Strange noise on Focusrite Clarett+ mic inputs

[asphaltradler]

After months of reading in this very helpful forum, I now need some help for a problem that's driving me mad.

Depending on the mics connected to other inputs, my Behringer B5 microphones, and to a lesser extent also the sE Electronics se8, produce a lot of deep-sounding additional noise on a Clarett+ 4pre I recently purchased.

The same mics on several other interfaces and preamps (Focusrite Saffire, Behringer UMC, M-Audio Audio Buddy, Zoom H6) definitely do not produce this characteristic noise. It ONLY really disappears on the Clarett when 4 mics are connected together and one of them is a Rode NT 1-A large diaphragm mic! This means that the mics on input 3+4 strongly influence inputs 1+2 and vice versa. There is definitely something wrong with the phantom power!

This behaviour really drove me crazy. I wrote with Focusrite support and gave them several audio files showing the noise, with different mics plugged in at different inputs. They finally told me to contact my dealer. So I sent it to them with even more explanations in the hope they would reproduce the issue first, then test it on another Clarett. But they just went ahead and sent me a new one, and hey - it took me 2 minutes to recognize it has the same problem!

Although I had described the problem in detail, apparently no counter-test was carried out.

So now I decided to make videos of that. In all 3 videos I only took the sound from inputs 1+2, i.e. you can't hear the mics on 3+4 (and also not the mobile phone camera certainly). Nevertheless, connecting there or switching the phantom power on and off has an effect on the noise on 1+2.

I tried different cables, switched inputs etc. but nothing helps. You have to listen with headphones and then can hear the noise very distinctively; it's not the usual high hiss of bad microphones but a deeper, water-flow-like 'swoosh' that is immediately recognizable.

Please: can anyone check that out who owns a Clarett/Clarett+? Best with Behringer B5 but I'm certain it must be happening with other mics, too. The sE Electronics se8 shows the symptom, too, but it is less pronounced and I noticed it only when it disappered by plugging in the Rode NT1 in 3 or 4.

The videos are in german but I added english subtitles. The effect is heard best in the 3rd one:

and is explained in detail in the first:

Regards, Cosmas

 

[asphaltradler]

I'm surprised no one has commented on my findings. Has no one taken an eye & ear to my videos?

What could be happening with phantom power there? Focusrite technicians seem to know nothing about that. They offered another exchange but after 2 already I'm really quite sure now that every unit has that problem but no one really seems to listen to real recordings made with them...

That shows also that normal measurements with 150 Ohm resistance aren't sufficient because it happens only with real condenser microphones, and with phantom power on.

Instrument input (without 48v power naturally) is totally quiet by the way. So I expect will be dynamic mics without power, but I don't use any.

Really, are there no Focusrite Clarett users out there?

 

[Blumlein 88]

I will listen to it later when I'm home. Posting from phone now. Your videos are showing up here.

 

[amirm]

That is a very odd problem. The sound is that of electronics or a mic picking up fan noise. I trust the latter is not the case?

 

[asphaltradler]

No really not!
That noise is there in every surrounding, with every cable and so on. I noticed it first when I wanted to make a quiet recording in standalone mode, in my headphone. And then spent dozens of hours to exclude every possible aspect.

In the video the room is totally quiet except the wall clock behind me - you should easily hear that clock ticking.

The notebook I connected for capturing the sound is fanless and also totally silent.
Also, when that sound was from surrounding, how should it stop when I switch phantom power with the NT 1 connected?

It's there always I assure you. And depends on nothing else - it does not alter the slightest when moving microphones, cables or device itself. Tried also different power supplies. And this is the 2nd Clarett, sent to me from the store without any preselection...

 

[asphaltradler]

It's not so special after all - the dark swooshing noise is there with 4 of my 5 condenser mics, so I suspect everyone else must have that too except perhaps a few lucky ones... The shown types, Behringer B5 (which have switchable heads, cardioid and polar - and both have it) and sE 8 are quite common I think. The Behringer is the most sold at Thomann...

 

[Blumlein 88]

It should not be that way. I don't know what to tell you. There have been a few reports of excess noise with Claretts on gearspace. I wonder if it was related to this without anyone noticing it was connected to phantom power.

