well it would make things easier ? you have 2 coils solder one connection and ur done. you covered everything and both connection would be on the outside (thats a benefit) and add to the output instead of been dead weight everything is actually driven. so unlike adding foil that does not do a thing., in this case even the useless foil in between is driven if you like. you just have to shift is a tiny bit.
i would not go for 30 mic if you would make something like this unless it would be bass... so 12-15 would be a better choice.
and my problem is... i only got 30 mic as thinnest with adhesive for now... i do have 15 + adhesive but only 40mm wide.. with makes it hard if not impossible to make such a thing. i am still waiting on my 30 cm wide roles... import became a bitch... i need a Eori witch i dont have... so i guess they will be send back. and they send it again without my VAT number... it becomes like 4 rolls of 30 cm for only 500 euro... well lets say we have a heart for the hobby 🙂
i would not go for 30 mic if you would make something like this unless it would be bass... so 12-15 would be a better choice.
and my problem is... i only got 30 mic as thinnest with adhesive for now... i do have 15 + adhesive but only 40mm wide.. with makes it hard if not impossible to make such a thing. i am still waiting on my 30 cm wide roles... import became a bitch... i need a Eori witch i dont have... so i guess they will be send back. and they send it again without my VAT number... it becomes like 4 rolls of 30 cm for only 500 euro... well lets say we have a heart for the hobby 🙂
Last edited:
But based on my previous measurements, I do have to add the undriven foil, or else the membrane is not stiff enough. If I used a stretched membrane like you linked then no problem, the tension takes care of that and I can have raw kapton or mylar in the space between the traces.
With an untensioned corrugated membrane, I have found that unless the membrane is stiff, which I have made it so far by having undriven aluminum areas, then distortion is way worse and on top of that the top end frequency response is worse too.
you dont need to use undriven foil. as far as i know, it just needs any foil driven or not to make the stiffness work. why not drive it as well.
its the same like magnepans with rubber magnets. the width of the magnet determent what high frequencies it can produce due to break up (trace or wire distance to trace or wire) using smaller magnets is one option, or use more wide coil at the cost of efficiency since halve the coil is not in a good field. or fill it up with alu that does not do anything... adding even less..
you could even change the width on each side to make one dominant , wso i you want one coil that has the perfect fiel to do more make them smaller. and the filler wider. in this case the wider coil is fed by the smaller one in a good field and will dominate. this is done in ribbons. often the wider traces are on the outside near the magnets where the field is stronger. to make the whole more even. (so wider traces are driven less hard)
i am not sure if you mean the undriven area is needed or the weight is needed there for stiffness. it might as well be driven intsead of been dead weight. dead weight is never good i think if you can avoid it
by the way we should talk once on Skype of Fb or Discord, same goes for Solhaga if he wants. typing is rather slow and we might be able to help each other out allot in our hobby !
its the same like magnepans with rubber magnets. the width of the magnet determent what high frequencies it can produce due to break up (trace or wire distance to trace or wire) using smaller magnets is one option, or use more wide coil at the cost of efficiency since halve the coil is not in a good field. or fill it up with alu that does not do anything... adding even less..
you could even change the width on each side to make one dominant , wso i you want one coil that has the perfect fiel to do more make them smaller. and the filler wider. in this case the wider coil is fed by the smaller one in a good field and will dominate. this is done in ribbons. often the wider traces are on the outside near the magnets where the field is stronger. to make the whole more even. (so wider traces are driven less hard)
i am not sure if you mean the undriven area is needed or the weight is needed there for stiffness. it might as well be driven intsead of been dead weight. dead weight is never good i think if you can avoid it
by the way we should talk once on Skype of Fb or Discord, same goes for Solhaga if he wants. typing is rather slow and we might be able to help each other out allot in our hobby !
Last edited:
I did test that, here is efficiency of 5-9 mm wide traces:
So in terms of effiicency, there is minimal difference between 6 mm wide traces with undriven areas, and 9 mm wide traces with no undriven areas since the magnetic field is stronger in the middle of the magnet gap.
