[soundoholic]

Thanks, cjfrbw. To your 507: I'm very curious about the stretch process - how did you managed to achieve the suitable tension of the mylar?? The smallest changes cause large change in sound (Stax makes permanently experiments!), furthermore the points where you stretch and the different force at that particular points are very important. Or not? I´ve made my experiments too...

 

[cjfrbw]

Hi, soundoholic,
 
I am sure my method is not precise.  I of course have no direct knowledge of the Stax industrial process.  
 
I aimed for right to left consistency and just stretched until no sags then tamped onto glue.  I stretched the top edge of the mylar sheet with tape onto a glass refrigerator shelf, then laid down the bottom half of the metal bias/spacer, and pulled the mylar over it, when the tightness seemed right, let the edges hit the glue.
 
This is the ultimate DIY, just eyeing things as I went along, it is not a precise method.  The 1.4 micron mylar is pretty fragile and tears easily, it will only stretch so much.  So, just a combination of trial and error, instinct, and observation.
 
After they are glued, I spray them with anti static spray and let dry to as even a consistency as possible. I finished the drying process with a small hair drier.  I then apply the top spacer, use a scalpel to trim, and put them back into the headphone. They sound similar right to left with a little tap to the membrane.  As I say, I can only aim for right to left consistency, not industrial precision.
 
So far the ad hoc membranes are extremely high definition, much more so than the original 507, but also smooth and without any harshness.
 
 However, I need to with hold any judgment until they have been in there a while without problems.  If problems arise, I might try a thicker mylar.
 
It is an intricate process and I am sure that my method is quite flawed compared to the Stax process.  However, I am surprised at the incredible detail of the new membranes, something I wasn't really expecting. 

 

[soundoholic]

Wow! I´m very impressed. If you managed to achieve a correct tension in the first step, at once - that´s great!
You´re right - first wait and see (hear), what happens in ordinary practice. I wish you could say the same good words to your work after a couple of months... 

 

[cjfrbw]

Hi, soundoholic,
 
Well, this is just an extended experiment.  I had a pair of expensive headphones that decided to commit suicide in various ways.  I am lucky they are built in a way that allows intervention to some extent for somebody with modest DIY skills.
 
The original 507 was designed to be a "better Lambda", and it was successful in that regard.  Better detail, deeper fuller bass, better high frequency extension, colorful tonality and smooth response.
 
However, they were never as detailed as my 3n/5n vintage normal bias head sets. Some guests actually preferred the 3n/5n due to the better detail.
 
The new membranes are extremely transparent, much more than either the original 507 or the 3n/5n.  They don't have the bass slam, but the bass is not deficient, either.  They are less like improved Lambdas and more like "super 3n/5n" with greater detail, much larger soundstage, and much more robust bass.
 
Whether this holds up with time only time can tell, but it a very interesting headphone in the short run, I am kind of reveling in the detail and transparency.
 
However, I would never trust a modestly equipped DIY guy stretching generic mylar on his dining room table over the Stax industrial product, either, so as I say, it is just an ongoing experiment at present and I am glad to have a functioning set of head phones that I  can like for a while and that I can continue to play with as needed.

 

[cjfrbw]

To follow up:
 
It is pretty clear now that by changing out the mylar and using a different electrostatic agent, one changes the variables in the headphone.  
 
The amount  of electrostatic agent affects the sensitivity.  After these recent membranes dried and stayed in the headphones, the pro socket had way too much bias. I probably used too much electrostatic agent, which appears to be some kind of graphite in a slightly oily base.  Stax must be very exacting in the kind and amount of electrostatic they use.
 
Basically, the 507's became normal bias headphones.
 
 
I guess when you change your membranes from the proven industrial standardized product, you also need to recognize that the ideal bias will change as well.
 
The pro socket makes them crinkle and sound way too sensitive and charged up.
 
However, the normal socket works very well with them.
 
With the normal socket, the new membranes are still incredibly lucid and mesmerizing. They seem to have an extra octave of bass compared to the normal bias traditional Lambda.  Not a great deal of slam, but deeper.  They are so clear, you can hear the bits of sampling hair on compressed files, but they still sound great. Vinyl records are  expansive.  High frequencies have shimmer and tremolos I can't hear from normal bias Lambdas or 3n/5n. 
 
So far I like them, I just don't know if they are stable with time.
 
They sound generally better than normal bias traditional Lambdas to me, still with more detail than any  Stax or other headphone I have heard so far.  Normal bias Lambdas are still the midrange champs, however.
 
I will probably redo the membranes again at some point and try to refine the techniques.
 
Anyway, the message is the DIY approach very much alters the variables of the headphone, it in no way resembles the original Stax product, and the bias charge changes due to the different kind and amounts of electrostatic agent applied to the mylar.
 
I guess for those who like to fool around with different things, this is OK, but not the original 507 by any stretch.  I will be happy when I get a stable, repeatable, good sounding result, and that seems do-able.

 

[chinsettawong]

Hi cjfrbw,

To go with higher voltage, everything has to be done right. It is actually very possible for you to go pro bias.

What kind of coating material are you using? How conductive is it? I normally coat the diaphragm with a very light coating. The coating is so light that it's virtually transparent. I had once measured its resistivity, and it's around 100 Mega Ohm per square. Just remember that high resistivity is good. It makes your diaphragm voltage stable.

With high bias voltage, you need to increase your diaphragm tension. A good enough tension gives you stable diaphragm position. Otherwise, the diaphragm may be sucked to one side of the stators.

If you are interested, you might want to try some thicker Mylar. It's easier to give it higher tension while maintaining good bass. 3 microns Mylar is pretty good.

Wachara C.

