Hi! I want to build a noise indicator that lights up LEDs dependig on the noise level. The signal from the mic is a filtered through a highpassfilter and then amplified with the amplification 1+ R1/R2. I'm not sure if this circuit design is viable. Have I forgotten something or should I replace something?
I can't test the circuit right now because I think the OP might be broken.
Thanks in advance!

 

There are LED bar graph audio displays available that would give a better display than your 3 LEDs.
https://www.aliexpress.com/item/1005003777401102.html

 

The Op-Amp probably can't supply that much Current,
and the LEDs will flicker so fast that You can't see them.

You're also putting AC-Voltage into your LEDs,
which they probably won't withstand for very long.
Duplicate LEDs wired with the opposite polarity, in parallel, may work.

The LEDs need a "Sample & Hold" delay Circuit ahead of them which will
keep them temporarily illuminated long enough for your Eyes to see them.
.
.
.

 

In principal the circuit will work, but I’d recommend a few modifications.

To allow for some sensitivity adjustment, you should have some gain adjustment using a pot for resistor R2.
Rather than using differing resistance values for R4 & R5, I’d recommend that instead you replace the diodes in each with differing zener voltage diodes which will give a better indication of differing output voltage levels.

You might need to experiment with different zener voltages along with supply voltages to achieve the best results. To make a 1.7V ‘zener’ you could use 2 diodes in series.

The circuit would need minimal modification to operate from a single rail supply, since when the op-amp output is negative it has no effect on the LEDs.

 

I would use a LM3915 led bar graph chip , it's more suitable for led display.

 

Have I forgotten something or should I replace something?

This modification is unrefined but with sensitive LEDs and the correct component values does work using a LM358 at 9 volts.

 

The modified circuit solves all of the problems, but finding the values of R6, R7, and R8 will probably take some experimenting. With the driver transistor you will not need the extra 2 diodes, unless they are there to set the illumination point of the yellow and red LEDs

 

The LM3915 is not made anymore. Do the Chinese stores make their own copies?

 

The LM3915 is still listed as "Active" at DigiKey, but they are out of stock.
https://www.digikey.com/en/products/detail/texas-instruments/LM3914VX-NOPB/367096

It's no longer available in a PDIP Package,
and is now only available in a SMD Package ( 20-PLCC (9x9) ).
.
.
.

 

If the TS is not making measurements other than Loud, too loud, and WAY TOO LOUD then there is no need for a display calibrated in DBs. Use a quad comparator to provide fairly repeatable switching points, and with a bit more logic there can be a believable transition between the three LEDs. As shown, the green will stay on as the yellow and red illuminate

 

I bought the last LM3915 that was made. It measures in 3dB steps. The LM3914 is still made but only in a surface mount package and its steps are linear volts, not logarithmic dBs.

An electret mic has a maximum input loudness and does not work properly when the loudness is Way Too Loud.

 

I bought the last LM3915 that was made. It measures in 3dB steps. The LM3914 is still made but only in a surface mount package and its steps are linear volts, not logarithmic dBs.

An electret mic has a maximum input loudness and does not work properly when the loudness is Way Too Loud.

Certainly that is correct about electrets, and some other kinds of microphones, as well. BUt since this would be a device to signal loud noise levels, there are many ways to reduce the sound level getting to the microphone A consistent 40 dB reduction should not be difficult at all. And looking at the three indicated levels, stability of calibration will be much more important than accuracy. DBs were not mentioned. four steps from a quad comparator will be enough.

 

https://www.thatcorp.com/datashts/4301data.pdf
contains a rms to DC converter with an output that is linear in dB, plus an op-amp that could be used for the mic amplifier.
The output could drive a LM3914

 

https://www.thatcorp.com/datashts/4301data.pdf
contains a rms to DC converter with an output that is linear in dB, plus an op-amp that could be used for the mic amplifier.
The output could drive a LM3914

Products from "THAT" company would work, but If the noise sensor is to be more than a one-off construction I suggest avoiding single-sourced components. I learned that from bitter experiences. Since this is not intended as a precision measuring system though, an op-amp followed by a simple rectifier/filter, driving a quad-comparator to control the LEDs will be the way to go.. And the whole assembly will probably cost less than that single 20-pin IC from THAT company. UNLESS prices have dropped a whole lot.

 

Products from "THAT" company would work, but If the noise sensor is to be more than a one-off construction I suggest avoiding single-sourced components. I learned that from bitter experiences. Since this is not intended as a precision measuring system though, an op-amp followed by a simple rectifier/filter, driving a quad-comparator to control the LEDs will be the way to go.. And the whole assembly will probably cost less than that single 20-pin IC from THAT company. UNLESS prices have dropped a whole lot.

Agreed. Your paying for a very good VCA which is surplus to requirements (unless you wanted to use it as an attenuator when things got too loud). THAT used to sell the rms-to-dc converter on its own in a 9-pin SIL package for a competitive price, and I have used it very successfully, without any supply problems (but that was a while ago).
The linear in dB outputs gets round the LM3915 being discontinued.

 

REpeating myself here, one more time: This is not an instrumentation type of measuring device. It is to indicate three noise levels: "OK," "Loud", and "TOO LOUD." so while the trigger points need to be stable, they are not specified levels yet.

 

The circuit proposed in post#6 has a measuring range of only 16dB. I suspect that with only 5dB between steps it isn‘t going to be too useful.

 

The circuit proposed in post#6 has a measuring range of only 16dB. I suspect that with only 5dB between steps it isn‘t going to be too useful.

THAT is the reason that I keep suggesting the use of a quad comparator. Much greater range.

 

Circuit Specialists has LM3915 on clearance.
LOGARITHMIC BAR/DOT DRIVER (circuitspecialists.com)

 

THAT is the reason that I keep suggesting the use of a quad comparator. Much greater range.

Not the display IC that’s limiting the range.
The minimum peak-to-peak voltage is 1.3V (because of the rectifier) and the maximum is 9V.