But as I wrote before while the efficiency great, and the distortion is great also as long as I don't push the xmax. When I push the xmax it does not perform well at all since the magnetic field is not that linear, thus tall order harmonic distortion skyrockets and it sounds bad. Technically I could just accept this and cross higher such that I use less xmax, but I want to build a midrange + tweeter capable capable planar so I want to be able to cross as low as possible.
Solhaga did great simulations in this post, and my measurements confirm his simulations:
This simulation especially which plots the difference in magnetic field strength, across +- 0.5 mm xmax, with different trace widths. The orange line, which hugs the 0 line so closely, is the 3 mm trace width. And it performs great in practice.
So as long as I have a trace / coil width of 3 mm, then there is minimal distortion due to differences in the magnetic field throughout the stroke length. But then I have to get creative since I still want good efficiency.
single sided ? so there is whatever you chose in between as separation lets say 0.5 mm or whatever. why fill in the backside ? if thats not a problem why use 2 alu foils.
On top of a magnet you can leave it open like BG, but if you want a flat top end (but less efficiency) you would not. you can either do this with useless weight. or minimize that.
i dont know why 3 mm traces is key. it depends on the the gap. and if more are used the distance between them. lets not forget what ever impedance you want.
if 9mm traces did not make a difference why back the foil completely ? no need for that extra weight i would say
Last edited:
Yep, single sided. And all area outside of 0.5 mm from the traces was undriven fill. The yellow is the driven traces, green is the undriven fill and the grey transparent rectangles are the magnets.
Exactly.
I looked back in my measurement folder and found distortion curves of 6 mm wide traces, one with kapton + undriven areas and one with mylar but empty mylar between the magnets, and only a small border to make it stiff at the suspension.
Here is a photo of the membrane with the empty mylar areas + border:
I'll have to backtrack on my saying distortion was way worse, in fact it was comparable. Where the undriven and exposed mylar membrane fell apart was in the high frequency response.
Sadly I can't find the measurements, looks like I didn't save them 😢 but the problem was that the 0-90 polar response was in no way smooth, above 5 khz. In fact, it was all over the place and very inconsistent. You can see part of the notch on the top end but the polar responses were way way worse than just that little notch. And since part of my requirement is that it also has to be a good tweeter, I discarded that idea.
what i dont get is, you use traces on one side and only a backing mylar or kapton(like everyone does), or another foil on the back ? to cover the rest on the back side as you suggested.
i myself for tweet/mid i would say single ended and use foil thick enough to retain corrugation .. and cover as much as possible.. so going over the magnet and leave a 1-2 mm space above the middle of the magnet. why ? since the top end looks better at the cost of output. yours falling off at 10khz. if you want pure mid range... you could do exactly what i mentioned above. making it heavy and still use all the weight to drive it. since that hurts the top end but for a mid... who cares it does not reach 6-10khz.. any ribbon or whatever tweeter can pick it up from there. you dont even have to cover as much as possible but chose whats most efficient. since the break up you get is above of where you would use it. and get more efficiency
sorry just read it was gone be a CBT, is that still the plan ?
i myself for tweet/mid i would say single ended and use foil thick enough to retain corrugation .. and cover as much as possible.. so going over the magnet and leave a 1-2 mm space above the middle of the magnet. why ? since the top end looks better at the cost of output. yours falling off at 10khz. if you want pure mid range... you could do exactly what i mentioned above. making it heavy and still use all the weight to drive it. since that hurts the top end but for a mid... who cares it does not reach 6-10khz.. any ribbon or whatever tweeter can pick it up from there. you dont even have to cover as much as possible but chose whats most efficient. since the break up you get is above of where you would use it. and get more efficiency
sorry just read it was gone be a CBT, is that still the plan ?
Last edited:
Up until pretty recently, all my membranes were of type 1, which by convention is the default. I copied this from your excellent videos when you did corrugated membranes so everything cut form a single sheet of aluminum, glued to the mylar / kapton and then done.
2. was a test to see if I can have 3 mm wide traces without murdering efficiency. My solution was to run double coils which in practice is twice the driven alu thickness relative to the outer filled areas. The currently best performing of this type has 20 um driven traces on both sides, and 20 um undriven alu fill. All glued with the 3M 77 spray.
