Help diagnosing wind fluttering/rustling noise in D-87/U87i DIY build

[dchang0]

I just finished building two U87i clones using poctop's D-87 boards. One turned out great--let's call this Mic 2. The other, Mic 1, has a low fluttering/rustling noise that sounds like a light wind blowing all the time and occasionally gusting. It is worst in Figure 8 and Omni modes but is much quieter in Cardioid mode. The front capsule seems stronger than the back capsule when in Figure 8 or Omni modes.

I've seen this exact problem mentioned a few times in other threads including the giant build thread and also threads on other sites, but no solution.

A few persons suggested checking the capsule and the Styroflex caps, but none of these were confirmed as the cause.

A side note that may be related: one of the capsules had the clear rubber tubing used to hold the wires together stuck to one diaphragm inside the case. I was able to shake the rubber tubing off without touching it, and there is no visible damage, but it is possible that some damage may have occurred.

So, here's what I've done so far to Mic 1:

• I tried swapping front and back of the capsule in Mic 1. No obvious difference. Front still appears strong, back appears weak. This suggests it is the circuit, not the capsule.
• I tried swapping the capsules between Mic 1 and Mic 2. The problem SEEMED to follow the capsule, but I am not so sure now. See the next diagnostic test outcome...
• I took the capsules completely out and put in the test capacitors (100pF silver mica; I didn't have 72pF as these are hard to find!). The result is certain: the wind noise is present in Mic 1, even without the capsule installed. This does not mean the capsule that came from Mic 1 isn't bad--it may still be weaker on one side than the other. I will test the capsules in Mic 2 next.

Anyway, my goal is this: I want to diagnose and fix this problem precisely with the least cost and least risk. For example, ordering a ton of spare polystyrene caps and swapping them in and out is both expensive AND risky (each resoldering may damage  possibly good polystyrene caps--let's say the one bad cap is the 3rd one, and while swapping the 1st and 2nd in and out I damage them. I end up with three bad caps when I originally had only one bad one).

I'd like your help figuring out exactly which one or two components are bad. Right now we know Mic 1's noise is in the mic circuit without the capsule. We know the noise is quietest in Cardioid mode and loudest in Figure 8 and Omni. That probably narrows down the problem component to one of a handful that are along the Figure 8 or Omni paths. I don't understand the circuit well enough to know what each component does.

After I test the capsules and confirm they are both good, I'll try swapping one of the two boards (probably the board with the most polystyrene caps) from Mic 2 into Mic 1. That'll rule out half the components (the problem might not be a polystyrene cap--who knows?)


Any and all help is appreciated. Thanks! I plan on posting the solution for others who encounter the same problem to see.

 

[Matador]

Just as general advice, start at the output of the mike and work backward.  Ground the input of the transformer, and confirm the microphone is perfectly silent.  Then ground the gate of the JFET, and confirm the microphone is perfectly silent.  Then move backwards through the circuit until you get to the capsule.  You should be able to isolate that way.

 

[Tim Campbell]

Very often this type of noise is caused by contamination within the capsule, on the pcb or in the switches.

 

[dchang0]

Thanks for the tips.

Re: contamination

Would I be able to see this contamination, or might it be so small or hidden that it can't be seen? I definitely washed the board as often as I could with isopropyl alcohol (until the Styroflex caps were installed, after which washing was only done with a Qtip dipped in isopropyl alc., but there were definitely times when solder flux or fingerprints could have gotten on the board during installation), and I handled the capsule only while wearing disposable powderless gloves.

Right now, I can't see any obvious contamination under bright light, but I have no idea what to look for.

I'll definitely try working backwards through the circuit as Matador suggested.

UPDATE:

Also, I have an oscilloscope that could be used to search for the noise in the signal at various points along the signal path. Thing is, being totally new to microphones, I am having a hard time tracing the signal path on the schematic (it is much easier with valve guitar amps).

 

[Tim Campbell]

I doubt you would see the contamination. What capsule are you using? I recently had a chinese ck12 type capsule with this very problem. I opened it only to find the backplate had never been thoroughly cleaned and was still coated with some form of residue.
A small hole in a membrane will also cause this problem. That can happen very easily when a capsule is delivered with cables attached. All it takes is for the cable end to come into slight contact with a membrane.

 

[dchang0]

Thanks, Tim.

These are a matched pair of RK87 capsules from M-P. I expect they were well-made and passed QA at M-P, but somehow one of the pieces of rubber tubing used to hold the capsule in the case got stuck to one of the diaphragms before I opened the case. M-P has graciously offered to replace the matched pair once I send them back for testing.

