Is there a way to reduce cell phone interference in DIY KM-84?

[joulupukki]

Slipping a ferrrite bead onto that gate lead is that much more worth a shot, then.

Could be worth a shot, but why would the mic exhibit more noise when the cell phone is closer to the XLR connector instead of near the capsule?

 

[joulupukki]

I see a possibly unpopulated cap position by the JFET - maybe part of the problem? Also, you could put very small caps to ground from your hot and cold outputs, there's room.

The missing C2 cap is intentional to achieve more sensitivity out of the mic and is omitted in the DIY and the MP builds.

Not sure I understand the part of inserting caps from hot and cold outputs to ground. Is there a schematic of this somewhere and what's the purpose of this usually? Something like 10nF?

 

[joulupukki]

This appears to be another case of audio rectification where an external signal is being demodulated within the mic, and the connections and case shielding are good. Ferrite beads may be the answer. You should get ferrite beads that are optimized for the UHF frequencies; these beads slip over the conductor. First, try a bead over the connection to the gate of the input FET, test with the cell phone, and compare the audio with/without bead. Next, try two beads, one on the lead to pin 2 of the XLR and the other to pin 3 of the XLR. This is a low impedance connection and should not affect the audio. The output transformer should stop RF but treating connections with ferrite beads to the "outside world" should help. I make my living with RF. Hopefully this works.

Interesting. Can you send me a link or manufacturer/part number to something that may work?

 

[ccaudle]

Not sure I understand the part of inserting caps from hot and cold outputs to ground.

The idea is to form a filter for RFI currents which get onto the signal lines to return to the chassis/shield. Tie those caps to pin 1 or directly to the metal shell.
Typical values would be 100pF to 1nF.

 

[Cqwet Dbdfte]

It most definitely DOES have a JFET input.


That sounds kinda risky, with exposed components/terminals...

Not really, I'd cram that carbon/plastic foam into the cavity.
Sometimes ESD black carbon foam can work. It is not a solid conductor. Any RF getting into the body would be attenuated. That JFET gate should an RF bead or maybe even a nH range inductor, an air wound coil may work.
That screen should low resistive conduction to the case at several points.

 

[Soundwaves]

This may be way out there, but has the old remedy of interference using a 52ohm capacitor across the leads stop being used. In the early days of TV when a ham radio was nearby the ham interference was stopped by the 52 ohm capacitor.

Just a thought for you. (If I missed seeing it in the previous posts, I apologize.)

 

[Khron]

52ohm capacitor

Ummmm...?

 

[Soundwaves]

Could have been a resistor, but I am remembering it being talked about 40 or 50 years ago. The 52 ohm is what stands out. And the ones' Dad used did look flat and clay colored. Not like a resistor.

 

[JohnRoberts]

This may be way out there, but has the old remedy of interference using a 52ohm capacitor across the leads stop being used. In the early days of TV when a ham radio was nearby the ham interference was stopped by the 52 ohm capacitor.

Just a thought for you. (If I missed seeing it in the previous posts, I apologize.)

capacitors don't come in ohms...

I recall as a young puke seeing small ceramic disc capacitors placed across outlet/plug leads. I recall disassembling one plug-in fixer and the small ceramic disc was all that was inside it. Caveat, this was back in the 1950s.

JR

 

[mrgrooves666]

The idea is to form a filter for RFI currents which get onto the signal lines to return to the chassis/shield. Tie those caps to pin 1 or directly to the metal shell.
Typical values would be 100pF to 1nF.

Jules takes this approach with his OPA alice design, as Rogs with his Opic, If I remember 10-22 nf from pin 1-2 and pin 1-3. Altough these are traffo-less designs, is this also a viable practice for traffo-coupled mics, or any kind of mic for that matter?

 

[Ike Zimbel]

The idea is to form a filter for RFI currents which get onto the signal lines to return to the chassis/shield. Tie those caps to pin 1 or directly to the metal shell.
Typical values would be 100pF to 1nF.

This is what the Neutrik EMC XLR connectors have going on inside the shell. It's not a fix for the mic itself, but you could get a couple of those and keep them with these mics.

 

[Khron]

Altough these are traffo-less designs, is this also a viable practice for traffo-coupled mics, or any kind of mic for that matter?

U47FET, 2x 100pF
U77, 2x 10nF
U87A, 2x 330pF
M269, 2x 2nF
U67, 2x 2nF
C414EB P48, 2x 22nF
C414B-ULS, 2x 10nF
C414EB, 2x 22nF

 

[cyrano]

TBH, I've never heard cell phone noise like that. Is that perhaps because of CDMA?

 

[Khron]

It definitely has to do with the (type of) network, especially if the signal's wonky and the phone / cell tower needs to fall back to the lower frequencies of 3G, they more readily down-modulate or intermodulate into the low-kHz area, which results in the all-too-familiar(?) "ti-tiri-ti-tiri" loop.

I've experienced that with my monitors a few years ago, but not since the 4G (and now 5G) stuff became more widespread.