3. is when I got the thought, hey, double sides are complicated, what if I just double the trace thickness on one side? I believe this will measure pretty much the same as 2. The main benefit would be easier manifacturing since there is less to align and with only 1 side I can piggy back on the kapton pre applied adhesive.
And 4. is the idea bounced by me and solhaga where what if we dont cut any gaps for the traces in the undriven aluminum foil but rather just place it on the other side? Not having the gaps makes the foil a lot stiffer which reduces distortion. And with no cuts on the thinner (and easier to rip) aluminum foil for the undriven areas, the manifacturing will be greatly simplified. This would probably be a bigger benefit for solhaga than me since he uses crazy thin aluminum that will rip if you breathe on it 😆. But even jokes aside this one will have nothing to align either, since one side is just a completely filled plane, so no risk of shorting the traces to the fill if I apply the traces slightly too much in the wrong direction.
Oh yeah, still a CBT. And I still want it to be a full range (from 300-400 and upwards) driver so no tweeter. After I nail down what specific membrane layout I will have, alu foil thicknesses and glues and such, the next step is to calculate the trace coil routing such I have shading, in the membrane.
The problem with a dedicated tweeter is that if I cross at 3-5 khz, then I will have lobing in the off axis response. Even 1 khz is pushing it, and if I have a tweeter that can play down to 1 khz, why not just go all the way and make it play down to 400 hz and skip the lobing problems altogether?
I was in the process of calculating and modeling that but then I tested 3 mm wide traces which measured better than I thought. I then decided to postpone the shading network until I figure out the membrane stuff, figure out exactly if I go single or double sided coils, single our double coils on each side and so on. Especially since how I wire the coils and if they are double sided or not greatly impact how I should wire the shading.
Oh i might miss understood your method then !! sorry 🙁 so you had a dual coil. i mean driven coil on the backside !! wehh my idea would be Nr 2 but then use more traces. and have front and back overlap. in this way you do not have useless weight ! also since you plot you do not have adhesive everywhere. but i get its easier with the drawing.. lets be fair you even take the time to make one so kudo's to you !! musdt say al of your posts so far look really well done !!
so lets say you use the same adhesive alu, you can make multiple traces that are a bit smaller covering more and you can dismiss the fillers.. not even sure needed the backsides of the foil as well... then again to me its a question how much that helps... compared to the amount of work. 🙂.
Yeah going cbt you going musch smaller foil like 1/5 the foil (or smaller)? or gone do the same i did, like with magnets ? less field more shading ? etc. i mean it works... but you dont want high eff since that wont happen of course with all the -12 -9 -6 -3 db parts 🙂
That is where the clever tricks come in 🙂
I think I can route the coils such that it is space inefficient, 150 cm of planar will have the total output of a 100 cm planar, but the efficiency per watt will be the same as said 100 cm planar within -0.5 dB or so, and probably even less of an efficiency hit than that if my calculations are correct.
I am looking at tricks like having multiple coils, but not running all coils all the way to the top of the speaker. Since the force is the current through the field, less coils means less force thus less output which equals shading but with practically no efficiency loss.
Here is a shortened model I did when I assumed I would have 6 mm wide traces where I can fit 4 individual traces at the bottom. When I get to the top where I only have a single trace then I need cut in parallel resistors but there I think I have solutions to waste as little output as possible by routing the resistors in clever ways.
Thats a super good idea !!!, its one of those why did not i think of it !!! nice talking to you as well !!! have a nice new years eve everyone. and i will follow this for sure !
its a bit like my idea with a ribbon, except the other way around 🙂 more coil less width higher output vs wide single coil on the other end 🙂 instead of less magnetic field 🙂
Since I have to wait for my next roll of tape with actual 50 um aluminum foil thickness, I have time on my hands. And I believe that I can use 30 um aluminum foil as as a substitute for a set of tests.
What I want know is how should I layout the kapton and the undriven alu?
A is the conventional setup where the driven traces are in between the undriven fill, with some small gaps.
B is the idea of having undriven fill over the whole plane, and placing the driven traces on the bottom.
