Head-Fi.org › Forums › Equipment Forums › Headphones (full-size) › The Koss ESP950 thread (ESP owners, please contribute! :)
New Posts  All Forums:Forum Nav:

The Koss ESP950 thread (ESP owners, please contribute! :) - Page 3

post #31 of 346
Thread Starter 

ESP phones info list

Does anybody know what the ESP10's bias voltage is?

Lets make a list of all the ESP's bia voltage:

ESP6: 360v (should be the same as the ESP7 )
ESP7: 360v (confirmed by zener voltage measurement )
ESP9: 720v (info)
ESP10: ???
ESP950: 600v (Koss manual)

Some ESP950 info (from Koss manual):
*Diaphragm: Ultra-low mass 1.5 micron C grade polyester film electronbeam coated with a proprietary semiconductive material 250 A thick.
*Acoustic Design principal: Circumaural open air design incorporating minimal acoustic damping.
*Diaphragm radiating surface area: 7 sq. in. (45.6 sq. cm)


E90 (amp for ESP950) info:
*Frequency response: 1.6 Hz to 50,000Hz (-3dB points) 100 vrms differential output.
*Input Impedence: 100k ohms
*Input level: 1 Vrms (for full power output)--- (then why did a SDS headamp connected to the input help the sound?)
*Voltage Amplification: 60 dB
*Audio output voltage: 600 Vrms differential, 2300 Vrms push-pull with soft limiting --- (wow, 2300V?? )
*Channel Separation: - 80 dB at 1 kHz, 100 Vrms differential output
*Signal to Noise: -100 dB at100 Vrms output
*Power consumption: 320 mA DC current at normal listening levels.
post #32 of 346
Thread Starter 

ESP950 pads

The ESP950's pads are too long, don't really provide an air seal around my ears. The pads dimensions are, measured between the longest points and widest points, 5" x 3.25". Thickness is about 15mm.

According to the info posted in the Stax thread (thanks Krmathis and Birgir), the Stax Lambda pads are 5" x 3", but the thickness vary from 0.3" to 1"; the Sigma pads are 4.6" x 3", uniform thickness of 8mm.

It looks like the Lambda pads might be able to made to fit the 950, the difficult part is they require using double sided tape to install....

Is there any aftermarket/upgrade pads available for the 950's?
post #33 of 346
Quote:
Originally Posted by AudioCats View Post
Is there any aftermarket/upgrade pads available for the 950's?
Could these pads fit the ESP950?
eBay.co.uk -> 90mm-95mm ear pad cups for dunu stax dac koss jbl akg
post #34 of 346
Thread Starter 
that is 4"x3.25", while it is wide enough, the length is too short by 1"......thanks though.

I think the Lambda pads can be a good candidate, the varying thickness is actually a good thing now I see.



.
post #35 of 346
Quote:
Originally Posted by AudioCats View Post
Thanks for the O2 picture, but I am not sure I saw the dust cover film..... is it the haze on top of the gold disc?
That is it.

Quote:
Originally Posted by AudioCats View Post
The ESP950's pads are too long, don't really provide an air seal around my ears. The pads dimensions are, measured between the longest points and widest points, 5" x 3.25". Thickness is about 15mm.

According to the info posted in the Stax thread (thanks Krmathis and Birgir), the Stax Lambda pads are 5" x 3", but the thickness vary from 0.3" to 1"; the Sigma pads are 4.6" x 3", uniform thickness of 8mm.

It looks like the Lambda pads might be able to made to fit the 950, the difficult part is they require the of using double sided tape to install....

Is there any aftermarket/upgrade pads available for the 950's?
The only other pads I can think off are the He90 pads. They are 12cm*8.5cm and 2.5cm thick but they cost more then your headphones...
post #36 of 346
I'm pretty sure the E90 doesn't use any opamps. The number of parts in this amp is staggering. The power supply is on the bottom half and the amplifier on the top. The top has a few identifiable ICs, but mostly resistors and transistors. The main amplification is done by 24 MPS 4123 and 24 S1VP 0550. Six per phase and two channels makes 24. The ICs on the top include two white, 4-DIP Motoroloa H11 G1 and a bunch more Motorola parts labeled MG34082P. MC33077P, 33077, etc. The power supply contains a few more ICs, the most prominent being TI TLC24M2CP and TLC24M?? and some marked M9030 LM393N...all these a 8-DIP form. I'd guess there's around 500 parts in the amp.

The stock power supply is 6V 1A, and the battery pack option uses 6 C cell batteries.

Problem is, I cannot open it all the way up to get pics. My guess is that the front is an adhesive aluminum that reveals four more screws, as I can see them from the back side but cannot access them.

audiocats: If you're willing to pay shipping, you can borrow it for a while with a license to poke around in there. Perhaps you can stick a washer back on the the front pot that came off from the back side
post #37 of 346
Thread Starter 

They work!!

Yes, these phones work! (why am I not surprised ). I couldn't wait until my 50:1 ESL trans come in, so I put everything I got together and created a makeshift ES transformer unit . The trans I am using for right now are my 300v/150vA toroids, and since I don't have any Koss 5 pin socket, I had to make one using computer power supply pins on a perf board. Cable is a quad-braid using surplus Alden coax guitar cable that I have laying around...... There was a very faint hiss/whistle from the left driver, at all volume setting ( I can unplug the signal cable and it was still there), but after 10 minutes of playing it disappeared. No hiss, no humm, nothing bad, apparently the bias supply circuit is working well
These 950 suckers are loud, add a pair of woofer you can probably use them as small desktop/PC speakers





Close up of my "for now" micky-mouse "energizer tray". The bias power come from a 9V/100mA AC wallward, then step back up using a 12V trans, to about 90v. This voltage then go throught a 9x voltage multiplier (~~> 800V DC), and goes to the zener board. The zener board uses 12x 51v zeners diodes, so I can select bias voltage from 0 to 612V, in 51v increments..... Guess that should cover most electrostatic phones, from ESP6/7 to Stax pro and ESP950's.

post #38 of 346
Thread Starter 
Quote:
Originally Posted by spritzer View Post
That is it.



