In search for Ultra-high quality electret mic interface circuit

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Member
Joined 2004
Paid Member
Mems microphones are good. They are extremely consistent. However there are a few tricks. First get bottom port mikes. They are supposed to work better. Second keep in mind that there is a port and is small but of finite length. That may affect the HF response. There seems to be some art in setting them up on a product. For this application some experimentation in the orientation will help figure out what would be right in terms of port and surface impact on the impinging sound waves. The usual measurement mike is modeled as a cylinder with a flat face. A mems mike could be used edge on and have much less impact on the sound field.
 
Member
Joined 2004
Paid Member
All good questions. Most of the mike preamps used for measurement are not that good. I standardized on some ancient HP preamps because their distortion is below .005%. The B&K equivalents were at .05% at best. Measuring the distortion of a microphne is dificult since all you have is a speaker device. Two tone IM is a nice option and worth exploring. There are also issues of air's nonlinearity.

Time for a literature search.

0.1% at 100 dB is not bad at all. I would guess you need a compression driver on a plane wave tube to get much lower at high SPL's. Maybe a tuned tube so it suppresses harmonics.
 
These questions arose from doubts about how it is possible to take valid THD measurements of modern low-distortion drivers capable of performance in the range of 0.1% and below just using an average mic with say 1% overall (capsule + preamp) distortion?

When taking near-field measurements at 2.83Vrms excitation signal with mic about to touch driver's membrane it puts very high demand on the mic's quality at large SPL level.

That is why I'm so desperate for quality circuit design and 3-terminal capsules.
 
After long read of mentioned forum I didn't come up with any beneficial conclusions.

I am reading TI technical note about transimpedance amplifier as it gathers my most interest:

http://www.ti.com/lit/ug/tidu765/tidu765.pdf

So I redraw the scheme using typical 2-terminal capsule:
attachment.php


But here comes the trouble of poor linearity at high SPL levels regarding such a connection.

3-terminal capsule wired the same way would look like this:

attachment.php


So now the question arose how to connect 3-terminal capsule to this circuit properly? Referring to EM173 datasheet accessible here:

http://www.micbooster.com/documents/EM173%20Oct%202011.pdf
I did scheme below:

attachment.php


Now lots of questions:

1) I am trying to figure how the voltage on electret changes in relation to applied bias current? Or another put: They stated 0.6 mA at 5V. But how much mA will be at say 8-9V?

2) Is the latest scheme correct? Is electret capsule being biased correctly this way?

3) How applicable are Demian (1audio) attempts to linearize the FET inside of electret in regards to rest of the circuit drawn here? Is it necessary or not?
Hint: Transimpedance stage can have ultra-high linearity but raw output from internal FET not necessarily...

Exact values of elements are not relevant at this stage of discussion but rest of questions wait in their order. For example:

- Input of source resistor somewhat sets the signal gain of the 3-terminal mic capsule. On the other hand we can alter the gain of transimpedance amplifier stage. So how properly balance both 'sinal gains' in order to preserve lowest possible noise and also to reach specific design goals:

1. nearfield THD measurements
2. typical SPL @ 1m measurements
3. ultra-low noise far-field pick-ups of random sound sources.

Please help me...
 

Attachments

  • mic pre.jpg
    mic pre.jpg
    81.9 KB · Views: 954
  • mic pre1.jpg
    mic pre1.jpg
    82.4 KB · Views: 886
  • mic pre4.jpg
    mic pre4.jpg
    82.5 KB · Views: 721
Last edited:

PRR

Member
Joined 2003
Paid Member
...transimpedance amplifier as it gathers my most interest:...

Why?

That plan is for high impedance sources. The capsule in your link is a 600 Ohm source.

Power it with 9V. Take the output to a low-gain booster of reasonably low OSI, such as NE5532.

Capsule current is rated 0.6mA max at 5V. It won't rise as high as 1mA at 10V. The following amp will probably use more.

135dB SPL is VERY loud. I do not think you have any "linearity" issues in any sane sound field. (If you measure at the throat of a 2" compression horn, you might.)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.