Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › My DIY electrostatic headphones
New Posts  All Forums:Forum Nav:

My DIY electrostatic headphones - Page 72

post #1066 of 1688
Quote:
Originally Posted by chinsettawong View Post
 

That's a nice and big active area.  I still don't understand how you will insulate your stator.  Will you powder coat it or will you spray paint it with some plastic paints?

 

The spacer thickness of 0.8 mm is too much, IMO.  If  I were you, I would try 0.6 mm or 0.7 mm.  You'll also need to increase your bias voltage higher. 

 

 

For the metal itself, I was just going to leave it as is. I've not had a problem with exposed metal on the stators in my previous builds. If I have any issues, I suppose I'll spray paint the inner-face as I've done on some PCB stators. I don't anticipate any safety issues with sweat/dust screens and meshes on both sides of the driver. 

 
I'll start out with 0.8 as I've already cut the parts, but if it's not efficient enough I'll keep those numbers in mind!
post #1067 of 1688
Thread Starter 

I have to confess that stretching Mylar to the right tension is extremey difficult.  When I give it too much tension, the bass is gone.  With too little tension, the diaphragm loses its stability.  Using the inner tire stretcher, even though its very easy to work with, it's rather difficult to make it just right, since you need to stretch the mylar so little and it's very difficult to measure the exact elongation.  Using the clip and weight method, you'll need a helping hand, otherwise the diaphragm can break easily.

 

I've redone the diaphragms on my new open buffle headphones so many times that I have lost count.  I decide that I would go with a little less tension to get the bass and increase the spacer thickness just a little to make the diaphragms stable.  I find it very easy to add the extra thin spacer in between the stator and spacer.  I cut a thin plastic sheet which measures around 0.1 mm and put it in between the stator and spacer.  So the new spacer thickness now becomes 0.6 mm.  The phones are a little less efficient now, but the bass is good, so I'll leave it like this.

 

Wachara C.


Edited by chinsettawong - 9/27/13 at 8:40pm
post #1068 of 1688

Nice looking stators, Dude. Have you thought about perhaps making the segments equal areas? Although of course, this might just make the system have more pronounced resonances. It's a bit of a problem not having access to a body of research on these subjects:blink:.

 

I'm having trouble ATM trying to figure out how to make diaphragm spacers/supports. 0.5mm PCB is virtually unobtainable here, and shipping from the US is prohibitive. Maybe I can mill the 1.6mm stuff down to thickness. I still have to face the mylar stretching problem.

 

On a brighter note, I got LinuxCNC (EMC2) up and running, it's head-and-shoulders above Mach3 IMO, and free too. If anybody's contemplating trying CNC, turn over a hard disk to a Linux install, forget Mach3.

 

w

post #1069 of 1688

What do you like more about LinuxCNC?  From what I gather the UI is like a normal computer program whereas mach3's UI is made to mimic a DRO on a machine.  I'm getting used to mach but I'll admit it's not the most intuitive program to use. 

 

Those stators are pretty cool, but in the interest of consistency I'd make the areas equal and I'd use plastic bracing so that the electrostatic field isn't altered by the spokes.

 

I never got around to posting when it was being discussed, but the main advantage of wire stators is immunization to corona discharge.

post #1070 of 1688

I hadn't even considered the implications of that geometry. There's no way I could make the pockets the same shape since it's an oval, but I suppose I could do same area. I can't imagine the difference being perceptible, but who knows.

 

 

Anyways, I tested the flatness of the stator I machined and it's not really going to cut it. Turns out the pocketing milling forces caused the bottom to punch out since the surface was pulled up into the end mill while cutting. I may be able to solve that with a straight-flute end mill, or alternatively I'm also considering getting it done on EDM.

 

I'm going to turn the stainless stators into a long term project. So that I can get my mind off these headphones and back to studying, I just cut some FR4 stators and built the headphones this weekend (besides, I'd want a direct comparison anyways after putting all this work into the stainless). They sound beautiful compared to my 007 clones. The sound stage is huge in comparison, and they really fill in the treble response which my 007's were a bit weak in. Also the bass is incredibly dynamic and punchy, unlike anything I've heard in any other headphones. I'd guess almost surely just due to their size.