 

[cjfrbw]

Hi, Wachara,
 
Thanks, you have been a big help.
 
Coating is spray can of Staticide ESD 6500.  Seems to be quite potent.
 
My first attempt painted a very thin layer of ESD 6500 and no glue to hold the mylar, and it worked OK with pro bias, but wasn't tight enough, so I tried the second set stretched a bit more.
 
The present set I just sprayed the ESD 6500 material on and let it dry, probably too much on there.
 
I will probably try a different stretching technique and thin application of ESD 6500 again for the next set.  
 
For the time being, I am enjoying the present set quite a bit on normal bias.

 

[chinsettawong]

Staticide ESD 6500 is good.  I would spray it on a soft sponge and use the sponge to wipe onto the diaphragm.  That way the coating can be quite thin.

 

[soundoholic]

Quite interesting your impressions, thanks, cjfrbw and Wachara. I have no experience in preparing of mylar, but a little bit in changing of BIAS voltage. I do think, with a slight change up you can result more transparency, punch and even improve the soundstage as well. I switch the voltage in my LRT to 660V when sparkling and fresh sounds are to achieve on non vivid recordings. It' s not at all significant improvement but for me enough to hear it. I don' t suppose such an increasing in bias could destroy the driver, are you thinking so?
Spritzer once wrote, his experiments went up to about 900V without destroying the foil. Hmm... In my case I don't hear very loud and 580V plus ca. 14% cannot be destructive...?

 

[cjfrbw]

Hi, soundoholic,
 
I agree that it interesting.  The DIY result of the 507 with normal bias with the current home made membranes is more transparent than anything I have ever heard with headphones. It also sounds good, very nice high frequencies and very substantial, deep bass, no harshness at all.  
 
The bad:  the membranes are frisky, can be noisy when left on without a music signal. As soon as the music signal is applied, though, they become completely quiet and the music is incredibly clear with a huge soundstage. Some sounds I have thought were coming from the environment, when they were coming from the headphones.
 
The source of the transparency is a mystery.  Is it bias, the electrostatic solution used, the stretch?  I don't know.
 
One could speculate that Stax aims for a robust commercial solution, but does not aim for maximum transparency, or just adjusts the transparency to price point since electrostatic headsets have transparency to spare to begin with.
 
I think I am just going to listen t them like this for a while, and break them open once in a while to see if the membranes are OK.  I will eventually redo the membranes with a better technique.
 
I am finding that I prefer the modded 507 with vinyl in the location where I use them over speakers.  I have never thought the headphone experience would be better than speakers, but now I am thinking that sometimes it might.
 
I cannot vouch for either longevity or commercial robustness with these DIY membranes, or even appropriate function. I would imagine a variable bias that allowed customization would be ideal, but I don't have access to that, either.
 
But it is certainly fun and interesting to play with, and it is nice to know I can just put in new membranes periodically or as desired.
 

 

[spritzer]

The bias shouldn't be calculated like that as the electrostatic force diminishes by the square of the distance.  That's why Stax use 230V bias for the 0.3mm gap and 580V for 0.5mm and if you keep increasing the spacers the bias voltage grows quite fast.  Sound-Lab speakers run at 12kV with teflon coated mylar and stators... 

 

[NamelessPFG]

You actually rebuilt the driver diaphragms yourself? I'm impressed, even if the non-glued nature of SR-507 drivers presumably helps a lot.
 
I'm still wondering if I should hold out for a spare SR-303/404 driver, or have someone just attempt to rebuild both drivers in the set I have to a matching sound since Yama's/STAX USA won't do it and I don't trust myself with matters this delicate that require extreme precision. Not after finding out that just gluing the driver back together isn't as simple as it looks.

 

[cjfrbw]

I wouldn't call it trivial to attempt the mylar replacement on the 507, but it is do-able because of the way it comes apart and goes back together.  I gather some of the other older Stax types are not so easy to get apart and put back together, so it seems to be an individual thing.
 
It seems with some of the home use, computerized milling devices, hobbyists can mill their own custom parts if they know the dimensions.  That might make producing mylar/bias diaphragms a bit easier, but I wouldn't know the first thing about using such a machine.
 
I did glue the second and third set I made, not too difficult with a small amount of glue, and it is easy to clean off the glue if it doesn't work. My concern was the glue might keep the parts from seating firmly together, but that was not the case with the 507 parts.

 

[soundoholic]

Quote:

The bias shouldn't be calculated like that as the electrostatic force diminishes by the square of the distance.  That's why Stax use 230V bias for the 0.3mm gap and 580V for 0.5mm and if you keep increasing the spacers the bias voltage grows quite fast.  Sound-Lab speakers run at 12kV with teflon coated mylar and stators... 

Yes, agree. One cannot supply NB Stax continuously with 580V plus strong signal - without the danger to destroy them. There are 350V - 152%! more. The foil could fry very quickly. But I had a DIY Amp with 280V bias for my NB Sigma and it went quite good, very good. But, as we know on contrary, there is a poor sound outcome with Pro Stax at 230V bias.
What is your experience with more bias voltage for a stax pro version; any sound alteration or no? How does BIAS increasing/dimin. hang together with different signal-strenght?
Thanks for your reply...

 

[spritzer]

No real benefit from increasing the Stax bias.  I usually keep it at 560-580V but other sets see a lot of benefit from messing with the bias a bit.  Koss ESP950 being one and the HE90 needs 500V and not a single volt higher. 
 
The rule of thumb is that the max input voltage a driver can see is bias*2 so if you lower the bias you can increase the drive voltages.  This is what Sony did with the ECL800 but this causes more issues IMHO.  Higher drive voltages means more gain and that is never easy to do properly.