiffness. As result You get a higher Fs and a rather lower Qt. The higher stiffness is beneficial in that way that You can drive the panel with higher signal voltages and the point when it comes to an overload situation where the diaphragm hits the stator is raised.
What it doesn´t do:
It doesn´t kill Fs or high Qts. To lower the Qt You need damping. That is one of the reasons why most fullrangers are not transparent. The damping mats would look rather ugly.
The silicone dots used by Audiostatic introduce not much damping, because silicone has rather low damping values.
So, with regard to Fs and Qt the two ways -using high tension, or using lower tension and dots- lead to similar results.
When placing the dots it should be kept in mind, that they reduce the maximum diaphragm travel and output. This isn´t a real problem with hybrid panels, but could be one with fullrangers. So a certain minimum distance between the dots has to be supplied for. This minimum distance could be reduced only if the dots feature some compressability, like foam does.
jauu
Calvin
You can add wings to lower the frequency where Phase-Cancellation starts, but You´d need unpractical wide wings. With a 10" wide panel P-C effects the range below ~800Hz, a 5" wide Panel will be effected below ~600Hz. So You won´t come around using some kind of equing.
Using less mechanical tension lowers the resonance freq, but doesn´t help with P-C. P-C is a phenomenon of panel size/baffle size only.
Since it is not difficult to build a panel the size You wish, be it larger or smaller, it is always beneficial to build as large as possible. Wings are just dead area (acoustical wise). Having active membrane area instead is far better. When You build ´transparent designs´ a larger panel looks better than a solid wing too.
What the dots do:
Putting dots between stator and membrane raises stiffness. As result You get a higher Fs and a rather lower Qt. The higher stiffness is beneficial in that way that You can drive the panel with higher signal voltages and the point when it comes to an overload situation where the diaphragm hits the stator is raised.
What it doesn´t do:
It doesn´t kill Fs or high Qts. To lower the Qt You need damping. That is one of the reasons why most fullrangers are not transparent. The damping mats would look rather ugly.
The silicone dots used by Audiostatic introduce not much damping, because silicone has rather low damping values.
So, with regard to Fs and Qt the two ways -using high tension, or using lower tension and dots- lead to similar results.
When placing the dots it should be kept in mind, that they reduce the maximum diaphragm travel and output. This isn´t a real problem with hybrid panels, but could be one with fullrangers. So a certain minimum distance between the dots has to be supplied for. This minimum distance could be reduced only if the dots feature some compressability, like foam does.
jauu
Calvin
In theory, phase cancellation becomes more complete at low frequencies. It seems impractical to EQ, but there is a good word for dipoles here http://www.linkwitzlab.com/MJ Dijkstra said:
I think I'd like to try an infinite baffle by extending the wings across the room corners and put a lid over it, which I can put my mid/top panels onto.
Hi Calvin
I can't agree with you.
The silicone dots DO a lot of damping. They do this by spreading (breaking up) the single large resonance to a more distributed resonance.
Tap on the membrane without dots, you hear doommmmmmmm
Now add the dots on the same panel, you hear domm.
This is a simple experiment anyone can do and you can hear the difference very easy.
With regarding phase cancellation I have yet to see measurements which show me that a panel of 30 cm width with some wings of 15 cm at both sides, need equalizing above 300 Hz. I think they don't unless beaming of mids and highs are uncontrolled and make bass relatively weak.
I can't agree with you.
The silicone dots DO a lot of damping. They do this by spreading (breaking up) the single large resonance to a more distributed resonance.
Tap on the membrane without dots, you hear doommmmmmmm
Now add the dots on the same panel, you hear domm.
This is a simple experiment anyone can do and you can hear the difference very easy.
With regarding phase cancellation I have yet to see measurements which show me that a panel of 30 cm width with some wings of 15 cm at both sides, need equalizing above 300 Hz. I think they don't unless beaming of mids and highs are uncontrolled and make bass relatively weak.
Hi,
I doubt it First, silicone´s damping values are quite low and second should You notice a considerable raise in Fs. This might sound as if the high Q is lowered, but in fact it isn´t.
With regard to measurements. I did that with a 10" panel and 4" Wings and there is still a considerable drop out at 300Hz.
jauu
Calvin
I doubt it
First, silicone´s damping values are quite low and second should You notice a considerable raise in Fs. This might sound as if the high Q is lowered, but in fact it isn´t.With regard to measurements. I did that with a 10" panel and 4" Wings and there is still a considerable drop out at 300Hz.
jauu
Calvin
Hi,
I doubt it First, silicone´s damping values are quite low and second should You notice a considerable raise in Fs. This might sound as if the high Q is lowered, but in fact it isn´t.
With regard to measurements. I did that with a 10" panel and 4" Wings and there is still a considerable drop out at 300Hz.
jauu
Calvin
I doubt it
First, silicone´s damping values are quite low and second should You notice a considerable raise in Fs. This might sound as if the high Q is lowered, but in fact it isn´t.With regard to measurements. I did that with a 10" panel and 4" Wings and there is still a considerable drop out at 300Hz.
jauu
Calvin
Calvin said:Hi,
I doubt it
With regard to measurements. I did that with a 10" panel and 4" Wings and there is still a considerable drop out at 300Hz.
jauu
Calvin
Hi
The damping properties of silicone dots are irrelevant. You could get the same thing with hard dots. If it would seriously raise Fs, all the full range esls of audiostatic would have a big problem. They wouldn't be able to generate low bass and they would resonate at a higher freq. This is not the case.
I would suggest that you try this experiment yourself. Seeing and hearing it is better than 1000 words.
The measurement you did, was it an electrically divided esl?
The mirror drive of audiostatic (used to compensate phase cancellation) starts to operate below 300 Hz. Final 1.2 starts to drop below around 300 Hz as well.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.