Impedance matching question

[tattee]

I've already tried using 1uF but I still wasn't able to detect it. I even used 470uF capacitor but still the same result.

 

[Mikebits]

From the data sheet, distortion seems to go to poop beyond 32 ohm load.

 

[tattee]

Hearing from the tones coming out from it, the 150 ohm sound better than the 8 ohm speaker. The 8 ohm speaker sounds very distorted with a lot of hiss coming out. But still I can detect the tones coming out from it.

 

[Mikebits]

This may be due to the 150 ohm speaker attenuating much of the intelligence in the signal. It may sound better to your limited hearing, but Perhaps the signal needs a fuller frequency response in order to function correctly.

 

[tattee]

Do you have any suggestions on how I can come up with that fuller frequency response?

 

[audioguru]

The output power of the Japanese amplifier IC is very low when its supply is only 3.3V.
The output power is much too low when its load is 150 ohms.
Put the two together then the output power is almost nothing.

Why do you have a resistor in series with your 150 ohm speaker that reduces its power even more?
Why do you have a battery across the speaker?

The gain is set to less than 2 so it will not work from a microphone.

I think the "hiss" when it had an 8 ohm speaker was the amplifier oscillating because its negative feedback is too high and maybe some output was coupled to the input. The datasheet mentions that a "snubber" should be used at the output to stop oscillation.

 

[tattee]

Hi audioguru! Actually that is not a battery but a capacitor. the actual circuit is actually on the first page. that is just a revised circuit I've made which adds the resistor and capacitor in the speaker acting as a low pass filter suggested by mneary.

Up to now, I still can't decode the DTMF tones coming out from the 150 ohms. I'm still tryng to figure out where the problem is.

 

[Nigel Goodwin]

Up to now, I still can't decode the DTMF tones coming out from the 150 ohms. I'm still tryng to figure out where the problem is.

As has been repeatedly mentioned, you've got no gain in the amplifier, and the speaker is far too high an impedance - it's most probably just not loud enough.

 

[tattee]

No, the problem is not on the volume or gain. I tried to lower the volume of the 8 ohm speaker comparing it to the loudest volume on the 150 ohm but still I could detect the dtmf tones coming out from the 8 ohm and not on the 150 ohm which is way louder.

 

[audioguru]

The 8 ohm speaker uses much more current than the 150 ohm speaker. They both have the same voltage swing so the power in the 8 ohm speaker should be 19 times higher than the power in the 150 ohm speaker. Then the 8 ohm speaker should be much louder than the 150 ohm speaker.

Your resistor/capacitor lowpass filter without any values might reduce the level too much or cut high frequencies too much. Maybe the high frequencies should be boosted instead of cut.

DTMF decoders have a certain amount of allowable "twist". Twist is the level difference of low tones to high tones.
Your 150 ohm speaker might have a poor frequency response which causes the twist to be too high.

If your amplifier is clipping (because its supply voltage is too low or because its input and output levels are too high) then the DTMF decoder might reject signals with the added harmonics.

 

[mneary]

Good morning Uncle $crooge. Time to read the past few posts in the thread, and maybe even the speaker data sheet. Yes the twist is too high. Late last night I suggested a low pass filter on the 150 ohm speaker because the data sheet showed it might need to be tamed in the 1k-2kHz range.

The 8 ohm speaker is probably bigger so it has a better near field response at low frequencies. I suggested foam coupling for the 18mm speaker might help that.

 

[audioguru]

I didn't see the datasheet for the tiny "speaker".
Its frequency response is horrible with a huge peak at 1.6kHz. No wonder it does not produce DTMF properly. A lowpass filter will not notch out the peak at 1.6kHz.

It would be a challenge to design a notch filter with the correct amount of depth and bandwidth to make the tiny speaker have a flat frequency response.

 

[tattee]

I have attached a screenshot of the speakers signal generated. notice the shift of A & B tones between speakers.

 

[tattee]

The 8 ohm speaker uses much more current than the 150 ohm speaker. They both have the same voltage swing so the power in the 8 ohm speaker should be 19 times higher than the power in the 150 ohm speaker. Then the 8 ohm speaker should be much louder than the 150 ohm speaker.

You're right, given same condition! But I do have a potentiometer installed, so i can adjust the volume. That is why I could set the 8 ohm speaker to its lowest volume and the 150 to a higher volume. The thing is, even if the 8 ohm has lower volume it can still be detected. So I think it is not about the volume or gain issue.

 

[audioguru]

For DTMF to work properly, the low frequency tones must be close to the same level as the high frequency tones.

Your tiny "speaker" has a huge peak at 1.6kHz that messes up its production of DTMF.

I think your amplifier is clipping (producing many harmonics) when you try to make it play your tiny "speaker" loud enough.

 

[tattee]

Wow, this is such an expensive speaker and I'm having so much problem with it. I appreciate all your help guys! Thanks!

The reason by the way on why I chose this over the 8 ohm speaker is because it takes lesser energy from my battery. Seems to me, it's not worth it!

Is there no other way I could possibly fix this thing? By the way, the DTMF input is coming from a DTMF chip which is controlled by a microcontroller.

 

[audioguru]

You could sweep frequencies into your tiny "speaker" and note the amount of peaking, the frequency and bandwidth of the 1.6kHz peak. Then make a notch filter circuit with the exact opposite.

 

[tattee]

Your tiny "speaker" has a huge peak at 1.6kHz that messes up its production of DTMF.

Does this mean that only the "A" "B""C" "D" tones of the DTMF are messed up since they are the only one under the 1633 Hz frequency?

I think your amplifier is clipping (producing many harmonics) when you try to make it play your tiny "speaker" loud enough.

How will I figure this out? Do I have to use an oscilloscope for this coz I don't have all the equipment really..

 

[audioguru]

Most frequencies are messed up because the tiny "speaker" has not enough output from 500hz to about 1200Hz and too much output from 1200hz to the 1633 highest frequency.

You should be able to hear distortion if the amplifier is clipping.

 

[mneary]

Does this mean that only the "A" "B""C" "D" tones of the DTMF are messed up since they are the only one under the 1633 Hz frequency?

Since I didn't expect you to be testing ABCD with the phone company I was hoping the low pass filter would be enough to correct the rest of the twist.

I was also hoping that the clipping (if any) was mostly third harmonic which the filter would attenuate and the phone company should ignore anyway.

As I read the fine print in your speaker's specs, I'm reminded of something else I wrote a few days ago. I suggested to try a foam donut to couple the speaker to the microphone. Lo and behold, I find that the frequency response was taken when coupled with an "IEC 318 Coupler". It's like putting the speaker right next to your ear and it makes a big difference.

This expensive speaker needs a proper enclosure (possibly a ring of high density foam will help) between it and a phone. Once you've done that, you can probably begin trying 50 cent speakers.