Electronic stethoscope ambient noise suppression

[Dr.VPot]

Hi all,
I'm trying to implement ambient noise suppression using dual mic in stethoscope, I want my primary mic close to chest piece hearing heart sounds and the secondary mic for listening to ambient noise. I looked into TI far field noise suppression technology, but not available for prototyping. Could any one help me with a circuit or a dev board or soc already available for this purpose. Thanks in advance.

 

[dknguyen]

Wouldn't you just take the ambient microphone's signal, phase shift it 180 degrees and add it to the chest piece signal? Seems like the crudest way to do it. Not sure how well it would work. Would need adjustable gain so you could tune it to the right magnitude to cancel out what the chestpiece is picking up.

 

[gophert]

Hi all,
I'm trying to implement ambient noise suppression using dual mic in stethoscope, I want my primary mic close to chest piece hearing heart sounds and the secondary mic for listening to ambient noise. I looked into TI far field noise suppression technology, but not available for prototyping. Could any one help me with a circuit or a dev board or soc already available for this purpose. Thanks in advance.

Put the microphone into a parabolic cup to make the microphone directional then add closed-cell foam to the edge of the cup to seal out ambient noise. This makes a great first approximation. Then add 180-degree phase shift and feed it in appropriately to cancel remaining ambient. This is done with a simple op amp.

After proper cup design and experimenting, If noise cancelling is needed, As a first experiment, I would put both microphones in the cup. One toward the heart and one 90-degrees to the heart (ambient mic). Connect the ambient to the (+) input of op amp as shown. Then use the pot to attenuate the mic to the right cancelling level.

 

[Dr.VPot]

Put the microphone into a parabolic cup to make the microphone directional then add closed-cell foam to the edge of the cup to seal out ambient noise. This makes a great first approximation. Then add 180-degree phase shift and feed it in appropriately to cancel remaining ambient. This is done with a simple op amp.

After proper cup design and experimenting, If noise cancelling is needed, As a first experiment, I would put both microphones in the cup. One toward the heart and one 90-degrees to the heart (ambient mic). Connect the ambient to the (+) input of op amp as shown. Then use the pot to attenuate the mic to the right cancelling level.

View attachment 110465

hi, thank you so much for your reply.
would you suggest uni directional capsules or omni directional or noise cancelling electret condenser mic capsules? also in your diagram which is primary mic and which is secondary?

 

[Colin]

What sort of microphone are you using.

 

[audioguru]

A unidirectional electret mic already has a vent at its rear to cancel rear and ambient sounds.

The opamp circuit here has 2.2k resistors from unfiltered 9V that will smoke the mics (their maximum voltage is only a few volts) since 2.2k gives the 0.5mA mic a voltage of 7.9V and it loads down the output level. I use 10k fed from filtered 8.5V.
The low value coupling capacitors block low frequency heart sounds. A heart sound booms, it does not "tick".
One mic has a 10k load and the other mic has a 20k load.
The preamp has no gain and is biased to use a dual polarity supply.

My electronic stethoscope uses an ordinary cheap omni-directional electret mic mounted in a plastic lid from a jar of peanut butter. The lid keeps the mic from rubbing on the chest and it blocks ambient sounds.

 

[Dr.VPot]

What sort of microphone are you using.

I have 3types...omni, uni, noise cancellation. all are electret condenser mics

 

[Dr.VPot]

A unidirectional electret mic already has a vent at its rear to cancel rear and ambient sounds.

The opamp circuit here has 2.2k resistors from unfiltered 9V that will smoke the mics (their maximum voltage is only a few volts) since 2.2k gives the 0.5mA mic a voltage of 7.9V and it loads down the output level. I use 10k fed from filtered 8.5V.
The low value coupling capacitors block low frequency heart sounds. A heart sound booms, it does not "tick".
One mic has a 10k load and the other mic has a 20k load.
The preamp has no gain and is biased to use a dual polarity supply.

My electronic stethoscope uses an ordinary cheap omni-directional electret mic mounted in a plastic lid from a jar of peanut butter. The lid keeps the mic from rubbing on the chest and it blocks ambient sounds.

Hey AG,
Even you try to block ambient noise passively with good design, mic is still picking up ambient sounds, i want to eliminate that so i'm inclined to wards this dual microphone suppression idea.
what values would you recommend in the above circuit to make it work?
Also would you recommend a omni mic instead uni in this scenario?

 

[audioguru]

A lowpass filter will reduce sounds from people talking and many other sounds but will pass low frequency heart sounds.
A directional mic at the patient will have its front blocked by the patient but its rear vent will still pickup ambient noises. A second directional mic close to but pointing away from the patient will cancel some ambient noises when simply mixed in-phase with the patient's mic.
An Omni mic at the patient can have a second Omni mic picking up ambient noises and they will cancel some of the noises when mixed out-of-phase.
I do not think two mics will reduce much ambient noise.

 

[Dr.VPot]

A lowpass filter will reduce sounds from people talking and many other sounds but will pass low frequency heart sounds.
A directional mic at the patient will have its front blocked by the patient but its rear vent will still pickup ambient noises. A second directional mic close to but pointing away from the patient will cancel some ambient noises when simply mixed in-phase with the patient's mic.
An Omni mic at the patient can have a second Omni mic picking up ambient noises and they will cancel some of the noises when mixed out-of-phase.
I do not think two mics will reduce much ambient noise.

but how do you think littman 3200 is able to use dual mics in their stethoscope to suppress ambient noise? also in mobile phones they use dual mics to suppress ambient noise
If at all i had to use a circuit to implement this. would you be ok with the above circuit if i replace 2.2k with 10k and 100nf with 10uf? will that work

 

[audioguru]

I have never seen nor used a Littman stethoscope.
Why not calculate resistor and capacitor values instead of guessing? 10uF into 10k ohms produces a cutoff at 1.6Hz earthquake frequencies.
The opamp circuit has a very low input resistance so if you use 10k to power the mics then the 10k input resistance of the opamp will cut mic levels in half.
The 10k resistors powering the mics directly feed power supply noise and possibility of low frequency motorboating oscillation into the inputs of the opamp.
The opamp is designed for a dual polarity supply instead of being biased properly to use a single polarity supply.
The opamp has no gain and no gain adjustment.
The circuit is missing auto gain control and/or a compressor.

