Author Topic: Microphone bandpass with unity gain design question.  (Read 2851 times)

0 Members and 1 Guest are viewing this topic.

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Microphone bandpass with unity gain design question.
« on: July 24, 2018, 07:32:10 pm »
Hi All, I am designing audio signal processing circuit, the idea is that signal from microphone will go through band pass filter, then be full wave rectified in the end. I broke it into two tasks and task1 is to filter the signal and make it ready to be received by full wave rectifier which will be task2, I am focusing on task1 at the moment
I am thinking to start with this basic circuit (illustration 1) and change components a bit:

It is important that gain of this circuit will be 1 (unity gain), I want signal from mic remain not amplified, just filtered. High Pass Frequency >= 1000Hz, Low pass frequency <= 15000Hz
Illustration 1 - circuit reference from internet:
Illustration 2 - modified circuit with changed component values and different op amp

R1 and R2 have matching values so that gain will be -1
C1 value chosen so that combined with R1 it forms high pass filter with cutoff about 1000Hz
C2 chosen so that conbined with R2 it will form low pass with cutoff frequency about 15000Hz
OpAmp of the choise is: MCP601 - datasheet here: http://ww1.microchip.com/downloads/en/DeviceDoc/21314g.pdf

Microphone will be biased so that in silence it can swing rail to rail with loudest sounds, I suppose it won't be exactly rail to rail some clipping will happen close to 0 and Vcc but the idea is the bias it to squeeze the best Vpp swing of the signal.

Once I can get this circuit to work I will further send it into Op Amp based full wave rectifier.

Does it look okay?


« Last Edit: July 24, 2018, 07:44:08 pm by alexg »
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1507
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #1 on: July 24, 2018, 08:16:58 pm »
An LM358 is much too noisy (rumble and hiss) to be a mic preamp. It also produces high amount of crossover distortion that might mess up your measurements.
It has trouble with all audio frequencies above a few kHz.
 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #2 on: July 24, 2018, 08:20:08 pm »
An LM358 is much too noisy (rumble and hiss) to be a mic preamp. It also produces high amount of crossover distortion that might mess up your measurements.
It has trouble with all audio frequencies above a few kHz.
I mentioned that I am replacing it with MCP621, will it be better choice?
 

Offline HB9EVI

  • Frequent Contributor
  • **
  • Posts: 722
  • Country: ch
Re: Microphone bandpass with unity gain design question.
« Reply #3 on: July 24, 2018, 08:41:46 pm »
still a good and affordable choice for audio is the venerable NE5534, or the LM833... the microchip opa are very good for digital proposes (adc, dac handling), but I wouldn't use them for audio
 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #4 on: July 24, 2018, 08:48:06 pm »
still a good and affordable choice for audio is the venerable NE5534, or the LM833... the microchip opa are very good for digital proposes (adc, dac handling), but I wouldn't use them for audio
Thank you, but there is a catch in my case, some criteria I want to meet for this circuit is side and power consumption, well, besides necessary filtering. My choice has fallen to MCP601 is because it is available in small package, and its power consumption is about 230uA. MAX4466 seems to be tailored specifically for Audio application and is very small and very efficient but I am not sure it can be used for the circuit above, i.e. as op amp in inverting configuration, can it?
 

Offline Audioguru

  • Super Contributor
  • ***
  • Posts: 1507
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #5 on: July 24, 2018, 09:15:13 pm »
The MAX4466 can be an inverting opamp if you want.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #6 on: July 24, 2018, 09:35:09 pm »
Microphone will be biased so that in silence it can swing rail to rail with loudest sounds, I suppose it won't be exactly rail to rail some clipping will happen close to 0 and Vcc but the idea is the bias it to squeeze the best Vpp swing of the signal.

Once I can get this circuit to work I will further send it into Op Amp based full wave rectifier.

Does it look okay?
You do know that common electret mic outputs 20-30 mV at max? No gain, what kind of rail-to-rail you expect?
MCP601 is a bad choice,  I like MCP6021, 6022 series - low noise, low THD. Have a look.
Why inverting configuration?  I'd select non inverting config, with RC frequency shaping in the negative feedback path. Plus you could keep RC on the input, so second order at least overall.
 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #7 on: July 24, 2018, 11:45:14 pm »
You do know that common electret mic outputs 20-30 mV at max? No gain, what kind of rail-to-rail you expect?
MCP601 is a bad choice,  I like MCP6021, 6022 series - low noise, low THD. Have a look.
Why inverting configuration?  I'd select non inverting config, with RC frequency shaping in the negative feedback path. Plus you could keep RC on the input, so second order at least overall.

Yes, I'm aware that electret mic outputs very low voltage, however as far as I know that is at regular speech, music, etc, levels around whats that 50 - 80dB SPL? I am building above mentioned circuit as a first stage into DB analyzer for industrial sound levels, etc, so I can capture loud sounds up to 130dB, at that volume mic output seems easily go rail to rail. All the way up to 3.3v upper half and approximately 300mV lower half. So audio clarity is not that tremendously important per se. Please feel free to correct me if I am doing something wrong, I want to learn along the way as much as I can.

