Greetings!
I assemble a lot of my own amps for projects. For instance, lately I've been using a sort of "all-in-one" class D amplifier from TI, the TPA3110D2, for simple portable stereo projects. For the most part, I'm very pleased with its performance.
One thing I noticed on my last build, is that is seems to have more background noise than previous projects. Not enough to be annoying, but enough that you can hear it between songs. (Interestingly, on other builds with this chip, I would notice a little white noise when turning the amp way up with no input signal, this time however, it seems to be just a steady quiet level of noise, no matter the volume level).
I'm curious if you guys have any advice or anecdotes on general amp PCB design and overall design of audio projects that might apply?
For instance, on this most current PCB design, here are some things that I did that I wonder if could affect performance:
1. The speaker outputs on the bottom of the PCB have a large positive power supply trace (+12V) crossing them on the top side of the PCB.
2. The left and right input traces on the top of the PCB have a similar large power trace crossing them on the bottom side of the PCB.
As for overall design:
1. The two conductors for the power switch for the amp are actually the shielding from the two speaker output cables. (I feel like this could be a no-no, but I wasn't sure when I built it.) This was done because it was a small project that had the on/off switch right by the two speakers, and limited space and cables. So this means that whenever the amp is on, there is +12V running through the shielding of the speaker cables.
I'm sure there are many other things that could be causing the noise, but any of these jump out at you?
I assemble a lot of my own amps for projects. For instance, lately I've been using a sort of "all-in-one" class D amplifier from TI, the TPA3110D2, for simple portable stereo projects. For the most part, I'm very pleased with its performance.
One thing I noticed on my last build, is that is seems to have more background noise than previous projects. Not enough to be annoying, but enough that you can hear it between songs. (Interestingly, on other builds with this chip, I would notice a little white noise when turning the amp way up with no input signal, this time however, it seems to be just a steady quiet level of noise, no matter the volume level).
I'm curious if you guys have any advice or anecdotes on general amp PCB design and overall design of audio projects that might apply?
For instance, on this most current PCB design, here are some things that I did that I wonder if could affect performance:
1. The speaker outputs on the bottom of the PCB have a large positive power supply trace (+12V) crossing them on the top side of the PCB.
2. The left and right input traces on the top of the PCB have a similar large power trace crossing them on the bottom side of the PCB.
As for overall design:
1. The two conductors for the power switch for the amp are actually the shielding from the two speaker output cables. (I feel like this could be a no-no, but I wasn't sure when I built it.) This was done because it was a small project that had the on/off switch right by the two speakers, and limited space and cables. So this means that whenever the amp is on, there is +12V running through the shielding of the speaker cables.
I'm sure there are many other things that could be causing the noise, but any of these jump out at you?
Power traces are unlikely to cause noise in the normal meaning of noise (hiss).
Continuous noise is usually a result of an inappropriate choice of resistor values, most probably at the input, but possibly in a feedback loop. A quick look at the datasheet for the chip you are using shows a very simple-to-use device with few ancillary components, the first thing I would check is that no resistors of an incorrect value have been inadvertently substituted.
When laying out a PCB it is best to keep inputs away from outputs and both away from power traces. This would mean, for example, that inputs would come in from the left with outputs on the right and power traces coming in from top or bottom. Power traces are, in general, best kept as close together as possible, even superimposed on a double-sided board. Power traces are most likely to contribute to low-level distortion in class B designs. I've never laid out one of these class D chip amplifiers, so there are probably considerations that I'm not aware of. Perhaps someone else will comment.
I see that there is an EVM (evaluation module) available from TI. If you haven't done so, look at the datasheet for this, it shows the PCB layout TI use to get the device to work at it's specified performance. You could do a lot worse than simply copy this. I know it seems like cheating, but professional engineers copy things all the time to get results.
Continuous noise is usually a result of an inappropriate choice of resistor values, most probably at the input, but possibly in a feedback loop. A quick look at the datasheet for the chip you are using shows a very simple-to-use device with few ancillary components, the first thing I would check is that no resistors of an incorrect value have been inadvertently substituted.
When laying out a PCB it is best to keep inputs away from outputs and both away from power traces. This would mean, for example, that inputs would come in from the left with outputs on the right and power traces coming in from top or bottom. Power traces are, in general, best kept as close together as possible, even superimposed on a double-sided board. Power traces are most likely to contribute to low-level distortion in class B designs. I've never laid out one of these class D chip amplifiers, so there are probably considerations that I'm not aware of. Perhaps someone else will comment.
I see that there is an EVM (evaluation module) available from TI. If you haven't done so, look at the datasheet for this, it shows the PCB layout TI use to get the device to work at it's specified performance. You could do a lot worse than simply copy this. I know it seems like cheating, but professional engineers copy things all the time to get results.
- Status
- Not open for further replies.