Interference spikes in studio gears

[mattenylander]

Hi,
I have a problem that I haven’t been able to fix… it appears as a static spikes that is equally spaced between each spike usually from about 875Hz, 2kHz, etc.

This is how it can look on the spectrum analyzer. This is just a cable from my Duet 2 out -> and back in Duet 2 In.

Strange thing is, if I hold my hand on the duet 2, just about where the logo is on the glass, the spikes disappears.


This pattern/or noise, whine is apparent in almost every gear from Microphones, audio interface, some Eurorack synths, Warm audio Bus Comp and even with a portable zoom Hn1 on battery in my home studio (living room)

When I use the Hn1 with headphones and walk around in my home studio the interference sound increases and decrease in some spots. In my bedroom however the interference noise/whine is gone…

At first I hade the same problem with my Rode NT1 as well but have managed to reduce it with the help of the new Neutrik EMC XLR connectors.

I can’t really grasp my head around it cause it seems like an interference noise… but in that case, why would touching the duet 2 reduce it? That to me is a sign that this is a ground loop problem.

So please some wise out there give me some ideas what it could be causing this spikes.

 

[Bo Deadly]

If you move it farther away from your computer does it make a difference?

What sort of lighting do you have?

It is very common to have noise at mains frequencies. But these are not mains frequencies. So my first thought would be that you have some device that is emitting EMI. One such device might be lights and especially lights on a dimmer were the dimmer crudely rectifies the mains AC.

Another source of such EMI might be a large tower PC that has after-market components like LED lights and fans.

Or something is switching power in a bad way like some 90's gear (the dark ages of digital) with a really crappy switching mode power supply.

Turn off *everything* including gear, lights, computers, network switches, vibrators, TVs, phones, etc ... except a laptop with one loopback cable attached and watch the spectrum. Take a snapshot of the spectrum and then switch one thing on like the lights and compare the spectrum with and without. Then repeat for each device one by one.

 

[Ike Zimbel]

How close are you to the nearest cellular antenna site? This is increasingly becoming a problem, not just with RF gear like wireless microphones, but also straight up audio equipment. The fact that the Neutrik EMC connector sorted this problem on your mic is a clue that RF might be the issue.

 

[warpie]

How close are you to the nearest cellular antenna site? This is increasingly becoming a problem, not just with RF gear like wireless microphones, but also straight up audio equipment. The fact that the Neutrik EMC connector sorted this problem on your mic is a clue that RF might be the issue.

What are some common frequencies that this problem occurs? Isn't 1-2kHz a bit it too low?

 

[Ike Zimbel]

The cell service starts at 618 MHz and goes up to 652 MHz from there (if all cell blocks are active in that market). In any case, it's kind of a "power trumps frequency" issue...there's just a ton of RF level coming off these things and it will get into anything that is even remotely susceptible. A big part of the issue is that the Telcoms are building out their networks and a lot of cell sites are getting physically closer to dwellings and people in general. I talk about that in this article.

 

[ccaudle]

What are some common frequencies that this problem occurs? Isn't 1-2kHz a bit it too low?

It is the signal modulation being demodulated by the equipment silicon junctions. I don't know what the newest modulation schemes sound like, but that interference looks similar to the old GSM interference that gear used to deal with (sounded kind of like a rough buzz that turned on and off when a nearby GSM phone rang).

why would touching the duet 2 reduce it? That to me is a sign that this is a ground loop problem.

I thought you said that spectrum was just the Duet output back to the Duet input. A "ground loop" is when you have shield current flowing between two devices that then gets into the audio reference at some point. A device connected to itself can't form a loop.
More likely that when you touch the device you change how effectively it acts as an antenna.

Make sure all of your cables have the shield connected at both ends. If the device has pin 1 problems you may be able to modify the device layout, you could use the EMC connectors as you pointed out, or you can use some copper tape to connect the cable connector pin 1 to the shell tab.

 

[mattenylander]

Thanks for all wise answers!

This is what I see from my window ( same room as my studio) not sure what type of antenna that is? The source of my problem?




I build a simple pick up coil that I plugged into the Hn1 and recorded it and this is how it sounds/looks like.


This is the same recording setup but in the bedroom, no noise.

 

[mattenylander]

It is the signal modulation being demodulated by the equipment silicon junctions.

Ok so just so I understand it better, do you mean that cell antenna sending out RF in the 600MHz - 700MHz area that gets picked up by the transistors in the circuit of my devices? And that gets demodulated in the transistor into a static/spikes interference type sound?

 

[mattenylander]

A big part of the issue is that the Telcoms are building out their networks and a lot of cell sites are getting physically closer to dwellings and people in general. I talk about that in this article.

Thanks! I’ll read that article!

 

[Bo Deadly]

I somewhat doubt that pickup is working very well. The uninsulated part is doing nothing. You need thin insulated wire like "magnet wire".

Note that metal, even thin metal, will almost completely block RF noise. So if the walls and roof of your building are metal, they will effectively block RF from sources like cell transceivers. Although any gap may let the RF pass. The shorter the wavelength, the smaller the hole that can let that RF in. You can see something like 5GHz might get through even a 1mm of a gap:



But glass will not block RF much. So the problem could be that you have RF coming in through the windows. Fortunately you can buy silver window film that will pretty effectively block RF. Search "silver window film" on Ebay or similar. Note that it does NOT have to be grounded. Just make sure you go to the edges.

