Preamp IC for hearing aid

[KlausST]

Hi,

the first to test:
* voltage (DC) across the mic. Try to adjust RMIC to see about half VCC ( if not already done)

then:
* CIN2 should be high audio quality. No electrolyte, no X7R, no Z5U best no ceramics at all. Best to use foil.
* add a cap in the range of 10nF in parallel to R4. Use X7R here. (i know it might increase the power up popp...)
* to get lower noise i´d go to 10k Ohms or less for R3, R4
* Cout2 .. best a foil capacitor.

Gain is about 30 (varying with mic impedance) with the 10MHz GBW it should be OK

What load did you connect?

Klaus

 

[aetosa]

Voltage across Mic is 2.3V.
All my caps are SMD ceramics, so I don't really have a choice. (anyway, in the 'complicated' version all caps are ceramics and works fine, so I 'll stick to them).
The load of this preamp is an...amp. High input impedance.

Anyway, I'll revert to the 'complicated' version because, finally, it's less complicated and has a far better sound quality.

Edit: I added a cap across R4 and...oscillates! That's strange...

 

[KlausST]

Hi,

with the given voltage it seems that the microphone has an impedance of 6k5. So increasing your resistor to 6k5 doesn´t make sense sinse your input to opamp is somehow low ohmic. (2k2).
So an high input impeadance non inverting amplifier might be more suitable here.

**********

The used OPAMP doesn´t look that bad - from the datasheet.

Good luck

Klaus

 

[Audioguru]

In my post #12 I said that the opamp is inverting with an input impedance too low. Obviously the opamp should be non-inverting with a high input impedance so the mic is not overloaded.

The popping is called motorboating (putt, putt, putt etc) because the opamp's fluctuating supply current causes the 3V supply voltage to jump up and down and the jumping voltage goes directly to the (+) input of the opamp through R3. Then the opamp amplifies the pulses on the supply voltage. C22 helps smooth the 3V supply but the (+) input of the opamp also needs a filter capacitor to ground.

The complicated circuit uses C44 to filter the bias voltage to the (+) input of the opamp.

 

[aetosa]

So an high input impeadance non inverting amplifier might be more suitable here.

In my post #12 I said that the opamp is inverting with an input impedance too low. Obviously the opamp should be non-inverting with a high input impedance so the mic is not overloaded.

You 're both saying that the non-inverting has a high input impedance.
Of course, compared with the inverting.
Why is this happening?
I thought that both inverting and non-inverting inputs of an opamp have the same circuit.

Edit: I found a conversation about this.

 

[KlausST]

Hi,

With the inverting circuit there is constant voltage at In-.
So with the input resistor you push I = U / R current into this pin.
This input resistor is the impedace that a source sees.



With a non inverting circuit the voltage at In+ is forced by the input voltage.
The input current is dominated by the opamp's input current. Very high ohmic.

Klaus

 

[aetosa]

I converted it to the non-inverting version, the 'complicated' one.
Works fine.
Only one draw-back: I must have a 100uF electrolytic capacitor bettween opamp's positive supply and gnd.
It's bulky and my goal is to minimize it as possible.
Is there any way to avoid it? (with less capacitance the sound quality falls, e.g with a 10uF small smd ceramic which would be ideal for the case))

 

[Audioguru]

I converted it to the non-inverting version, the 'complicated' one.
Works fine.
Only one draw-back: I must have a 100uF electrolytic capacitor bettween opamp's positive supply and gnd.
It's bulky and my goal is to minimize it as possible.
Is there any way to avoid it? (with less capacitance the sound quality falls, e.g with a 10uF small smd ceramic which would be ideal for the case))

Please post your non-inverting preamp schematic and give details about "the sound quality falls".
Distortion? Poor low frequency response?

When the supply bypass capacitor is 100uF then it holds up and smooths the supply voltage if the battery has a high internal resistance. Then the low frequencies are not distorted. A 100uF electrolytic capacitor is a high impedance at high frequencies.
When the supply bypass capacitor is 0.1uF ceramic then it is a low impedance at high frequencies that prevent the opamp from oscillating at a high frequency which distorts and reduces the level of audio frequencies.
Use both capacitors.

 

[aetosa]

It's this. C10 is now 100uF.


(the opamp & the amp gets power from a small dc-dc converter. It converts 1.5V of the battery to 3V)
When I say 'quality falls' I mean that the output power falls, and distortion appears when C10=10uF. Like if it lacks current.
The output ripple of the converter is small and anyway not in the audio spectrum.
I 'll try again with a 10uF to verify this.

 

[D.A.(Tony)Stewart]

As astosa recommended, MAX9814ETD, after looking at the design, I wouldn't look for anything else.

I designed a throat mic for noisy environments with AGC and fast attack : decay ratios of 1:100 . Similar used for jet fighter pilots. The same problems can occur in hearing aids. Very poor SNR from mono omnidirectional feature.

Problems that must be overcome are;

- threshold for speaker hysteresis or shutdown with variable amplitudes and low or high SNR.
- time delay from feedback of inverted audio causing phase shift and oscillation if coupling loss is less than loop gain. Think Bode plots.
- low hiss amp.
- high gain directional mic to block background ambient.. This cannot be achieved by quality mic with back vent to equalize front/back pressure if implanted in ear. ***
-gain control on mic. 2k pullup is low gain, 10k pullup is 5x more.
- 30dB AGC is essential
- fast attack, slow decay is essential 1:500 ms max , pop filter even better.
- Mic supply filter 1k+10uF is adequate then 6+3k to mic. For more gain, less noise and small cap. 100ohm part is to minimize drop from power amp , don't share with mic bias, but no need with part as low noise current source regulated to V/2 and amp filter is included.




