Hi jaxboy.
I noticed on my next post to the one that you quoted ,showed two exciters closely mounted.
As you can see they are definitely not working in tandem 🫣
Even though they are operating in the same phase.
Positioning is very important.
I only used Evo-stik as my joint gun filler had gone off 🤗
https://www.wickes.co.uk/Wickes-Dec...XXNJgALCE8_CF_Fa8qwaAsGzEALw_wcB&gclsrc=aw.ds
I have used this in the past with good filling properties on large spaces on exciter and spines .
You can use silicone for a softer smaller mountings , the choice is yours.
As long as it does no run and sag down while drying.
I usually put about four or so blobs on the exciter to allow the exciter to breathe, and not fill up the whole area.
Unless I drill a hole in the spine.
Steve.
Jaxboy.
For the life of me , I could not remember which panel I had the caulk on.
Just remembered it was on my rigidly mounted ply panel.
See pics.
This is not such a great idea on 3mm ply panels as they tend to bend over time.
But the rigid mounting seems to have prevented this panel from bending too much ?
Steve.
For the life of me , I could not remember which panel I had the caulk on.
Just remembered it was on my rigidly mounted ply panel.
See pics.
This is not such a great idea on 3mm ply panels as they tend to bend over time.
But the rigid mounting seems to have prevented this panel from bending too much ?
Steve.
Attachments
Hello Steve,
I made one pair of panels with a 3mm plywood (already described). The membrane got a serious warp after applying the varnish. The warping completely disappear after gluing the membrane on the peripheral suspension (some foam). They have been working for more than a year and a half now, no visible bending (location : France, Paris area = not highly warm/humid)
Christian
Steve,
I read in the PE forum about that exciter, and it said it was an M4. That is way too big. I just tried an M3 and it would not work either, but the size of the hole was that size. It's possible that there are no threads in the hole. I can't tell because it's recessed on both sides of the exciter and my eyesight isn't all that great. I was going to run a screw through the spline and into the exciter instead of filling the space with foam, which keeps falling out. Please straighten me out here: Does the filler need to be firm or soft?
I read in the PE forum about that exciter, and it said it was an M4. That is way too big. I just tried an M3 and it would not work either, but the size of the hole was that size. It's possible that there are no threads in the hole. I can't tell because it's recessed on both sides of the exciter and my eyesight isn't all that great. I was going to run a screw through the spline and into the exciter instead of filling the space with foam, which keeps falling out. Please straighten me out here: Does the filler need to be firm or soft?
Starting from this page on AudioCircle I am posting some interesting posts by crackie:
https://www.audiocircle.com/index.php?topic=70541.3000
« Reply #3005 on: 27 Sep 2017, 07:57 pm »
Hello all........what a fascinating thread.
Its interesting to see just how far it is possible to get using empirical development. I was first involved with NXT in 1997 and 1998 when I attended their training courses in Huntingdon. In 1999 I moved to Huntingdon and worked for an injection moulding company who carried out most of NXT's & Mission's prototyping work; Mission / NXT were owned by the same people at the time. The moulding company also supplied various components to Amina technologies and in 2012 I began work in their design department.
The best impirical experimenters on this thread have reached similar conclusions to some extremely sophisticated FEA modelling. When understanding DML behaviour its best not to compare it with conventional moving coil speakers at all. Moving coil drivers are pistonic or at least they aspire to be whereas DMLs are incoherent devices with multiple areas of the panel being in or out of phase with each other at any given time. DMLs oftem struggle at high volumes or when there is a lot of LF in the programme material. This is because the nodal behaviour of the panel, at the exciter drive point, is often very different / contradictory to the input signal being fed to that exciter. Hope that makes sense.......
Panel suspension / termination is important too.........less is more. Good old fashion elastic bands are a simple and very effective solution.
