Annoying tone added to Clarett noise floor as soon as second mic is connected

moongate

New member
Hi,

I recently purchased a Clarett 4Pre and connected my Rode NT1 and NTG3 to it (initially inputs 3+4 with P48).

With gain dialed to 6 for the NTG3 I immediately recognized an annoying tone added to the noise floor of that microphone.
I'm not used to this kind of background noise when using the NTG3 before.
That noise remains audible when listened to closely as part of my recordings on the Clarett.

I moved the NT1 to input 2. That way I am able to turn P48 on and off separately for both microphones.
When I disable P48 for the NT1 the tone in the noise floor of the NGT3 is gone. Recordings turn out as to be expected.
But as soon as I turn on P48 for the NT1 again to be able to use it the noise floor on the NTG3 gets this annoying tone again.
The situations remains when I unplug the USB cable of the Clarett for testing.
If I add a Rode VXLR adapter to either one of the remaining free inputs (1,3) the situation is slightly better but not as good as in single microphone mode.
In any case I did not get a 4 input interface to only use a single mic at a time ;-)

After a lot of playing around I connected both microphones using the same cables to my old Zoom H4n (also XLR inputs).
On the H4n I do not hear this issue even with the gain for the NT1 at 80% and NTG3 at 60%. Recordings turn out just fine.

I described the problem to Focusrite support a few days ago but so far no reply.
I would really like to do something this weekend and would appreciate your thoughts.

Best Moongate

PS. Added recordings of NTG3 only and the change as soon as I plug in the NT1 as the second mic. Solo it sounds as expected. With the NT1 the Clarett gives a jet plane like noise floor on the NTG3 input.
 

Attachments

  • NTG3 only with gain 6 via Clarett normalized.mp3
    311.7 KB · Views: 12
  • NTG3 with gain 6 via Clarett and NT1 attached gain 0 normalized.mp3
    445.9 KB · Views: 8
Last edited:
What would be useful is to record a short bit of audio, then a bit of silence. Normaling the noise gives no indication of its absolute level. So, for example, record a 'test one two three' kind of thing.

Howeverr, I do not the difference in quality of the noise between the two samples.

Also handy wold be just the NT1. The NT1 has (if working properly) extremely low self-noise, so it would give a god indicatin of the Claret's noise.
 
Thank you very much for confirming that you also note a change in noise quality.

It's pretty late but I tried my best to get some more samples together on my Clarett.

1st file - is NT1 only at gain 80 on input 2 (P48 only on input 1+2) - the signal is as clean as I am used to from my former Rode AI-1
2nd file - is NTG3 only at gain 80 on input 4 (P48 only on input 3+4) - You can hear the higher but still pleasant self noise of the NTG3 at this rather high gain.
3rd file - is the NTG3 at gain 80 on input 4 (P48 on input 1-4). The NT1 is connected at input 2 with gain 0 and not recorded - you can hear the change in noise tone, reminding me of sitting in an airliner.
4th file - is the NTG3 at gain 80 on input 4 (P48 on input 1-4). On input 2 I plugged in a rode VXLR adapter with nothing attached to it. Still you can hear the change.

This time I did not normalize the files as the change from 60 to 80% gain on the NTG3 already makes the noise tone change quite apparent. Normally I would not record this mic so hot.

I also tried powering all 4 inputs while only recording the NTG3 with gain 80. But when nothing else is connected it virtually makes no difference to the noise floor tone if all or just two inputs receive P48. For the change at least something needs to be plugged into one of the inputs in parallel in addition to the NTG3.

I am totally puzzled to see this behavior with the Clarett as both microphones record just fine on my Zoom H4n in comparison.
 

Attachments

  • 01 Rode NT1 only on Clarett input 2.mp3
    541.3 KB · Views: 5
  • 02 Rode NTG3 only on Clarett input 4.mp3
    541.3 KB · Views: 7
  • 03 Rode NTG3 only recording on Clarett with NT1 plugged in but gain 0.mp3
    541.3 KB · Views: 14
  • 04 Rode NTG3 only recording on Clarett with Rode VXLR adapter plugged in.mp3
    541.3 KB · Views: 10
I had a listen to the four tracks.

1 Rode NT1 only track has very low noise.
2 Rode NTG3 only track has very low noise
3 & 4 both exhibit substantially more noise.

I note that you are talking quietly. However, that was very useful in getting an idea of the comparative level.

I have no explanation for this behaviour. However, if you have effectiveyl noise free on AI-1, and with either mike plugged into Claret, but noise with both plugged in, it does suggest an issue with the Claret (and not with the microphones).

Perhaps someone with Claret experience may come along and shed some light. I use a Presonus interface which is pretty much invisible.
 
Last edited:
What happens if you put both mics only in 1&2?

I have had a weird problem with my old Safire but it seemed like it was related to the 3/4 inputs and with condenser mics. Didn’t happen all the time it seemed but I never puzzled it out because I was lucky enough to have something on 1/2 that covered the track. It was a constant tone.
 