The only other pads I can think off are the He90 pads. They are 12cm*8.5cm and 2.5cm thick but they cost more then your headphones...
It turned out the seal is not that super criticle with these phones. I mentioned these pads don't seal all the way around my ears, but when I pressed the phones down to complete the seal, the bass didn't seem to get better....
post #39 of 346
Dude, you are a mad scientist.
post #40 of 346
Thread Starter 
Quote:
Originally Posted by luvdunhill View Post
I'm pretty sure the E90 doesn't use any opamps. The number of parts in this amp is staggering. The power supply is on the bottom half and the amplifier on the top. The top has a few identifiable ICs, but mostly resistors and transistors. The main amplification is done by 24 MPS 4123 and 24 S1VP 0550. Six per phase and two channels makes 24. The ICs on the top include two white, 4-DIP Motoroloa H11 G1 and a bunch more Motorola parts labeled MG34082P. MC33077P, 33077, etc. The power supply contains a few more ICs, the most prominent being TI TLC24M2CP and TLC24M?? and some marked M9030 LM393N...all these a 8-DIP form. I'd guess there's around 500 parts in the amp.

The stock power supply is 6V 1A, and the battery pack option uses 6 C cell batteries.

Problem is, I cannot open it all the way up to get pics. My guess is that the front is an adhesive aluminum that reveals four more screws, as I can see them from the back side but cannot access them.

audiocats: If you're willing to pay shipping, you can borrow it for a while with a license to poke around in there. Perhaps you can stick a washer back on the the front pot that came off from the back side
Great info, thanks. Now knowing the E90 is so complicated, I have lost interest in attempting to mod one....
-- just did a search on the part #'s you provided above,
*LM393: comparator,
*CM33077: dual opamp (11V/uS slew rate, PS voltage +/-2.5v ~ +/-18v, rail to rail output)
*CM34082: dual opamp (30V/uS slew rate, wow, that is fast for the 80's standard, no wonder they were so expensive! supply voltage +/-5v~+/-22V, rail to rail output)
*motorola H11G1: darlinton opto-isolator (that makes sense, only 4 pins).
*Couldn't find any info about TLC24M, even the Ti site doesn't show any. Possibly a wrong #?
* The strange thing is: the power input is only 6V DC, which, in theory, isn't enough for the CN34082 to work properly (will need at least 10V or +/-5). Maybe the TLC's are voltage doublers to provide a +/- power supply?

By the way, I went back to the Stax thread and found your 950 pictures, the drivers seemed to be the same as mine, except you already removed the thin black foam layer in front of the driver, exposing the dust cover film.

Have fun
post #41 of 346
Those pictures make me proud!

The hiss you are mention is the bias leaking somewhere. It could be a hair or some other debris that is touching the diaphragm and is creating a pathway for the bias. It could also be arcing on the driver connectors since they are uninsulated.

Btw. You know you want some HE90 pads for them...
post #42 of 346
Quote:
Originally Posted by AudioCats View Post
300v/150vA toroids
are these Antek toroids? I'm still trying to figure out where to buy my custom wound toroids.. I need something in the dual secondaries with 350 VAC / 0.3 A each.
post #43 of 346
Thread Starter 
Quote:
Originally Posted by luvdunhill View Post
are these Antek toroids? I'm still trying to figure out where to buy my custom wound toroids.. I need something in the dual secondaries with 350 VAC / 0.3 A each.
Antek, yes. I thought they do have 360V/400vA toroids for pretty cheap.

By the way, these power toroids don't sound good. bass doesn't go deep, and rolled off highs. The sound is mid-rangy and doesn't have much details. Good thing I didn't buy these toroids for step-up purposes

-- update: these Antak toroids deserve more SQ credit. what I found out about the initial set-up:
*the bass problem is partly due to the IC-- I didn't realize I had the 3' cotton copper quad-braid tester IC pluged in (thought it was the silver teflon quad braid I usually use). For some reason the cotton made the bass very dry, it hits hard but not bouncy.
*the resolution had something to do with the bias, I was using the 450V (originally intended to test the ESP10). Setting to a more appropriate 612V helped the details and sharpness.
*Do the transformer requires burn-in? I let the phones run as small speakers (playing back ground music), after two days the sound seem to be a bit better. Not sure how much of it was due to the higher bias voltage and the silver IC though.
*After all, these power toriods are not too bad, the sound is actually quite enjoyable at this point


Have fun
post #44 of 346
Quote:
Originally Posted by AudioCats View Post
The ESP950's pads are too long, don't really provide an air seal around my ears. The pads dimensions are, measured between the longest points and widest points, 5" x 3.25". Thickness is about 15mm.
I don't believe a better sealing earpad will make much difference to these phones because the last time I took them apart there were 4 slotted ports in the pannel holding the transducer. These phones are essentially open to the outside. So unless Koss has changed the design, a better seal at the ears won't actually give you the type of air-tight seal that you get with the Stax lambdas.

Of course you could try blocking the ports and using a better sealing earpad. I would expect some sonic differences then.
post #45 of 346
The phones should improve quite a bit with a sealed baffle/earpads. Just look at the lengths Stax went to to seal up the SR-007. Damn it... now I have to buy two f these...
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Headphones (full-size)
Head-Fi.org › Forums › Equipment Forums › Headphones (full-size) › The Koss ESP950 thread (ESP owners, please contribute! :)