If you keep volumes levels sensible and/ or keep extreme LF away from the panel then they can be superb, just like any well designed cabinetless speaker ( electrostics, ribbons, Linkwitz dipoles etc etc )
« Reply #3041 on: 13 Oct 2017, 09:33 am »
Quote from: Bendingwave on 30 Sep 2017, 09:41 pm
Thanks for the welcome(s). Apologies for the delay responding.
For the sake of discussion, we can think of a moving coil driver's voice coil and an exciter's voice coil behaving in the same way and responding faithfully to the AC applied at its input. The cone motion on an 'ideal' moving coil driver would faithfully reproduce the input AC signal and apply pressure ( or not ) to the air in front of the cone. State of the art moving coil drivers from people like Accuton, Raidho, SEAS, etc variously use metal alloys, ceramics diamond and are notionally pistonic at the lower end of their bandwidth; they all cease to be pistons at the high end and become non linear. The point at which this happens is easy to determine by reviewing their CSD or distortion characteristics. A good old impedance curve also gives away many clues about non linearity.
Regarding your point about bass drivers needing to be pistonic and midranges/tweeters not needing to be pistonic, I think each transducer should behave pistonically ( over the chosen operating bandwidth ). IMHO the primary reason that different driver materials sound the way they do is due to their breakup behaviour when they are no longer behaving like pistons. When they are not pistonic they become modal, like a DML.
Regarding best panel materials; internal damping is the killer of HF so softer cores are poor. The best foam cores, such as Rohacell, have very high compressive strength and work well. Honeycombe cores ( aluminium and Nomex ) are better still. Amina and Podium use honeycombs for a reason; I've met Shelley Katz from Podium several times and did some work for him early in the 'layered sound' project; the big Podium products sound great and have received some great reviews. They are very simple things though. High quality Nomex honeycomb cored panels and 4 decent tulip exciters; exciter technology has since moved on and, with a little research, its possible to source similar quality panel materials too.
DMLs modal panel behaviour cause them to be very labour intensive to measure. Small microphone position and/or distance changes can have significant changes in the measurement. It is conventional to measure a moving coil at 1W/1m somewhere between HF and mid driver axes, this will only give you a small snapshot of a DMLs true output. A polar plot or some form of averaging is needed to really see how a panel is behaving.
« Reply #3047 on: 21 Oct 2017, 07:46 pm »
Quote from: Bendingwave on 19 Oct 2017, 01:15 am
For a DML the standard industrial grade (IG ) is fine; the numbers in their product range identify are the material's density. 31 IG is the lowest density and therefore the highest sensitivity for 'our' application. IG51 would fine too. There are many other low mass thermal insulation foams which can work extremely well too.......unfortunately it is often difficult to find them in the optimum thicknesses though. 4-6mm is optimal for large full range panels.
There are alternative honeycomb panels which cost less than similar Nomex and Aluminium alternatives.
http://www.coretexgroup.co.uk/Honeycombcore.html
Amina's panel standard panel size dictates the response at the low end. There is a high pass to protect the panels and improve power handling; low end is deliberately limited to -3dB at 90Hz. The exciters used are Amina's own designs, are assembled in-house and they aren't available to the public. The specific HF exciter extends HF to above 40Khz. As I mentioned in earlier, excessive internal damping i.e soft cores are the killer of HF.
http://aminasound.com/file-directory/product/loudspeakers/mobius/datasheet/Mobius5_Datasheet.pdf
I don't have any white papers, sorry.
« Reply #3049 on: 24 Oct 2017, 07:18 am »
Quote from: Bendingwave on 23 Oct 2017, 01:24 am
The best PMI foams like Rohacell are still behind honeycombs when is comes to compreesion strength; consequently their HF rolls off. Amina's HF exciter does not extend to 40kHz on a Rohacell panel.
