DC-DC CMOS (HEX INVERTER) REVERSE BIAS?

[rogs]

Ah ok, that makes sense.

In your vocal sample, how high did you set your mic preamp?

The preamp was set with 20dB of gain.

As Khron has pointed out, I should have compared the mic preamp with the input of the Tascam loaded.
With the input pad setting the gain to 0dB, the mic circuitry adds about 3dB of extra noise. With the pad removed, the gain of the preamp is increased by 20dB. And so of course is the noise from the mic preamp!
The Tascam preamp itself adds very little to the noise level.

Without access to a calibrated 94dBSPL reference signal, it is probably best to use a microphone with a known noise level to compare. As already mentioned, I use my Rode NT1 as a reference. It has a published noise level of 4.5dB(A). I would estimate that an OPIC circuit has a noise figre about 3 or 4 dB worse than that. The differential audio output of the Alice OPA circuit will add another 3 dB or so to that figure.
So you should expect a noise figure of around 10 or 12dB (A) from your Alice preamp. Still pretty good.

You might like to glance though these papers from Neumann and Sound Devices, which give some insights on how noise is measured in mics and preamp?
.
https://www.neumann.com/en-en/homestudio-academy/what-is-self-noise-or-equivalent-noise-level
https://www.sounddevices.com/microphone-preamp-noise
It's not always an easy subject to understand .... I'm still learning all the time!

 

[AUDIO FREQ]

Here it is rog. There is a loud pop when I plug the mic in, so please becareful.

 

[rogs]

That's seems about right for the circuit you are using...

I have realised that my original 16 bit noise file does not accurately record the low level noise, and so I have re-recorded it as a 24 bit file.
I've also made the recording full range (no input 80Hz HPF applied).

That more accurate file adds about 10dB of noise when the mic is plugged in.

Using my Tascam, the noise floor is around -95dB with 0dB of gain applied, and that rises to about -85dB when the mic is plugged in.

A copy of my new recording here, with 10 seconds at each level: OPIC.noise.24 bit.wav
(plug in 'glitch' removed)

Your Apollo recording has a more impressive -104dB noise floor, with the mic signal then adding around 20dB of noise.

With the differential configuration of the Alice OPA circuit, I believe that is about what you should expect....

• The second differential opamp output will take the signal, and invert it.... That action will increase the signal - and the noise - level by 6dB.

• The noise generated by the 1st opamp - plus the noise added by the second opamp, and the 2 x 2k2 resistors introduced into the signal path - will add a further 3 or 4dB of noise to the output signal. That will account for the extra c.10dB of noise present in your recording.

Remember though that the extra 6dB of signal presented at the differential ouput will enable the preamp gain to be reduced by 6dB, to get the same signal level output that a 'single sided' audio configuration supplies.

So the actual noise floor will only appear to be about 3 dB worse -- which is about right......
Doesn't change the absolute noise figure of course, but allows for 6dB less gain being required.

There are a number of well known manufacturers who use single sided audio - with a passive impedance balanced output on pin3.
Neumann, Rode and dbx are those that immediately spring to mind. I'm sure there are many others.
Applying any extra gain required from an external high quality preamp, rather than from an internal inverting audio output will probably result in a lower perceived overall noise floor....

It might be worth applying an A weighted EQ curve to the recorded audio file? ...... That is what is normally used to quote noise figures, and will probably appear to 'improve' the noise floor by some 5 or 6 dB ......

My suggestion that there was something wrong in post #50 was incorrect.... I had made 2 errors with my first recording...
• I had included an 80Hz HPF, which removed a lot of the LF noise, and gave an artificially low noise figure, when compared to a 'flat' response
• The accuracy of my 16 bit recording at those low levels did not give a true represention of the actual noise present.

I can only apologise for any confusion I've caused in making those errors....

 

[AUDIO FREQ]

That's seems about right for the circuit you are using...

I have realised that my original 16 bit noise file does not accurately record the low level noise, and so I have re-recorded it as a 24 bit file.
I've also made the recording full range (no input 80Hz HPF applied).

That more accurate file adds about 10dB of noise when the mic is plugged in.

