Amplifier (TPA311xD2-Q1) Bad Gain

Hello,

I am using an Audio Codec (VS1053b) to generate audio signals, which are to be amplified and sent to a 4 Ohm speaker. I measured the audio codec output as a clean 2Vpp wave, which is good. As I send the Audio through the amplifier, it is being amplified; however, not at the value I have set it to.

  • I have provided 24V to the Amplifier
  • I have configured R1 and R2 following Image 1 to provide it a Gain of 36dB
  • I have grounded the P_Limit pin
  • For the rest, I basically follow Image 2

A few things that require clarification:

  1. It’s a bridge amplifier so with a supply of 24V, its max output should be 48V in an ideal world?
  2. 36dB = 20log(V/2) = 126Vpp … which contradicts statement 1?
  3. The gain I’m getting is not even close to 48V (moreso around 20Vpp) so am I setting it up incorrectly?

How are you measuring peak-to-peak voltage? Do you have an oscilloscope?

Does your oscilloscope measure differentially? If you are measuring relative to ground, 20V is about right.

A multimeter will measure RMS and 2V peak-to-peak is about 0.7V RMS (assuming a sine wave) and 48V peak-to-peak is about 17V RMS.

  1. 36dB = 20log(V/2) = 126Vpp .. which contradicts statement 1?

Not really - It just means your amplifier is [u]clipping[/u] (driven into distortion). Of course, you should have a volume control to adjust/reduce the overall gain.

If you want to measure gain (which is voltage gain), measure at a more-reasonable signal level... Maybe at ~1W out.

P.S.
Be careful with high power testing when a speaker is connected. A "100W" speaker is designed for 100W of program material which hits 100W on the peaks with an average of about 10W and you can fry the speaker with continuous 100W test-tones. In a 2-way or 3-way speaker system the tweeter (and midrange) are rated for even less because most of the energy (in normal program material) is in the lower frequencies.

Amplifiers are normally tested with a "dummy load" ( a high-power 4 or 8-Ohm resistor).

DVDdoug:
How are you measuring peak-to-peak voltage? Do you have an oscilloscope?
Does your oscilloscope measure differentially? If you are measuring relative to ground, 20V is about right.

Yes, I'm measuring with an oscilloscope. For the input voltage (from the Codec), I am measuring between Signal and GND. For the output voltage (from the amplifier), I am measuring between the two outputs of the amplifier (+ and - ). IF I was measuring relative to GND, then yes I would expect 20V. I'm not sure if you've worked with the amplifier, but did you flash firmware onto it? I'm not sure if there are specific registers that I need to modify to attain max gain.

At the end of the day, the goal is to make my speakers louder. I am not sure why they are so quiet.

At the end of the day, the goal is to make my speakers louder. I am not sure why they are so quiet.

Hmmm... 20V peak-to-peak into 4-Ohms is 12W so I wouldn't expect it to be "quiet". ...Unless you are sending very-low or very-high frequency tones. ...A 12W 1kHz or 2kHz tone should be rather loud if you are close to the speaker. Maybe your speaker is defective?

I am measuring between the two outputs of the amplifier (+ and - ). IF I was measuring relative to GND, then yes I would expect 20V.

Usually the ground on the 'scope probe is earth-grounded so you have to be careful... You could be shorting-out the (-) output. You might want to double-check by measuring both outputs separately, relative to ground.

Have you checked the power supply to confirm it's "holding up"?

I'm not sure if you've worked with the amplifier, but did you flash firmware onto it? I'm not sure if there are specific registers that I need to modify to attain max gain.

No, I have not. You can check to see if it's a gain problem or a clipping problem by lowering the input level and re-calculating the gain. (The [u]TPA3116D2[/u] doesn't have firmware or registers. The gain is set with resistors.)

...One more thing to check - Did you build the amplifier from the chip, or did you buy an assembled amplifier board? The TPA3116D2 chip has differential inputs so the (-) input should be grounded (probably through a capacitor) to a ground shared with the CODEC.

