Questions about electret microphones

[uoficowboy]

Hi - I'm working on a design that will use an electret microphone. I've never used one before, so I am a little lost here.

First off, I'm trying to understand the sensitivity of these devices.

1. It seems the standard is to rate them in dba, where 0dba = 1V/pascal at 1KHz. So a 0dba microphone would output a 1V sinusoid (I'm assuming that is peak to peak) if given a 1KHz 1 pascal (again, assuming peak to peak) input. I think this is correct. However, there is also the unit dba SPL. I think this refers to 20 micro pascals (so 0dba SPL = -94dba). So a 0dba SPL microphone would produce a 1V sinusoid with a 20 micro pascal (at 1kHz) input. Is that right?

The **broken link removed** appears to have the standard 0dba unit.
The PUI POM-2738P-C33-R appears to use the same unit, but it doesn't specify it as carefully as Panasonic.
The **broken link removed** once again uses the same unit, and has amazingly flat frequency response.
The CUI CMA-4544PF-W uses this unit as well.

2. Regarding biasing, I'm not exactly sure how this should be done. Take a look at the Panasonic part mentioned earlier. It shows a 2.2K resistor being put in series with the positive supply. It then quotes a max voltage of 10V and a typical voltage of 2V. Since electret mics are essentially JFETs, I'm assuming that they're just saying that you need to keep the voltage across the part between 2V and 10V and it will operate in its normal manner. However, how do they get the 2.2K value? Since I believe in this case the JFET will be acting like a voltage controlled current sink, couldn't you double the signal amplitude coming out of the device by doubling the resistor?

Further, they quote a -3db sensitivity change at 1.5V across the microphone. But they also quote the sensitivity with 2V across the microphone and resistor combination, and they quote the current through the microphone as 0.5ma maximum. That suggests to me that in the worst case scenario (in the way they tested the device's sensitivity) it has 2V - 0.5ma*2.2K = 0.9V across it. Surely I'm missing something? Maybe typical current is an order of magnitude less than the max?

Finally, do these devices have any significant sensitivity or noise or any other changes when you change the supply voltage? I understand that if the voltage is too low the device will stop working, but what if you increase the supply voltage from 2V to 4V?

3. Regarding SN ratio - I'm going to refer to the CUI datasheet, as it gives more information. It says that it is 60db at f=1KHz, 1pa. Since the sensitivity is -44dba, that suggests about 6.3mV signal at 1KHz, 1 pascal. SNR of 60db would then mean that the noise would be three orders of magnitude less than that, or 6.3uv. Does that seem correct?

4. What about max pressure levels? I don't see anything in these datasheets quoting max pressure levels. Is it just that they are high enough that they are not particularly sensitive to this?

5. All of these devices spec their output impedance. Why do I care about it? Is that just for filtering purposes (ie if you're capacitively decoupling the output you would use that resistance so that you set your low end frequency low enough?)

6. On the Panasonic datasheet, there is a sensitivity graph with two lines plotted - one marked 0 degrees and the other marked 180 degrees. What does that mean?

That's it for now. I would really appreciate any help you all can give me with this!!

Thanks!!

 

[goldsmith]

First off, I'm trying to understand the sensitivity of these devices.

Hi uoficowboy
It depends on quality of electret microphone which you're going to buy and input impedance and SNR ration or your preamplifier .

So a 0dba SPL microphone would produce a 1V sinusoid with a 20 micro pascal (at 1kHz) input. Is that right?

It depends on what kind of voice you're applying through that . if your voice is like a sine wave then it would deliver a sine wave !

Let me confess something ! i will never take care of things which you said . because when i'm going to design an amplifier i will consider it may be used with several types and brands of microphones . so it is enough that the Zin be around 20 k ohms and gain be as much as enough so there won't be any need that be wondered of those things , i think .

2. Regarding biasing, I'm not exactly sure how this should be done. Take a look at the Panasonic part mentioned earlier. It shows a 2.2K resistor being put in series with the positive supply. It then quotes a max voltage of 10V and a typical voltage of 2V. Since electret mics are essentially JFETs, I'm assuming that they're just saying that you need to keep the voltage across the part between 2V and 10V and it will operate in its normal manner. However, how do they get the 2.2K value? Since I believe in this case the JFET will be acting like a voltage controlled current sink, couldn't you double the signal amplitude coming out of the device by doubling the resistor?

