Replacement Diodes

[mudihan]

Hi!

I want to rebuild my Koss ESP 6 electrostatic headphones, and, specifically, I need to find replacements for two types of diodes, the descriptions of which, as indicated in the schematic, are here:

1, Diode, rectifier, 600v, 1 amp, reverse avalanche characteristic, Power Components, 1EM6S (Only), <1K ohm forward, >20 Mohm reverse.

2, Diode, zener, 185V, 1 watt, Solitron type 1R185A; Power Components TZ185A; Centralab HW 160A or 180A, <1K ohm forward, > 20 Mohm reverse.

None of the exact models is available any more and I want to use Mouser to find the proper replacements. I did find
http://mouser.com/ProductDetail/Vish...IzDutgF8dKk%3d
for No. 1
and
http://mouser.com/ProductDetail/ON-S...lB%2fnwzejA%3d
for No. 2,

But I am not sure they are suitable.

Thanks!

[mudihan]

Anyone with the faintest idea?

Thanks!

[DavidJanszen]

I think those should work fine. You could even just use 1N4007 for the rectifiers. You've not stated the motivation for your rebuild, but FWIW, I wouldn't bother replacing the diodes if the originals have not failed.

[mudihan]

Originally Posted by DavidJanszen

I think those should work fine. You could even just use 1N4007 for the rectifiers. You've not stated the motivation for your rebuild, but FWIW, I wouldn't bother replacing the diodes if the originals have not failed.

Many thanks.

The phones are having some noises (while still sounding great). Since the parts count is low, (5 resistors, 5 caps, and 7 diodes), I want to refresh them with new parts.

I have just attached the schematic. The 3 together in the left corner are the avalanche ones, and the 4 together to the right are the zeners.

I hope that this will help confirming the choices, or even finding more suitable (higher performance) ones.

[DavidJanszen]

Didn't realize the design was self-biasing. Based on that, I'd say 1N4007 is too leaky. I'd shoot for the lowest reverse leakage, high voltage diodes you can find. For the rectifiers, I'd use 1N4936 or MUR180. I have never tried to find low leakage Zeners. That's sort of a conundrum, since they way they regulate is sort of by leakage, so probably that is not an issue, and the generic ones you were planning to use will be fine. Maybe just try to find one with as sharp a knee as possible.

Curious:

When the music is very quiet, have you noticed any tendency to not fully bias the transducers or take a long time to do so?

What kind of noises? Under what conditions? In one or both transducers?

[mudihan]

Originally Posted by DavidJanszen

Curious:

When the music is very quiet, have you noticed any tendency to not fully bias the transducers or take a long time to do so?

What kind of noises? Under what conditions? In one or both transducers?

When the music is quiet, the sound level does drop slightly and gradually (Is this an effect of not fully biasing?). From what I read, this may be due to the small caps being used (0.0068uf). I am planning to upgrade them to 0.015uf film ones. Given the self-bias design, will good power from a power amp and caps of decent size solve the problem of dropping volume in a prolonged quiet passage (such as those in a Mahler symphony)?

The noises are statics and are randomly but frequently present on both sides.

[DavidJanszen]

Bigger caps will discharge more slowly, but also *charge* more slowly, causing longer delays in bias voltage when first energized,and if you go too big, they may never become fully charged except on very loud passages. I am pretty sure that the Koss engineers struck an optimal balance in their choices of parts.

Intermediate improvements may occur if it is possible to substitute caps and diodes that have less leakage, and using some other way of regulating the voltage rather than zener shunt regulation. The zener string will leak even when the voltage is below target.

Consistent levels will only be obtained using a bias supply with a separate power source, rather than "thieving" power from the audio. I'd recommend that as the crux of your rebuild project.

On the other hand, I think the noise you describe will be unaffected by this effort. I think it is probably caused by mere airborne particles built up on the wiring and creating tiny arcing paths for the bias voltage (and hopefully not within the transducers), and maybe for the signal as well. A careful cleaning of the circuit board and wiring using Q-tips dampened with ammonia might do the trick.

[DavidJanszen]

I mean the very plainest ammonia that does not have any detergent added, which will leave no residue. Any detergent residue would be too conductive at those voltages and would thereby degrade or ruin operation.

I'd suggest alcohol or Windex, but that might be a bad idea, since there may be coatings or materials that would be attacked.

[mudihan]

Thank you for the detailed the explanations.

Designing a power supply section for the given self-bias design is something that's far beyond my capability -- I am not trained as an engineer myself. I will certainly try cleaning the units though I may also parallel caps of different values (one large and one small, for example) to see if that helps both faster charge and slower discharge.

[DavidJanszen]

Adding capacitance will slow the discharge, and it will also slow the charge. It may limit its ultimate value to something lower than what it should be. In other words, it will take longer to come up to full voltage, and may never get there. Engineering has a lot of these trade off situations, and even though you are not an engineer, you may enjoy confirming or invalidating this prediction as you vary the capacitance. You might ask Russ Knotts at Just Real Music about an external supply for your headphones. www.justrealmusic.com