• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Mod input sensitivity TPA3255

DiegoAlm

Member
Joined
Jul 7, 2023
Messages
7
Likes
0
Hello everyone, this is my first post on the forum, sorry if it's not in the right subforum.

I built an active speaker with a 24V battery some time ago, a DSP ADAU1701 and two TPA3118 and it sounded quite good and loud. I wanted to increase its power and I bought two TPA3255 and a 36V converter but the problem is that it has not increased as much power as it should and investigating I have discovered which is because the input sensitivity of the amplifier is 2Vrms and the output of the DSP is 0.9V is there any way to decrease the input sensitivity or increase the output of the DSP? Would it help to change the OP Amps or the resistors? The specific amplifier is this
https://a.aliexpress.com/_mOGPHeg
 
Welcome to ASR!

The easy option IMHO is to add a headphone amplifier with a gain of at least 2 between your DSP board and speaker amp. If you want to DIY, you can build yourself a battery powered CMoy, which should be relatively easy. The design as given by Chu Moy (RIP) had a gain of 11, so you may want to adjust the resistor values to give your desired gain.

If you are more adventurous, not afraid of mucking with tiny surface mount components, and want to do surgery to your TPA amp, you will need to figure out the input buffer schematic. Absent of that here, I am using the TI TPA3255 eval board schematic as example.

The TPA3255 uses differential inputs, and thus there are 2 op-amps for the input buffer per channel. Since the manufacturers usually use dual packaged op-amps, that means 1 chip per channel.

In the TI TPA3255EVM, the buffer is inverting and the gain is given by -R8/R7 and -R41/R42 (the two sets needs to be the same as each other). In the TI EVM design, the buffer is unity gain and therefore R7 = R8 = R41 = R42. [Edit] You may double R8 and R41 or reduce R7 and R42 by half (which will reduce the input impedance of the amp to 5 kohm) to double the gain of the buffer.
TPA3255_design.png

 
Last edited:
Welcome to ASR!

The easy option IMHO is to add a headphone amplifier with a gain of at least 2 between your DSP board and speaker amp. If you want to DIY, you can build yourself a battery powered CMoy, which should be relatively easy. The design as given by Chu Moy (RIP) had a gain of 11, so you may want to adjust the resistor values to give your desired gain.

If you are more adventurous, not afraid of mucking with tiny surface mount components, and want to do surgery to your TPA amp, you will need to figure out the input buffer schematic. Absent of that here, I am using the TI TPA3255 eval board schematic as example.

The TPA3255 uses differential inputs, and thus there are 2 op-amps for the input buffer per channel. Since the manufacturers usually use dual packaged op-amps, that means 1 chip per channel.

In the TI TPA3255EVM, the buffer is inverting and the gain is given by -R8/R7 and -R41/R42 (the two sets needs to be the same as each other). In the TI EVM design, the buffer is unity gain and therefore R7 = R8 = R41 = R42. [Edit] You may double R8 and R41 or reduce R7 and R42 by half (which will reduce the input impedance of the amp to 5 kohm) to double the gain of the buffer.
View attachment 297584
With the SE input, you must not change R41/R42 under any circumstances, as these belong to the subsequent inverted OPAmp.

@DiegoAlm
You may only raise R8/R7. Just set the value to 20k like the AIYIMA A07
 
Sorry for the wait. Do I have to change R7 and R8 or just one of them? How would you locate which ones they are on this board? I can guess the area but there are several with the same value.
Imagen de WhatsApp 2024-09-25 a las 22.36.38_ce5fb996.jpg
 
Don’t do anything, as that’s how it’s meant to be. The difference between 36V and 24V is only 3.5dB, so you won’t notice any significant increase in loudness. These are the laws of electrical engineering. Take a look at the technical data for the TPA3255 and the graph. Similarly, the difference between 130W and 60W is also 3.5dB. I’m sorry, but I don’t have better news for you.
TPA 3255 bei 36V u 24V.jpg
 
Sorry for the wait. Do I have to change R7 and R8 or just one of them? How would you locate which ones they are on this board? I can guess the area but there are several with the same value.

