Electret mic and phone for heart and lung sound for telemedicine

[Buk]

Our first intention is for doctor to hear live breathing sound from covid patients.

See this article (warning: long read) to understand why transmitting heart or breath sounds over a live cellular call will not work.

I was going to try and summarise it, but it is easier to simply quote the part salient to this discussion:

Yet even if your cellphone distills crisp, noise-free speech, there’s no guarantee it will arrive at the listener intact. The next threat comes when the phone transmits the call to a base station. Modeled after standard wire-line phones, most mobile phones today digitize audio frequencies from 300 to 3,400 hertz. But unlike landlines, which provide each caller with a dedicated, full-capacity channel, cellphones must share a limited amount of wireless spectrum. So they compress the voice data to let more users connect.

Standard compression rates vary from 12.2 kb/s to 4.75 kb/s, depending on the volume of voice traffic and the strength of the wireless signal. Calls compressed to speeds as low as 7.95 kb/s can still sound almost as good as a landline connection. But beyond that, “you start to hear compression artifacts,” including missing syllables and distortions such as ringing or warbling, says Jerry Gibson, a wireless-engineering expert at the University of California, Santa Barbara.

If you’re making a local call to a mobile user on your own carrier network, count yourself lucky. The compressed data will likely travel to the receiving cellphone without further manipulation, and so voice quality may not be half bad. But say you’re talking to someone across the country or on a different carrier. In those cases, your local network will typically direct the call into the backbone telephone network, which was designed to carry landline traffic at 64 kb/s. So transcoding equipment at the exchange point must convert the mobile voice data to the higher wire-line rate.

A standard landline phone can decode that signal without losing more information. But if your call is sent to another cellphone, voice quality will take another nosedive when the base station serving the phone recompresses the data to fit into a cramped wireless channel.

Sending a .wav or .acc file for playback on the doctor's phone bypasses these problems completely.

 

[rjenkinsgb]

The "modulator" could as simple as a 555 with a bit of filtering to clean up the square wave.
The frequency can be modulated via pin 5, after preamplification with a suitable opamp.

The demodulator would be a bit more complex, but not many needed - or it could be done via a PC program.

That method could also be used for such as ECG waveforms.

 

[doldett]

See this article (warning: long read) to understand why transmitting heart or breath sounds over a live cellular call will not work.

I was going to try and summarise it, but it is easier to simply quote the part salient to this discussion:

Sending a .wav or .acc file for playback on the doctor's phone bypasses these problems completely.

Many thanks Buk. I guess using WIFI for video call is gonna encounter the same problem?

 

[Buk]

Many thanks Buk. I guess using WIFI for video call is gonna encounter the same problem?

I know next to nothing about video calls. The best I can offer is a link to a question about audio quality on video calls. From there, it depends upon which random stranger you choose to believe. Some report "crystal clear audio" others "sound aweful"; and in large part the problem seems to be that the user has no control over which codecs or bitrates any given call will use, as it chooses dependant upon the quality and bandwidth conjestion it finds when setting the call up.

 

[doldett]

I know next to nothing about video calls. The best I can offer is a link to a question about audio quality on video calls. From there, it depends upon which random stranger you choose to believe. Some report "crystal clear audio" others "sound aweful"; and in large part the problem seems to be that the user has no control over which codecs or bitrates any given call will use, as it chooses dependant upon the quality and bandwidth conjestion it finds when setting the call up.

I see. Best option is to record sound and then send it to doctor.

 

[Buk]

doldett: Is it your intention to develop apps for this?

 

[Buk]

The "modulator" could as simple as a 555 with a bit of filtering to clean up the square wave.
The frequency can be modulated via pin 5, after preamplification with a suitable opamp.

The demodulator would be a bit more complex, but not many needed - or it could be done via a PC program.

That method could also be used for such as ECG waveforms.

I get that the FM encoding would permit transmission of below range frequencies via voice-limited calls; but is there any advantage to doing the modulation in hardware, rather than in software?

Also, from my vague recollections of my playing with 555s a long time ago, don't you need a carrier frequency 10x higher than the highest fundemental frequency in order to retain sufficient harmonics to reproduce the original signal with any fidelity?