In fact capacitive microphones are nothing more than a variable capacitor so needs to be biased . for driving them there are several ways . using a resistor which will be connected to that and another head will be tied to the supply rail . or perhaps using a j fet amplifier with high input impedance .

since electret mics are essentially JFETs

Who told this ? Jfet is kind of semiconductor . kind of transistor .

Finally, do these devices have any significant sensitivity or noise or any other changes when you change the supply voltage? I understand that if the voltage is too low the device will stop working, but what if you increase the supply voltage from 2V to 4V?

Somehow ! because when you are decreasing the supply voltage the amplitude of out put voltage will be decreased so the noise might have more effect on the signal . if you increase the PSU voltage then the delivered out put will be higher . do you know how capacitive microphones are working ? consider a variable capacitor in series with a source and resistor . if you change the value of capacitor then the voltage across it will have some variations . ok ?


I think most of these confusions and questions will be solved if you but one of those microphones and then test it via a battery or PSU perhaps , and then tie it via a resistor into the supply rail and then use a coupling capacitor and then connect it to the oscilloscope and then say something nearby it and see what the oscilloscope can show . so measure it and then you'll see most of your questions are being answered automatically !


A question , i can answer all of your questions but before that a question , what kind of aim you're looking for by this microphone ? a pretty precise project or perhaps geophysics projects ? or just sound amplifiers ? if just sound amplifiers these considerations have nothing important to deal with !

Best Wishes
Goldsmith

 

[Audioguru]

In fact capacitive microphones are nothing more than a variable capacitor so needs to be biased.

No.
Look in Goggle and you will see that the electret mic (that we are talking about) is a condenser mic with a high voltage built into its electret material. It makes a variable low value capacitor with the moving diaphragm as one plate so it has an extremely high output impedance. Therefore there is a Jfet impedance-reducer inside that must be powered by a few volts.

 

[BradtheRad]

I made an electret microphone. The battery was a single 1.5 V cell. (I forget the resistor and capacitor values, but they were typical of schematics I'd seen.)

The mic worked like a champ.

Unfortunately after a period of disuse, the battery voltage deteriorated. I was unaware of this next time I used the mic. The playback exhibited shatter (raucous distortion) during moments of loud speech. I believe the signal was being clipped at the ground rail or the supply rail.

Besides the fact that I should have checked the battery voltage, I think I ought to have selected a resistor which resulted in a signal level halfway between the ground and supply rail. In other words, a resistance whose value was effectively the same as the average resistance within the electret element. It cannot be read directly, since it is constantly changing. The waveform must be observed on an oscilloscope.

 

[uoficowboy]

A question , i can answer all of your questions but before that a question , what kind of aim you're looking for by this microphone ? a pretty precise project or perhaps geophysics projects ? or just sound amplifiers ? if just sound amplifiers these considerations have nothing important to deal with !

Right now I'm just trying to understand how to read the datasheets of these parts. Once I understand the datasheets better I will have a better chance at selecting one that fits my application. My application is not overly high end. I do understand that I could answer my questions with some good calibrated audio equipment, a lab setup, and a lot of time. But I do not have the equipment for that. So I'm asking here!

 

[goldsmith]

Look in Goggle and you will see that the electret mic (that we are talking about) is a condenser mic with a high voltage built into its electret material. It makes a variable low value capacitor with the moving diaphragm as one plate so it has an extremely high output impedance. Therefore there is a Jfet impedance-reducer inside that must be powered by a few volts.

Hi all

Audioguru
What if you tell us difference between condenser and capacitor ? both are exactly the same together . but the word condenser is a bit older than capacitor as i know and as i can remember it has been discussed here in this forum about two years ago . and i was one of the contributors of the discussion too .

Therefore there is a Jfet impedance-reducer inside that must be powered by a few volts.

I know some of the microphones have something auxiliary , internally but it is not have anything with the term that that microphone is like a Jfet . The basic of that is a capacitor .
Could i understand your meaning correctly ? or perhaps you're referring to something else ?