Rohacels sound has slightly detail relative to the honeycomb cores; its compressive strenght is not as good. The sound signature of a panel cannot be attributed to the choice of core material alone; a huge number of factors contribute. Core, skin and adhesive all make significant differnces as do the size, mass, aspect ratio, exciter type and exciter position on the panel, suspension method, edge treatment. etc . It would be possible to make a honeycomb panel less detailed than a Rohacell if you deliberately wanted to engineer it to behave that way.
Amina chose the best panel materials they could find and then used some tricks to further enhance and response shape the panels response. These tricks were first developed by NXT and were used in most of Mission's DML products; response shaping can easily be achieved by using a small miniDSP 2x4 today but it is possible to achieve a very flat in room acoustical power response by using this response shaping method. DMLs are already very good in this respect and imho this is why DMLs sound appeals so much.
« Reply #3050 on: 24 Oct 2017, 07:36 am »
Quote from: Bendingwave on 23 Oct 2017, 01:24 am
Apologies, I forgot a couple of points you raised. The coretex panels' compressive strength is behind the aluminium and Nomex but not by much; the coretex type panels had Mylar skins with a thinn damping scrim in between. I thought they sounded great. Nomex core with Mylar type skins ( Podium ) is excellent and probably my favourite subjectively.
Amina's panels are designed to work in wall, between studs. The size restriction limits the number of modes at the low end; there would be very poor perfomance between 55Hz and 90Hz on a panel that size. Consequently they need subs to work at there best; taking LF away froma DML is always a good thing. I developed the ALF80 and ALF120 subs when I was working there. The 80 is a conventional bandpass but the 120 was fun to do and has some unique stuff going on inside. A home cinema setup with 5 or 7 Amina panels with a 120 is genuinely full range system and sounds great. http://www.insideci.co.uk/reviews/amina-alf120-in-wall-subwoofer-review.aspx
« Reply #3054 on: 30 Oct 2017, 07:32 am »
Quote from: Bendingwave on 24 Oct 2017, 08:18 am
The adhesives used for commercial panel construction are typically supplied in sheet form and require the correct combination of heat, time and pressure during the curing process. Different adhesives are needed for different cores and skins; for example aluminium oxidises in air very quickly and becomes difficult to bond to and some 'polymer' skins have low surface energy and are inherently difficult to bond. This link explains in more detail https://compositesuk.co.uk/system/files/documents/Adhesive bonding of composites_0.pdf
Regarding mylar sheet for skins
http://www.katco.eu/uk/site__1344/
https://www.ebay.co.uk/i/1620591235...=1007017&device=c&campaignid=856243387&crdt=0
« Reply #3195 on: 9 Jan 2020, 07:07 am »
Not visited the thread for quite a while so just catching up......
Unless the exciter design is very poor, a DML's HF extension is determined by the panel material's compressive strength and internal damping properties, not the exciter performance. Aluminium honeycomb panels are excellent and Nomex honeycombs are also a good compromise for high comp strength and low panel mass. The low panel mass of, paper skinned, aluminium honeycomb panels gives 93/94dB @1W sensitivity when used with a decent exciter. Aluminium honeycombs with high quality skins give flat HF extension to 25kHz and decent output above 30kHz.
DML panels have to be VERY big to make bass flat bass below 150Hz. Panels are packed with modes at higher frequencies but there are only 3 fundamental modes at low frequencies ( length, width and the lowest diagonal mode ). Panels are sparsely modal at LF and densely modal at HF.......consequently LF performance is seriously lumpy.
When developing a DML, try measuring the off axis polar response and at multiple different mic distances too; DMLs are strange things and don't behave anything like conventional moving coils. DMLs are simultaneously in phase and out of phase with themselves at the same time!! Small changes in mic position can have a huge effect on measured performance.
https://www.audiocircle.com/index.php?topic=70541.3000
Re: NXT.......rubbish??....THINK AGAIN!
« Reply #3005 on: 27 Sep 2017, 07:57 pm »
Hello all........what a fascinating thread.
