Electret mic and phone for heart and lung sound for telemedicine

[Buk]

The Jfet inside an electret mic is an impedance converter with an input resistor to ground that is 1G ohms. Operating at about 0.4mA (0.5mA max) and with a low value drain resistor is voltage gain is not high. Its output level is about the same as a dynamic mic (coil and magnet). A preamp is still needed.

I know all of that. That is not an answer to the question I asked.

which was:

Can't you make some pretty good low-noise, low-gain amplifiers using j-fet(s)?

Are you stating that it is categorically impossible to preamplify the signal from a electret mic capsule without using two op amps and 12v?

Are you stating that it is impossible to create a suitable low-noise, low gain preamp for an electret mic using a j-fet and 3V?

Because if you aren't saying either of the above, you are saying nothing useful.

"Learning about electronics"

No idea what that is, but I haven't seen/read it; so the put down attempt is lost on me.

 

[doldett]

I modified the circuit by putting trrs sockets and changed R7 from 1M to 100k. The gain of the preamp circuit gain is then 5.

Below are DIY steth head with hot glue applied and test setup for video recording. Note that all audio cables were used.



Here is the recording of the heartbeat and its spectrum.

This is the recording of the breathing and its spectrum.

My friend said she can clearly hear the heart and lung sound from both recording. However, I still think the breathing sound is not really clear to me and I can barely hear it.

If I compare the spectrum of lung recording with that of very good one I found online, the differences are clearly seen T__T.

 

[audioguru]

Why don't you post the audio files here attached to your post? I must login to CMU with a password to listen to your recordings. I never did.
Before this thread, I have never seen your colored streaks spektrum. I always viewed waveforms like this:

 

[audioguru]

Two Jfets I have used have VGS(off) specs that are 10:1 and 12:1 which need a fairly high supply voltage and a selected bias voltage to match the specs of the Jfet that is used.
Many years ago I made mic preamps with two transistors. Later all the mic preamps were opamps.

 

[Buk]

Two Jfets I have used have VGS(off) specs that are 10:1 and 12:1 which need a fairly high supply voltage and a selected bias voltage to match the specs of the Jfet that is used.
Many years ago I made mic preamps with two transistors. Later all the mic preamps were opamps.

So, what do you think of this:

using K170BL/V's and J74BL/V's?

 

[audioguru]

I have never used Japanese jfets. These ones have a low VGA(off) but it is still a very wide range.

 

[doldett]

Why don't you post the audio files here attached to your post? I must login to CMU with a password to listen to your recordings. I never did.
Before this thread, I have never seen your colored streaks spektrum. I always viewed waveforms like this:

Sorry AG, I never realized it was like that. Next time I am gonna attach the recording into the post.
Here are the recordings of heart and lung sound. Bummer! can only attach photos.

Let's try this online service.

Heart sound

Breathing sound

AG, what software do you use to get audio waveform like that?

 

[rjenkinsgb]

Are you stating that it is impossible to create a suitable low-noise, low gain preamp for an electret mic using a j-fet and 3V?

For a bipolar low-noise preamp, look at ZTX851 & ZTX951; they are pretty much as good as you can get for low noise audio.
Obviously not ideal for ultra high imnpedance - but an electret mic with a low value feed resistro is not ultra-high impledance, it's quite low.

(And as far as mic preamps not using transistors any more - take a look at the preamp inside any professional condenser mic; I've not seen one so far that uses an IC preamp).

The topology of the FET circuit posted above reminds me of some John Linsley-Hood preamp designs? Look at the left half of this, for example; not identical but a similar balanced concept, with exceptionally good characteristics.

 

[Buk]

Here are the recordings of heart and lung sound

They aren't bad at all.

The heart sounds can be greatly improved with a little noise reduction: Your heartbeat recording with some cleanup and a little amplification applied, but you'd need to ask your doctor freind if anything important is missing from this version.

A visual comparison between your original and the modified version:

As you can see by the contrast between the TFFT spectral analysies, the noise is wildly distributed throughout the spectrum, so it is not easily removed with any of the normal filter circuits -- low/high/bandpass or notch.

There is a kind of filter called a noise gate, that suppresses sounds below an given volume. Ie. anything below -30db is squashed completely; but I do not know how easy this is to do in hardware; or even if that is possible at all?

I'll have a go at doing something similar for the breath sounds, but I think that may be harder to get good results.

 

[Nigel Goodwin]

There is a kind of filter called a noise gate, that suppresses sounds below an given volume. Ie. anything below -30db is squashed completely; but I do not know how easy this is to do in hardware; or even if that is possible at all?

I'll have a go at doing something similar for the breath sounds, but I think that may be harder to get good results.

Easily done in hardware, it's fairly simple - and of course that's how it was always done in the past.