There are good quality $10 hearing aids in China, but if it were for me, I would modify a Walkman with earbuds a stereo external cable mic to bear in stereo with switch to cancel common mode background noise *** in both channels with high attack/decay ratio AGC and then switch to mix in MP3 music. Poor ear bud ,& mics are no comparison to even cheap stereo electret mics external protected with sock filters for wind and friction and good ear buds cheap from www.skullcandy.com no comparison to poor audio quality hearing aids that also howl when you remove them from positive feedback (no phase margin from delay)

 

[Audioguru]

It's this. C10 is now 100uF.

You posted the inverting opamp circuit again that has a very low input impedance that attenuates the signal from the mic.
Please post the non-inverting opamp circuit you used.

Since the ripple is low on 3VDC then remove R9.

 

[aetosa]

You posted the inverting opamp circuit again that has a very low input impedance that attenuates the signal from the mic.
Please post the non-inverting opamp circuit you used.

Since the ripple is low on 3VDC then remove R9.

Oh my God! You 're right! I posted wrong image, here is the correct one:



I'll remove R9.

SunnySkyguy all you 're saying sounds very interesting.
I want to make it my self so that I 'll be able to improve it constantly.
So, you 're saying that I should go for MAX9814?

But also, if there is some good hearing aid with 10$ which you can suggest me, I will also buy it.
Do you have something specific in mind?

(my experience till now with cheap hearing aids from China is:
1.low amplification
2.loudness falls as the battery voltage falls. So it sounds weaker and weaker till you change the battery)

 

[Audioguru]

Your schematic shows R5 as a capacitor that might cause the opamp to oscillate at a high frequency. Why does it have R5? A resistor to ground is not needed at the output of an amplifier. But if the amplifier has crossover distortion like in an LM324 or LM358 it reduces the crossover distortion and increases the battery current drain.

Your circuit has C6 that cuts all high audio frequencies. Since the circuit produces acoustical feedback howling without it then maybe the muffled sound is better than continuous squealing.
C6 would not be needed if the earphone had a better seal in the ear canal.

My son bought a very cheap Chinese "Amplified Ear" product. It did not amplify anything since its output sound level was the same as the input sound level.
It cut all high audio frequencies so everything sounded like a cheap AM radio.

 

[aetosa]

Your schematic shows R5 as a capacitor that might cause the opamp to oscillate at a high frequency. Why does it have R5? A resistor to ground is not needed at the output of an amplifier. But if the amplifier has crossover distortion like in an LM324 or LM358 it reduces the crossover distortion and increases the battery current drain.

It was a mistake (in the schematic only).

So, I guess, I should remove R5.

Your circuit has C6 that cuts all high audio frequencies. Since the circuit produces acoustical feedback howling without it then maybe the muffled sound is better than continuous squealing.
C6 would not be needed if the earphone had a better seal in the ear canal.

I have omitted C6 in the real circuit, you 're right.
In fact, it's the 'channel of the sound to mic' that has to be restrict sound coming from any direction.
E.g putting the mic in a tube so that sound enters only from the inlet of the tube.
Earphone is an ordinary commercial earphone, so it's 'seal' to the ear canal can't be anything special.

Now I have another problem, the dc-dc converter suddenly stopped working!
It only has 2 resistors, 2 capacitors 1 inductor and 1 IC, everything looks ok and I can't find the fault!
I replaced the IC nothing.
I replaced the output cap, seemed to work, and now it doesn't!
No short-circuits found!

Edit: the IC gets burned. I allready burned 3 of them (MCP1640), I don't know why! I find no short circuit!
And, again, this IC is supposed to have short circuit protection, so why do they get burned??

 

[Audioguru]

Since you are using an ordinary cheap earphone then acoustical feedback howling will occur when the mic can hear the earphone so the mic and earphone are too close together or the voltage gain is too high.

Maybe you accidently connected the battery with reverse polarity to burn the MCP1640?

 

[aetosa]

No I didn't, anyway I can't, it's the battery holder so that I can't connect it with reverse polarity.

 

[aetosa]

Well...finally the problem was found.
It was a tiny ccrack in...the ground plane...
Works fine.

BUT sound awfull, I think worse than LM324. Which is strange.
I 'll take a look again, maybe I've made some wrong connection of some part.
I don't see anything till now.

 

[aetosa]

I was wrong, the sound is fine, it was just a noisy environment.

- - - Updated - - -

I removed C10, the bulky electrolytic capacitor I wanted to get rid off. It works fine!!!

Then I replaced R9, which is now 100 ohm, and started 'motor-boating'.....
Why 's that?

 

[Audioguru]

I think our ears are stereo and are shaped so that we hear sounds differently from different directions then we can concentrate on somebody's speech from one direction in a noisy environment.
A hearing aid DONT DOO DAT.

 

[aetosa]

Nobody said the opposite.

I put back R9 (100 ohm), works fine.
I guess I can't get rid of it.
Anyway I don't mind.
The important thing is I got rid of the bulky C10!
(I must clarify this: C10, 10uF ceramic SMD, remains. What was removed, was a 100uF electrolytic capacitor which I had put in parallel with C10. So the bulky electrolytic has gone, only the tiny SMD remained, which is exactly what I wanted!)