Right now I'm just trying to understand how to read the datasheets of these parts. Once I understand the datasheets better I will have a better chance at selecting one that fits my application. My application is not overly high end. I do understand that I could answer my questions with some good calibrated audio equipment, a lab setup, and a lot of time. But I do not have the equipment for that. So I'm asking here!

Well , let me tell you something ! don't limit yourself to the equipments . many years ago which i didn't have any oscilloscope i found someway to handle my aim without it . if i don't have anything it can't stop me ! of course reading the datasheet is good but sometimes you should know that there is some differences between datasheets and real results . so it is possible that you design something based on datasheet and see the difference result in practice ( however a bit ) for instance you can take two NPN transistor with exactly the same type datasheet tell you something about the hfe . but in practice you can see for example two BD139 are dealing with difference hfe in compare together .

Ok i'll help you in both manners . one of them practically and the other one theoretically .

Have you ever seen circuits which has been called level indicator ? they can be simply designed by op amps or by transistors . they are having some LEDs and when your input signal is going to be higher ( it's amplitude ) number of LEDs which are turned on will be increased ( usually 10 LEDs ) if you want i can give you one of the simplest circuits for this . you can drive a microphone and lead it's out put signal through it and thus you can compare some microphones together ( according to the frequency response or sensitivity or quality )
You see , there are too many ways for everything and a problem should not stop anybody .
( i can remember that when i was a little boy i converted my TV into an oscilloscope easily ( from its aerial path ) ! ha ha ! )
Ok , now choose which way . are you like to use practical ways which can be more precise than relying on datasheets , or just datahseets ? it is up to you i just will try to help you in your desired path .

Good Luck
Goldsmith

 

[Audioguru]

A condenser is a capacitor but the word condenser is very old. A condenser mic used an external 48V supply and a high input impedance vacuum tube preamp.
An electret mic has the electret material permanently charged to 48V and has the Jfet with a high input impedance.

The terminals of an electret mic are connected to the Jfet, not to the extremely high impedance condenser.

An LM3914, LM3915 and LM3916 are level indicator ICs that drive up to 10 LEDs. They are accurate. I made one with two LM3915 ICs joined to drive 20 LEDs over a range of 60dB.

 

[goldsmith]

A condenser is a capacitor but the word condenser is very old.

Hi Audioguru
Yes , of course . as i mentioned it too .

A condenser mic used an external 48V supply and a high input impedance vacuum tube preamp.
An electret mic has the electret material permanently charged to 48V and has the Jfet with a high input impedance.

Agreed . both with the same mechanism ( approximately )

The terminals of an electret mic are connected to the Jfet, not to the extremely high impedance condenser.

And i didn't say this too .

An LM3914, LM3915 and LM3916 are level indicator ICs that drive up to 10 LEDs. They are accurate. I made one with two LM3915 ICs joined to drive 20 LEDs over a range of 60dB.

I have used LM3914 . it was brilliant ! 10 LEDs as a simple result and 100 LEDS with 10 LM3914 and some filters as a simple spectrum indicator ! :wink:

 

[Audioguru]

I have used LM3914.

It is linear so it does not have the logarithmic response to levels needed for audio like an LM3915 has.

 

[goldsmith]

It is linear so it does not have the logarithmic response to levels needed for audio like an LM3915 has.

Hi Audioguru
Thanks for that i didn't know LM3915 is logarithmic . it's cool !

 

[Audioguru]

Hi Audioguru
Thanks for that i didn't know LM3915 is logarithmic . it's cool !

The LM3914 is linear. Each step is 1/10th of the total. If it is used for audio levels then the lower outputs are spaced far apart and the upper outputs are crammed together.
The LM3915 is logarithmic so each step is 3dB more than the next step for a total range of 30dB. 3dB is double the power which is heard as a small increase of level.

The LM3916 is a "VU" meter but I have never seen one and I do not like the spacing of its levels.

I made two level indicators. One had two LM3915 ICs in series for a total of 60dB. The other has a peak detector and an automatic gain circuit that adds 20dB to the LM3915 for a total range of 50dB.