If the highest fundementals in a breathy weeze are say (a guess) 1kHz, wouldn't that need a carrier of at least 10kHz in order for the doc to have a chance of hearing it?

 

[doldett]

doldett: Is it your intention to develop apps for this?

no, it is not.

We initially tried using lavalier mic, placed in DIY stethoscope (funnel), plugged into my phone. During video call from Line application, my friend heard the heartbeat but not the breath sound. So we thought if the signal from mic could be amplified and lung sound were audible, we could use it for telemedicine. How naive we were ? Anyway, not giving up yet and willing to continue and explore ideas proposed in this thread further.

 

[Buk]

How naive we were ?

I'm coming to the conclusion that the circuits presented in this thread are naive. Rather than amplifying a lot and then low-pass filtering, I think the signal from the mic needs to be low-pass filtered and then moderately amplified.

The idea of filtering first being to remove as much noise from the signal before it is amplified.
Moderate amplification because lavalier mics are already well matched to the input stages of cell phones.

Heartbeats and lung sounds recorded with a lavalier mic will be quiet relative to ambient noise and need amplification, but if you amplify before filtering, the amplified noise still drowns the amplified signal.

 

[audioguru]

I think an ordinary old acoustical stethoscope was put on my back for the doctor to hear my breathing.

 

[Buk]

I think an ordinary old acoustical stethoscope was put on my back for the doctor to hear my breathing.

It can be done from both front and back, and often comparisons between the two.

See https://www.registerednursern.com/lung-auscultation-points-sounds/

 

[doldett]

I'm coming to the conclusion that the circuits presented in this thread are naive. Rather than amplifying a lot and then low-pass filtering, I think the signal from the mic needs to be low-pass filtered and then moderately amplified.

The idea of filtering first being to remove as much noise from the signal before it is amplified.
Moderate amplification because lavalier mics are already well matched to the input stages of cell phones.

Heartbeats and lung sounds recorded with a lavalier mic will be quiet relative to ambient noise and need amplification, but if you amplify before filtering, the amplified noise still drowns the amplified signal.

To low-pass filter and then amplify, I am wondering whether I could swap the 1st and 2nd stage opamp of AG's circuit. The 2nd opamp is low-pass filter with 2.5 kHz cutoff frequency with gain of 1.6.

Just got water cooling pipe insulation. Plan to mount the mic and hopefully it could accoustically isolate the handling noise.

 

[Buk]

To low-pass filter and then amplify, I am wondering whether I could swap the 1st and 2nd stage opamp of AG's circuit.

I think that only the second stage is required, though it needs the cut-off frequency to changed and to be made adjustable. Any more amplification than ~2x the output of the electret overdrives the input stages of the phone and causes the clippiong we've already seen.

I found a (very long and detailed) technical (not medical) assessment of 13 different (10 FDA approved class II) electronic stethoscopes done by an American company.
They ALL have switchable Bell/Diaphram modes.

(background: the traditional steth head has two sides; one an open bell used (mostly) for listening to heart sounds; the other, a flexible diaphram used (mostly) for listening to breath sounds.)

The 'Bell' modes filter the frequencies from 20Hz (one goes down to 15Hz) upto 200/250/350/420/500/650/1000 Hz.
The 'Diaphram' modes generally start higher; 70/100/200/350Hz and go upto 500/800/1200/1500/2000 Hz.

Many also have an 'Extended' mode which covers their lowest Bell mode frequency upto their highest Diaphram mode frequency.

However, of those that provide a recording facility, they all record all freqencies 20(15) .. 20kHz and filters can then be applied in software on an attached PC or proprietary hardware. Many provide upto 20/24x amplification of the mode frequencies, but this is done in the receiving unit not at the mic end.

Most of them have active ambient noise suppression of varying types.

I think it will be very difficult to achieve a usable result with a mic/amp/filter hardware-only plugged into a cell phone solution. my reasoning is that the low volume of heart and especially breath sounds relative to ambient and handling noise is such that you cannot boost those frequencies enough without overdriving the input stages of the phone.