C is same as B, but with the unconventional thought that hey, what if we only use the kapton as an electrical insulator and remove it elsewhere to save weight?
D was my even crazier idea that maybe the acrylic adhesive is enough insulation to not short the driven traces to the fill. And I tried it and no, it doesn't work. It most definently does short to the fill. The following membranes would have the same resistance if it didn't short, but in the pure aluminum membrane on the left the resistance is way too low.
My plan is to build A, B, C in 2 versions each, 14 um undriven alu and 14 um undriven alu, all with identical self adhesive 30 um alu driven traces.
I will then take the best performing variant and reproduce it with the 50 um self adhesive aluminum tape and compare that one to my previously best performing membrane which is a dual sided coils of 20 um alu, 20 um undriven alu with gaps around the coils on one side.
I tried to cut some kapton for variant C, it is quite fragile but it does cut. It was a nightmare to weed though
But I succeeded! Or at least I would have if I didn't weed the wrong parts 😆
I decided to call it a day after that...
I don't expect to make any progress tomorrow since I have to work but I plan to continue on tuesday and re cut some kapton. Although to be honest I hope that B turns out the best because it is the easiest to manifacture. Cutting and weeding the kapton needed for C is very time consuming and it is extremely fragile. To make C feasible to manifacture in large membranes I'd probably have to buy kapton with pre applied adhesive and a release liner, but that is even more expensive than plan tape or non adhesive kapton.
What I want know is how should I layout the kapton and the undriven alu?
A is the conventional setup where the driven traces are in between the undriven fill, with some small gaps.
B is the idea of having undriven fill over the whole plane, and placing the driven traces on the bottom.
C is same as B, but with the unconventional thought that hey, what if we only use the kapton as an electrical insulator and remove it elsewhere to save weight?
D was my even crazier idea that maybe the acrylic adhesive is enough insulation to not short the driven traces to the fill. And I tried it and no, it doesn't work. It most definently does short to the fill. The following membranes would have the same resistance if it didn't short, but in the pure aluminum membrane on the left the resistance is way too low.
My plan is to build A, B, C in 2 versions each, 14 um undriven alu and 14 um undriven alu, all with identical self adhesive 30 um alu driven traces.
I will then take the best performing variant and reproduce it with the 50 um self adhesive aluminum tape and compare that one to my previously best performing membrane which is a dual sided coils of 20 um alu, 20 um undriven alu with gaps around the coils on one side.
I tried to cut some kapton for variant C, it is quite fragile but it does cut. It was a nightmare to weed though
But I succeeded! Or at least I would have if I didn't weed the wrong parts 😆
I decided to call it a day after that...
I don't expect to make any progress tomorrow since I have to work but I plan to continue on tuesday and re cut some kapton. Although to be honest I hope that B turns out the best because it is the easiest to manifacture. Cutting and weeding the kapton needed for C is very time consuming and it is extremely fragile. To make C feasible to manifacture in large membranes I'd probably have to buy kapton with pre applied adhesive and a release liner, but that is even more expensive than plan tape or non adhesive kapton.
Last edited:
So i received my tape... finally, if i measure it with adhesive its 48 micron. quite thick., i tried removing the adhesive and it was around 15 micron aluminium. not the 13 but still as light as one can find an alu tape. (halve the thickness i had before in larger rolls) its 30 cm wide so you can make more interesting foils. ill measure some 30 mic alu later this week. to see what the weight difference is. there is a point the glue is more heavy then the foil... i had a roll with the same thickness that i liked for tweeters but was only 40 mm wide. so a 300 mm wide roll will make new things possible.
then again super77 is thinner. but less reliable as well as more easy to peel; off. (and does not take heat as well) lets see what we can make with it. and if you need any let me know. it aint cheap though 🙁
then again super77 is thinner. but less reliable as well as more easy to peel; off. (and does not take heat as well) lets see what we can make with it. and if you need any let me know. it aint cheap though 🙁
Last edited:
Is there a reason you are not using 30NF which I saw you use in a previous video? It looked like it worked great, but maybe there are problems with that adhesive you have found?
its hard to get reall thin, and surface tension might be a problem if thinned , i use 30Nf for all other things. mylar to frame adhesion etc. its possible but everything needs to be really clear etc.. i found Super77 be rather forgiven. but thickness is also hard to controll 🙂
I have been inspired by @solhaga to order some EPDM strips, I should have them by the weekend so I can play around with them and see if I can echo his findings of lower distortion.