Attached is a spectrum of just the noise (clip 04) and it is only some 40dB below the average level of the speech. That is pretty poor.

I suspect the noise is being generated by the phantom power DC-DC converter but first check that there is an identical voltage on both pins 2 and 3 of the XLR (mic end) ref pin 1.

A couple of dynamic mics, spook juice off and with a similar reference voice level would pin this down for sure?

Dave.
 

Attachments

  • scarlet noise.png
    scarlet noise.png
    104.5 KB · Views: 9
Hello All,

thank you for your input. I just tested the following.

Testing NTG3 on all inputs
I moved the NTG3 between all inputs while probing with the NT1 on any other input.
I am able to hear this change in background noise tone in any combination.
This can be on a single input pair (1+2 or 3+4) with both mics on P48 - other pair not on P48.
Or each mic on a different input pair with with input pairs on P48.
Again using the NT1 as the other input source or the passive Rode VXLR adapter both trigger the issue.

Tone change directly after activating second P48
Testing this I recognized that the Clarett starts out with the same airliner like noise floor tone when P48 is activated on the other input pair. The tone is then leveling off after a few seconds when nothing else is attached on that input pair.

I have tried to capture this in the attached file.
We start out with the NTG3 only on input 4 with gain 80 - everything fine so far.
I then press P48 on input 1+2 - you should be able to shortly hear the change in the noise floor tone after the click of the switch - the tone is then leveling off as nothing else is attached to that input pair.
I then add the passive Rode VXLR adapter to any of the remaining 3 inputs, it really doesn't matter which.
This permanently brings back the annoying noise floor tone change.

Zoom H4n comparison
I have tested the same steps with my H4n.
NTG3 only at gain 80 then adding a second input source.
On the H4n the noise floor remains absolutely the same with NTG3 only or NTG3 + NT1 attached at the same time.

Multimeter tests
The Clarett delivers 48.3V on the microphone end of the cable on pins 2 and 3 after sitting for a while.
Initially when activating P48 it starts out on 48.9V.
There is no change in voltage on the measured input when P48 is added to the other input set.

Compared to that my Zoom H4n only delivers 45V on pins 2 and 3 while battery powered.
Still this seems to be working fine, resulting in recordings without this issue the Clarett is showing here.

My thoughts
The Clarett is not the cheapest audio interface and my 4Pre also not the entry level model of it.
It is above the entry level Scarlett line.
I am currently afraid if I send this unit back for replacement I get a brand new one back which is then showing the very same issue.
 

Attachments

  • NTG only then P48 on other inputs then rode adaptor plugged in.mp3
    1 MB · Views: 2
I would definitely send the unit back as defective. This isn't normal behavior, and the Clarett has been around for a long time, so if it was inherent in the design, it would have surely been noticed. The probability of a second unit having the same issue seems very small to me.
 
It's disappointing, but I agree it should be returned. Whether you opt for the same unit or get something else is your call.

I digress now...

I was *this close* to getting that very unit because there were not a lot of available interfaces recently (everyone decided they needed an entry interface to get online, oddly). I ended up with a different brand (and lower price) Steinberg UR44C, even though the Clarett has the best specs on paper for the number of pres and price. The only thing that tugged at me a little is that I know the Clarett been around for some time, initially coming out as a Thunderbolt (2?) model, but then they introduced the USB version, and dropped the Tbolt one, instead of updating it for Tbolt 3, or even USB-C/3.1, which means it's only imperceptibly better than the now 3rd gen Scarlett 18i8 at almost 2x the price. I really think Focusrite needs to pay a little more attention to the space between their Scarlett and Red lines, because it just seems a bit behind the curve. (If it had had the LED meters of my old Saffire I might have gotten past my hesitation, but it's too much like the 18i8 right now.)
 
One final test I would do it to make up at least two XLR plugs with pins 2 and 3 shorted. Plug straight into the interface then do a 24 bit recording at minimum and maximum gain and with phantom power on and off.

I can tell you the min gain recording should give you a noise floor of at least -100dB fs and I would expect max gain to be better than -80dB fs. IMO P48 should make no difference at all.

Dave.
 
It's disappointing, but I agree it should be returned. Whether you opt for the same unit or get something else is your call.

I digress now...

I was *this close* to getting that very unit because there were not a lot of available interfaces recently (everyone decided they needed an entry interface to get online, oddly). I ended up with a different brand (and lower price) Steinberg UR44C, even though the Clarett has the best specs on paper for the number of pres and price. The only thing that tugged at me a little is that I know the Clarett been around for some time, initially coming out as a Thunderbolt (2?) model, but then they introduced the USB version, and dropped the Tbolt one, instead of updating it for Tbolt 3, or even USB-C/3.1, which means it's only imperceptibly better than the now 3rd gen Scarlett 18i8 at almost 2x the price. I really think Focusrite needs to pay a little more attention to the space between their Scarlett and Red lines, because it just seems a bit behind the curve. (If it had had the LED meters of my old Saffire I might have gotten past my hesitation, but it's too much like the 18i8 right now.)

