Building Noise Dosimeter's integrator circuit using analog

[kxl359]

Hi All,

I am working on a noise dosimeter design. The hard part is that I can only use analog circuit...
I got stuck in the time integrator part. My circuit went through amplifier, A-weighting filter, RMS, exponet, and threshold all using analog.
So, I can take out all the signal below 90db.

Now I don't know how to build (a noise dose indicator) 8-hour time integrator to add up all the noise larger than 90db using only analog circuit...

Any idea?

Thanks!!!

 

[FvM]

What's your specific problem? Analog instruments like this have been made for decades before small and powerful microprocessors have been available.

Interesting project!

I presume the integration is performed energy-wise, so the dB level measurement has to be converted to a linear energy quantity before integration. Selecting a suitable energy scale is a key point, I think.

 

[crutschow]

It will be very difficult to design an 8-hour integrator in analog. You will need a high impedance CMOS op amp (perhaps an auto-zero type) and a very low leakage capacitor such as a polypropylene film type. Even then, it will require careful adjustment of offset voltages to keep the drift within your error limits for 8 hours. You may also need to build the circuit on Teflon stand-offs to minimize circuit board leakage.

 

[kxl359]

What's your specific problem? Analog instruments like this have been made for decades before small and powerful microprocessors have been available.

Interesting project!

I presume the integration is performed energy-wise, so the dB level measurement has to be converted to a linear energy quantity before integration. Selecting a suitable energy scale is a key point, I think.

In fact, the spec is simple... use only analog circuit to build a noise dosimeter.
So, I need help on the time clock and the integrator...

 

[Audioguru]

If an "average" person is exposed to 90dB for 8 hours for a few days then their hearing is destroyed but they might be able to hear 300Hz to 3kHz speech, not 20Hz to 20kHz music. All the important consonants of speech and the higher harmonics of music are gone.

 

[ebclr]

If you can use a analog swith , it's a question to turn on and off this switch at the wanted sample time and integrate the result in an op amp integrator.

The clock for the swich you can made with a op amp oscilator.

One question Why you must use analog only in 2013 digital world ?

 

[FvM]

The integrator has to be reset at t=0 and read-out at t=8h. I don't think that you'll prefer an analog electronic timer implementation. If simple digital elecronic isn't acceptable for this design part, an electromechanical timer would be appropriate, I think.

As an additional comment, if you have problems to implement an accurate long time integrator, an electromechanical counter could be used to store integer energy quantities, only keeping the fractional part as analog signal in the integrator.

 

[crutschow]

All in all your project is a very difficult one to do in all analog. I doubt that you will ever get it to satisfactorily work. It should really be done with some digital parts.