Its interesting to see just how far it is possible to get using empirical development. I was first involved with NXT in 1997 and 1998 when I attended their training courses in Huntingdon. In 1999 I moved to Huntingdon and worked for an injection moulding company who carried out most of NXT's & Mission's prototyping work; Mission / NXT were owned by the same people at the time. The moulding company also supplied various components to Amina technologies and in 2012 I began work in their design department.
The best impirical experimenters on this thread have reached similar conclusions to some extremely sophisticated FEA modelling. When understanding DML behaviour its best not to compare it with conventional moving coil speakers at all. Moving coil drivers are pistonic or at least they aspire to be whereas DMLs are incoherent devices with multiple areas of the panel being in or out of phase with each other at any given time. DMLs oftem struggle at high volumes or when there is a lot of LF in the programme material. This is because the nodal behaviour of the panel, at the exciter drive point, is often very different / contradictory to the input signal being fed to that exciter. Hope that makes sense.......
Panel suspension / termination is important too.........less is more. Good old fashion elastic bands are a simple and very effective solution.
If you keep volumes levels sensible and/ or keep extreme LF away from the panel then they can be superb, just like any well designed cabinetless speaker ( electrostics, ribbons, Linkwitz dipoles etc etc )
« Reply #3041 on: 13 Oct 2017, 09:33 am »
Quote from: Bendingwave on 30 Sep 2017, 09:41 pm
Thanks for the welcome(s). Apologies for the delay responding.
For the sake of discussion, we can think of a moving coil driver's voice coil and an exciter's voice coil behaving in the same way and responding faithfully to the AC applied at its input. The cone motion on an 'ideal' moving coil driver would faithfully reproduce the input AC signal and apply pressure ( or not ) to the air in front of the cone. State of the art moving coil drivers from people like Accuton, Raidho, SEAS, etc variously use metal alloys, ceramics diamond and are notionally pistonic at the lower end of their bandwidth; they all cease to be pistons at the high end and become non linear. The point at which this happens is easy to determine by reviewing their CSD or distortion characteristics. A good old impedance curve also gives away many clues about non linearity.
Regarding your point about bass drivers needing to be pistonic and midranges/tweeters not needing to be pistonic, I think each transducer should behave pistonically ( over the chosen operating bandwidth ). IMHO the primary reason that different driver materials sound the way they do is due to their breakup behaviour when they are no longer behaving like pistons. When they are not pistonic they become modal, like a DML.
Regarding best panel materials; internal damping is the killer of HF so softer cores are poor. The best foam cores, such as Rohacell, have very high compressive strength and work well. Honeycombe cores ( aluminium and Nomex ) are better still. Amina and Podium use honeycombs for a reason; I've met Shelley Katz from Podium several times and did some work for him early in the 'layered sound' project; the big Podium products sound great and have received some great reviews. They are very simple things though. High quality Nomex honeycomb cored panels and 4 decent tulip exciters; exciter technology has since moved on and, with a little research, its possible to source similar quality panel materials too.
DMLs modal panel behaviour cause them to be very labour intensive to measure. Small microphone position and/or distance changes can have significant changes in the measurement. It is conventional to measure a moving coil at 1W/1m somewhere between HF and mid driver axes, this will only give you a small snapshot of a DMLs true output. A polar plot or some form of averaging is needed to really see how a panel is behaving.
« Reply #3047 on: 21 Oct 2017, 07:46 pm »
Quote from: Bendingwave on 19 Oct 2017, 01:15 am
For a DML the standard industrial grade (IG ) is fine; the numbers in their product range identify are the material's density. 31 IG is the lowest density and therefore the highest sensitivity for 'our' application. IG51 would fine too. There are many other low mass thermal insulation foams which can work extremely well too.......unfortunately it is often difficult to find them in the optimum thicknesses though. 4-6mm is optimal for large full range panels.