Might help to use REW or Multitone and post an FFT of the noise in both conditions. Otherwise sounds like a design flaw of some sort if two of them are the same.

You also might try connecting resistors you have put across an XLR. Say 100 ohms or 15o ohm and some higher value like 470 ohms to see in REW or Multitone how the noise varies with those loads.

Finally just a thought. If I remember this uses USB C doesn't it? If so are you connecting to a USB 2 port on your computer with an adapter? Maybe a USB2 Port struggles to supply enough current and this causes issues. Just a guess on that.

 

[asphaltradler]

Here are the sound files and description what is happening in them complete with Screenshots of the DAW but unfortunately only in German. Deepl will be of use for non German speaking persons here - I no longer have the strength to translate this now, so much effort has already been put into this problem:-(

Microsoft OneDrive

1drv.ms

 

[asphaltradler]

If so are you connecting to a USB 2 port on your computer with an adapter? Maybe a USB2 Port struggles to supply enough current and this causes issues. Just a guess on that.

It's there also standalone without any connections except mics and headphone :-( Clarett 4 has its own power supply (as I wrote, i tried several different), is not bus powered and needs no computer. I just connected the fanless small laptop for the videos to capture the sound from the interface, but the easiest eay to detect the problem is in standalone mode when configured before like on this picture, so that headphone output listens only to the first 2 inputs.

 

[asphaltradler]

Might help to use REW or Multitone and post an FFT of the noise in both conditions.

I thought that too before, but the problem is, the swooshing is easy to hear but makes no great impact on FFT. Because naturally even a quiet room is quite (i always mix up those two) loud - the noise floor when recording is -40db or so and buries the swooshing totally. In some cases ears are better than optical measures it seems.

It's really easy to spot acoustically as in the short video with switching phantom power. Far from suble. Or the 'Spur 3' mp3 of my sound files, link above:
At 0:22 I plug in the NT1 instead of the se8 and the dark dirt is gone.

 

[AnalogSteph]

Last time I encountered a case of excessive noise (on a Swissonic UA-2x2), the most likely explanation was a bad (leaky) coupling capacitor in one leg of the #1 channel, which of course happened to be the one I'd meant to be using (Murphy). Figured this sort of parts quality issue could happen on a 65€ 2x2 interface. This sort of thing basically breaks the symmetry in the balanced input (and hence CMRR) and introduces a noise source on one leg.

In this case the issue may be further upstream, in the actual phantom power supply itself. Putting more load on it seems to quieten it down a lot. Is the boost converter getting very noisy with low load? Are they possibly following it up with a linear series regulator that struggles to keep in regulation at very low currents and lets more noise through? It has been found that there are differences between 78 series regs between different manufacturers in this regard, for example. The same could happen if it's just a follower in a capacitance multiplier. (There's a reason why shunt regulators have remained somewhat popular for supplying miserly but sensitive loads.) Or perhaps it's just an RC with the capacitor just barely running out of spec? (Then I would expect a lot more variability between two interfaces though.)

How good is your soldering? You could make a dummy microphone that does absolutely nothing but draw current with 2 resistors soldered to an XLR connector, between (3 and 1) + (2 and 1). Maybe 3.3 kOhms which would draw about 9.5 mA, close to the maximum allowed. Should be about two typical microphones' worth of current or even a bit more. (You may or may not want to connect 2 and 3 together, the difference in noise between both versions on that input could be interesting.)

I have to wonder why the issue is affecting these SDCs particularly strongly. It's not like they have high output impedance (a nominal 75 ohms for the Behringers; I spy 4 transistors in there so perhaps a Schoeps circuit clone of sorts?). I can't imagine that a major manufacturer like Focusrite wouldn't have pretty closely matched P48V resistors either. Mind you, a Behringer C-2 in particular is not the loudest condenser mic ever at -41 dBV/Pa... my own noise problem was bothering me a lot on the Marantz MPM-1000 (-38 dBV/Pa) while a t.bone SC400 (-32.5 dBV/Pa) was virtually fine.

 

[DVDdoug]

Phantom power specs seem to be a bit fuzzy. Some mics need more current than others and the more mics you're powering the more current you will "pull". It probably can't provide enough current. That could make the power supply noisy or the voltage could drop to where the head amp in the microphones isn't working properly.

