Weird Noise in Only Some Situations!

[Hexo]

Hello every body,
My friend has this high pitched noise on inputs of 3 different interfaces with 3 different mac/win-laptop/iMac computers(all cross-tested with even different cables). Some sound cards are bus powered and some work with their respective power adapter...
https://drive.google.com/file/d/1PMycvW ... sp=sharing
I tested 2 of his cards in my studio and there were no noise at all. We checked ground electricity and apparently that's not the case. We also checked his laptop running on batteries with his bus powered...nope!
Although the noise is very quiet(-95dB) it is increased with preamp gain boost.

What do you think I'm missing here?

Edited to show link to image - JP

[resistorman]

If the sound only occurs in one location, it's possible that there is some external source of interference. It can be motors, lighting, etc.

[Hexo]

If the sound only occurs in one location, it's possible that there is some external source of interference. It can be motors, lighting, etc.

Thanks, but unfortunately he brought his laptop to my studio and we re-tested everything, but the noise is still there!

[Tim Gillett]

...Although the noise is very quiet(-95dB) it is increased with preamp gain boost.

What do you think I'm missing here?

Maybe that -95db below FS is more than good enough for many situations. Preamps make noise. Noise doesn't have to be totally eliminated. It has to be seen in context and managed depending on the specifics of the recording situation.

[Wonks]

What exactly is the testing arrangement?

Are these mic inputs or line inputs?

Is anything connected to the inputs?

What's the noise like with a cable connected and the signal connections shorted to ground?

[Hexo]

What exactly is the testing arrangement?

Are these mic inputs or line inputs?

Is anything connected to the inputs?

What's the noise like with a cable connected and the signal connections shorted to ground?

Nothing is connected to inputs. Can you please explain your last question more?

[Hexo]

...Although the noise is very quiet(-95dB) it is increased with preamp gain boost.

What do you think I'm missing here?

Maybe that -95db below FS is more than good enough for many situations. Preamps make noise. Noise doesn't have to be totally eliminated. It has to be seen in context and managed depending on the specifics of the recording situation.

I totally agree. But that -95dB is at gain=0(or the minimum possible gain on the interface) and our signal to noise ratio is still thoroughly compromised.

[Wonks]

Inputs can pick up significant noise if they aren’t connected. You don’t record with nothing connected to an input.

Your lowest noise on a mic input will be measured with a 300 ohm resistor connected between pins 2 and 3 of an XLR cable. But shorting those pins to ground (pin 1) will give a very similar figure (within 3dB)

Likewise, with TRS cables into a line input short the T and R and S connections together.

You should then get a noise figure very close to the manufacturer’s figures.

If you don’t then there is something else very wrong. But I think you should find that you get very acceptable figures.

[resistorman]

Nothing is connected to inputs.

There is your problem.

[Wonks]

Please note I should have said ‘most representative’ noise, not ‘lowest’, for the 300 ohm resistor test.

Without a load attached, the signal wires’ voltages can float, and any noise picked up by the short lengths of wire and PCB track between the physical input connector and the preamp can cause the sort of noise signal you showed in your screenshot.

[Hexo]

Inputs can pick up significant noise if they aren’t connected. You don’t record with nothing connected to an input.

Your lowest noise on a mic input will be measured with a 300 ohm resistor connected between pins 2 and 3 of an XLR cable. But shorting those pins to ground (pin 1) will give a very similar figure (within 3dB)

Likewise, with TRS cables into a line input short the T and R and S connections together.

You should then get a noise figure very close to the manufacturer’s figures.

If you don’t then there is something else very wrong. But I think you should find that you get very acceptable figures.

Thanks for your comprehensive explanation. I'll do as suggested once I get a multimeter. But the weird thing is the sound cards don't make that noise when they're connected to my iMac while nothing is connected to their inputs. So it looks like they behave differently in almost the same circumstances!

[Hugh Robjohns]

... the weird thing is the sound cards don't make that noise when they're connected to my iMac while nothing is connected to their inputs. So it looks like they behave differently in almost the same circumstances!

Not weird at all.

With nothing connected to the inputs, the connectors act as aerials, picking up whatever RF might be in the local environment. Subtle differences in the circuit board layout and grounding will give different channels different sensitivities to different RF interference.

It would appear that there is more RF at your friend's place than yours...

But the bottom line is that inputs aren't intended to be left disconnected. If you want to measure preamp noise performance you should use a 'dummy load' in a screened case to provide a suitable source impedance for the preamp.

For a mic input, solder a 1% 150 Ohm metal oxide resistor between pins 2/3 of an XLR plug. Some manufacturers cheat their specs by shorting pins 2/3 directly without the resistor, which improves the EIN figure by a few decibels.

[Hexo]

... the weird thing is the sound cards don't make that noise when they're connected to my iMac while nothing is connected to their inputs. So it looks like they behave differently in almost the same circumstances!

Not weird at all.

With nothing connected to the inputs, the connectors act as aerials, picking up whatever RF might be in the local environment. Subtle differences in the circuit board layout and grounding will give different channels different sensitivities to different RF interference.

It would appear that there is more RF at your friend's place than yours...

But the bottom line is that inputs aren't intended to be left disconnected. If you want to measure preamp noise performance you should use a 'dummy load' in a screened case to provide a suitable source impedance for the preamp.

For a mic input, solder a 1% 150 Ohm metal oxide resistor between pins 2/3 of an XLR plug. Some manufacturers cheat their specs by shorting pins 2/3 directly without the resistor, which improves the EIN figure by a few decibels.

I totally agree and I will test them as suggested as soon as I can. But please consider that we've already tested them with different microphones connected and not only the noise is not gone, it's increased relative to gain boost. One more thing: he brought his laptop to my room and we re-ran all tests. The noise is there in his laptop, and absent in my iMac in all circumstances!

[resistorman]

I totally agree and I will test them as suggested as soon as I can. But please consider that we've already tested them with different microphones connected and not only the noise is not gone, it's increased relative to gain boost. One more thing: he brought his laptop to my room and we re-ran all tests. The noise is there in his laptop, and absent in my iMac in all circumstances!

Well, this is different than what you've been saying

[Drew Stephenson]

Does his laptop have a mechanical disk or an SSD?
If the former it might be picking up some physical vibration.

[Wonks]

Microphones pick up noise from the room. It's what they do. You're in it and your computer's in it and there's probably wind, traffic and street noise at low level. Unless you are in a soundproof and acoustically well-treated room and your computer(s) are in a separate room, there will be a lot of ambient noise you aren't aware of being picked up on a microphone that's way above the electronic noise level.
In real life your brain tunes a lot of that noise out, but when you hear it on a recording, it doesn't. It's weird the way it happens, but it's a fact.

Average room noise is between 30 to 50dB(A), and 30db is a very quiet room. A loud singer (not shouting) is around 80dB(A), maybe 90dB (A) when close-up. So you may only have between 30dB and 60dB between your vocal and the background noise in a standard untreated and non-soundproof room.

Which is why you need to judge the equipment noise purely based on using terminated cables (as has been said, shorted to ground cables will be slightly quieter).

Test like that to work out the interface preamp noise. You can't do it any other way. If that still measures high, then you have a problem with the set-up.

Otherwise any noise above that is noise is almost certainly going to be noise picked up from the room. Capacitor mics do make a little bit of electronic noise, but that is still way down compared to any ambient noise. Dynamic mics shouldn't have any self-noise.

[Hugh Robjohns]

Please consider that we've already tested them with different microphones connected and not only the noise is not gone, it's increased relative to gain boost.

I don't think you mentioned that. Nevertheless, the situation remains the same. RFI is getting into the interface somehow, and there is more RF interference at your friend's house than yours.

The noise test with the dummy plug might give some further clues as to how the interference is getting in. But it's not unknown to get in through some microphones and/or cables. And interface grounding can also play a significant role.

One more thing: he brought his laptop to my room and we re-ran all tests. The noise is there in his laptop, and absent in my iMac in all circumstances!

More helpful news.

So, it might be that the laptop or its power supply are generating the interference. Or it may be that the laptop is not providing a good ground for the interface. Or it might be that the laptop is injecting noise via the interface connection. Sometimes different USB ports can have more/less noise than others.

This kind of problem can be tricky to resolve but careful, logical testing usually gets there.

[Hugh Robjohns]

Dynamic mics shouldn't have any self-noise.

Of course they do. They generate thermal noise due to the resistive element of their output impedance. In most modern mics this is nominally 150 Ohms and results in a self-noise level around -131dBu (at 20C and measured between 20Hz and 20kHz).

A capacitor mic has a similar output impedance so generates a similar thermal noise, but the active circuitry in the impedance converter inevitably adds a few dB more noise.