Koss ESP/950 electrostat thread

[HeadphoneAddict]

after a bit more listening today I feel the smegma pros still have the edge on them for overall balance. The 950 is a bit shouty in the upper mids and are pretty bright. All from the E90 amp and SRD-7Mk2. I'll wait until the eXstata gets here for proper full evaluation but so-far the Stax still have it.

In the end I sold the ESP-950 and kept the Smegma Pro. You should be proud.

[Smeggy]

heh, well the Beyer ET1000 is beating both right now except in comfort.

[mypasswordis]

heh, well the Beyer ET1000 is beating both right now except in comfort.

Uh, you still running it from the E/90?

[Smeggy]

Uh, you still running it from the E/90?

Nah, SRD-7

[manaox2]

My latest pair of ESP950 I am glad to say have no channel imbalance. After whipping up a quick adapter, these kick so much ass out of the T1. Anyone know where I can grab a pro Stax cable?

[spritzer]

Try calling Yama's and tell them the cable on your Lambda is foobar. I'm guessing though that it would be cheaper to get a SRE-725 extension cable instead.

[manaox2]

Try calling Yama's and tell them the cable on your Lambda is foobar. I'm guessing though that it would be cheaper to get a SRE-725 extension cable instead.

Thanks, I'll stay alert for the next one to come around at $50.

[cetoole]

So, I know that it is pretty much public knowledge that the ESP950 bias is 600v, and Koss even says this, but has anyone actually looked at the bias circuit? I got my ESP950 back from RMA today, and noticed that is has the squeal that so many of us know just too well out of my SRM1/mk2, but is dead quiet out of the stock E/90. This has been really, really bugging me, so I finally went ahead and tore the E/90 open to get a look at the bias supply. Well, its pretty simple, a 10M resistor from the 600v (measured 612v) rail to the pin connecting to the bias, and another 10M resistor from that to ground, in parallel with a 0.022uf 630v film cap. My measurement of the bias voltage, using a DMM with a 10Mohm rated input impedance, gave me 202v, just what would be expected from a 10M-(10M||10M) divider. I dunno guys, this sure looks like the E/90 is really just putting out 300v bias, and if that is the case, it would really explain the all-too-common squeal these tend to develop and we have been attributing to dust. Thoughts?

[GPH]

I'm sorry to break your theory, but the two ESP950 I've had squealed out of the E90.

[cetoole]

Not if my theory is that the ESP950 bias is ~300v, not ~600v.

There are definitely construction issues in the headphones, so I really have no problem believing that there is something going on beyond just the bias voltage.

[cclragnarok]

My ESP950 also seems to squeal louder and more often from my Headamp KGSS.

[spritzer]

Well Koss does state the bias voltage to be 600V:

SPECIFICATIONS

ESP/950

Physical Specifications

Size of earcup: Approximately 150 mm h x 100 mm w x 40 mm d (including cushions)

Cushions: Large foam filled leatherette type for high ambient noise isolation and conformance to facial structure.

Headband: Extendable, stainless steel band with self adjusting, pivoting yokes and headband release mechanism.

Headband Cover: Wide leatherette style vinyl with 12 mm thick polythylene foam padding.

Yokes: Horizontal and vertical pivoting with integral headband lock and detach mechanism for convenient storage

Weight: 12.45 oz. (353 gm)

Cord Length: 47 inches (182.9 cm)

Extension cord length: 72 inches (182.9 cm)

Electroacoustical Specifications

Frequency response: 8 Hz to 35,000 Hz

Sensitivity: 104 dB, 100 Vrms differential input a 1 kHz

Diaphragm radiating surface area: 7 sq. in. (45.6 sq. cm)

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.