 

[Blumlein 88]

I tried this with a couple of usb powered interfaces I have on hand and some CAD M179 microphones. The M179 is known to draw more current than most mics for phantom. Specs listed it at 8 milliamps (4 or 5 ma is more common). Did not make one bit of difference in the noise whether on or off. There was a momentary change when phantom was turned on or off, but the noise level settled back down to the same value within the nearest one tenth db. Made no difference even while turning on or off in other inputs.

This doesn't mean anything for the Clarett which still has a problem.

 

[restorer-john]

How good is your soldering? You could make a dummy microphone that does absolutely nothing but draw current with 2 resistors soldered to an XLR connector, between (3 and 1) + (2 and 1). Maybe 3.3 kOhms which would draw about 9.5 mA, close to the maximum allowed. Should be about two typical microphones' worth of current or even a bit more. (You may or may not want to connect 2 and 3 together, the difference in noise between both versions on that input could be interesting.)

I think this is a good idea and it will rule out any acoustic contribution from the 4 microphones in close proximity being mixed and monitored.

It's hard to work out what is exactly happening. I presume the gain pot increases/decreases the noise when present? Don't you have the gain pots at zero when a connected condenser is not phantom powered?

 

[asphaltradler]

I think this is a good idea and it will rule out any acoustic contribution from the 4 microphones in close proximity being mixed and monitored.

It's hard to work out what is exactly happening. I presume the gain pot increases/decreases the noise when present? Don't you have the gain pots at zero when a connected condenser is not phantom powered?

Acoustics don't matter - only the first two mics are selected in the headphone mix or in the DAW. And you're forgetting that the dark swoosh is already there with only 1 and/or 2 connected. I hear it always. And only connected the other two because it then disappears! Or becomes weaker.

The 'problem noise' becomes stronger depending on the gain at 1 and 2 (but of course also the normal micro and background noise). Gain on 3+4 plays no role in this, just that there are mics connected and phantom power for them is active (1+2 and 3+4 share each a switch). If 3+4 don't have phantom their gain is irrelevant because they act as if not connected at all, no difference. And the gain on 3/4 doesn't affect 1/2 in any way.

Übersetzt mit DeepL

 

[asphaltradler]

It probably can't provide enough current.

Well why does it then get better with the large mic which supposedly draws more current?
The 2 B5s don't draw much I suspect. It's even already there with only 1 connected. But almost disappears with 4...

 

[asphaltradler]

tried this with a couple of usb powered interfaces I have on hand and some CAD M179 microphones.

I also don't have it with the same mics when using a Zoom H6, Focusrite Saffire 24 Pro Firewire, Behringer UMC 202hd, MAudio Audio Buddy Mic Preamp...

 

[asphaltradler]

How good is your soldering? You could make a dummy microphone that does absolutely nothing but draw current with 2 resistors soldered to an XLR connector, between (3 and 1) + (2 and 1). Maybe 3.3 kOhms which would draw about 9.5 mA, close to the maximum allowed

Well I'd rather not... I'm more the software type, electronics is a bit out of my reach. I built a pair of diy Hypex nCore mono core amps though but it was some struggle.

Besides, I think Focusrite should get that straight. My interest here is to verify first if the phenomenon can be reproduced by anyone and, if so, get the Clarett fixed by Focusrite themselves (instead of sending me further damaged ones).

For my future usage I've ordered a Motu in the meantime. If I was getting mad this would exhibit the same fault, but I suspect this because not even the cheap UMC202 has it. Perhaps I just should have stayed with that?

 

[asphaltradler]

Mind you, a Behringer C-2

They are B5, not C-2. Or what do you mean?
But the SE8 has it, too only a bit weaker. Didn't wanted to make it even more complicated by adding videos with se8 on 1/2 and all other combinations

 

[AnalogSteph]

hey are B5, not C-2. Or what do you mean?

Sorry, missed that. Those would be at -38 dBV/Pa (cardioid) / -40 dBV/Pa (omni) then. Pretty average for SDCs, I guess. sE8s are rated -32 dBV/Pa, which would explain why they are less affected. I find it rather concerning that the problem would still be audible with mics this comparatively hot.