Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Electronic Stethoscope, Electret condenser mic distortion issues

Status
Not open for further replies.
VPot has already emailed me and I told him I had a circuit that has no background noise and you can clearly hear the blood being pushed through and I have had sold 30,000 units and none has complained. But he does not listen and has no idea what he is doing.
I have tried to sell these detectors to Doctors, but they don't buy anything. They would not spend $50.00 on something that is ahead of its time, because they do not understand electronics. And this has been the case for the past 40 years.
 
The lowpass filter has a gain of 1.6 at low frequencies so that the filter has a Butterworth frequency response and that the capacitor values can be the same. If the gain of the filter is one then the feedback capacitor value must be double the value of the capacitor to ground but capacitors with double values are not available so two capacitors in parallel must be used when the gain is one.

You are measuring "just after R5" that is in the middle of the lowpass filter circuit. The input of R5 is the input of the lowpass filter circuit and is the output of the preamp.
Here is the response again:
 

Attachments

  • lowpass filter again.png
    lowpass filter again.png
    23.9 KB · Views: 290
VPot has already emailed me and I told him I had a circuit that has no background noise and you can clearly hear the blood being pushed through and I have had sold 30,000 units and none has complained. But he does not listen and has no idea what he is doing.
I have tried to sell these detectors to Doctors, but they don't buy anything. They would not spend $50.00 on something that is ahead of its time, because they do not understand electronics. And this has been the case for the past 40 years.
Hello Colin,
I've 300$ digital stethoscopes i bought and used from big players like littman and eko devices. I've absolutely no issues trying yours if you have kept the purchasing simple, rather sending money on paypal in advance. I'm sure you might have an online inventory on your website if you have already sold 30000. Could you please redirect me to your website or something where I could buy it with my card. thanks
 
Replying to your conversation where I could not attach anything:
A filter response test is made with a sinewave input that sweeps the frequencies with a constant input level.
I do not know why the Litmann recording has so many sharp sounds and I do not know why the frequency graphs have a linear horizontal frequencies instead of the normal logarithmic.
Here is the schematic modified for using a single 9V battery:
 

Attachments

  • electr. steth 9V.png
    electr. steth 9V.png
    24.8 KB · Views: 296
Replying to your conversation where I could not attach anything:
A filter response test is made with a sinewave input that sweeps the frequencies with a constant input level.
I do not know why the Litmann recording has so many sharp sounds and I do not know why the frequency graphs have a linear horizontal frequencies instead of the normal logarithmic.
Here is the schematic modified for using a single 9V battery:
Thank you John for the schematic.
1) why did you need 20k pot before lm386 instead 10k?
2) can I just use a resistor value by calculating the clipping point for lm386 instead of pot to output one sound level all the time?(just trying to eliminate pot in prototype)
 
I changed the pot to 20k so that the capacitor coupling to it would not need a huge value.
The pot is a two-resistors voltage divider. Adjust the pot to what levels you want then measure its two resistances and use two fixed resistors to replace the pot. Do the same with the gain adjustment if you want.
 
View attachment 110588
After Talking with you on the phone yesterday, This is What I would Recommend.
hello gary, I tried your circuit, for 800hz. but for some reason its attenuating low frequencies like 100hz or below as well. its kinda acting bandpass. please look at following scope output attachments of raw signal. and lowpass output for 80hz and 400hz
do you have any clue?
 

Attachments

  • 80hz_raw.JPG
    80hz_raw.JPG
    579 KB · Views: 276
  • 400hz_raw.JPG
    400hz_raw.JPG
    587.3 KB · Views: 290
  • 400hz_lowpass.JPG
    400hz_lowpass.JPG
    617.1 KB · Views: 276
  • 80hz_lowpass.JPG
    80hz_lowpass.JPG
    567.3 KB · Views: 271
I changed the pot to 20k so that the capacitor coupling to it would not need a huge value.
The pot is a two-resistors voltage divider. Adjust the pot to what levels you want then measure its two resistances and use two fixed resistors to replace the pot. Do the same with the gain adjustment if you want.
Ok John will give it a try.
 
Yes, Gary's circuit had its coupling capacitor values too low so it cut low frequencies. I simulated it and fixed it but it does not make as much as a second-order (12dB per octave) lowpass filter.
 

Attachments

  • Garys lowpass circuit.png
    Garys lowpass circuit.png
    27.9 KB · Views: 273
  • Garys fixed lowpass.png
    Garys fixed lowpass.png
    35.4 KB · Views: 274
I received another "conversation" that I cannot make an attachment for:

Changing the 33k of R5 and R6 to 43k makes the -3dB cutoff frequency 80Hz, not 800Hz. Changing C3 to 4.7nF makes the filter a single order with a cutoff at 52Hz then C3 increases the slope to second order above at above 1700Hz.

