well not because of the trapped air, that might even be somethign working against the idea i did not test yet. but there is trapped air, so if i push with one finger on the foil in the midle the other foil will move outwards , but not at the spot i pushed on but overal.. and i am not sure yet if that is beneficial or not. was thinking of a way to test that , but i need it to be leaky somewhere if i can..... not sure how to do that, since i think if its half in between it might be even more shitty then not at all. bit like a leaky closed speaker.
reason why distortion is lower is what i explained with foils moving in tandem, one foil nearing the magnets the other will go away from it. and if it was acting as one perfect driver... witch it wont it would see a more constant field over its entire travel
Hi!
Just a friendly reminder why push pull maybe is not the best option.
The only distortion component that is "eliminated" is the second harmonic. This harmonic is usually dominant in tube amplifiers and in loudspeakers. There is so many listening tests on tube amplifiers with single ended configuration, and also some Nelson Pass transistor amps, where the second harmonic is very high. And people tends to like them. They can have 5% 2nd harmonic distortion and listening fatigue is non existent. I contrast you have a "modern" high power amp with 0.01% distortion, mainly 3rd harmonic, and some spikes of higher order, but very low. And the reviews is not that nice reading. Hard harch aggressive sharp, barb wire, etc.. why? THD is much better in the modern high power amp.
If we go to the literature and dig in to MP3 encoding and listening tests that has been done we learn that you can remove lots of harmonics of higher order IF you keep some harmonics that has been "masking" the ones that you removed. If we turn this around, it seems that 2nd harmonic is masking the 3rd harmonic that is masking the 4th harmonic etc... It also seems that even harmonics like 2nd 4th 6th etc is always present in music, and 2nd harmonic is also generated in our own ears. To summarize 2nd harmonic is wanted! Nelson Pass has made pre amps with high amount of 2nd harmonic that is adjustable. Turn a knob on the front and get extra 2nd harmonic! I have built some Dipole subwoofers and experimented with facing both drivers in same direction or in opposite direction. When you let the woofers play in opposite direction you eliminate the 2nd as I said and it does sound less realistic, more synthetic, hard.
If you can find a way to address 3rd harmonic or 5th harmonic, that would be interesting!
Just a friendly reminder why push pull maybe is not the best option.
The only distortion component that is "eliminated" is the second harmonic. This harmonic is usually dominant in tube amplifiers and in loudspeakers. There is so many listening tests on tube amplifiers with single ended configuration, and also some Nelson Pass transistor amps, where the second harmonic is very high. And people tends to like them. They can have 5% 2nd harmonic distortion and listening fatigue is non existent. I contrast you have a "modern" high power amp with 0.01% distortion, mainly 3rd harmonic, and some spikes of higher order, but very low. And the reviews is not that nice reading. Hard harch aggressive sharp, barb wire, etc.. why? THD is much better in the modern high power amp.
If we go to the literature and dig in to MP3 encoding and listening tests that has been done we learn that you can remove lots of harmonics of higher order IF you keep some harmonics that has been "masking" the ones that you removed. If we turn this around, it seems that 2nd harmonic is masking the 3rd harmonic that is masking the 4th harmonic etc... It also seems that even harmonics like 2nd 4th 6th etc is always present in music, and 2nd harmonic is also generated in our own ears. To summarize 2nd harmonic is wanted! Nelson Pass has made pre amps with high amount of 2nd harmonic that is adjustable. Turn a knob on the front and get extra 2nd harmonic! I have built some Dipole subwoofers and experimented with facing both drivers in same direction or in opposite direction. When you let the woofers play in opposite direction you eliminate the 2nd as I said and it does sound less realistic, more synthetic, hard.
If you can find a way to address 3rd harmonic or 5th harmonic, that would be interesting!
well that might be true but second order in a single ended is extreme high. it makes up 95% percent going far beyond 5% from a tube amp.
besides that, the idea is also been able to trade excursion for efficiency. since i dont care about efficiency
class D can easily put out some power. but i do understand your point !!and you might be right. in this case its not push pull. at least not with the costs coming with push pull. this contraption costs less then a single ended. thats why i wanted to see what it does.
besides that, the idea is also been able to trade excursion for efficiency. since i dont care about efficiency
class D can easily put out some power. but i do understand your point !!and you might be right. in this case its not push pull. at least not with the costs coming with push pull. this contraption costs less then a single ended. thats why i wanted to see what it does.
I played 3 tunes with less damping just enough to get rid of the huge hump. and its the first time i have to much lows.. compared to mid top well a first for everything i gues. it is a binaural recording so mics in my ears omni to be precise so it picks up low end really nice... bit to nice in this case but they must be heared with a headphone, cant use speakers (wel you can but should not)
well a first for everything i gues. it is a binaural recording so mics in my ears omni to be precise so it picks up low end really nice... bit to nice in this case but they must be heared with a headphone, cant use speakers (wel you can but should not)
How can they move in tandem if they are not coupled magnet gap wise?
