Again, it was TNT that pushed me to re-evaluate the EPDM sealing strips!
And again again, I really love the way we are pushing each other in these threads.
And again again, I really love the way we are pushing each other in these threads.
So do I!
I have now cut 3 membranes with the 51 um aluminum self adhesive with 33 um acrylic glue.
1. First the traditional membrane: 51 um traces & 20 um for the fill outside the traces.
2. Next one of the weird ones. A solid sheet of 20 um backing with kapton just under the traces. 50 um alu for the traces of course.
3. And then the easiest one to manifacture. A complete layer of 12 um self adhesive kapton, with the glue sticking it to a solid sheet of 14 um alumium. Then cut traces of the self adhesive 50 um alu glued to the back. Because only the traces are cut, and the kapton + 14 um alu can be bigger and cut afterwards, this one has centimeters of margin rather than fractions of a milimeter, so very easy to make.
I also did some uh... very scientific audio tests of the membrane noise:
I am not sure it actually means anything, but I think the 3. has the least noise when flapping. Number 2. has quite a high pitch too. Although I have notices that after corrugating, they don't make the same kind of noise when I flap them around.
Tests of the foam vs the EPDM profile:
First of all it is very easy to cut consistently. I run a ball point pen in the groove where I want to cut for extra visibility, then cut it off with a scissor. Worked great!
But now for what you are actually interested in, the performance!
All measurements use the samee membrane, which is kapton tape + solid 14 um alu & traces from self adhesive aluminum applied to the bottom. I.e. number 4 from this image:
I have mounted the same membrane in first in red with a foam surround ,then with the EPDM rubber surround in green.
But now first the efficiency. Both are fed the same signal.
I can echo @solhaga s findings in that the efficiency is a lot higher, although my 8khz+ efficiency is lower than his. But since most of the music is in the lower frequencies that is no problem, as long as the polar response is good which I will check tomorrow.
But distortion is very interesting. Keep in mind that while the response is EQed to not be a mountain curve, they are fed the same signal, which means that in on the low end, where the efficiency is 3 dB greater with the EPDM foam, even at this higher output, the distortion is lower! And not only that, the distortion profile is nicer since we are now 2nd harmonic dominant rather than 3rd harmonic dominant.
In conclusion, foam suspension is out and EPDM foam is in!
I started to measure the different membrane configurations but then I got a problem. Or rather I got a problem with the EPDM rubber that I could gloss over with the foam. On the top and bottom, where only the sides are supported by the suspension, I get buckling of the membrane:
This is obviously a problem, so to fix it I am printing some top and bottom caps with horizontal foam to lock the membrane in place. The plan is to add horizontal suspension like the red bordered strip in this image shamelessly stolen from solhagas thread:
But the keen eyed among you will think, but Ollboll, if you got problems with buckling and are printing the solution now, how did you get the working distortion measurement? And the answer is that that measurement was with the membrane type 3 with soild 14 um over the whole membrane, which makes it pretty stiff in the horizontal direction compared to the conventional which has gaps where there is only kapton. Stiff enough that I believe it performed great and could to a greater extent resist the buckling forces. But when I tested with a conventional membrane, type 1. in the image then I got serious problems where the edge buckled closer to the magnets. It still measured pretty good, but had a large distortion peak in the 300 hz range when pushing xmax.
While I think it is fixable it is still one more advantage of the more easy to manifacture membrane of type 3, since it lets me get away with more mistakes.
Tomorrow I hope to test my fixes for the buckling membrane and measure again to see the differences of the membrane types.
First of all it is very easy to cut consistently. I run a ball point pen in the groove where I want to cut for extra visibility, then cut it off with a scissor. Worked great!
But now for what you are actually interested in, the performance!
All measurements use the samee membrane, which is kapton tape + solid 14 um alu & traces from self adhesive aluminum applied to the bottom. I.e. number 4 from this image:
I have mounted the same membrane in first in red with a foam surround ,then with the EPDM rubber surround in green.
