[spritzer]

This resonance issue crept up in the first SR-007Mk2 revision Stax did.  They increased the height of the earpads for extra comfort which in turn meant extra space and there were some midrange issues with that change.  As soon as I lowered the pads to their normal position the coloration was gone. 

 

[odanas]

I love this thread so much. Thank you everyone for all the information and experience you've gathered up on one thread. I am just about ready to build my own pair but I'm kind of stuck because I don't have an electrostatic amp. 
 
Does anyone know of a fairly quick and cheap design that I could use to just drive whatever headphones I make? Basically I just want to be able to hear sound coming out of it. I've seen a few simple amps for ESLs but I figure the high bias and perhaps other things would make those not so good for headphones.
 
I'm a mechanical engineer and I know some stuff about electronics but the simpler the design the better. 
 
Thanks guys and amazing job! I can't wait to make a few pairs.

 

[.Sup]

very interesting thread indeed! Keep it up lads!

 

[dude_500]

Odanas, the simplest way to drive electrostatic headphones is a pair of audio transformers. All you need in addition is a bias supply which can be made from a transformer connecting to 110/220 and perhaps a voltage multiplier.
 
 
As for pad material, I took a measurement just now of with and without the leather on the too-big pads. Blue is with leather, red is without. Doesn't have much impact:
 

 
 
I'll make some new frames that hold smaller earpads hopefully sometime this week to try out. Does anyone know any good software that is free/cheap for modeling geometries like the headphones we make? I've found quite a few pieces of software online, but they're all targeted at either modeling buildings or more aimed at analyzing various magnetic speakers and don't have a lot of support for custom raw geometries of drivers and enclosures. I found VA One, but I can't find a trial or any discussion of price, I can only imagine how much it costs.

 

[AmarokCZ]

odanas: just use power amplifier (speaker-amp or some powerful head-amp) two output transformers for tube amps (~1:50 vinding ratio) and some simple bias supply (Delon multiplier or any other).
dude_500: just for information: I tried thin rectangular earpads (just to test impact of earpads on sound) with inner dimensions of 9x10cm and final version of earpads is with 6.5cm circular hole and there is virtually no difference in responce below 3kHz (could be measured, but it's not audible since it's no more than 1.5dB in narrow bands - e.q. 1.5dB dip from 800 to 900Hz).
EDIT2: dude_500: just saw your last measurements, and I'm wondering what foam are you using. I tried similar yellow foam (polyurethane) as found in many headphones and bass response without leatherette was bad - almost no bass.
How does the free-air FR (this one) change if you change earpad for "good one" and if you remove earpad?

 

[arnaud]

Does anyone know any good software that is free/cheap for modeling geometries like the headphones we make? I've found quite a few pieces of software online, but they're all targeted at either modeling buildings or more aimed at analyzing various magnetic speakers and don't have a lot of support for custom raw geometries of drivers and enclosures. I found VA One, but I can't find a trial or any discussion of price, I can only imagine how much it costs.

I am unaware of free / cheap software at this time. As for VA One, you will not find a free trial version license easily and the cost is prohibitive at the amateur level. There is a cheaper education version but these are only sold to universities not private individuals afaik. If you give the dimensions if your headphone (driver, stator, earpad) including thickness and materials list, I can give a shot a simulating the response.

The only issue is that I am having trouble finding time for this and haven't finished the revisit of my own 009 and haven't started wachara's and amarakcz's models. On the other hand, these guys build seem finalized already so the need for simulation isn't really there . Your resonance issue is interesting because I see it with the stax headphones as well.

 

[dude_500]

Quote:

EDIT2: dude_500: just saw your last measurements, and I'm wondering what foam are you using. I tried similar yellow foam (polyurethane) as found in many headphones and bass response without leatherette was bad - almost no bass.
How does the free-air FR (this one) change if you change earpad for "good one" and if you remove earpad?

It is memory foam, which seems to me to have very good sealing qualities to air (after all, if air easily permeated through it it wouldn't do the memory foam thing very well).
 
Here are all the free-air responses. They are all about the same, which really reinforce the concept that the whole 1khz thing must be a resonance related to the new earpads, in a sealed state:
 
Blue: New large earpads, leather removed (1" thick memory foam)
Red: No earpad at all
Green: Old smaller good earpads that I'm going to replicate
 

 

[tigon_ridge]

Quote:

It is memory foam, which seems to me to have very good sealing qualities to air (after all, if air easily permeated through it it wouldn't do the memory foam thing very well).
 
Here are all the free-air responses. They are all about the same, which really reinforce the concept that the whole 1khz thing must be a resonance related to the new earpads, in a sealed state:
 
Blue: New large earpads, leather removed (1" thick memory foam)
Red: No earpad at all
Green: Old smaller good earpads that I'm going to replicate
 

The fact that they look almost identical tells me it has nothing to do with earpad material, as I had implied. It has everything to do with the shape/size of the earpads; as you said, resonance. There is no reason to have such a large and wide space between your ears and the drivers. You effectively create a chamber where sonic waves are bouncing around canceling each other. The more direct the sound transfer, the better.

 

[arnaud]

I am not sure the answer is as simple. Below the first acoustic resonance, the cavity acts like a mass and spring and restrain the diaphragm motion. As such, te low frequency extension / fundamental resonance of the diaphragm is affected by the effective volume and probably ear pad stiffness. It also prevents front / back wave cancellation but I think any combination of leather / foam thickness with a good seal can achieve the same effect.

 

[AmarokCZ]

I agree with tigon_ridge.
Also if you look on previous measurements (red curve) there are some HUGE swings below 400Hz. But on those three free-air responses you posted now there is nothing in 200 to 400Hz region. So you must have changed something between those two measurements.

 

[dude_500]

Quote:

I agree with tigon_ridge.
Also if you look on previous measurements (red curve) there are some HUGE swings below 400Hz. But on those three free-air responses you posted now there is nothing in 200 to 400Hz region. So you must have changed something between those two measurements.

Indeed, all my measurements (especially free air) lack precision. They should be taken as approximations with disregard to fine details. I have not built a test-jig to hold both the driver and microphone in precisely the same position each time, nor do I have any means of canceling ambient noise which can be fairly loud at times in my dorm room.
 
That said, there is absolutely no doubt the 1khz peak is huge and always present with the new earpads when sealed. It is also true that those bumps at 400hz you speak of are always there, and almost identical when sealed (naturally, sealed is easy to repeat and my measurements have very good repeatability even day to day. Free air, not so much.)
 
Since it is now completely clear that it is a resonance, I'm not wasting any more time with these large earpads. 

 

[schorsch]

Fantastic thread )
 
I read it for quite a while now.
You should try to make a Sigma housing for the sennheiser HE 90 clone! Should be an interesting try.
 
Regards Georg
 

 

[wiinippongamer]

@Dude-500 are you using your own compensation curve for measuring them?

 

[dude_500]

Quote:

@Dude-500 are you using your own compensation curve for measuring them?

No. I'm using an electret microphone element which is +-1db below ~8khz. Above that, I don't really care all that much anyways at this point since it seems the way we measure has issues in that frequency range. Maximum deviation within audio range is a +3db peak below 20khz so it should be fairly good even in that range anyways.

 

[dude_500]

Do any of you heat treat your mylar films? I've read that that is done in professional products to help keep the mylar from losing its tension.
 
I've had some drivers built for a couple days that I've been listening to, and I noticed today that they've gone unstable, and there are wrinkles and a lack of tension in the films. Has anyone else experienced this problem?