 

I'll have to end up increasing tension because as predicted with the 0.8mm spacers, they're not all that efficient. I'm only at 280V bias right now though so it should be fine with more voltage on that and a bit more tension to keep it stable (since I'm not exactly worried about losing some bass).

 

      

post #1071 of 1688
Thread Starter 

Hi Dude_500,

 

As you have noticed, the size of the drivers does make a difference.  :wink_face:  You definitely get a bigger sound stage and more dynamic. 

 

Are you biasing at 280V for your 0.8 spacer thickness?  That's way too low.  You won't have enough juice to feed the phones even with your KGSSHV.  At least you should try 580V, IMO.

 

By the way, your headphones look great.

 

Wachara C.

post #1072 of 1688

I was very grateful for the evaluation version of Mach3 initially. It allowed me to get some results quickly and I didn't have the enormous hassle of making a Linux machine see my Network Attached Storage disk, so I could just send CAD-originated gcode from anywhere.

 

With EMC2 I had to cope with the fact that my Linux is very rusty, but once installed and networked I found it more intuitive. It's possibly true that it's a computer user's program, rather than a shop-floor miller's program, because I found everything I needed just by looking in the menus. CNC has moved, anyway, from being a blue-collar machine-driver's game to being a creative outlet for designers.

 

Mach3 crashed the bit into the table, twice blew a fuse and a FET because it sometimes corrupts the axis positions when you load a new program going from metric to imperial (or the other way, I'm not sure), I found I had to check for that before every machine start.

 

EMC2 has several toolpath display options in 2D and 3D, showing table limits. Smooth graphics. The display in Mach3 is primitive, difficult to manipulate with keypresses + mouse to zoom and pan, and Mach3 frequently loads programs out of scale.

 

Mach3 hits you with a zillion on-screen functions many of which are really irrelevant and only run the risk of swamping the user with options. CNC is really all about gcode and writing gcode or creating it by exporting from a CAD program, not being able to override the current settings, such as feed rate, from the control panel. That's what we call '**** on a bull'. Most of the buttons on an industrial DRO stay clean. They're just there to baffle the apprentices.

 

EMC2 has very few on-screen controls. You can set the jog spacing to a variety of steps from 0.005 mm to 5 mm to continuous. Brilliant when you want to get an axis back to 0. That's about the only on-screen control, the jog, apart from homing the axes, start, emergency stop. I had to write a program to zero an axis for Mach3.

 

EMC2 has a much more reliable feel, more worked-out, more polished. Night-and-day is an overworked metaphor in audio, but that's what I thought when I changed over. I have a lot of complaints about Linux, and the guys who wrote it not having any consideration for users, but I really feel that EMC2 is function-focussed and built by guys who really knew what they were doing in terms of creating a useable, useful interface.

 

And then it runs 500+ lines of code for free, if you have a hard disk spare, which I had. I'd already run into the 500 line barrier when trying to rout PCBs. That's $150 to spend on bits or materials.

 

w

post #1073 of 1688
Thread Starter 
Quote:
Originally Posted by n3rdling View Post
 

What do you like more about LinuxCNC?  From what I gather the UI is like a normal computer program whereas mach3's UI is made to mimic a DRO on a machine.  I'm getting used to mach but I'll admit it's not the most intuitive program to use. 

 

Those stators are pretty cool, but in the interest of consistency I'd make the areas equal and I'd use plastic bracing so that the electrostatic field isn't altered by the spokes.

 

I never got around to posting when it was being discussed, but the main advantage of wire stators is immunization to corona discharge.

 

There are a few UI on LinuxCNC that you can choose to use.  The program running under Linux is very reliable and free.  I can't ask for anythng more.  :biggrin:

 

When you say that wire stators are immune to corona discharge, do you mean that because the wire is round and the there is no sharp adge so that there is no corona discharge? 