Using my Tascam, the noise floor is around -95dB with 0dB of gain applied, and that rises to about -85dB when the mic is plugged in.

A copy of my new recording here, with 10 seconds at each level: OPIC.noise.24 bit.wav
(plug in 'glitch' removed)

Your Apollo recording has a more impressive -104dB noise floor, with the mic signal then adding around 20dB of noise.

With the differential configuration of the Alice OPA circuit, I believe that is about what you should expect....

• The second differential opamp output will take the signal, and invert it.... That action will increase the signal - and the noise - level by 6dB.

• The noise generated by the 1st opamp - plus the noise added by the second opamp, and the 2 x 2k2 resistors introduced into the signal path - will add a further 3 or 4dB of noise to the output signal. That will account for the extra c.10dB of noise present in your recording.

Remember though that the extra 6dB of signal presented at the differential ouput will enable the preamp gain to be reduced by 6dB, to get the same signal level output that a 'single sided' audio configuration supplies.

So the actual noise floor will only appear to be about 3 dB worse -- which is about right......
Doesn't change the absolute noise figure of course, but allows for 6dB less gain being required.

There are a number of well known manufacturers who use single sided audio - with a passive impedance balanced output on pin3.
Neumann, Rode and dbx are those that immediately spring to mind. I'm sure there are many others.
Applying any extra gain required from an external high quality preamp, rather than from an internal inverting audio output will probably result in a lower perceived overall noise floor....

It might be worth applying an A weighted EQ curve to the recorded audio file? ...... That is what is normally used to quote noise figures, and will probably appear to 'improve' the noise floor by some 5 or 6 dB ......

My suggestion that there was something wrong in post #50 was incorrect.... I had made 2 errors with my first recording...
• I had included an 80Hz HPF, which removed a lot of the LF noise, and gave an artificially low noise figure, when compared to a 'flat' response
• The accuracy of my 16 bit recording at those low levels did not give a true represention of the actual noise present.

I can only apologise for any confusion I've caused in making those errors....

No worries Rog, im a noob and all of this is a learning experience for me. So thank you for taking the time to explain and educate me.

Question, regarding my test. Was I supposed to engage 48v? Because I did. I recorded the mic preamp with 48v engaged, and then plugged my mic in with the 48v already engaged. Was I supposed to do that? I was under the impression that the mic wouldn't be on without it.

 

[rogs]

No worries Rog, im a noob and all of this is a learning experience for me. So thank you for taking the time to explain and educate me.

Question, regarding my test. Was I supposed to engage 48v? Because I did. I recorded the mic preamp with 48v engaged, and then plugged my mic in with the 48v already engaged. Was I supposed to do that? I was under the impression that the mic wouldn't be on without it.

I left the 48v phantom power engaged, so that any noise on that supply would apply to both test input conditions. There shouldn't be any nosie from that supply of course, but some phantom power feeds are better than others! .... I would think the Apollo should be fine.
Leaving a mic which requires phantom power connected to the input when the power is not switched on is an interesting one..... As you say, the mic won't work, but I'm not sure what kind of load the active electronics look like with no power applied? ...
It shouldn't do any harm, but I'm not sure it will look like an ideal load in that state?...... Someone else maybe able to advise on that....

For this test I simply left the phantom power on with both the mic connected and disconnected.

In my latest audio file, I also fitted a 150R load resistor across the input with the mic disconnected. As I explained earlier, with the Tascam input pad switched on there was no difference whether it was there or not. but in the interests of 'correctness' I fitted it this time !
(It certainly makes a difference when the pad is not applied! )

What I did check was how much extra noise the Tascam preamp added to the signal path, when an extra 20dB of gain was applied...... It turned out to be less than 2dB. That's one of the reasons I prefer to add the gain in the preamp, and not from a differential audio output.

In reality, these noise levels will be swamped by ambient noise anyway. As the Alice OPA designer Jules Ryckebusch has commented -- the mic noise is usually 'down in the weeds'!
Very low noise requirements aren't that common. Often for specialised nature recordings - or scientific experiments..
Nice to get the noise as low as possible anyway though!