[Tjj226 Angel]

A friend of mine tried a similar idea, directly using low tension copper foil as estats diaphragm, using 2mm spacer、2000V+ bias and a step up transformer to drive it like an electrostatic speaker, it just works, but soon, he switched to stax pro bias driver.

Similar design electrostatic headphone is already commercially available. For e.g., Sonoma M1 and its subsequent models Aperio and Bravora，these estats use low tension metal-coated interlayer film as the diaphragm + flexible mesh stators, and their bias voltages are set at 1350VDC (M1 and Bravora) and 1800VDC (Aperio) respectively.

Just to clarify, did he switch back to probias because he liked the sound better or because it was easier to live with the pro bias?

 

[c12Windfury]

Hey guys when designing your stator what program do you use to draw it out before turning it into a cnc file?

 

[chinsettawong]

A very old and obsolete program called ArtCam is very good.

 

[lkos98]

Hey guys when designing your stator what program do you use to draw it out before turning it into a cnc file?

I'm using Fusion 360 for all parts (free personal license). Unfortunately Fusion becomes extremely slow with too many surfaces, so I'm using this: https://github.com/LinuxCNC/simple-gcode-generators/tree/master/grill
for the holes in the stator.

 

[c12Windfury]

Hey guys when cnc a part does the gcode determine the size of the part or does the machine software itself?

 

[lkos98]

Hey guys when cnc a part does the gcode determine the size of the part or does the machine software itself?

CNC machine you mean? Partly, because it's firmware configuration determines how many mm (or inch) will it move by a gcode command. Otherwise the CAD program determines the size, gcode just translates this size to a CNC movement. What I'm trying to say is, that if your CNC is configured wrong, a command for example G0 X200 (move fast X axis 200mm) might move on X lets say 300mm.

 

[lokhanglee]

Hi,

I've just joined this site today. I just want to say hi to everybody and show you my version of electrostatic headphones and amplifier. I've successfully made the phones and the amp for about a year now. I just love them.

I hope my work might inspire somebody here.

Wachara C.

It is absolutely brilliant… I am jealous of your skills

 

[lkos98]

Finally found some time and willingness to redesign my 90mm circles. I have bass boost (about 15-20 dB) at 30Hrz, which is overwhelming. I made some 74mm active area phones and the bass was much better, but still a bit boosted. Obviously this boost depends more on the active area, than the stretching. So I decided to modify the 90mm circles by dividing the area by segments. I made one with 3 segments and one with 2. The dynamics of the 3 segmented sounds worse than the 2 segmented. The attached measurements could not explain why.
Here are the FRs of both. I excluded the frequencies above 1.3kHz, because the graphs look the same, but depend a lot on the cavity where the microphone is placed.

Some bass boost on 2 segmented is still there - about 3dB, but first - I cannot count on my microphone setup and second - I like it more than the "flat" 3 segmented.
I would like very much to hear your thoughts on why dynamics of the 2 segments sounds better? By the way if I remove the 1/6 smoothing, 3 segmented has many resonances of about +-3dBs bellow 80-100Hz.

 

[chinsettawong]

I don’t quite understand your mention about 2 segments and 3 segments. Are you segmenting the stators? How?

 

[lkos98]

I don’t quite understand your mention about 2 segments and 3 segments. Are you segmenting the stators? How?

Yeah, sorry, I was not clear. Segments are on the frame that's holding the membrane. The pictures I showed is of the frame where membrane is glued to - on the blue 2 big circles (3 segments) and on the red - one line in the middle (half circle - 2 segments). The rest of the lines are not touching the membrane - they are just for support and as a safety grill. Martin -Logan ESL for example have such segments on the whole panel.
So, instead of moving the whole membrane at once, it's divided, separate areas are moving together. This is done to avoid high amplitude movement of the central area of the membrane, hence touching the stator. Pros are lower stretching, higher stability and somewhat flattening of the FR, but you lose bass (which I wanted anyway). A comparison is with lets say one 12" dynamic speaker (subwoofer) and few 6" speakers. The area is the same, but moved air volume is different. Bass is less, middles and highs are more. Bellow is the comparison FRs. By just halving the membrane (everything else is the same) there is 18dB lost in 30Hz.

 

[chinsettawong]

From the picture, it’s impossible to see how you segment the diaphragms. Are you using a silicone dot to segment the diaphragm? I actually though about that a long time ago but I’ve never tried it. Losing bass is what I’m afraid of.

 

[lkos98]

From the picture, it’s impossible to see how you segment the diaphragms. Are you using a silicone dot to segment the diaphragm? I actually though about that a long time ago but I’ve never tried it. Losing bass is what I’m afraid of.

Yes, I remember that you proposed that. I also never tried it and sure some bass will be lost. But placing a small 0.5-0.6 mm drop of silicone would be quite difficult and inconsistent. The idea of segmenting is the same - placing some braces between the membrane and stator. Elastic force of the membrane is stronger at the edges and weaker at the middle and by segmentation this force is better distributed.
My design with no segments tends to boost the bass a lot. I tried to equalize it with simple high pass RC filter inside the amp, and it works, but this means this amp will be married to these phones only.
Bellow is a simple picture of two segmented frame (spacer) to which the membrane is glued. Stator is the same, all the rest is the same - it's just now instead of one I have two moving parts of the same membrane. I hope it's clear now.
So far I made just one channel of the phone, so left is with 3 segments, right with 2 and they don't sound the same. I am now making the second channel with 2 segments, once done I will have a better view of how it sounds compared to the full (not segmented) phones.

 

[chinsettawong]

Thanks for showing the picture. It's clear now. With this design, I think you lose quite a bit of bass. Is it worth it?

 

[tib8]

Thanks for showing the picture. It's clear now. With this design, I think you lose quite a bit of bass. Is it worth it?

Please do not forget that lkos98 is using "inside out" drivers: there is one stator in the middle between two membranes, and it is driven by a single ended amp.
I have built this configuration myself. It works well except one thing: it has way to much bass compared to the highs.
If this segmentation solves this, I can just congratulate to Ikor98 for the idea.
I have seen this kind of segmentation only on electrostatic speakers (like: ).

 

[chinsettawong]

Unlike our headphones, Charlie, in the VDO, isn't making his Electrostatic loudspeakers full range. He has a woofer to help boost the low frequencies. So, his ESL's sound doesn't have to go down very low, and therefore his spacer is thin and the diaphragm width is narrow.