 

[jp8]

I have RFI tested several Schoeps style mics with the 10, 15 or 22nF wired capacitors mounted on the PCB and on the XLR. YMMV, but in my tests it made no difference as far as RFI sensitivity is concerned when I injected RF signals between 500 and 1200 MHz. Those capacitors act more like inductors the way they are mounted. If you want optimal RF interference suppression, the PCB and circuit design and the pin1 to mic body connections should be designed with RF in mind.

Although I'm afraid it will prove difficult to make a mic RF proof when the transformer and its wires are close to the XLR and there is hardly any space for beads and other stuff others have already mentioned in this thread. Yet, I have designed a PCB for the CM-60, based on the KM84 circuit and added a ferrite bead between pin 1 and signal ground. I just want to give it a try and don't want to rule out that it helps. A common mode filter + beads + XLR with built-in spring-loaded contact to the mic tube as in the CM-63 will for sure work better, but is not an option here.

Btw, talking about the ferrite bead: in the original KM84 circuit, there is a 10 Ohm resistor between pin 1 and circuit ground. Don't know whether this is for hum or RF suppression or both. Anyway, it is not on the MP and Graeme's PCBAs, maybe due to a lack of space or other reasons? But MP did include this resistor on the LDC version of the KM-84. Does anyone know why? What surprises me, by the way, is where Neumann decided to make the connection of the mic tube to the circuit: it is connected to signal ground at the impedance buffer side of R13 instead of to pin 1. This means there is an interruption of the Faraday shield, which should be contiguous.

Jan

 

[joulupukki]

Confirmed that 1nF caps from pins 1 - 2 and 1 - 3 does nothing. Perhaps higher-valued caps would work ... or ... the simplest solution is just to have people coming to record put their phones into airplane mode or leave them in the other room ... or use the MP mics. My original goal (other than having fun with a DIY KM-84 project) was to get these DIY mics sounding as good as the MP mics and sell the MP mics (since I only really need one stereo pair). They definitely sound as good as ... they just have this annoying sensitivity to cell phones. It'll be interesting to see what you discover @jp8 if you end up building that PCB for the CM-60 – and you definitely free up space by using SMD components.

 

[cyrano]

It definitely has to do with the (type of) network, especially if the signal's wonky and the phone / cell tower needs to fall back to the lower frequencies of 3G, they more readily down-modulate or intermodulate into the low-kHz area, which results in the all-too-familiar(?) "ti-tiri-ti-tiri" loop.

I've experienced that with my monitors a few years ago, but not since the 4G (and now 5G) stuff became more widespread.

That sounds all too familiar, but in joulupukki's case, it's just klak klak klak...

Now he's in beautiful Utah, so I'm guessing the cell phone system is CDMA? I think only Vodaphone uses GSM in the USA?

 

[jp8]

It definitely has to do with the (type of) network, especially if the signal's wonky and the phone / cell tower needs to fall back to the lower frequencies of 3G, they more readily down-modulate or intermodulate into the low-kHz area, which results in the all-too-familiar(?) "ti-tiri-ti-tiri" loop.

I've experienced that with my monitors a few years ago, but not since the 4G (and now 5G) stuff became more widespread.

Those "ti-tiri-ti-tiri" loops are 217 Hz bursts, AFAIK. Anyway, that's the frequency I use to modulate the 500-1200 Mhz oscillator in my RF jammer.

But even if the RF signal is not AM modulated and your mic picks up RFI, it may cause distortions. When a continuous RF signal is demodulated in BJT circuits, it will just create a DC offset that may increase THD without being immediately noticed. The audible AM modulation makes it easier to detect the RF interference, so I simply modulated the carrier by switching the power to the oscillator at 217 Hz, i.e., giving 100% AM modulation.

Jan

 

[joulupukki]

We’ve got 5G networks here, but my house, especially in the basement where my studio is, barely hangs on to a data connection via LTE.

 

[jp8]

Confirmed that 1nF caps from pins 1 - 2 and 1 - 3 does nothing. Perhaps higher-valued caps would work ... or ... the simplest solution is just to have people coming to record put their phones into airplane mode or leave them in the other room ... or use the MP mics. My original goal (other than having fun with a DIY KM-84 project) was to get these DIY mics sounding as good as the MP mics and sell the MP mics (since I only really need one stereo pair). They definitely sound as good as ... they just have this annoying sensitivity to cell phones. It'll be interesting to see what you discover @jp8 if you end up building that PCB for the CM-60 – and you definitely free up space by using SMD components.

FYI, at the frequencies of interest, smaller caps may even prove to work better than larger caps. It's all about parasitic inductance, which is often smaller with small capacitance values, depending on the capacitor geometry. But many Schoeps style circuits have those 10...22nF caps on the outputs, which are most likely there to suppress RFI from the microphone. Mics with a polarization voltage oscillator running at 2Mhz may turn into RF transmitters without these caps, because of a poor PCB layout. The caps remove this interference, which shouldn't have been there in the first place, btw.

Jan