 

Wachara C.

post #1074 of 1688

waki thanks for that write up.  My Mach3 is licensed, but I might check out EMC2 since the toolpath views look way better.  I don't know much about Linux though, so I will probably run into problems there.  What kind of speeds and feeds are you guys achieving on your PCB/MDF cuts?

 

Quote:

Originally Posted by chinsettawong View Post
 

 

There are a few UI on LinuxCNC that you can choose to use.  The program running under Linux is very reliable and free.  I can't ask for anythng more.  :biggrin:

 

When you say that wire stators are immune to corona discharge, do you mean that because the wire is round and the there is no sharp adge so that there is no corona discharge? 

 

Wachara C.

Yes precisely :)

post #1075 of 1688
Thread Starter 

Hi n3rdling,

 

Try EMC2 and you'll love it.  :)

 

My max feed rate is around 4800 mm/min.  When drilling holes on my stator I run the feed rate at around 3600 mm/min.  But when routing and cutting, I slow it down to around 1800 - 2400 mm/min.

 

My CNC mill has ball screws.  The pitch is 5 mm.  I use unipolar stepper drivers running at half step.  I've got the Gecko G540 driver but haven't had time to hook it up.  I'm sure that by using Gecko G540, biploar drivers, I can run a little faster.

 

Wachara C.

post #1076 of 1688
Quote:
Originally Posted by chinsettawong View Post
 

 

My max feed rate is around 4800 mm/min.  When drilling holes on my stator I run the feed rate at around 3600 mm/min.  But when routing and cutting, I slow it down to around 1800 - 2400 mm/min.

 

You cut FR4 at 1800-2400mm/min? I'm confused, I cut it at 100-200mm/min. Do you do tons of extremely shallow passes?

post #1077 of 1688
Thread Starter 
Quote:
Originally Posted by dude_500 View Post
 

 

You cut FR4 at 1800-2400mm/min? I'm confused, I cut it at 100-200mm/min. Do you do tons of extremely shallow passes?

 

With 1 mm FR4, it takes just 1 pass at that speed.  And I thought I was cutting too slow. :wink_face:

post #1078 of 1688
Quote:
Originally Posted by wakibaki View Post
 

So you apply the antistatic fluid to one side of the diaphragm only? I would have thought you could double the force on the diaphragm if there is twice as much charge distributed on it...

 

w

 

Quote:
Originally Posted by bidoux View Post
 

The equation that rules the way the diaphragm moves only considers the difference in voltage thus it is not necessary to double coat the diaphragm.

 

If you take a look at the section "DISTORTION IN ELECTROSTATIC LOUDSPEAKERS: Feb 1956" on this http://douglas-self.com/ampins/wwarchive/wwarchive.htm#diel page of Douglas Self's site (The Wireless World Archive), you will see why it is absolutely necessary to consider the charge on both sides of the diaphragm.

 

w

post #1079 of 1688

I don't think it's really necessary.  The displacement is miniscule at the diaphragm thicknesses we're using: 1-3 um.  If you wanted a perfect match of fields I guess you could coat both sides, but this has two disadvantages: a near doubling of the membrane weight and loss of the greatly improved environmental immunization of the uncoated side.

post #1080 of 1688

What I'm talking about is understanding the forces inside the driver. As long as you persist in the view held by bidoux, you'll never understand what's going on.

 

You may not think it necessary to consider the charges on both sides of the diaphragm, but perfectionists will never be satisfied as long as a potential source of distortion is ignored.

 

2 other things, I don't know where you got them. Doubling of the membrane weight. This would mean that the coating vastly outweigh the mylar. I don't know what numbers you have to justify this, but you should post them in the same post where you make the assertion, not just throw it in with no substantiation. 'Loss of the greatly improved environmental immunization of the uncoated side'. I don't understand what you mean by this, and I don't know which 'immunization' mechanisms you refer to.

 

w

New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › My DIY electrostatic headphones