Imagen de WhatsApp 2024-09-25 a las 22.36.38_ce5fb996.jpg



You should put a 8.2k resistor in parallel (on top of) the resistors in the blue circles (= R7-1 and R7-2 in this particular amp).

Reasoning: because the 0.9V out is too low and the gain of the amp is 12x (based on the datasheet of the TPA3255 and the input circuit having -1x gain) and the usage of the 36V power supply it requires more gain as 2V would just get it to clipping level at full power. You thus need 2.2x more gain if 2V is to be the clipping level.
This will increase the gain 2.2x (+6.8dB)

The consequence of this is that the input resistance of the amp drops from 10kohm to 4.5kohm. This should not be a problem for any run of the mill DAC to drive.

Hope this helps... adding (soldering) 2 small SMD resistors on top of the already present resistors is easier than having to replace the feedback resistor.
 
Last edited:
Sorry for the wait. Do I have to change R7 and R8 or just one of them? How would you locate which ones they are on this board? I can guess the area but there are several with the same value.View attachment 394793
The question is, why would you want to change something?
If it gets loud enough, no matter where the potentiometer is, then a change is unnecessary.
 
Because i need more power, never its loud enough in a Boombox.
The question is, why would you want to change something?
If it gets loud enough, no matter where the potentiometer is, then a change is unnecessary.
 
You probably need more gain and driven from a (too) low output voltage source never been able to reach the max. output power.
The change from 24V to 36V (TPA5251, 12.5dB gain)) or even 48V(TPA3255, 16dB gain) will not increase the max. output power simply because the gain of the amp remains the same and so will the amp never clip at 24V nor at 36 or even 48V. It will go equally loud unless you increase the input voltage or gain to benefit from the higher power supply voltage.
 
Last edited:
Basic question, but what's the sensitivity of your speakers? You may be pushing at a closed door (you may need much more power to make any difference, possibly beyond what your amp can provide without clipping).
 
I don't want to sound rude but I just want the input sensitivity of the amplifier and the output voltage of the DSP to match. What does it matter the sensitivity of the speaker or that it only gains 3db? I have a 36V battery and two TPA3255 amplifiers and I want to get the most power possible out of them, I have no problem soldering SMD, previously I mounted a preamplifier that multiplied the signal by 5 so I lowered the signal by 7db at the DSP output and it sounded good but I need to get rid of the preamplifier to make everything more compact
 
Additional question: What is the rated power of the 36V power supply? It should have a minimum of 300W.
 
Additional question: What is the rated power of the 36V power supply? It should have a minimum of 300W.
He wrote 36V battery
 
Additional question: What is the rated power of the 36V power supply? It should have a minimum of 300W.
When i wrote the first post of this thread i was using 24V battery and a 36V converter now im using a 36V Lifepo4 battery
 
Then comes the question what is the current output capability of the battery/converter?
Ideally a maximum of 0.5C, 1C continuously, 3C max. and 5C pulse. So it depends on the capacity. Even smaller cells, e.g. 26650 3600 mAh, can deliver 10A.
There are also cells that can do more, but they are expensive.
 
Ideally a maximum of 0.5C, 1C continuously, 3C max. and 5C pulse. So it depends on the capacity. Even smaller cells, e.g. 26650 3600 mAh, can deliver 10A.
There are also cells that can do more, but they are expensive.
Can 10C continuosly and 15C peaks
 
The powersupply is not an issue here nor the efficiency of the speaker.
Only the too low gain for this specific combination is the issue here.
Easy to solve with 2 SMD resistors.
 
The powersupply is not an issue here nor the efficiency of the speaker.
Only the too low gain for this specific combination is the issue here.
Easy to solve with 2 SMD resistors.
so you claim that there is a -6.8dB missing in his amplifier? That would mean he only makes 27W per channel. Are you sure?
 
Back
Top Bottom