The classic example was for recording drums - by using noise gates on the drum mikes you can reduce the various spurious noises you get from drum kits, giving a nice solid drum sound - it really improves the drums.

This is an example of a more complicated commercial one:

Boss NF-1 Noise Gate Guitar Pedal Schematic Diagram

Schematic diagram of the Boss NF-1 Noise Gate guitar effect pedal. This simple effect was sold from 1979 to 1988.

www.hobby-hour.com

 

[Buk]

The topology of the FET circuit posted above reminds me of some John Linsley-Hood preamp designs?

The circuit came from here: https://audioxpress.com/article/EB-604-410-All-JFET-Line-Amp-Starter-Kit, and teh name on it is different, but who knows how much influence your guy had on this one.

For a bipolar low-noise preamp, look at ZTX851 & ZTX951; they are pretty much as good as you can get for low noise audio.
Obviously not ideal for ultra high imnpedance - but an electret mic with a low value feed resistro is not ultra-high impledance, it's quite low.

Thanks, though you'd have to explain the benefits and downsides of moving from jfet to bipolar?

For my own purposes, I'm looking to interface both an electret and peizo units to a smartphone getting as much spectrum & signal with the least noise as I can, whilst (if possible?) driving them from the DC bias provided by teh phone itself. (I think that may not work for the peizo?)

I've downloaded the android development kit and am looking at putting together an app to capture the input and then apply filtering in software; which is more my domain than hardware. But the last time I developed an app was 10+ years ago and it was little more than a "Hello world" affair and much has changed since. Whether the notion will go anywhere I've no idea.

I wouldn't bother at all when there are apps like WavePad already available for free, except I have an extreme ad-adversion. (No. I don't stutter

 

[Buk]

This is an example of a more complicated commercial one:

Wow.

That looks very complicated. How much of that is dedicated to the variable decay time? (And could it be removed if all you want is anything below the sens level gone?)

 

[Buk]

I'll have a go at doing something similar for the breath sounds, but I think that may be harder to get good results

Doldett.

I can't work out what in your lung sounds recording is the signal and what noise.
You'd have to get your doctor freind to twiddle the knobs, or listen while you did. The software I used is WavePad which is free for none commercial use. it has a busy interface and a bit of a learning curve, but most of it you can ignore and I could point you at the 2 or 3 bits you might use.

 

[Nigel Goodwin]

Here's a simpler noise gate schematic, which is nice and easy to understand:

**broken link removed**

 

[audioguru]

OPA2134 low noise audio opamps do not produce all that background noise. Maybe lousy old 741 opamps would.
Did the noisy opamps come from ebay?

Just now, I held my Sound Level Indicator device (VU meter with an electret mic) to my chest and it showed my heartbeats. It has a range of 50dB and shows no output when there are no sounds. It does not have a lowpass filter. It does not use low noise audio opamps but its own noise is very low.

Like most people on these forums, I use the free LTspice program to see waveforms or a frequency sweep on my pc.

 

[rjenkinsgb]

That looks very complicated. How much of that is dedicated to the variable decay time?

VR2..

The bistable at the bottom is an electronic on-off switch, so the bottom third could be removed it you wanted it always on; everything across to Q9 and teh 4k7 resistor. Leave the 1uF cap in circuit.

 

[Buk]

Here's a simpler noise gate schematic, which is nice and easy to understand:

**broken link removed**

Really? You credit me with too much knowledge

Part of the problem is that these circuits are for stereo (speaker) output levels; I'm looking for mono input mic levels.

My (quite likely half-arsed) notion is that a j-fet has a Vgs(off) value. So in a simple preamp circuit like this:

If Rs was an 'electronically variable resistor' -- several ways of achieving this I believe -- then a 'detection circuit' measuring the voltage across Rg, could vary the value of Rs so that when the value across Rs was below some (adjustable) threshold, Vgs moved below Vgs(off) effectively suppressing output until the voltage across Rg rises back above the threshold. And (maybe) the afore mentioned 'detection circuit' could be as simple as a diode with a correctly chosen cutoff value feeding the gate of a second jfet in place of -- or series with -- Rs?

 

[Buk]

VR2..
Leave the 1uF cap in circuit.

Couldn't work out which cap you meant?

One of (2) labelled 1/16 or the one labelled 100/16. or...

 

[Buk]

OPA2134 low noise audio opamps do not produce all that background noise. Maybe lousy old 741 opamps would.
Did the noisy opamps come from ebay?

I think I'd suspect noisy electrolitic caps before the opamps.

 

[doldett]

OPA2134 low noise audio opamps do not produce all that background noise. Maybe lousy old 741 opamps would.
Did the noisy opamps come from ebay?

AG OPA2134 was purchased from Digikey, not from ebay.