If you added a software element -- developed a custom app -- then it ought to be possible to get enough of the required signal into the phone that could then be filtered and then amplified in software to produce a good result. Even then, I think that you would need some level of noise cancelling; which at minmum requires the addition of a second mic to the 'head' and hardware to subtract what the second mic hears from what the first mic hears. (I've do not know enough about electronics to know whether this is difficult to do.)

This is the front page of the 13 pages of cut-sheets for the commercial e-stethoscopes in case you are interested.

 

[audioguru]

The circuit I fixed 17 years ago was designed to use headphones, not to feed a cell phone. It use five lousy old 741 opamps with no negative feedback and did not work.
After correcting the missing negative feedback, the noise level from the 741 opamps was very high and they did not produce enough current to drive many low impedance headphones. I changed the opamps to a low noise audio dual opamp. I changed the preamp to non-inverting since the original inverting preamp had a low input resistance that attenuated the mic. I added the LM386 audio power amplifier to drive the headphones or a speaker. My new circuit was powered from a single 9V battery instead of the 2 batteries used for 741 opamps. Later I added a pot to the preamp for adjusting its gain from 1 to 22 times. The power amplifier has a volume control and a fixed gain of 20 times.

All microphones are amplified first before filtering. My choice of cutoff frequencies for the lowpass filter were guesses of 103Hz for heartbeats and 2.5kHz for breathing.
The Sallen-Key lowpass filter must be fed from a low impedance for it to have the correct sharp Butterworth sharp cutoff and no frequency peaking. Since each electret mic model has a different output resistance then if it feeds the filter, the input resistor of the filter must be reduced to the correct amount.

The very simple "mic in a lid" I used produced little or no handling or background noises.

It is interesting to see that today there are at least 13 different manufacturers of very expensive electronic stethoscopes that drive headphones or can be recorded.

 

[Buk]

The circuit I fixed 17 years ago was designed to use headphones, not to feed a cell phone

And I am quite sure it performed admirably for that purpose; but as you say, that purpose was not to drive a cellphone.

 

[Buk]

The very simple "mic in a lid" I used produced little or no handling or background noises.

Then you must have a applied some magic; because every patent I have read has passages like this:

The present invention describes a contact type electronic stethoscope with a noise interference resisting function for auscultation, and the stethoscope includes a chest piece for contacting a patient's body, a contact type microphone installed in the chest piece for receiving sounds of patient's organs, and an elastic member disposed on an external side of the contact type microphone, so that harsh sounds and irrelevant noises produced by rubbing the surface of clothes with the patient's body can be avoided during an auscultation, if the piece chest is pressed and contacted with the patient's body without reaching a predetermined pressure and maintains a predetermined distance from the patient's body.

 

[Buk]

All microphones are amplified first before filtering.

But wasn't it you that pointed out that many/most/all electret units have a JFET builtin that both (pre-)amps the signal and presents a very high output impedance at the pins?

 

[audioguru]

The datasheets for most electret mics (Digikey has over 800 of them) say that their output impedance is around 2k ohms. One I measured was 3.9k ohms and is in parallel with the 10k resistor powering it resulting in a source impedance of 2.8k ohms.

I never heard the word "auscultation" before so I looked it up. It is the act of using a stethoscope for a doctor to hear heart or breathing sounds. A definition says that the room should be quiet and should be heated so that the patient has no shivering that causes motion noises. Maybe an old doctor has hands that shake making motion noises? My doctor and I are lucky to have have steady hands.

The voltage gain of the Jfet inside an electret mic is small. It is a common source circuit that overloads when the mic is used in a drum or in a piano so some people modify the Fet wiring to be a source-follower.

 

[doldett]

The very simple "mic in a lid" I used produced little or no handling or background noises.

This gives me hope. I see that you used lid with mic mounted on rubber sleeve. I wish my insulation would do the same job as yours in reducing background noises.

 

[Buk]

Sorry doldett. I don't have the expertise to come up with a suitable circuit for your endevour and AG seems entirely consumed with defending his 17 y/o design for a circuit that is unsuitable for your purpose; despite that noone is attacking it, beyond agreeing with him that it is unsuitable for your purpose.

Maybe that was his peak and he cannot -- or cannot be bothered to -- design anything new.