1) I have a bad circuit in Mic 1, not yet diagnosed. We know it is the circuit because it flutters even with no capsule installed (test capacitors in place of both sides).

2) I have a bad capsule. We know this is bad because it has light wind noise in Mic 2 that tests silent with no capsule installed.

Making some progress... I'll test the good capsule on Mic 2 to make sure my tests got it right. I also want to test it with a known-good K47 capsule from a U47fet build that definitely does not flutter.

 

[dchang0]

I have tested the front diaphragm, suspected of being bad, of the bad capsule versus the rear diaphragm, suspected of being good, wiring each of them in as the front capsule and with the same test capacitor wired as the rear capsule. This is in Mic 2.

Just looking at the waveforms from recording each side shows that the front, suspected bad capsule, has roughly double the fluttering noise as the back capsule, which also has some fluttering noise.

I will now try the expected-good capsule, which I recall had no noticeable fluttering noise.

 

[dchang0]

Greaaaaaat. Now the expected-good capsule also makes fluttering noises on Mic 2.

Perhaps the "good" Mic 2 isn't good after all, yet, it is utterly silent if test capacitors of 100pF are installed in place of the front and back sides...

Now I'm basically lost. I'm going to test the K47 capsule in Mic 2 to see if I can break this deadlock.

Another important thing to note: the signal to noise ratio is very low. My speaking voice about 6in. from the capsule is not much louder than the wind/fluttering noise in any of these tests on either capsule. That seems very wrong.

 

[dchang0]

Okay, the known-good K47 capsule from the known-good D-47fet build works in Mic 2. Back side has the test capacitor. It too has some fluttering noise but slightly lower (visible on the waveform) than the RK87 capsules, and the voice signal is slightly higher (also visible).

So this seems to point to Mic 2 as having a problem that goes away when no capsule is installed.

I will now put the RK87 capsule (one side at a time) into the known-good D-47fet build to see if it becomes as quiet as the K47 capsule and test all four diaphragms in the matched pair.

 

[dchang0]

One side of the expected-good RK87 capsule tests fine in the D-47fet build. The noise floor is much lower and there is no fluttering. Also, the signal to noise ratio is much higher--my speaking voice signal is very high relative to the noise floor.

So this is looking like it is a problem within the D-87 builds, with one build better than the other but both pretty bad.

I will check the other three diaphragms on the RK87 capsules to exonerate them both. I still expect one side of the expected-bad capsule to be bad, since it did have much more noise than the other 3 sides so far.

 

[dchang0]

I just finished testing each of the sides of the two RK87 capsules in the known-good D-47fet build, and they all work perfectly. Nice and loud, low noise floor, and no wind noise.

So it is safe to say that the wind noise is coming entirely from the two D-87 builds, Mic 1 much worse than Mic 2, and Mic 2 behaves well when there is no capsule installed at all (two 100pF caps in place) and misbehaves slightly when a capsule is installed.

This is a very puzzling problem, but at least I can avoid returning the two RK87 capsules for now.

I will start first by checking the biasing of the two D-87s.

 

[dchang0]

Dang. Now Mic 2 also has noticeable fluttering with no capsule installed (only test capacitors).

So the act of swapping in and out the capsules and test caps resulted in noise. I am pretty sure there is no contamination, since I went over the hi-Z board with a Q-tip dipped in isopropyl alcohol. Oh well, time to start diagnosing the circuit...

 

[Tim Campbell]

Also try cleaning the switches and check the xlr wiring.

 

[dchang0]

Will do, Tim. Just measured resistance across pins 2 & 3 on the XLR and on both mics, they are 39.4 ohms, which seems to be as expected.

I visually checked the transformer wiring to the PCB and it appears to be wired correctly. I have the newer "D-87M" blue boards and the Cinemag CM-13113.

Yellow = S+ goes to C16
Orange = S-
Brown = P-
Red = P+ goes to C7

Well, according to the Cinemag schematic, if the dot represents the positive side, I really should swap Y and O and swap B and R. I just copied Andrew Schaap's wiring from his photos.

Also, I searched the main build thread and found that several other builders described having the same or similar problems as me:

Same:

hugo: exact same symptoms
OrganizedKonfusion: "sounds like I've got my mic out in some strong winds." Almost exactly the same symptoms. He got different results by tightening the saddle screws--considering that I was removing and replacing the capsule during testing, this is definitely a possibility, but his saddle fitment sounds like it is much worse than mine because I used the saddles that came with the RK87 capsules and were a perfect fit.

Similar:

Guitylerham: "stormy noise"

 

[dchang0]

It does look like it might be faulty switches.