This was also my thinking and based on that the current situation is even more disappointing.
I am currently trying to get a Scarlett 2i2 from a friend for testing. Should be able to get it early next week. Quite exited to see if the issue remains or not.
 
One final test I would do it to make up at least two XLR plugs with pins 2 and 3 shorted. Plug straight into the interface then do a 24 bit recording at minimum and maximum gain and with phantom power on and off.

I can tell you the min gain recording should give you a noise floor of at least -100dB fs and I would expect max gain to be better than -80dB fs. IMO P48 should make no difference at all.

Dave.

Hello Dave,

this sounds like an interesting test as well. It will need some preparation though.
I could easily short 2-3 on the ends of my XLR cables but by want to also rule them out first using the Scarlett from a friend.

I also tested a bit further and did a factory reset on the Clarett which also resets the control software.
I then compared Clarett and Rode AI-1 using the NTG3 only.

There is one thing I don't understand.
I think we can all agree the higher the gain needs to be the more noise will be captured on the recording.
I try to modify the gain to get to a level of -12dB with the NTG3 on each interface with a regular voice over reading voice.

The Clarett claims to have a gain of 57dB.
The Rode AI-1 only has 45dB.

But to get to a level of -12dB with the same text read in the same loudness...
- i only need to set gain to 25% on the AI-1
- and at the same time to 60-70% on the Clarett.

This already does not seem right.
 
Hello Dave,

this sounds like an interesting test as well. It will need some preparation though.
I could easily short 2-3 on the ends of my XLR cables but by want to also rule them out first using the Scarlett from a friend.

I also tested a bit further and did a factory reset on the Clarett which also resets the control software.
I then compared Clarett and Rode AI-1 using the NTG3 only.

There is one thing I don't understand.
I think we can all agree the higher the gain needs to be the more noise will be captured on the recording.
I try to modify the gain to get to a level of -12dB with the NTG3 on each interface with a regular voice over reading voice.

The Clarett claims to have a gain of 57dB.
The Rode AI-1 only has 45dB.

But to get to a level of -12dB with the same text read in the same loudness...
- i only need to set gain to 25% on the AI-1
- and at the same time to 60-70% on the Clarett.

This already does not seem right.

You have stumbled upon a problem I have banged on about from time to time...Gain specifications for AIs.

We don't really need a "gain 55dB" figure since that tells us jack. Much better would be "X" dBu/mV for -20dB fs. It is trivially easy to find the sensitivity of most microphones in fact all dynamics are withing a few dB of 1.0 mV or around -60dBV. The SM57/58s are about 1.5mV, most modern ones around 2mV.

A bald 'gain' figure only makes sense if we know the system output. A mixer say with 60dB mic gain would imply an input of 1mV would produce +4dBu at main outs with faders set at '0' (not 'nowt!')

'Domestic' 'Prosumer' audio specs are a mess and some pro gear is not much better. I like to mangle the quote.. "Lies, damned lies and SPECIFICATIONS"!!

I suspect you have just got a bad sample. I have an old Mk1 8i6 and that is pretty quiet. Comparable to my NI KA6. I also had a Behringer UMC204HD for a few weeks and that was at least as good as both of them and had maybe a tad more gain? 'It'Appen.


Dave.
 
Last edited:
OK, interesting.

After spending the whole weekend with the issues on the Clarett I was ready to go nuts.
I even opened my XLR cables and noticed that the Rode cable has 2 ground leads attached - on the female plug ground is attached to pin 1 + a separate ground lead to the button spring which connects to the outer metal casing of the plug. I don't see this second ground soldered on my second cable.

Anyway I have access to the Scarlett 2i2 now and it worked instantly without issues with both of my cables and microphones.

I have prepared the same setup for a test of both interfaces.

NT1 on P48 attached with gain 0
NTG3 on P48 attached with gain 95

I then ran a recording test with this setup on either one of the interfaces.

I hope you can clearly hear the difference in recording quality between the two attached samples.
I also note a significant lower level of the recording on the Clarett - and yes I speak with the same loudness on both recordings.
This also explains why I came across speaking so "silent" on all former Clarett samples.

It seems it is now confirmed that there is something wrong with this Clarett unit.
Still it was worth the effort as I do not like returning things lightly without making sure the problem is not sitting in front of the device ;-)
There is much to be configured on the Clarett that can screw up things I suppose. But in this case the issue remains even on factory defaults.

So I guess I will prepare the Clarett for a return now and have to make my mind up if I could still fall in love with a non-defective unit after this initial experience...

Best Moongate
 

Attachments

  • NTG3 recorded with 95% gain on Clarett NT1 attached gain 0.mp3
    590.8 KB · Views: 13
  • NTG3 recorded with 95% gain on Scarlett 2i2 NT1 attached gain 0.mp3
    674 KB · Views: 62
Back
Top