Just to see if it would help, try putting something metal in front of the windows. But again it really needs to be over the edges and right up against the window so that there are no gaps.

 

[ccaudle]

do you mean that cell antenna sending out RF in the 600MHz - 700MHz area that gets picked up by the transistors in the circuit of my devices? And that gets demodulated in the transistor into a static/spikes interference type sound?

Yes, exactly. And other frequencies, in some settings there are high powered handheld radios (walkie-talkie style) in lower frequency bands (more an issue at concerts or conventions, not usually in a house), TV broadcast in that range and higher, cell phones also broadcast in the 1800MHz-1900MHz range. The higher frequency interference like 1900MHz phones is harder to shield, because even small gaps in shielding let those wavelengths out and in, but normal audio wiring also becomes very lossy at those frequencies, so it varies whether there are problems above 1GHz or not. The old GSM phones used 900MHz and 1800MHz, you used to hear interference from those a lot, but vendors got better at filtering the inputs and outputs to keep phone frequencies out, and the newer signal modulation schemes are not quite as annoying as that GSM handshake at the beginning of a call.

Interference problems usually reduce quickly with distance, so it seems the opposite problem may have occurred for you, your interference problems increased greatly because someone decreased the distance by a large multiple between your flat and the nearest transmit antenna.

 

[Ike Zimbel]

Those do look like cellular antennas in your picture. Were they installed recently?

 

[mattenylander]

Those do look like cellular antennas in your picture. Were they installed recently?

Ok I understand, yeah I think they where installed about 6 months ago.

 

[Brian Roth]

I've "fought" the noise from cell phones/towers for years now. First I recall was in a PA system used in a "city" meeting, when all sorts of buzzes/pharts went through the system as the politicians were speaking.

More recently when working at a studio cutting lacquer for pressing.....

We never heard any probs when doing a cut, and only after the test pressing there were "urps" and "pharts" on the pressing. Our local cell phones somehow found their way into the cutting amp. Nowadays....all cell phones OFF when cutting.

Bri

 

[mattenylander]

Just to see if it would help, try putting something metal in front of the windows.

I’ll have a experiment with this at this weekend and report my findings!

 

[Brian Roth]

This sounds funny, but even a metal "cookie sheet" can make a difference while testing.

Bri

 

[mattenylander]

ground loop" is when you have shield current flowing between two devices that then gets into the audio reference at some point

Ah Ok I thought that one could create a ground loop within the circuit as well.

It is the signal modulation being demodulated by the equipment silicon junctions.

Interesting, Do you know a source I read more about this? To Learn some more about it.

Make sure all of your cables have the shield connected at both ends. If the device has pin 1 problems you may be able to modify the device layout, you could use the EMC connectors as you pointed out, or you can use some copper tape to connect the cable connector pin 1 to the shell tab.

I noticed yesterday that, if I use and external preamp (Sub Zero RS100) -> Duet input I don’t have any noise/spikes even with unbalanced cables, but if I patch Duet out -> Duet in I get the noise/spikes, Balanced cables makes it better but don’t remove it completely so it seems that that the Duet Output brings//transfer/pick-up the noise…

I also get the same noise/spikes if I patch the Behringer Mic 100 preamps Out -> Duet in, and and nothing patched in to the mic 100. The noise spikes get worse when the pad is engaged on the Mic 100.
So it almost seems like the outputs of those bring/transfer/pick-up the noise as well.

If I where to modify the device layout, are there any specific parts in the circuit design that’s more prone to cause this type of noise issues? Any specific places I should take a extra loook at?

 

[Newmarket]

Contrary to what has been stated above you do need to "Ground" metal for it to be an effective shield against rfi wrt the E-field component that is dominant here. This may well be the case by design for lightning protection. But worth checking.
Floating metal can cause havoc wrt rfi.

 

[ccaudle]

I thought that one could create a ground loop within the circuit as well

There are some effects with similar symptoms, but the cause is usually referred to by different terms (e.g. ground bounce, internal EMI, common impedance coupling).

Do you know a source I read more about this? To Learn some more about it.

There is not as much information online as I expected. Maybe I just wasn't searching for the right terms, but these look like some decent sources to get started:
EDN article on RFI (from 2008)
Analog Devices Op-Amp Applications Handbook ch 7 (see section 7-6)
chapter 2 of this thesis paper

If I where to modify the device layout, are there any specific parts in the circuit design that’s more prone to cause this type of noise issues?

Connectors. Look for info on "pin 1 problem," it has been discussed extensively in many places. The short version is that cable shields should be considered part of the chassis in terms of shielding, so cable shields should connect to the chassis, not to the circuit board.
The next step after than is filtering on signal connections, which could be as simple as a ferrite bead on signal connections instead of a straight wire connection. Sometimes clamp on ferrites which clamp around the exterior of the cable to increase common mode impedance (like often found on computer cables) can help. Those don't require modifying any internal connections so are an easy first step to try.

 

[cyrano]

These are tuneable GSM antenna, but I don't recognise the brand. Depending on where you are, it could be interesting to check who's responsible for them.

Over here, the operator is legally bound to fix that RF interference. This kind of antenna can be tuned in software, so there shouldn't be a problem to turn the wave away from your apartment.

It usually takes a bit of effort to get them going, but if you don't try now, they'll argue there wasn't a problem in the future, so don't wait a few years to complain.