But now first the efficiency. Both are fed the same signal.
I can echo @solhaga s findings in that the efficiency is a lot higher, although my 8khz+ efficiency is lower than his. But since most of the music is in the lower frequencies that is no problem, as long as the polar response is good which I will check tomorrow.
But distortion is very interesting. Keep in mind that while the response is EQed to not be a mountain curve, they are fed the same signal, which means that in on the low end, where the efficiency is 3 dB greater with the EPDM foam, even at this higher output, the distortion is lower! And not only that, the distortion profile is nicer since we are now 2nd harmonic dominant rather than 3rd harmonic dominant.
In conclusion, foam suspension is out and EPDM foam is in!
I started to measure the different membrane configurations but then I got a problem. Or rather I got a problem with the EPDM rubber that I could gloss over with the foam. On the top and bottom, where only the sides are supported by the suspension, I get buckling of the membrane:
This is obviously a problem, so to fix it I am printing some top and bottom caps with horizontal foam to lock the membrane in place. The plan is to add horizontal suspension like the red bordered strip in this image shamelessly stolen from solhagas thread:
But the keen eyed among you will think, but Ollboll, if you got problems with buckling and are printing the solution now, how did you get the working distortion measurement? And the answer is that that measurement was with the membrane type 3 with soild 14 um over the whole membrane, which makes it pretty stiff in the horizontal direction compared to the conventional which has gaps where there is only kapton. Stiff enough that I believe it performed great and could to a greater extent resist the buckling forces. But when I tested with a conventional membrane, type 1. in the image then I got serious problems where the edge buckled closer to the magnets. It still measured pretty good, but had a large distortion peak in the 300 hz range when pushing xmax.
While I think it is fixable it is still one more advantage of the more easy to manifacture membrane of type 3, since it lets me get away with more mistakes.
Tomorrow I hope to test my fixes for the buckling membrane and measure again to see the differences of the membrane types.
Last edited:
Yup. That was also my idea but apparently my fingers didn't get the memo so they wrote foam instead of EPDM profile 😆
I tried out the top and bottom caps with the horizontal EPDM profiles and they worked great!
I will reuse the type numbers from this image, I am comparing them all.
First number 1. I could not get it working with the new EPDM profiles, even with the buckling fix at the edges. I get the a distortion peak, I think it is buckling along the driver and nost just at the edges since the EPDM foam is much stiffer and requires more force to push against.
I did not make any EQed measurements for number 2 so lets skip that one for now.
Number 3. Probably the best one so far.
Number 4. Interestingly it has more distortion than number 3. It is more efficient, but only below 700 hz.
In conclusion, I was afraid that number 4 would perform the best but it turns out number 3 has the lowest distortion. This is great because it is also by far the easiest one to manifacture! And interestingly enough number 3 is the exact configuration that solhaga now uses, although I have thicker aluminum foil for both the traces and the solid tructural side.
But then I decided to compare the new current best EPDM compatible membrane number 3 to the previously best foam compatible membrane which was a number 2. Note that both drivers are fed with a 300 (or was it 400?) hz LR4 highpass, so focus on the differences rather than the exact roll off. The foam is smoother on the top end, but the EPDM driver has a lot more efficiency in the 300 - 200 hz range. At 400 hz it is 9 dB more efficient! It tapers off of course but that is still huge, considering most of the sound is in the lower frequencies.
While I didn't do any EQed measurements, the difference is large enough that in the low frequencies, the efficiency is higher and the distortion is lower.
I also did a 0-90 measurement of the new leader, EPDM with membrane of type 3:
It's not as good as the previous membrane on the topmost end, but it is passable and probably good enough. Especially since the reduced low frequency distortion will probably let me cross at 300 hz LR4 without any problems when scaling upp to a full sized driver.
In conclusion I have decided to go forward with both the EPDM profiles and membrane of type number 3. So I need to order more 50 um self adhesive aluminum tape + a roll of 14 um alumium foil.