I have prepared some membranes from https://www.diyaudio.com/community/...o-be-used-in-a-dipole-cbt.412132/post-7885587, I will test them also during next weekend. All of the following membranes use 30 um self adhesive aluminum driven traces.
First conventional membrane with 14 um outer fill and gaps
Next conventional membrane with 20 um outer fill and gaps
Solid 14 um fill with driven traces on the rear. Heaviest of the bunch
Now the unconventional membranes. Solid 14 um aluminum fill but only kapton under the traces
And in 20 um alu
An interesting detail here is that the solid fill solid kapton does not corrugate as easily, they were all the same length before corrugating.
I have prepared some membranes from https://www.diyaudio.com/community/...o-be-used-in-a-dipole-cbt.412132/post-7885587, I will test them also during next weekend. All of the following membranes use 30 um self adhesive aluminum driven traces.
First conventional membrane with 14 um outer fill and gaps
Next conventional membrane with 20 um outer fill and gaps
Solid 14 um fill with driven traces on the rear. Heaviest of the bunch
Now the unconventional membranes. Solid 14 um aluminum fill but only kapton under the traces
And in 20 um alu
An interesting detail here is that the solid fill solid kapton does not corrugate as easily, they were all the same length before corrugating.
The new tape has the correct thickness! 51 um foil + 33 um adhesive times two = 168 um, which is within the margin of error of my measurement.
I might as well cut some membranes with it too such that when I measure, I can measure them all.
I also bougt some EPDM P shaped sealing profiles of the same type that @solhaga used. It is 9 mm wide with a P profile. When I cut away the flat part such that only the D is left, turns out the D is pretty much exactly 6 mm wide which is awesome because it is exactly the width I want, and it is very easy to cut consistently.
If this EPDM suspension turns out as good as solhaga found it to do, amazing! But even if it is comparable then the ease of manifacturing would be benefits enough for me to use it. Even better is that the adhesive does not cover all the way to the edge, so I think it would be very easy to apply with the adhesive without having to worry about catching the magnets with the adhesive and making it not lie flat, which was a problem with the foam strip suspension.
And since then I would not need the plastic edge parts to hold the suspension in place, I could simplify the design and not split it top and bottom but print in a solid piece.
That would also let me be clever and help solve a small problem: When I mount the membrane it doesn't always want to lie exactly in the middle, where I want it but is slightly biased to either side. With a solid edge I could add small cylindrical holes, just large enough to let me push in small T shaped profiles to help push the membrane to the middle and keep it there when tightening the screws. And then when all the screws are tightened then I just pull them out since they are not needed anymore.
I might as well cut some membranes with it too such that when I measure, I can measure them all.
I also bougt some EPDM P shaped sealing profiles of the same type that @solhaga used. It is 9 mm wide with a P profile. When I cut away the flat part such that only the D is left, turns out the D is pretty much exactly 6 mm wide which is awesome because it is exactly the width I want, and it is very easy to cut consistently.
If this EPDM suspension turns out as good as solhaga found it to do, amazing! But even if it is comparable then the ease of manifacturing would be benefits enough for me to use it. Even better is that the adhesive does not cover all the way to the edge, so I think it would be very easy to apply with the adhesive without having to worry about catching the magnets with the adhesive and making it not lie flat, which was a problem with the foam strip suspension.
And since then I would not need the plastic edge parts to hold the suspension in place, I could simplify the design and not split it top and bottom but print in a solid piece.
That would also let me be clever and help solve a small problem: When I mount the membrane it doesn't always want to lie exactly in the middle, where I want it but is slightly biased to either side. With a solid edge I could add small cylindrical holes, just large enough to let me push in small T shaped profiles to help push the membrane to the middle and keep it there when tightening the screws. And then when all the screws are tightened then I just pull them out since they are not needed anymore.