Here is my original schematic with these changes:
 

Attachments

  • lowpass filter messed up.png
    lowpass filter messed up.png
    32.4 KB · Views: 267
I received another "conversation" that I cannot make an attachment for:

Changing the 33k of R5 and R6 to 43k makes the -3dB cutoff frequency 80Hz, not 800Hz. Changing C3 to 4.7nF makes the filter a single order with a cutoff at 52Hz then C3 increases the slope to second order above at above 1700Hz.

Here is my original schematic with these changes:
sorry i meant both c3 and c1 should be 4.7nf for 2nd order 800hz... I figured it out but just double checking because, you made a typo here
https://www.electro-tech-online.com/threads/electronic-stethoscope-ambient-noise-suppression.152839/
see post #15
 
Only a "Doctor" would say this:
"I've absolutely no issues trying yours if you have kept the purchasing simple, rather sending money on paypal in advance. "

I have sold over 100,000 items using Paypal and I have had 24 million visitors to my website and you are the first person in the world to dispute PayPal.

That's why, out of the 30,000 units sold, I have only sold a few to "Doctors."
 
Only a "Doctor" would say this:
"I've absolutely no issues trying yours if you have kept the purchasing simple, rather sending money on paypal in advance. ""
I don't know what "rather sending money on paypal in advance" means.

Mike.
 
I received another "conversation" that I cannot make an attachment for:

Changing the 33k of R5 and R6 to 43k makes the -3dB cutoff frequency 80Hz, not 800Hz. Changing C3 to 4.7nF makes the filter a single order with a cutoff at 52Hz then C3 increases the slope to second order above at above 1700Hz.

Here is my original schematic with these changes:
John,
I've following frequency spectrum graphs... one is my device with tl072 pre-amp-lowpass-lm386, the other is eko devices commerical steth.
1)look at eko's graph they are attenuating frequencies lower than 30hz, which is desired in my case too. can I just add a right cap to do so or do i've to have a band pass?
2)do you think their filter is 2nd order?
please advice
 
Where can I see the frequency response graph from "Eco"?
The capacitor C2 from the mic fed from the mic impedance in series with R2 is a first order highpass filter. A capacitor can be added feeding the LM386 input volume control (like in my circuit that has no negative supply) is another single order highpass filter.
The two filters add to make a second order highpass filter and you can set it to a 30Hz cutoff frequency if you want. It does not make a sharp Butterworth corner but makes a droopy corner. A second order Sallen Key highpass filter using one opamp can make a sharp Butterworth filter.
 
Where can I see the frequency response graph from "Eco"?
The capacitor C2 from the mic fed from the mic impedance in series with R2 is a first order highpass filter. A capacitor can be added feeding the LM386 input volume control (like in my circuit that has no negative supply) is another single order highpass filter.
The two filters add to make a second order highpass filter and you can set it to a 30Hz cutoff frequency if you want. It does not make a sharp Butterworth corner but makes a droopy corner. A second order Sallen Key highpass filter using one opamp can make a sharp Butterworth filter.
stupid me.. didn't attach pics..sorry,, here you go
what resistor is combining with cap before lm386 to form high pass?? pot??
I do not know why the Litmann recording has so many sharp sounds and I do not know why the frequency graphs have a linear horizontal frequencies instead of the normal logarithmic.
Its just a different way of plotting it in a different scale, done by python.
 

Attachments

  • eko_miral.GIF
    eko_miral.GIF
    164.2 KB · Views: 255
  • my_device.GIF
    my_device.GIF
    246.8 KB · Views: 249
Last edited:
Only a "Doctor" would say this:
"I've absolutely no issues trying yours if you have kept the purchasing simple, rather sending money on paypal in advance. "

I have sold over 100,000 items using Paypal and I have had 24 million visitors to my website and you are the first person in the world to dispute PayPal.

That's why, out of the 30,000 units sold, I have only sold a few to "Doctors."
musk and me don't like paypal anymore :p . Anyways 'll do that over the weekend.
 
I do not know why your frequency response graphs have lumps and bumps. They should be a perfectly flat line. I also do not know why the slopes are wrong.
Maybe they have actual heartbeat signals showing the levels of various frequencies as the input instead of constant amplitude sinewaves?
 

Attachments

  • filters.png
    filters.png
    395.8 KB · Views: 270
I do not know why your frequency response graphs have lumps and bumps. They should be a perfectly flat line. I also do not know why the slopes are wrong.
Maybe they have actual heartbeat signals showing the levels of various frequencies as the input instead of constant amplitude sinewaves?
so in audacity spectrum plot feature you have an option to choose size which sets" how many frequency divisions are used for the spectrum", which defines the resolution of the graph. when you decrease the size you get a perfect flat graph, above pictures are somewhere middle size. the highest size would have lot of spiky curves. i tested two commercial digital steths recordings both have such graphs... but they definitely have digital filters... but mine is total analogue, still it look like that.

Also do you think analog circuits in practical would give such smooth performance curves as they were shown in simulations?
 
An analog filter produces an absolutely straight line of level change per octave of logarithmic frequency spread, just like is shown in my simulations. I have never seen graphs like you showed and they are completely wrong.
 
Status
Not open for further replies.

New Articles From Microcontroller Tips

Back
Top