There are alternative honeycomb panels which cost less than similar Nomex and Aluminium alternatives.
http://www.coretexgroup.co.uk/Honeycombcore.html
Amina's panel standard panel size dictates the response at the low end. There is a high pass to protect the panels and improve power handling; low end is deliberately limited to -3dB at 90Hz. The exciters used are Amina's own designs, are assembled in-house and they aren't available to the public. The specific HF exciter extends HF to above 40Khz. As I mentioned in earlier, excessive internal damping i.e soft cores are the killer of HF.
http://aminasound.com/file-directory/product/loudspeakers/mobius/datasheet/Mobius5_Datasheet.pdf
I don't have any white papers, sorry.
« Reply #3049 on: 24 Oct 2017, 07:18 am »
Quote from: Bendingwave on 23 Oct 2017, 01:24 am
The best PMI foams like Rohacell are still behind honeycombs when is comes to compreesion strength; consequently their HF rolls off. Amina's HF exciter does not extend to 40kHz on a Rohacell panel.
Rohacels sound has slightly detail relative to the honeycomb cores; its compressive strenght is not as good. The sound signature of a panel cannot be attributed to the choice of core material alone; a huge number of factors contribute. Core, skin and adhesive all make significant differnces as do the size, mass, aspect ratio, exciter type and exciter position on the panel, suspension method, edge treatment. etc . It would be possible to make a honeycomb panel less detailed than a Rohacell if you deliberately wanted to engineer it to behave that way.
Amina chose the best panel materials they could find and then used some tricks to further enhance and response shape the panels response. These tricks were first developed by NXT and were used in most of Mission's DML products; response shaping can easily be achieved by using a small miniDSP 2x4 today but it is possible to achieve a very flat in room acoustical power response by using this response shaping method. DMLs are already very good in this respect and imho this is why DMLs sound appeals so much.
« Reply #3050 on: 24 Oct 2017, 07:36 am »
Quote from: Bendingwave on 23 Oct 2017, 01:24 am
Apologies, I forgot a couple of points you raised. The coretex panels' compressive strength is behind the aluminium and Nomex but not by much; the coretex type panels had Mylar skins with a thinn damping scrim in between. I thought they sounded great. Nomex core with Mylar type skins ( Podium ) is excellent and probably my favourite subjectively.
Amina's panels are designed to work in wall, between studs. The size restriction limits the number of modes at the low end; there would be very poor perfomance between 55Hz and 90Hz on a panel that size. Consequently they need subs to work at there best; taking LF away froma DML is always a good thing. I developed the ALF80 and ALF120 subs when I was working there. The 80 is a conventional bandpass but the 120 was fun to do and has some unique stuff going on inside. A home cinema setup with 5 or 7 Amina panels with a 120 is genuinely full range system and sounds great. http://www.insideci.co.uk/reviews/amina-alf120-in-wall-subwoofer-review.aspx
« Reply #3054 on: 30 Oct 2017, 07:32 am »
Quote from: Bendingwave on 24 Oct 2017, 08:18 am
The adhesives used for commercial panel construction are typically supplied in sheet form and require the correct combination of heat, time and pressure during the curing process. Different adhesives are needed for different cores and skins; for example aluminium oxidises in air very quickly and becomes difficult to bond to and some 'polymer' skins have low surface energy and are inherently difficult to bond. This link explains in more detail https://compositesuk.co.uk/system/files/documents/Adhesive bonding of composites_0.pdf
Regarding mylar sheet for skins
http://www.katco.eu/uk/site__1344/
https://www.ebay.co.uk/i/1620591235...=1007017&device=c&campaignid=856243387&crdt=0
« Reply #3195 on: 9 Jan 2020, 07:07 am »
Not visited the thread for quite a while so just catching up......
Unless the exciter design is very poor, a DML's HF extension is determined by the panel material's compressive strength and internal damping properties, not the exciter performance. Aluminium honeycomb panels are excellent and Nomex honeycombs are also a good compromise for high comp strength and low panel mass. The low panel mass of, paper skinned, aluminium honeycomb panels gives 93/94dB @1W sensitivity when used with a decent exciter. Aluminium honeycombs with high quality skins give flat HF extension to 25kHz and decent output above 30kHz.
