Koss ESP/950 electrostat thread

[Dreadhead]

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.

That was my experience too I believe.

[mypasswordis]

Ah okay, seems some of the drivers still had issues, then. Either way I don't think I'll ever own a pair.

[padam]

So I guess if this is true, best is to use the E/90 to provide the whatever bias and use a different whatever amp to drive the phones with a special cable that splits the drive and bias.

[n_maher]

It sounds like Colin's meter should have a high enough impedance to get an accurate measurement but I'm out of my comfort zone speaking about such things. What I'd do is call Koss and try to get someone who knows something on the phone. There isn't magic going on here so it shouldn't take much to get an actual, verifiable answer and if it turns out to be 300V that'll be pretty interesting considering that some have suggested running them over 600V for best results.

[grawk]

I know my srd7/ta running at 620v sounds pretty darn good with esp950s, and never squealed.

[Vodka]

How does a recabled ESP950 compare with the LNS/Omega2 when it comes to vocal realism? I'm considering upgrading from my SR-Lambda.

[cetoole]

I prefer my Omega2 to my ESP950. No idea about the LNS, havnt had any luck in my searches for one.

[luvdunhill]

I wonder if one possibility is that they've narrowed the problem down to too high bias, and when the headphones (and amps) come in for repair, they lower the bias on the amp. I remember them being unwilling to just fix the headphones, you had to include the amp as well, so that's sorta the basis of my crazy theory.

[spritzer]

You might be right there Marc but it could also be so that Koss can eliminate the amp as the source of the issue. With the SP amps then the phones were often blamed for something which was clearly the amp malfunctioning. OTOH the E90 has so much voltage output capability that a 300V bias should be more then enough.

I'm looking at the only pics I have of an E90 and the bias clearly comes off a large trace (which goes around the edge of the board) and into a voltage divider (can't read the values though as there are wires in the way) and with the bias wire connected between the ballast resistor and a cap. What bothers me is that Stax never used this setup even when they used voltage dividers off the B+. They always use a separate ballast resistor... Anybody with an ESP950 want to do some experiments?

[cetoole]

What sort of experiments?

[spritzer]

If you have a variable bias supply to see how they act at 300V and upwards while playing with ballast resistor values. When the first of the Meier HE90's shipped then users were experiencing noise issues with them running off a Stax 4M7 Pro output jack but they were better when fitted with an inline 5M resistor. Worth a shot...

[cetoole]

I do not have a variable bias supply yet, though adding a 5M inline resistor to my stax adapter did nothing for the noise running off pro bias.

[ericj]

a few points of data:

When i got my esp/950, it had a squeal in both drivers when driven by either the e/90 or my srd-7. at the time i had no pro-bias sources.

Koss insisted on getting their hands on the e/90 as well when i sent it in for repairs.

When i got them back, the work order said they'd replaced both 'elements' and the cable, and 'adjusted' the e/90.

They no longer squeal.

I haven't tried them on a low bias again, but they work fine with the 580v bias on my eXStatA amps, which are variable, but i haven't bothered to vary 'em.

[spritzer]

Are you using a 5M ballast resistor with them on the exstata? I never had any issues with the bias on 3 different sets and multiple bias supplies but something really strange is going on here. Long live Koss and their overly complicated ESP designs.

[cetoole]

Are you using a 5M ballast resistor with them on the exstata? I never had any issues with the bias on 3 different sets and multiple bias supplies but something really strange is going on here. Long live Koss and their overly complicated ESP designs.

No, SRM-1mk2. Still waiting on a few bits for my exstata.

[ericj]

Are you using a 5M ballast resistor with them on the exstata? I never had any issues with the bias on 3 different sets and multiple bias supplies but something really strange is going on here. Long live Koss and their overly complicated ESP designs.

i believe so

[cetoole]

For anyone who may care, here is the E.90 bias supply, front three components. Not visible is that the long leg of the resistor on the left and the right leg on the cap both go to the ground trace on the bottom of the board, while the long leg of the middle resistor goes to B+, which measured at around 612v.

[audionut]

Gentlemen, This is going to be a somewhat long post regarding the infamous noise/squeal problems of the 950's. After reading Cetoole's posts on the E90 bias circuit, I feel I needed to add my experiences with the problem over the past year, so here goes.

