#11 sE Electronics sE X1 rebuild

It's been a good few years since this mic (sE Electronics X1) came into my possession. As many others i have, i bought this one as faulty. The main problem was, it was noisy as hell (think wind-noise, or the sound of blowing straight into a microphone).

Once i had another capsule to test this with (a temporarily removed capsule from a Shure KSM27), i figured i'd hook it up to the X1 to see whether the capsule was the source of the noise. Nope, the noise was still there. After i ruled out the (stock) capsule, i replaced it with a 100pF styroflex capacitor, for minimal hassle (and risk of damaging a capsule) during subsequent testing of the electronics.

After that, i saw the two ceramic anti-RFI capacitors on the output connections. From experience, both personal and of others (<cough>Shure<cough>), i was aware that at least in certain cases, ceramic caps CAN develop into noise sources. Fine, let's swap'em out. I replaced them with a pair of 22nF film caps - pretty standard value, in the case of this Schoeps knock-off circuit design. No change, yet again.

Fine, what next? There's the two PNP transistors that drive the outputs. I replaced the stock ones (can't remember what they were, very likely some 2N5401 / MMBT5401 variant) with a pair of 2SC2406 transistors. That did no good either.

Ok, let's replace the DC-blocking (and high-pass) capacitors as well - through-hole ceramics, this time. Fat lotta good that did (read: didn't help at all). This is getting annoying, you know that?

Ended up replacing the JFET as well (for the second time, in fact). Stock it might've come with a 2SK170. I replaced that too, with a J305. Nope, still no change, still noisy as hell.

Getting a bit exasperated here. Just for the sake of thoroughness, i also swapped the two protection diodes across the PNP outputs, with a pair of 5.6v zeners (stock were a pair of MELF 1N4148s, most likely; no visible markings apart from the cathode band). That made precisely zero difference too.

I even replaced the high-value resistor - who knows, there might've been a lot of leakage across it. That didn't do squat either. This is unreal...

That's it, time to whip out the heavy artillery. In other words, i recreated the circuitry digitally (or virtually, if you will), scrambled together a single-sided PCB design, and even added a DC-DC converter, to increase the capsule bias voltage (or at least provide the option). The stock board is pretty cramped, with a mix of SMD and through-hole components, and making another PCB is easier and less stressful (or fiddly, at least) than struggling to replace the remaining stock passives.

Took me a good few hours (interspersed among some work, meals and YouTube), but i got it done. Just to make it clear right off the bat: most of the component and value choices were made based on the "because i had some" criterion.

As it turned out, the mounting holes positions from the real board weren't in 100% aligned to the grid in the CAD software (and i couldn't be arsed to tweak them to THAT extent), so those were drilled with the plastic-mounting-doohickey as a stencil / template.

Don't judge me on the aesthetics - it's a one-off and/or a "proof of concept", if you will. To prove that some of the passives left on the stock board must've been the source of the noise. And prove, it did.

The JFET source and drain resistors are the standard Schoeps 2.2k, but everything else was nearly arbitrary, and/or based on measurements. The DC-DC oscillator is SMD because, as pointed out above, i had some appropriate components for just such a circuit. I really went to town on the power filtering though - four 47u caps, just for the heck of it. And "because i can". That's one more than the stock circuit had. Wink-wink, nudge-nudge...

The JFET i ended up using was a 2SK30A, recovered from an AKG Perception 200 i had modded earlier. Same as the capsule, in fact - might as well put it to (good) use, right? The gate resistor is a 2Gohm one - can't remember what other mic that came from. The two output PNP's, even though i had initially designed the board for employing a pair of 2SA970, i ended up using two 2N5401s ('cause i had more of those). Other than that, i didn't have a 15v-ish Zener diode, so i series'd up three 5.6v ones. Just to "guild the lily", i even whacked in a 1Meg trimmer to bias the JFET gate - except for the fact i couldn't be bothered to mess with that (hook up a sig-gen, monitor distortion etc); maybe later...

Also, i chose to go ahead and peel off the copper from the area where the teflon stand-off pin sits, just to minimize whatever potential for leakage currents there may be. We're dealing with gigohms of resistance / impedance here, so any pollution or contamination is a leakage path.

I ended up having to add an extension wire for the capsule bias voltage. The wires of the capsule  at hand were quite short, so i just had to do what i had to do, to get it done. One extra 2mm hole, and a pretty stiff piece of wire out of a laptop battery, and that was that.

And whaddya know - it's FINALLY as quiet as you'd hope / expect. Take that, evil forces of mysterious component issues!..

I've yet to hook up the pad switch with the associated capacitor, but... Eh, that can wait as well. Not exactly an uber-critical feature. But i did hook up the high-pass, and that works fine too. Score!...