On Mic 1, with no capsule and two test capacitors installed, I removed the wires from only the pattern-selector switch, leaving the Low Cut and Pad switches wired in.

And the windy noise went away in all positions (I used tweezers to simulate Fig 8 and Omni). In terms of noise, Cardioid is the quietest, Omni has slightly more noise, Fig 8 has about the same noise as Omni but slightly higher-frequency.

I'll solder the switch back in, and if the noise comes back, then voila, problem diagnosed!

 

[dchang0]

Okay, confirmed that the pattern switch adds noise. However, I notice that the wind/fluttering noise is still present, just about -6dB when the switch is removed at the switch end (leaving the three wires running to the switch attached at the main board end).

Do you think that the twisted wires to the switch might be acting as a sort of capacitor? I wonder if straight wires, separate from each other, would be better. First, I'll remove the wires entirely. If all the noise is gone, then great.

 

[dchang0]

Confirmed. I removed all the switches AND the twisted-pair wires that were leading to them (the headers on the hi-Z PCB are unpopulated).

All the wind noise is gone from Mic 1, even at high gain.

So the twisted wires and the switches were at fault (well, we will know for sure once I add the capsule back in).

I guess I need to buy a new set of switches and run untwisted wires to them, but where does one get these switches (or preferably, better ones)?

 

[dchang0]

ALMOST SOLVED Mic 1:

And the winner is.... Tim!

Completely removing the switches AND the twisted pairs of wires going to them from the hi-Z board absolutely removed all wind noise. I put the "bad" capsule into Mic 1, and it was ultra-quiet even with the gain cranked, and the signal was nice and loud. I even soldered in a jumper and tested both Omni and Fig 8 modes, and they worked just as quietly.

I rewired all the same switches back in (including poctop's switch PCB) using straight wire runs (not twisted), and there is no noise.

So I don't have to buy anything except new switches (and I don't want to buy the same kind, given that they were so noisy).

Next question: what other switches fit poctop's switch PCB for the SYT-5 body? I recall someone asking about alternatives before--they wanted a shorter toggle...

UPDATE: There is now a different problem--the "just fixed" Mic 1 will work perfectly quietly for a while (with switches wired in with straight wires), and then a sudden rush of loud rustling will occur. I'm not touching the mic when this occurs, so it is not likely to be a loose cable or wire or something that would shake loose. This is a new problem I have not heard before.

 

[dchang0]

Uh oh. Mic 2, the one that is supposed to be good with the good capsule, does have some very light wind noise when no switches or wires are attached... So there is something else, perhaps a little contamination, going on... It may even be the "good" capsule. I'll attach the switches with straight wires and see how bad it is... I may need to put this capsule back into the known-good D-47fet and leave it in there for a longer time to see if it is the capsule. This is supposed to be the "bad" capsule.

UPDATE: Just hooked up only the pattern switch with straight wires in Mic 2, and the wind fluttering noise is back and as strong as before. The switch is likely bad. Guess I need to order an alternative switch after all.

UPDATE 2: Confirmed by removing the switch and the wires that the wind noise is coming from them. I also tried a short jumper to activate Fig 8, and there's the little bit of light wind noise that is probably due to contamination of the capsule. I am very sure this switch is dead.

So, the current list of problems:

1) Mic 1 that is mostly repaired has weird intermittent loud rustling noises that happen once per 3-4 minutes, even when the mic is left alone and nothing is touching or moving it. Shaking the mic on purpose in case it is a loose connection does not trigger the problem.

2) Mic 2 has constant light wind noise even with no switches or switch wires attached. May be bad capsule.

3) Mic 2 definitely has bad pattern toggle switch--any time it is hooked up the loud wind noise comes back.

So there's a lot of improvement, but the mics are still not quite usable.

 

[dchang0]

No luck figuring out the intermittent loud rustling noise in Mic 1, and while trying to figure it out by pushing on things with a plastic spudger and then reflowing nearly all the solder joints, I somehow added back the continuous light wind noise, even without any switches or switch wires attached! I also managed to make the loud rustling noise continuous, but then I was able to make it go back to intermittent by reflowing the solder joints yet again.

So the problem of wind noise can be created by something other than the switches. That was apparent anyway, because the switch and switch wires are not attached in Mic 2, and that one had light wind noise too.

That said, I get the least amount of intermittent or continuous noise with the switches and switch wires totally removed. They add to the problem for sure, and by a lot.

I am at a loss here. I've been soldering for decades and make nice shiny joints, so I don't think it's because of poor soldering techniques.


At this point, what I may do is order a whole set of new switches (different design) and replace the Styroflex caps with silver mica and film and see what happens.