DML panels have to be VERY big to make bass flat bass below 150Hz. Panels are packed with modes at higher frequencies but there are only 3 fundamental modes at low frequencies ( length, width and the lowest diagonal mode ). Panels are sparsely modal at LF and densely modal at HF.......consequently LF performance is seriously lumpy.
When developing a DML, try measuring the off axis polar response and at multiple different mic distances too; DMLs are strange things and don't behave anything like conventional moving coils. DMLs are simultaneously in phase and out of phase with themselves at the same time!! Small changes in mic position can have a huge effect on measured performance.
Use an M3 screw in that driver. The thread starts halfway down into the hole.
Jaxboy.
I stuck this on years ago, rather quickly, so it doesn't look that great, but it has worked perfectly for all these years.
As you can see from this photo I left gaps in the glue to allow The exciter to vent out.
if you place the canvas on a flat surface when interesting the glue filler the exciter will be fixed in the correct position when dry.
the canvas also has quite a bit of play in it, so the coil will be in its ideal place.
It will be difficult to adjust and setup if you use the bolt hole.
Hope this helps.
Steve.
I stuck this on years ago, rather quickly, so it doesn't look that great, but it has worked perfectly for all these years.
As you can see from this photo I left gaps in the glue to allow The exciter to vent out.
if you place the canvas on a flat surface when interesting the glue filler the exciter will be fixed in the correct position when dry.
the canvas also has quite a bit of play in it, so the coil will be in its ideal place.
It will be difficult to adjust and setup if you use the bolt hole.
Hope this helps.
Steve.
Attachments
Maybe the high-tensile M3 hex-cap screws I used did cut their own threads in that case 🤦♂️.Maybe those four ears are supposed to be screwed on to a mounting plate?
Hello Jaxboy
If it can give ideas, here is a picture of my last panels (rear view)
The exciter (DAEX25FHE) is mounted on a square plywood with its 4 peripheral holes and screws (model for wood).
Then this assembly is glued (original 3M tape) to the membrane (be careful here to keep the right orientation for the square piece, the next step allows some error)
Then screw the 2 spines on the frame (see one of the screws on the picture, bottom right)
Then put silicon on the sides of the square (upper, lower side) in contact with the spines. Do to that, the panel was horizontal on my bench (the membrane facing down) with some blocks between the membrane and the bench to keep the exciter at its natural position.
Christian
PS : risk of that = to have some noise from loose screws. In my plywood panels, it is about same principle but with bolts. I glued 3mm diameter nuts on the exciter (same 4 holes) to get threads. The advantage is then the screw comes from the rear direction (not from the membrane like in the picture) so you can screw them again if necessary.
homeswinghome,
So you did not glue a small thin rounded piece of wood to the membrane, and instead made a frame for attaching the body of the exciter to the spline? It seems that you put the nuts on top of the mounting holes and the frame on top of the nuts. Is that correct? Wouldn't it be simpler to just glue the body to the spline and eliminate the nuts and frame? From what I've seen on here, the small rounded thin wooden piece glued to the membrane is instrumental in achieving better tone and low frequency extension. Why did you not do that? How is the freq response of your DML? Please do not take any part of these questions as denigrating your design.
So you did not glue a small thin rounded piece of wood to the membrane, and instead made a frame for attaching the body of the exciter to the spline? It seems that you put the nuts on top of the mounting holes and the frame on top of the nuts. Is that correct? Wouldn't it be simpler to just glue the body to the spline and eliminate the nuts and frame? From what I've seen on here, the small rounded thin wooden piece glued to the membrane is instrumental in achieving better tone and low frequency extension. Why did you not do that? How is the freq response of your DML? Please do not take any part of these questions as denigrating your design.