I will probably play around with some wool and damping material and see if I can smooth out the topmost end somewhat, remove the small 9 khz dip and get the response closer to that with the conventional membrane with foam surround but if I can't then that is fine. I am OK with not having the perfect planar tweeter, my full range heart is still beating so I value avoiding a crossover in the 800 - 3000 hz region higher than having perfect response at 8-20 khz.
I will reuse the type numbers from this image, I am comparing them all.
First number 1. I could not get it working with the new EPDM profiles, even with the buckling fix at the edges. I get the a distortion peak, I think it is buckling along the driver and nost just at the edges since the EPDM foam is much stiffer and requires more force to push against.
I did not make any EQed measurements for number 2 so lets skip that one for now.
Number 3. Probably the best one so far.
Number 4. Interestingly it has more distortion than number 3. It is more efficient, but only below 700 hz.
In conclusion, I was afraid that number 4 would perform the best but it turns out number 3 has the lowest distortion. This is great because it is also by far the easiest one to manifacture! And interestingly enough number 3 is the exact configuration that solhaga now uses, although I have thicker aluminum foil for both the traces and the solid tructural side.
But then I decided to compare the new current best EPDM compatible membrane number 3 to the previously best foam compatible membrane which was a number 2. Note that both drivers are fed with a 300 (or was it 400?) hz LR4 highpass, so focus on the differences rather than the exact roll off. The foam is smoother on the top end, but the EPDM driver has a lot more efficiency in the 300 - 200 hz range. At 400 hz it is 9 dB more efficient! It tapers off of course but that is still huge, considering most of the sound is in the lower frequencies.
While I didn't do any EQed measurements, the difference is large enough that in the low frequencies, the efficiency is higher and the distortion is lower.
I also did a 0-90 measurement of the new leader, EPDM with membrane of type 3:
It's not as good as the previous membrane on the topmost end, but it is passable and probably good enough. Especially since the reduced low frequency distortion will probably let me cross at 300 hz LR4 without any problems when scaling upp to a full sized driver.
In conclusion I have decided to go forward with both the EPDM profiles and membrane of type number 3. So I need to order more 50 um self adhesive aluminum tape + a roll of 14 um alumium foil.
I will probably play around with some wool and damping material and see if I can smooth out the topmost end somewhat, remove the small 9 khz dip and get the response closer to that with the conventional membrane with foam surround but if I can't then that is fine. I am OK with not having the perfect planar tweeter, my full range heart is still beating so I value avoiding a crossover in the 800 - 3000 hz region higher than having perfect response at 8-20 khz.
Last edited:
I just realized I picked the wrong membrane diagram image!
The correct image is:
Where:
1 -> A
2 -> A (but traces on both sides)
3 -> B
4 -> C
Where 3 / B performed the best. Also, B happens to be the one which is the easiest to manifacture so double yay!
The correct image is:
Where:
1 -> A
2 -> A (but traces on both sides)
3 -> B
4 -> C
Where 3 / B performed the best. Also, B happens to be the one which is the easiest to manifacture so double yay!
I just got one of my completely crazy ideas, and I can't get it out of my head! I think I might just have to build and try it out...
First the background as context:
I was looking at the 0-90 response of the current best membrane which has a solid backing of 14 um aluminum to add stability. This makes it very stiff which is nice but it has those two dips that the conventional membranes that are less stiff, do not.
And I started thinking about ways to solve this. One idea I want to try is to add some wool yarn as damping in the two outer holes, in the hope that it solves it, but I don't think it will.
But then I got this gut feeling. I think I have seen those dips elsewhere. I think it looks kinda similar to the dip that say metal cone tweeters have, right before the breakup. Kind of like this which is stolen from Seas. That dome is 26 mm wide, and the dip is at ~ 23 khz. My membrane is 55 mm wide, and the breakup is at 9 khz. If we calculate (26/55)*23 = 10.9, we don't get the magical 9 but close enough I think.
There exists of course tweeters that do not have this dip, they are usually made of coated fabric (or sometimes paper) or some other soft non metallic material like this Seas version of the same dome with a coated fabric dome:
Which finally sets the stage for my crazy idea. What if I, instead of having 14 um aluminum foil as my structural backing layer, used paper? More specifically kraft paper? I don't think I have read about someone having a paper planar but it should work right? 80 g / m2 kraft paper is also very cheap so it would not break the bank to try it out.
Another candidate would be to use 50 um thick mylar, if I could heat form it in some way to hold the corrugations but that is probably a lot more complicated to manifacture.
First the background as context:
I was looking at the 0-90 response of the current best membrane which has a solid backing of 14 um aluminum to add stability. This makes it very stiff which is nice but it has those two dips that the conventional membranes that are less stiff, do not.
And I started thinking about ways to solve this. One idea I want to try is to add some wool yarn as damping in the two outer holes, in the hope that it solves it, but I don't think it will.
But then I got this gut feeling. I think I have seen those dips elsewhere. I think it looks kinda similar to the dip that say metal cone tweeters have, right before the breakup. Kind of like this which is stolen from Seas. That dome is 26 mm wide, and the dip is at ~ 23 khz. My membrane is 55 mm wide, and the breakup is at 9 khz. If we calculate (26/55)*23 = 10.9, we don't get the magical 9 but close enough I think.
There exists of course tweeters that do not have this dip, they are usually made of coated fabric (or sometimes paper) or some other soft non metallic material like this Seas version of the same dome with a coated fabric dome:
Which finally sets the stage for my crazy idea. What if I, instead of having 14 um aluminum foil as my structural backing layer, used paper? More specifically kraft paper? I don't think I have read about someone having a paper planar but it should work right? 80 g / m2 kraft paper is also very cheap so it would not break the bank to try it out.
Another candidate would be to use 50 um thick mylar, if I could heat form it in some way to hold the corrugations but that is probably a lot more complicated to manifacture.
Last edited:
I saw no significant differences when I tested paper as backing layer, it was really thin though.
I eventually ruled out paper as backning on the basis that it is not stable when it comes to moisture absorption and overall durability.
Perhaps make some corrugation tests on the intended paper before making a complete membrane?
I eventually ruled out paper as backning on the basis that it is not stable when it comes to moisture absorption and overall durability.
Perhaps make some corrugation tests on the intended paper before making a complete membrane?
Status update!
New corrugator with springs worked great! I could apply more force and make the corrugations deeper in my kapton + solid aluminum backing membrane where the pre-set distance didn't make them deep enough:
The top membrane is the one I re-corrugated with the new springed corrugator. I also tried to corrucate 80 g/m2 and while it does corrugate, somewhat, it is not springy at all. I believe it would make a pretty poor planar membrane, so I am not planning to try it for real.
New corrugator with springs worked great! I could apply more force and make the corrugations deeper in my kapton + solid aluminum backing membrane where the pre-set distance didn't make them deep enough:
The top membrane is the one I re-corrugated with the new springed corrugator. I also tried to corrucate 80 g/m2 and while it does corrugate, somewhat, it is not springy at all. I believe it would make a pretty poor planar membrane, so I am not planning to try it for real.
the flapping does tell you something about damping. most noticable when there is foil without any trace. no you filled everything up so it might not be a huge deal
nice corrugator !!! would love to print the same but a bit wider, if you ever want to share the design. (my 3d cad skills are like 0.0) maybe i can buy the design ? loooks super clean !
You (and anyone else who wants to) are of course welcome to use the corrugator for free, it is generated with a parametric CAD lib in python (cadquery) and is stored in this git repo that should be publicly available. In theory, making it wider should not be more than bumping some numbers.
If you have problems getting the libraries to work then I can generate the STL files for you in your desired dimentions.