E/90 Energizer/Amplifier

Physical Specifications

Dimension: 65mm h x 110 mm w x 155 mm d

Weight: 17 oz. (482 gm)

Electrical Specifications

Frequency response: 1.6 Hz to 50,000Hz (-3dB points) 100 vrms differential output

Distortion (THD+NOISE): 0.001% at 1kHz and 100Vrms differential output

Input Impedence: 100k ohms

Input level: 1 Vrms (for full power output)

Voltage Amplification: 60 dB

Audio output voltage: 600 Vrms differential, 2300 Vrms push-pull with soft limiting

Channel Separation: - 80 dB at 1 kHz, 100 Vrms differential output

Signal to Noise: -100 dB at100 Vrms output

Bias voltage supply: 600VDC

But I've been working after similar lines that the Stax bias supplies simply aren't compatible. 10M/10M will half the voltage but it is strange that there is no ballast resistor. I'm going to ask KG...

[cetoole]

I know what the Koss spec is, and I am wondering if it is simply wrong. Both my analysis and measurements of the E/90 I have right here show 300v. Pretty obviously this isnt a recent change, as my PCB is dated 7-11-90, and has no obvious modifications made to it. I tried adding in a 5M ballast resistor inline with the ESP950 driven off the Stax amp, no change. Its just fine plugged into normal bias. I wonder if just some drivers are able to handle the voltage and some arn't. Also explains why Smeggy didnt have catastrophic results when he ran his ET1000 off the E/90.

[spritzer]

Could be but could you do one thing for me, measure the amp output. If the amp sits at -300V then a +300V bias would result in +600. I seriously doubt that this is the case but it was something which KG did with his DC coupled tube amp.

[Dreadhead]

I know what the Koss spec is, and I am wondering if it is simply wrong. Both my analysis and measurements of the E/90 I have right here show 300v. Pretty obviously this isnt a recent change, as my PCB is dated 7-11-90, and has no obvious modifications made to it. I tried adding in a 5M ballast resistor inline with the ESP950 driven off the Stax amp, no change. Its just fine plugged into normal bias. I wonder if just some drivers are able to handle the voltage and some arn't. Also explains why Smeggy didnt have catastrophic results when he ran his ET1000 off the E/90.

Good find man. I never thought to measure this since I assumed it was fine but yes mine most definitely squealed out of the KGSS until I guess I lucked int a pair that worked after 3 rmas. All pairs except the first one were dead quiet on the E/90. This would explain an awful lot.

[mypasswordis]

This would explain a lot. I was wondering how pretty much every ESP950 had problems, when Koss isn't otherwise known for QC issues. Nice discovery, Colin.

[spritzer]

Since I don't have an E.90 here I asked Kevin to test his and see what's what. For the record I've never had any problems with the ESP950's I've owned over the years, on Stax amps, adapters, Gilmore amps and that crazy DC-DC supply Marc came up with.

[luvdunhill]

damn Colin. Nice work there.

[cetoole]

Could be but could you do one thing for me, measure the amp output. If the amp sits at -300V then a +300V bias would result in +600. I seriously doubt that this is the case but it was something which KG did with his DC coupled tube amp.

Nope, all outputs measure under 1v offset.

[nikongod]

did you account for the input impedance of the volt-meter loading down the output impedance of the bias supply?

Every STAX amp i have measured (attempted to...) measure bias on at the jack says ~200-260V (its been a while)

[cetoole]

My meter is rated for 10M input impedance while measuring DCV, so that in parallel with the 10M resistor to ground in the bias circuit would give 5M. 600v through a 10M/5M divider should give 200v, which is what I measure when doing this. Also, the 300v figure is just from looking at the circuit (10M resistor to bias pin from ~600v rail, 10M||0.022uf from bias pin to ground.

[Kabeer]

What does this effectively mean in the big scale picture?

[deepak]

This would explain a lot. I was wondering how pretty much every ESP950 had problems, when Koss isn't otherwise known for QC issues. Nice discovery, Colin.

Both my ESP950s squealed on the E90. I wasn't able to try other amps. The third one was fine.

[BoilermakerFan]

What does this effectively mean in the big scale picture?

It's potentially part of the problem with Koss driver failures and I'm not sure I would use them off a Stax Pro jack on the SRD7.

[Salt Peanuts]

If my memory serves me right, my ESP950 squealed on both E90 and SRM-1 Mk2. I sent them in to Koss service and when I got them back, they didn't squeal on E90 but continued to do so intermittently on SRM-1 Mk2.