 

[Dr.VPot]

I found a circuit reading a literature article... Ive attached the circuit. let me what's your take on it.

EDIT: also another article on far field suppression
https://file.scirp.org/Html/3-1730410_69043.htm#f5

 

[audioguru]

The new circuit you found has a very poor design. It simply has the preamps connected with opposite phase so that they cancel all sounds, unless a sound is closer or is louder at one mic.
1) The LM324 opamp has noise (hiss) and crossover distortion therefore is never used for audio.
2) The capacitors from the mics are calculated wrong and pass earthquake frequencies down to o.34Hz (3 fluctuations per second) which is far below 20Hz and higher frequencies we can hear.
3) The opamps are missing an important supply bypass capacitor.
4) The mics are fed 12V at 0.5mA through 2.2k resistors from the noisy supply without filtering so they are destroyed by the resulting 10.9V which is far more than the few volts they are designed for.
5) Its lowpass filter passes frequencies up to 2853Hz including most voice sounds but cuts all important consonant sounds important for understanding speech. Mine passes up to only 103Hz so voice sounds are mostly gone. I added a switch to change the cutoff frequency to 1030Hz so that breathing sounds can be heard.

The "bandpass" filter in the Chinese article is so simple and poor that I laughed when I saw it. I also laughed at the translation errors.

 

[Dr.VPot]

The new circuit you found has a very poor design. It simply has the preamps connected with opposite phase so that they cancel all sounds, unless a sound is closer or is louder at one mic.
1) The LM324 opamp has noise (hiss) and crossover distortion therefore is never used for audio.
2) The capacitors from the mics are calculated wrong and pass earthquake frequencies down to o.34Hz (3 fluctuations per second) which is far below 20Hz and higher frequencies we can hear.
3) The opamps are missing an important supply bypass capacitor.
4) The mics are fed 12V at 0.5mA through 2.2k resistors from the noisy supply without filtering so they are destroyed by the resulting 10.9V which is far more than the few volts they are designed for.
5) Its lowpass filter passes frequencies up to 2853Hz including most voice sounds but cuts all important consonant sounds important for understanding speech. Mine passes up to only 103Hz so voice sounds are mostly gone. I added a switch to change the cutoff frequency to 1030Hz so that breathing sounds can be heard.

The "bandpass" filter in the Chinese article is so simple and poor that I laughed when I saw it. I also laughed at the translation errors.

I see a switch but where does that go in the circuit in LP filter? I would definitely be interested in having a cuttoff frequency upto 1030hz, because heart murmurs can go upto 800hz

 

[audioguru]

The frequency and parts values are easy to calculate. With 47nF capacitors then R5 and R6 should be 43k for an 800Hz cutoff. Here is my switched circuit:

 

[Dr.VPot]

The frequency and parts values are easy to calculate. With 47nF capacitors then R5 and R6 should be 43k for an 800Hz cutoff. Here is my switched circuit:

Hey AG,
thanks for the circuit. I don't have much familiarity with electrical circuits or audio engineering, my objective of the project is beyond stethoscope, for that first I need a good electronic stethoscope with loud heartsounds and and ambient noise suppressed. I'm gonna implement your circuit and also the passive noise cancellation techniques you mentioned over the weekend. But i'm also looking forward to implement dual mic suppression as well. But you mentioned few issues with the circuit posted by one of the member, you mentioned opamp is designed for dual supply, but he never mentioned any particular opamp in his post.
1)did you mean to recommend a dual supply opamp for this purpose or can i also use a single supply one.
2) could you recommend a best opamp model for this purpose
3) there is MAX9814 microphone amplifier with AGC, can i use that in this scenario.
4) if so could you get me a working schematic of the circuit with right values for me?

 

[Colin]

We sell a steth to doctors that transmits to an FM radio. There is no background noise and the heart sound is very clear. We use 2 transistor amp. Your problem is overloading the mic and a noisy amp. We use a piezo diaphragm. No-one here has the slightest idea what they are talking about . . . because they have never produced the product.

 

[Dr.VPot]

We sell a steth to doctors that transmits to an FM radio. There is no background noise and the heart sound is very clear. We use 2 transistor amp. Your problem is overloading the mic and a noisy amp. We use a piezo diaphragm. No-one here has the slightest idea what they are talking about . . . because they have never produced the product.

hi colin,
I've used cm-01b piezo based sensor, but the sounds were feeble and we couldn't amplify much, since there is lot of whitenoise associated with that sensor. Could you post the circuit you were talking about. It would help thanks

 

[audioguru]

ANY opamp can use a single supply if it is biased properly. A single supply opamp also must be biased properly the same.
An OPA4134 is a good quad audio opamp with the same pins layout as the LM324 in that circuit. It works with a supply voltage from 5V to 36V.
I do not know how well the AGC in the MAX9814 works, it might mess up heart sounds like it messes up music.
I will correct that circuit so that it works better.

I wonder why Colin did not provide a link to his steth product or give one for you to try?

 

[dknguyen]

I wonder why Colin did not provide a link to his steth product or give one for you to try?

I assume NDA.