And, if they are not coupled, how does the distortion get lower?
I mean, these are not stiff dynamic bass speaker membranes.
Because the loadspeaker is a dipole and the distortion cancellation happens in the air before the back sound and front sound hits the microphone?
I am not doubting you, WrineX. But I'd like to know what's going on.
Last edited:
The trapped air provides a very tight coupling between the two membranes, but it will not point-to-point coupling. The membrane excursion is near point-to-point parallel at very low frequencies, because the both of them has near identical law of force, except the driving force. The driving force and the provided excursion can summarize totally, because the tight of coupling "overwrites" their elemental nonlinearities, like an output transformer primary halves at PP tube amp (at low frequency). As we follow the analogy, the coupling will be the weaker (at trafo and dual membraned speaker, too) at higher freqencies. Fortunately, the membrane excursion will be microscopical in this case + crossed over the woofer.
+1 analogy: the "leaky" air (between membranes) is not a problem, because provides huge time constant. It functiones quasi dynamic coupler and DC-decoupler, like the hydraulical valve lifter in modern engines (there is a little leakage of oil, which cancels the slow thermal expansion, but couples the dynamic strokes).
+1 analogy: the "leaky" air (between membranes) is not a problem, because provides huge time constant. It functiones quasi dynamic coupler and DC-decoupler, like the hydraulical valve lifter in modern engines (there is a little leakage of oil, which cancels the slow thermal expansion, but couples the dynamic strokes).
I guess, at higher frequencies appear a lot of disadvantages: the membrane acts as new reflecting surface and the two membranes interact, because the distance of radiating surfaces makes phase lag and interference. Further, the dual membrane provides doubled moving mass/radiating area (this is true only above a few kHz range). Of course, Wrinex's implementation has steep X-over, eliminated these problems.
Sometimes ideas need time to mature. I like your project. This is barebone-best DIY !!!
Two questions about this project:
1. In analogy to ESL stator optimal opening ratio - is there data available for best ratio in this case? I imagine that the optimal ratio might depend on the strength of the magnetic field.
2. Could you please describe in a brief how-to your way of producing the coil and then to bond it onto the membrane?
OFF-TOPIC:
One of the common misconceptions about 2nd harmonics is that this order of distortion just adds a distinctive and single signal one clean octave above the main signal, thus adding some brio. Like an organ player adding some high-pitched vox angeli to it's registration.
You may have a look here for an more analytical viewpoint about 2nd harmonic distortion. This paper of Uli Brüggemanns is worth every cent and has the potential to radically dismantle the widespread cherishing attitude towards 2nd harmonics. In real-world ANY, and therefore also 2nd harmonics distortion does not only add vox angeli, but also leads to a whole nasty spray of IM products polluting all over the audio spectrum. So in reality, your additional register becomes rather vox crap, so to say. For an enlightening subjective experience, you might selectively listen to synthetized and isolated 2nd harmonic products of any audio track you like. Do it, it's strongly (!) recommended, and for shure you will dyslike it.
As for the maybe true logic applied to endorse lower harmonics masking the higher ones, then in analogy you may add a steamdrill to your audio gear as well. It works quite decently. You will not be disappointed in terms of masking uneven harmonics.
ON-TOPIC AGAIN:
So wrinex, please ahead iterating with, and reporting from your project. It would be really great to finally end up in a planar design producing less than -80dB (!) of distortion products.
Addendum - Edit time has elapsed:
Therefore another question: did you already tried FEMM analysis to maybe optimize the magnetic field homogenity of the magnets assembly?
In these terms, any approach promising a minimizing of distortions is favorable. And as such, this dual-membrane, acoustically push-pulled symmetric construction seems to be a favorable design to me in terms of potential drive asymmetry cancellation. In the same logic, any approach linearizing the magnetic drive drive will contribute to a reduction of distortion products.
Therefore another question: did you already tried FEMM analysis to maybe optimize the magnetic field homogenity of the magnets assembly?
Last edited:
I did a small comparison to a known driver. since when playing with the after a while im not sure if its any good...
Hmm today i was playing some more with the dual foils. i changed the coil to cover more of the foil. not sure yet if i like it, it performed as good as the old one where i added some extra alu to vover the bare mylar.. (makes it move less random there and adds some weight for lower res) but i am in doubt non stop is it worth the effort or not ? i thought i just do a measurement with the panel and a known panel, the Magnepan LRS. they are not known for bass., (i did chose the one with the lower resonance) but the LRS is roughly twice as big as my test panels.
picture
picture 1 The setup (the lrs has besides more surface area also more baffle)
picture 2 FR Brown LRS bass (i used a higpass to not blow anything) Green dual membrane panel.
picture 3 Distortion LRS
picture 4 DIstortion Dual membrane... most visable is 2e order distortion but also the 3th is a bit lower and the whole panel can reach lower. dB not calibrated, but it was on the edge i want to push the LRS
the small panel has some weird peaks, in 3th still, i might be able to find out what those are in the end.. i had to increase the output of my panel +3 dB to be the same level. then again my foil has an xmax of 6mm and the maggy 1-1.5mm. so i might reduce mine a little unless i want to reach lower like 40 hz or something (halving resonance is 4 times the excursion)
Hmm today i was playing some more with the dual foils. i changed the coil to cover more of the foil. not sure yet if i like it, it performed as good as the old one where i added some extra alu to vover the bare mylar.. (makes it move less random there and adds some weight for lower res) but i am in doubt non stop
is it worth the effort or not ? i thought i just do a measurement with the panel and a known panel, the Magnepan LRS. they are not known for bass., (i did chose the one with the lower resonance) but the LRS is roughly twice as big as my test panels.picture
picture 1 The setup (the lrs has besides more surface area also more baffle)
picture 2 FR Brown LRS bass (i used a higpass to not blow anything) Green dual membrane panel.
picture 3 Distortion LRS
picture 4 DIstortion Dual membrane... most visable is 2e order distortion but also the 3th is a bit lower and the whole panel can reach lower. dB not calibrated, but it was on the edge i want to push the LRS
the small panel has some weird peaks, in 3th still, i might be able to find out what those are in the end.. i had to increase the output of my panel +3 dB to be the same level. then again my foil has an xmax of 6mm and the maggy 1-1.5mm. so i might reduce mine a little
unless i want to reach lower like 40 hz or something (halving resonance is 4 times the excursion)
Last edited:
Well its not an ESL so i dont know
even for this contraption its pure trial and error. i noticed if i damp it screws it up. so it might benefit maybe more open area then it is right now. and damp from the outside or electrical (-13dB at 50 hz right now haha) if you remove that.... its allot of 50 hzabout how i made my foils... i made 400 videos of witch at least 50 show how it works
in short i plot my coil from a huge roll of alu tape. remove the backing add the foil. remove the foil and coil. remove whatever did not stick because i left the backing. roller it flat and you got mylar with alu coil with adhesive that can handle high temps.
These are very learningful graphs! Especially two things: the quasi PP woofer panel makes magnitude better distortion and both of panels knows physics. Decreasing frequency, the response starts fall at resonance point with 18dB/oct. rate, because the membrane compliance determines the excursion.
i thought any panel should drop 18dB below res. 6 from baffle and 12 from playing under res?
im cooking up a third panel slightly bigger slots on the sides of the outer magnets. and slightly wider foil to get resonance down a little more. ... hmm or should i go much wider just for fun ? could make it 60 high 50 wide isntead of the 60 high 38.6 wide. and add maybe an extra 2 rows of magnet..
going wider has benefits i think. since the floor is an endless baffle so the biggest loses are sides and top. making it wider might gain some and gets the rest down at the same time, since having 2 foils makes it shoot from 34 to almost 50Hz, for the same foils used
im cooking up a third panel slightly bigger slots on the sides of the outer magnets. and slightly wider foil to get resonance down a little more. ... hmm or should i go much wider just for fun ? could make it 60 high 50 wide isntead of the 60 high 38.6 wide. and add maybe an extra 2 rows of magnet..
going wider has benefits i think. since the floor is an endless baffle so the biggest loses are sides and top. making it wider might gain some
and gets the rest down at the same time, since having 2 foils makes it shoot from 34 to almost 50Hz, for the same foils used
i used femm for this setup as well, i dont want to twist them , since it gives me a really narrow space to put coil. but thanks ! ur right about the learning curve haha. its the most anoying piece of software to work with
especially 1 change to ctrl Z , after that... you can start all over haha
Yes, all of panels does it, except if the resonance is very undamped. We can see the 18dB/oct. slope in this case too, but well below the resonance point only.
well i only damped it with the DSP right now.. and i am not liking that to much, but its fast for now - 12 dB or something at 50-52Hz
and another 8 dB at listening position for my room haha the LRS goes down to 50 in my room but only because my room is that bad
here a measurement on my couch
Blue LRS and Red dual panel... the res of the panel is spot on the room mode of my room... not ideal. i did not change a thing compared to the distortion measurements
and another 8 dB at listening position for my room
haha the LRS goes down to 50 in my room but only because my room is that badhere a measurement on my couch
Blue LRS and Red dual panel... the res of the panel is spot on the room mode of my room... not ideal. i did not change a thing compared to the distortion measurements