I agree that the 0.22uF capacitors in the photos appear too small. Some are marked 224.
Simple arithmetic shows that 0.22uF into the 30k input impedance of the TDA7297 amplifier causes -3dB (0.707 times) at 24Hz. Then frequencies below 121Hz begin to be attenuated.
Simple electronics shows that a series capacitor feeding a resistance forms a "highpass filter" Please look it up.

Most cheap digital multimeters are accurate for 50Hz and 60Hz but have poor accuracy above.

 

You measured 3.3W with your cheap DMM but the datasheet shows about 5.5W if the 12VDC supply is correct and the speaker plus its wiring is actually 8.0 ohms. Maybe the 12V drops to 10.5V when it produces the amplifier power?
But the difference between 5.5V and 3.3W is only 1.67 times which is a very small loudness difference. 10 times the power sounds twice as loud because the sensitivity of our hearing is logarithmic.

Your cheap little speakers without an enclosure will also sound weak and produce no low and no very high audio frequencies.

 

That explains the lower output at 60hz but the signal out should be higher.

The chart in the datasheet shows it down 1db at 60Hz. So, no, it does not explain the low output.

 

The external input capacitor is the only part in the amplifier circuit that cuts low frequencies. With a huge input capacitor then the low frequency response of the amplifier circuit will be almost DC (less than 1Hz).

 

The chart in the datasheet shows it down 1db at 60Hz. So, no, it does not explain the low output.

If the input cap is 0.22 uf then the impedance at 60hz is appx 12K. The input impedance of the chip is typically 32K.
That would produce appx a 27% loss of input voltage to the chip. The output power would be appx half that measured at 1000hz .

 

So he chart is lying? Where did you get the input impedance?

 

So the chart is lying? Where did you get the input impedance?

From the data sheet but corrected to 30K with a typical voltage gain of 32db or 40

 

Okay. I calculate that the output voltage of the high pass filter is .944 of the input with a phase shift of -28°

So going from 1.34V at 1000Hz to 0.08V, a loss of 95% is certainly not explained by the input filter.

Either the circuit is wired wrong or his measurements are nonsense. I suppose it could be the meter.

 

The input capacitors appear to be a much smaller physical size than 0.22uF, maybe 0.022uF.

 

The impedance of the input capacitor varies with frequency and is higher at lower frequencies. This forms a voltage divider with the input impedance of the chip which is typically 30K ohms according to the data sheet.
A 0.22 uf cap has an impedance of appx 12K at 60 hz. Without going into the math a 0.06 signal will be reduced to 0.042 volts hence the output voltage is lower. What doesn't add up, from your measurements, is the output voltage vs the gain of the chip which is typically 32db or appx a gain of 40.
Multiplying 40 X 0.042 volts = 1.68 volts not 0.08 volts, if the caps are actually 0.22uf as you say.

The previous time I think it was the VOM bug, I set it at 20 VAC, and when I read the input, it says 0.06 or stuff. But this time, when I set it to 2 VAC, it says only 0.003 VAC at 60Hz, and the output is 0.11 VAC, which now makes sense to what you have said above. When I increase the frequency, the gain is increase, at 1000Hz, the input is 0.031 VAC, and the output is 1.37 VAC, which means the gain is 44. Your information is so helpful. Thank you so much

 

You guys' observations are so insane, I just checked the capacitors that I bought, and it's only 220pF. Which means the input impedance is 12M at 60Hz

 

Max output power at 12 volts requires appx 7 volts across the speaker. The input signal after the cap needs to be 0.160 volts at 1000hz.

 

Well, the speaker is really loud now, after I temporarily use a 0.1uF capacitor for the input signals. But besides the enclosure of the speaker, what can I do to improve the sound of the speaker to make it more clear and the noises disappear or dimishes?

 

Take it off the breadboard and wire up the chip using a perfboard. Keeping all component leads and wiring as short as possible.

 

Take it off the breadboard and wire up the chip using a perfboard. Keeping all component leads and wiring as short as possible.

Sure, I mean to make a PCB design for it, but I wonder should I add up any filter for the output to make it sound better

 

xoksined

Well, the speaker is really loud now, after I temporarily use a 0.1uF capacitor for the input signals. But besides the enclosure of the speaker, what can I do to improve the sound of the speaker to make it more clear and the noises disappear or dimishes?

What noises? Do you get noise when there is no input signal? If not, the noise is either coming from the source or is actually distortion.

 

Sure, I mean to make a PCB design for it, but I wonder should I add up any filter for the output to make it sound better

No. I recommend using the pc board layout shown in the data sheet with those exact parts listed.

 

Bad sound is usually caused by clipping at the amplifier output because the input signal level is too high. Turn it down!
Clipping is when the output voltage swing is as high as it can go then the output waveform approaches being a squarewave with many added harmonic frequencies that sounds bad. In Rock "music" clipping is called fuzz or overdrive.

 

Sure, you can add a filter to reduce the distortion but the filter will also remove high music frequencies and remove high speech frequencies.

 

xoksined

What noises? Do you get noise when there is no input signal? If not, the noise is either coming from the source or is actually distortion.

Well, I think it's distortion.