When I connect the Behringer ECM8000 to the pre-amp and apply phantom power the spectrum noise floor rises from -110 dB to -60 dB at 20 Hz. And then falls off at 6 dB/oct. I don't understand that. I have very little experience using microphones so I don't know it this is normal behavior for an ECM8000. One of the things I don't quite understand is that the noise level decreases when I knock on the surface the microphone is standing on. Or some minutes after I switch on the phantom power the noise also decreases.
The absolute power levels are not relevant I think. I am using the sound card of a PC for measurement. And a spectrum analyzer application.
The schematic of the pre-amp is attached.
This spectrum when the pre-amp is connected to the audio input and no phantom power applied to the microphone. Noise floor at approximately -110 dB. The 50Hz hum at -80 dB I can understand.
s the
When I switch on the phantom power the spectrum rises to -60 dB
The pre-amp output on the scope. Because the trace is very slow I had to use a long exposure time. And the noise of the camera shutter is also recorded.
The noise decreases when I knock on the table surface. After some time (minutes) the larger noise may return.
Pre-amp connected to sound card input but phantom power removed.
Microphone mounting.
I have also tried to hang the mic at its cable, or at a rubber elastic band. No difference.
The measurement function of the microphone seems normal. When I apply an arbitrary tone to my speaker it is displayed correct in the spectrum. Up to about -10 dB no noticable harmonics.
I don't understand what is happening here.
The absolute power levels are not relevant I think. I am using the sound card of a PC for measurement. And a spectrum analyzer application.
The schematic of the pre-amp is attached.
This spectrum when the pre-amp is connected to the audio input and no phantom power applied to the microphone. Noise floor at approximately -110 dB. The 50Hz hum at -80 dB I can understand.
When I switch on the phantom power the spectrum rises to -60 dB
The pre-amp output on the scope. Because the trace is very slow I had to use a long exposure time. And the noise of the camera shutter is also recorded.
The noise decreases when I knock on the table surface. After some time (minutes) the larger noise may return.
Pre-amp connected to sound card input but phantom power removed.
Microphone mounting.
I have also tried to hang the mic at its cable, or at a rubber elastic band. No difference.
The measurement function of the microphone seems normal. When I apply an arbitrary tone to my speaker it is displayed correct in the spectrum. Up to about -10 dB no noticable harmonics.
I don't understand what is happening here.
When I first started to use a microphone for testing, I was surprised at how much low frequency wind noise was present in a quiet room.
Just moving your arm will move the air, which is picked up by the microphone.
Then there are draughts around the doors and windows.
Inaudible, but measurable.
Just moving your arm will move the air, which is picked up by the microphone.
Then there are draughts around the doors and windows.
Inaudible, but measurable.
Interesting. Never thought about that. 50 dB above noise level in the ball park for you as well?
These are the days without any wind. I do have an air conditioner running. I switched it off, no observable difference. But then again, for what I observe it seems the air conditioner noise seems to be higher in frequency.
I wrote something else and then decided I ought to do a quick check before mouthing off. 🤣 I had a bad experience with an ECM8000 that turned into a very effective low frequency noise generator after a couple of years of ownership. I now have UMIK-1, UMIK-2, and Isemcon EMX-7150s all of which have the benefit of being calibrated.
This is a UMIK-2 in my listening room. There is definitely some ambient LF noise present, but I would guess a significant percentage has to be microphone (1/f) self noise.. These measurements were taken a few minutes apart (blue is later) and are calibrated dBSpl. I generally do acoustical measurements in the 80dBSpl+ range.
I should repeat the measurement at some point with my Isemcon EMX-7150 for a sanity check.
This is a UMIK-2 in my listening room. There is definitely some ambient LF noise present, but I would guess a significant percentage has to be microphone (1/f) self noise.. These measurements were taken a few minutes apart (blue is later) and are calibrated dBSpl. I generally do acoustical measurements in the 80dBSpl+ range.
I should repeat the measurement at some point with my Isemcon EMX-7150 for a sanity check.
Last edited:
@kevinkr Well what I see in your curve is that the noise at 20 Hz is also about 30 dB higher than at 1 kHz. This it the first time in my life I actually look at the spectrum of a (hopefully) good measurement microphone.
Looking at your curve it might be normal what I observe with my ECM8000. Really I had no idea I could expect this.
Looking at your curve it might be normal what I observe with my ECM8000. Really I had no idea I could expect this.
I've seen noticeable if not huge differences in self noise of my ECM8000's, I've also seen excessive noise on a malfunctioning M2010's preamp section, and your ECM's response looks kind of like the M2010's bad pre-amp. Maybe a new ECM is in order? They are going for $20 new at the moment.
Here's two of my ECM's, each level calibrated before wrapping in a plastic bag and buried in a big bucket of sand in a quiet corner of my basement.
In contrast, here's an NTi Audio M2010 with a bad pre-amp section and with a good pre-amp section.
Here's two of my ECM's, each level calibrated before wrapping in a plastic bag and buried in a big bucket of sand in a quiet corner of my basement.
In contrast, here's an NTi Audio M2010 with a bad pre-amp section and with a good pre-amp section.
Ok, I admit it. I hadn't considered 1/f noise.
My observations arose when building a servo-sub; did you know that a loudspeaker can act like a microphone?
I used an oscilloscope to view the feedback signal, and just moving my arm could be observed on the oscilloscope trace.
This might not be relevant to issues with your microphone measurements, but I became acutely aware of low frequency wind noise during the servo-sub's development.
A recording of Musical Fidelity's Anthony Michaelson, playing the clarinet in a church, sent my subs crazy.
My observations arose when building a servo-sub; did you know that a loudspeaker can act like a microphone?
I used an oscilloscope to view the feedback signal, and just moving my arm could be observed on the oscilloscope trace.
This might not be relevant to issues with your microphone measurements, but I became acutely aware of low frequency wind noise during the servo-sub's development.
A recording of Musical Fidelity's Anthony Michaelson, playing the clarinet in a church, sent my subs crazy.
I had the same experience. Another mic cured the problem.
A lot of venues have massive amounts of subsonic noise, air handlers in HVAC systems are notorious as are trams and underground (subway) in city recording venues. Air blowing across the mic. Wind screens help to an extent. Air currents in large spaces drive the need for HPF filters at around 15 - 20Hz..
I have subs and have observed some interesting behavior below 20Hz.
Even a perfectly working condenser microphone in perfect silence, without any kind of wind, will usually have a lot of low-frequency noise rolling off at -20 dB/decade. That's just the thermal noise of the resistors that bias the input of the built-in amplifier and the shot noise of the gate leakage of the input FET filtered by the capacitive impedance of the microphone capsule. It should not change when you knock on something, though.
Some time ago I repaired a Rode Broadcaster condenser microphone that had fallen on a hard surface and produced scratching sounds since. It turned out to be due to a ceramic multilayer SMD capacitor that had become a sort of short circuit, but with poor contact.
Some time ago I repaired a Rode Broadcaster condenser microphone that had fallen on a hard surface and produced scratching sounds since. It turned out to be due to a ceramic multilayer SMD capacitor that had become a sort of short circuit, but with poor contact.