I bought my 950's new in December of '08. No problems with noise whatsoever with E90. After about 2 months, decided to build a KGSS P2P, which I did. The dreaded noise showed up in right channel only with KGSS bias set at 600V. Funny thing was, I could take the bias supply from my opened up E90, and use it for bias while using just the amp section of my KGSS, and all was well. I had many emails between KG and myself, trying to "fix" the bias problem with my KGSS, so I didn't have to use the E90 bias. He ended up telling me to try a using a zener string instead of the adjustable bias, with some extra filtering. The zener string worked, but I had to use 450V string to not have the noise issue. Anyway, after a few months, I decided to experiment by adding different zeners in to gradually get the bias up to 600V, and guess what, it worked, with no noise. This combo worked flawlessly for over 8 months, 600V bias with my KGSS. Fast forward to a couple of months ago, and my phones fell off my table and broke the headband where it attaches to earpiece. Sent them to Koss for the repair of headband assembly, and when they came back, the noise was back. I've sent them back to Koss 2 times in the last month, and both times, they still are making the noise in right side only. I got to thinking back to my original KGSS build and all the experimenting I did, and remembered after Cetoole's observations and posts, that my E90 read 597V before the resistor, but it read very low at the actual bias point. Yes, I know, my 10m Fluke 83 will load it down, but on my KGSS and eXStata, it loads it down way less. I never got around to measuring or looking at the actual values of the resistors used in the E90, but I, like Cetoole, think that the actual Koss bias is significantly lower than the 600V that are used in the KGSS and eXStata. I'm not really qualified enough to back this up, but I really feel there's something really different about how the E90 bias is developed, and that it isn't "a true" 600V bias supply as compared to the ones in the DIY amps like my KGSS. Tonight, I'm going to experiment yet again, and vary my bias down on my eXStata down to 350V or so. I do know for a fact, that when I hooked it up to the low bias circuit on the eXStata, and varied it up as hi as it would go (340V or so), no squeal or static. Really sounds like Cetoole maybe right, and the actual bias for the 950's should be closer to 300V. Again, hopefully someone (KG, Spritzer) can confirm this or finally figure this out.

[HeadphoneAddict]

audionut, that's great information. I want to add that for the year that I owned ESP-950 with stax cable (bought from Spritzer) that I never had any squealing issues when used with SRM-1 Mk2 Pro, Woo GES, SRD-7 Pro, or E90 with adapter. It only seems to be some ESP-950 that do this, but not all of them. I really wonder why that is?

[spritzer]

Since I've never heard the squealing noise it's hard to know what's going on but the phones having a physical shock causing the squeal could point to dust issues. The dust covers on the ESP950 are suspect so it ic certainly a possibility in combination of the lack of an output capacitor on the KG bias supplies.

[audionut]

QUOTE=HeadphoneAddict;331848]audionut, that's great information. I want to add that for the year that I owned ESP-950 with stax cable (bought from Spritzer) that I never had any squealing issues when used with SRM-1 Mk2 Pro, Woo GES, SRD-7 Pro, or E90 with adapter. It only seems to be some ESP-950 that do this, but not all of them. I really wonder why that is?

[audionut]

Since I've never heard the squealing noise it's hard to know what's going on but the phones having a physical shock causing the squeal could point to dust issues. The dust covers on the ESP950 are suspect so it ic certainly a possibility in combination of the lack of an output capacitor on the KG bias supplies.

Yes, let me clarify about the "noise". Mine actually makes 2 different kinds, one that I would describe as the squealing noise, and one that' more of a static noise (think TV snow noise, only not as loud). Since mine are fairly new, don't think it's a dust problem, because I've lightly blown them out on occasion, with no difference. Frustrating to say the least. Spritzer, you mentioned adding another 5M resistor. I can easily do this tonight, but exactly where are you talking about adding it? Were you saying to add it in line with the bias pin, which is very easy for me to do as both my KGSS and eXStata are not cased up? Let me know, as I can easily do some experimenting to see if this can be worked out.

[spritzer]

Yes, let me clarify about the "noise". Mine actually makes 2 different kinds, one that I would describe as the squealing noise, and one that' more of a static noise (think TV snow noise, only not as loud). Since mine are fairly new, don't think it's a dust problem, because I've lightly blown them out on occasion, with no difference. Frustrating to say the least. Spritzer, you mentioned adding another 5M resistor. I can easily do this tonight, but exactly where are you talking about adding it? Were you saying to add it in line with the bias pin, which is very easy for me to do as both my KGSS and eXStata are not cased up? Let me know, as I can easily do some experimenting to see if this can be worked out.

The 5M resistor would go in line with the bias making the ballast a total of 10M.

What did Kevin tell you to do with the KGSS bias supply to increase the filtering? A cap to ground after the ballast resistor is what Stax did prior to the SR-Omega project so something similar to that could work.

[audionut]

The final circuit ended up like this C:\Documents and Settings\Tommy\My Documents\Tina Bias 2.htm However, my KGSS no longer has this bias supply, as I recently went back to an adjustable one that's identical to the eXStata bias supply. Wouldn't you know it, my phones last night wouldn't act up (first time they've been good in over a week since getting them back from Koss) with my KGSS and bias adjusted to 500V. Didn'nt inrease it anymore as I wanted to listen noise free for once! Anyway, guess I'll increase to 600V tonight, and see if they'll act up. If they do, I'll add the 5M in to make it a total of 10M, and see if that helps.

[audionut]

Looks like my link above for the bias didn't work. Will try to figure it out and post later.