DVDdoug:
Hmmm... 20V peak-to-peak into 4-Ohms is 12W so I wouldn't expect it to be "quiet". ...Unless you are sending very-low or very-high frequency tones. ...A 12W 1kHz or 2kHz tone should be rather loud if you are close to the speaker. Maybe your speaker is defective?

Not to say it's necessarily quiet, but I was fully expecting a 40Vpp into 4Ohms -> significantly louder than it is right now.

DVDdoug:
Usually the ground on the 'scope probe is earth-grounded so you have to be careful... You could be shorting-out the (-) output. You might want to double-check by measuring both outputs separately, relative to ground.

My oscilloscope is floating (I've cut the ground prong from the power chord), but I'll try to measure separately.

DVDdoug:
Have you checked the power supply to confirm it's "holding up"?

I believe so, the PS is capable of providing 90W.

DVDdoug:
No, I have not. You can check to see if it's a gain problem or a clipping problem by lowering the input level and re-calculating the gain. (The [u]TPA3116D2[/u] doesn't have firmware or registers. The gain is set with resistors.)

...One more thing to check - Did you build the amplifier from the chip, or did you buy an assembled amplifier board? The TPA3116D2 chip has differential inputs so the (-) input should be grounded (probably through a capacitor) to a ground shared with the CODEC.

I bought the assembled amplifier board. I ensured that the Codec output and the (-) input of the TPA3116D2 are sharing the same ground. I'm looking into the datasheet and I can't seem to find this value but at what voltage would I expect to see clipping?

I'm looking into the datasheet and I can't seem to find this value but at what voltage would I expect to see clipping?

I didn't find that either but Fig. 14 shows 75W (with 10% distortion) at 4 Ohms with a 24V supply.

...Plugging that into my handy-dandy spreadsheet says that's about 48V peak-to-peak which means there is very little voltage-loss through the amp.

Or, 60W at 1% distortion is about 43V P-P.

You seem to have selected slave mode, so it isn’t going to do anything at all without a SYNC clock.

With the 2V p2p input I’d select MASTER, 32dB. That’ll give a full 48V p2p with 1.2V p2p input
(17Vrms, 420mV rms, to use the much commoner rms rating).

17Vrms into 8 ohms is 36W, into 4 ohm is 72W

MarkT:
You seem to have selected slave mode, so it isn’t going to do anything at all without a SYNC clock.

I do not understand what you mean by this. Would you be able to elaborate?

MarkT:
With the 2V p2p input I’d select MASTER, 32dB. That’ll give a full 48V p2p with 1.2V p2p input
(17Vrms, 420mV rms, to use the much commoner rms rating).

17Vrms into 8 ohms is 36W, into 4 ohm is 72W

I’m not entirely sure I understand what it means to be in Master vs Slave mode.

In terms of my application, I an using an MCU to “tell” the encoder when to send Audio signals to the Amplifier.
In that sense, the Amplifier is a slave?

Would I harm my circuit by setting it as a master?

Muhmmbles:
I do not understand what you mean by this. Would you be able to elaborate?

I'm not entirely sure I understand what it means to be in Master vs Slave mode.

Then you have no reason to use slave mode at all since you won't be using master/slave setup.
In slave mode the amp requires another, master, amp to tell it when to switch. In master mode
it is autonomous - use master mode.

Master/slave setup is for paralleling multiple amplifier chips.

MarkT:
You seem to have selected slave mode, so it isn't going to do anything at all without a SYNC clock.

With the 2V p2p input I'd select MASTER, 32dB. That'll give a full 48V p2p with 1.2V p2p input
(17Vrms, 420mV rms, to use the much commoner rms rating).

17Vrms into 8 ohms is 36W, into 4 ohm is 72W

I've set it to MASTER, 32dB and the output increases to 30Vpp with a 2Vpp input :confused:

Excellent. You'll be needing ear protection now(!)