As for "why inverting config" is because I am not exactly sure how to do it with non inverting config, for unity gain.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #8 on: July 25, 2018, 12:03:41 am »
130 dB? Have a part number?
There are a few datasheet on my hard drive, saying 110 dB maximum for electret and 120 for mems. 
What is the purpose to have amp with gain 1? You can put a cap to ground, to have LPF. Same with non-inverting config, cap to ground at the input.
 If intentions is measurements, wouldn't it be wise to have PGA, so you can get bigger dynamic range?
 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #9 on: July 25, 2018, 02:10:40 am »
130 dB? Have a part number?
There are a few datasheet on my hard drive, saying 110 dB maximum for electret and 120 for mems. 
What is the purpose to have amp with gain 1? You can put a cap to ground, to have LPF. Same with non-inverting config, cap to ground at the input.
 If intentions is measurements, wouldn't it be wise to have PGA, so you can get bigger dynamic range?

Unfortunately I do not have mic's datasheet, mic can go up to 110dB that's fine, what I am saying is that the loud environment that I am going to monitor has sound loudness up to 130dB, more less, who knows, but the message it that in my application mic will swing way beyond 20-30mV
I'm sorry but I am not sure what PGA is.
 

Offline MasterT

  • Frequent Contributor
  • **
  • !
  • Posts: 851
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #10 on: July 25, 2018, 02:25:36 am »
You can't get meaningful measurements results if mic is pushed out of specification. Output becomes very non linear, and reliability is in question.
PGA - programmable gain amplifier. What I think, you need some kind of acoustic damping material, that would ensure SPL never goes above 110 dB. Same time if loudness is low, uCPU would change PGA settings, driving gain up.
What is connected after amplifier? 24-bits good audio quality ADC may get you 114 dB dynamic range even w/o PGA.
 

Online floobydust

  • Super Contributor
  • ***
  • Posts: 7676
  • Country: ca
Re: Microphone bandpass with unity gain design question.
« Reply #11 on: July 25, 2018, 02:42:45 am »
Why are you rolling off the low end at 1kHz? There is significant sound energy below there, in both music and industrial noise.

There's different types of microphone capsules - cardioid, omnidirectional, noise-cancelling, so have some awareness of what your capsule and housing need to do to pickup what you want. Sensitivity is around -45 dB Pa.

You'll need to RC filter 3.3V power to the mic and 1/2Vcc divider otherwise all noise on that rail will get amplified.
 

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4321
  • Country: us
  • KJ7YLK
Re: Microphone bandpass with unity gain design question.
« Reply #12 on: July 25, 2018, 03:00:52 am »
It is important that gain of this circuit will be 1 (unity gain), I want signal from mic remain not amplified, just filtered.
Standard practice is to boost mic-level signals to something around "line-level". That is because mic level is "down in the mud" and has a poor signal-to-noise ratio. There is no apparent reason to want to process the signal down at such a low level.

Quote
Microphone will be biased so that in silence it can swing rail to rail with loudest sounds,
This appears to be completely contrary to your requirement that the "signal from mic remain not amplified".
"Rail-to-rail" implies amplification.  Unless your power rails are measured in micro-volts which of course is absurd.
 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #13 on: July 25, 2018, 03:19:33 am »
Why are you rolling off the low end at 1kHz? There is significant sound energy below there, in both music and industrial noise.

There's different types of microphone capsules - cardioid, omnidirectional, noise-cancelling, so have some awareness of what your capsule and housing need to do to pickup what you want. Sensitivity is around -45 dB Pa.

You'll need to RC filter 3.3V power to the mic and 1/2Vcc divider otherwise all noise on that rail will get amplified.
About 1kHz, that is a very good question, honestly I cannot justify it reasonably, the way I come up with that number is that I simply was triggering some noise around microphone while capturing waveform and was not particularly happy how lower frequencies sending ripples allover and when played with software filter in the scope I saw that once you trim off anything under 1kHz waveform is nice and symmetrical, so that is how I come up with 1k lower end roll off, for now I just want to make it work and see that I can sort of have control over it but I do admit that trimming off 1K is a bit too much, probably need to back it up to 500 or even 200Hz

As for power, I am using MIC94310, which is an LDO with ripple blocker technology they say,  it is super quiet and clean supply.

 

Offline alexgTopic starter

  • Regular Contributor
  • *
  • Posts: 90
  • Country: us
Re: Microphone bandpass with unity gain design question.
« Reply #14 on: July 25, 2018, 03:23:23 am »
This appears to be completely contrary to your requirement that the "signal from mic remain not amplified".
"Rail-to-rail" implies amplification.  Unless your power rails are measured in micro-volts which of course is absurd.

Sorry, probably worded it wrong and used wrong terminology, in simple words what I want is to be able for microphone to swing all the way up to 3v as close as possible and down to 0v as close as possible if that makes sense.
 

Offline Richard Crowley

  • Super Contributor
  • ***
  • Posts: 4321
  • Country: us
  • KJ7YLK
Re: Microphone bandpass with unity gain design question.
« Reply #15 on: July 25, 2018, 03:40:34 am »
This appears to be completely contrary to your requirement that the "signal from mic remain not amplified".
"Rail-to-rail" implies amplification.  Unless your power rails are measured in micro-volts which of course is absurd.

Sorry, probably worded it wrong and used wrong terminology, in simple words what I want is to be able for microphone to swing all the way up to 3v as close as possible and down to 0v as close as possible if that makes sense.

No. You used exactly the right terminology.  However your requirements are impossible.  Since there are only microvolts of signal coming out of the microphone capsule, you will need to amplify the signal to produce a 3V peak-to-peak signal.

But it is equally curious why you think you need a 3V PP audio signal?  Virtually all audio gear typically uses internal signal levels around 1V PP. Mic-level (unamplified) signal level is impractical because of poor SNR (signal to noise ratio).  And 3V PP is also impractical because it wastes power boosting the signal to unnecessarily high voltage.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf