converting mike output voltage to SPL

[MCPete]

My understanding is that the output voltage of most microphones is directly proportional to sound pressure. First assume that this is true for the microphone that I want to use to measure relative SPL by reading the output voltage of the mike. Is it true that

SPL difference = 20*log(e2/e1)

where e1 and e2 are the voltage outputs of the mike where I make some change and I want to know the resultant change of SPL?

In effect I'm asking how a sound level meter converts the voltage output of its mike to a dB reading without going into the required conversion circuitry or how to assign an absolute dB reading for a given output voltage.

I've constructed an electret condenser mike with a preamp using the NE5534 opamp with ACL = 50 to do some special type of SPL reading nearfield to a driver that at the moment I don't want to reveal (top secret! for now). We all have our delusions, or maybe not.

Thanks to all for any knowledgeable input about this.

-Pete

 

[Svante]

My understanding is that the output voltage of most microphones is directly proportional to sound pressure. First assume that this is true for the microphone that I want to use to measure relative SPL by reading the output voltage of the mike. Is it true that

SPL difference = 20*log(e2/e1)

where e1 and e2 are the voltage outputs of the mike where I make some change and I want to know the resultant change of SPL?

Yes, for a linear microphone, voltage is directly proportional to the sound pressure. However, many electret microphones are not linear when it comes to high levels. This is a bit tricky, because it is har to notice. There is no sharp clipping, but a graph of voltage level vs sound level is not always a straight line.

Anyway for the linear microphone, the formula is correct.

In effect I'm asking how a sound level meter converts the voltage output of its mike to a dB reading without going into the required conversion circuitry or how to assign an absolute dB reading for a given output voltage.

Hmm... It has to have some circuits that does the conversion of the voltage to dB. Furthermore it must be calibrated.

I've constructed an electret condenser mike with a preamp using the NE5534 opamp with ACL = 50 to do some special type of SPL reading nearfield to a driver that at the moment I don't want to reveal (top secret! for now). We all have our delusions, or maybe not.

Thanks to all for any knowledgeable input about this.

-Pete

Ok, nearfield measurements can result in high levels, high enough to provoke the non-linear behaviour I mentioned above.

My recommendation would be to compare your design with a "known good" level meter. It is quickly done by holding both meters near a loudspeaker cone and varyling the sound level.

 

[bjorno]

Hi,

Is this the kind of circuits you’re looking for? See 1(1) or:

http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1154,C1086,P7526

http://www.linear.com/pc/downloadDocument.do?id=24931

http://www.ortodoxism.ro/datasheets/analogdevices/26000412AD636_b.pdf

b

1(1)

 

[MCPete]

grateful for replies

Thank you, Svante, for confirming my formula for calculating SPL difference and indicating how to go about checking linearity of my mike. Actually at this point I am happy with measuring SPL change only and don't want to tackle the more difficult pursuit of building the circuits supplying an absolute dB reading.

Also thanks to bjorno for the references to circuits and ICs for constructing a sound level meter. Given the available free time that would be interesting to pursue.

Regards,
Pete

 

[hozone]

hello,
i've also trying to convert voltage reading from an electrect mic to SPL db.

Se
For my test, actually i'm using a ZX-sound like pre: http://www.inexglobal.com/downloads/ZX-sound_e.pdf with R5=33K and WM61A
my reading are:
min output noise @ ~0.2mV RMS
min reading ~60db @ ~120mV RMS
max reading ~85db @ ~2.5V RMS

I've buy a calibrated SPL meter to compare read values to real sound values.

Applying your formula (for 85 db reading 2.5v)
20*log10(2500/120) = 26db
infact 85-60 = 25 ~ 26.

so it works!

Basic question: is the above application of your formula right?

Question 2
my ZXsound like pre has maximum output @ 2.5V, but a low range (almost 85-60db=25db).
i would like to have at least 40db range, is this possible? any suggestion?

thanks!

 

[JMFahey]

You must use far less gain, because at 85dB SPL it's almost clipping, per your description.
You *must* modify your electret to skip the internal preamp, which saturates easily.
As usual, you'll find it in the excellent Rodd Elliott pages.
Recording and Measurement Microphones
Per courtesy, please don't post his schematics here, simply the link to the proper page, as I just did.

 

[hozone]

tks!

no way of having from 50db to 90db SPL (i would like to read this range) without Linkwitz mod?
MCPete use a 50x gain opamp if i've understand, i also would like something like this, better if powered on 5v, and biased on 2.5v. Because i would like to read signal with a microcontroller, and the micro ADC reads from 0 to 5v.

 

[Dissi]

no way of having from 50db to 90db SPL (i would like to read this range) without Linkwitz mod?

Your sound detector circuit has a very poor rectifier stage (IC1/2). Accuracy and operating range are therefore considerably limited. Use a better rectifier to get a dynamic range of 40 dB:

ttp://sound.westhost.com/appnotes/an001.htm

 

[cT equals piD]

tks!

no way of having from 50db to 90db SPL (i would like to read this range) without Linkwitz mod?
MCPete use a 50x gain opamp if i've understand, i also would like something like this, better if powered on 5v, and biased on 2.5v. Because i would like to read signal with a microcontroller, and the micro ADC reads from 0 to 5v.

My previous nickname was MCPete; not liking that name, I changed it with my new email address.

That Linkwitz mod really is difficult to do. But I did it (twice), so you can do it too.

Yes, your application of the formula is correct. Yes, closed loop gain, ACL, of the opamp of my mike amp is about 50.

The power supply for my powered mike is a 9V battery. There is no rectification included with my powered mike. I use the mike for measuring relative SPL of the reproduction of sine wave signals, measuring AC voltage.

The maximum peak output voltage of an opamp is always less than half of the supply voltage for the opamp. With an adequate power supply voltage to the opamp, conceivably you could get a maximum 5VRMS output from the opamp, depending on the output voltage of the WM-61A and a realistic ACL of the opamp.

Regards,
Pete

 

[JMFahey]

better if powered on 5v, and biased on 2.5v. Because i would like to read signal with a microcontroller, and the micro ADC reads from 0 to 5v.

Well, looks like everybody wants to do everything on software these days 😉
Truth is, that the ADC reads "from 0 to 5V" is a minor point.
You need first to amplify/attenuate/filter/process/rectify/whatever your mic signal so you end up with a varying DC voltage, representative of the magnitude you are measuring (in this case SPL) , which is amplified/attenuated/shifted so it moves between 0 and 5V , and that is fed to the ADC to do what you want.
Yes, probably you can simulate a diode, a capacitor, and many other things.
Personally, being old school (literally 🙁 ) , I feel comfortable with actual diodes, capacitors and Op Amps.
But of course, do it your way.
It's just that, as noticed above, a +5V powered Op Amp will never provide 5V PP audio.
OK, maybe a couple can get close, but you are limiting yourself a lot by working that way.
JM2C .

 

[hozone]

Well, looks like everybody wants to do everything on software these days 😉

It's not that i want to do it on micro, i would like to get SPL values and send to COSM.com to log an ambient noise, so software is the only way.

I've open a post here, cause this post was not about the preamp.
http://www.diyaudio.com/forums/analog-line-level/229875-electret-mic-preamp-micro-adc.html
If anyone could read, it will be helpfull 🙂

--
My SPL measure could event be low accurancy (+-6db is accettable) even not weighted, and it should be at least from 40/50db to 90/100.
Micro side i'm using simple RMS, or radix-4 fft, depends on the signal precision and on how much resources i've got on micro.

Because i would like to post this project, i prefer NOT to do Linkwitz.
Said this, my electronics skills are not as good as my it skills, so i ask you expert if you've a working circuit, i've try some as you can see, but nothing as good as i need 😱
--