Hi all,
I've been working on a TPA3251 PFFB project lately. I've made a TPA3251 board, compatible with TI's TPA3251EVM some times before, but I gave up using it because of the noise (around 100uV when the inputs are shorted). In this sense, I'm a bit obsessed with noise.
Anyway, my new project will be with PFFB and the analog supply part will be lower noise. I'll use the TPS7A4901, which I've used and liked before, as a linear regulator.
The issue that's on my mind is;
According to PFFB application guide SLAA788A, if each SE output of TPA3251 has a dc offset of VCC/2 (+18v) (which it is) and the offset in the buffer circuit is Va/2 (which is 6v), wouldn't we be giving the 18v offset that we will get from the filter output directly to the output of the buffer opamps with the 18K-2K7 voltage divider? Isn't there a possibility of a problem in this case?
I've been working on a TPA3251 PFFB project lately. I've made a TPA3251 board, compatible with TI's TPA3251EVM some times before, but I gave up using it because of the noise (around 100uV when the inputs are shorted). In this sense, I'm a bit obsessed with noise.
Anyway, my new project will be with PFFB and the analog supply part will be lower noise. I'll use the TPS7A4901, which I've used and liked before, as a linear regulator.
The issue that's on my mind is;
According to PFFB application guide SLAA788A, if each SE output of TPA3251 has a dc offset of VCC/2 (+18v) (which it is) and the offset in the buffer circuit is Va/2 (which is 6v), wouldn't we be giving the 18v offset that we will get from the filter output directly to the output of the buffer opamps with the 18K-2K7 voltage divider? Isn't there a possibility of a problem in this case?
Hi,
From SLAA788 the input buffer as per figure 2 page 3 does not have an offset, hence if the negative input (input pin to 10k resistor to negative opamp input) is left floating, then the opamp output should be 0volts - given that the positive opamp input is at ground.
As you have stated, the series resistance of 2.7k and 18k will allow for a current of 400uA for the TPA3251 quiescent output of 18volts. I expect that this current is sunk by the opamp output. There may be a slight offset in the opamp output to compensate for the small current in the opamp negative input, but i expect this offset to be rather small.
For the noise, the SLAA788 states that the output noise is 28.3uV A-weighted noise. For your noise of 100uV, what bandwidth is this across, and what does the spectrum look like ?
Regards,
Shadders.
From SLAA788 the input buffer as per figure 2 page 3 does not have an offset, hence if the negative input (input pin to 10k resistor to negative opamp input) is left floating, then the opamp output should be 0volts - given that the positive opamp input is at ground.
As you have stated, the series resistance of 2.7k and 18k will allow for a current of 400uA for the TPA3251 quiescent output of 18volts. I expect that this current is sunk by the opamp output. There may be a slight offset in the opamp output to compensate for the small current in the opamp negative input, but i expect this offset to be rather small.
For the noise, the SLAA788 states that the output noise is 28.3uV A-weighted noise. For your noise of 100uV, what bandwidth is this across, and what does the spectrum look like ?
Regards,
Shadders.
Thanks for the answer Shadders,
I guess I didn't explain the subject properly.
My first application was not based on SLAA778. I only took TPA3251EVM as a reference. There, the + inputs of the opamps are connected to 1/2Va as you know. Therefore, there is a +6v offset on the opamp outputs specific to that application.
On the other hand, my noise measurement was made with Quant Asylum 403, but AES17 filter was not applied. Since PFFB was not applied in the circuit, the output noise was obtained at these levels. Because the datasheet also says that noise should be expected around 60uV. I think mine was a little higher.
In my new project, I will apply PFFB. However, I will not feed the opamps symmetrically, but I will feed them with single 12v supply like in TPA3251EVM. Therefore, there will be a +6v offset on the opamp outputs.
I guess I didn't explain the subject properly.
My first application was not based on SLAA778. I only took TPA3251EVM as a reference. There, the + inputs of the opamps are connected to 1/2Va as you know. Therefore, there is a +6v offset on the opamp outputs specific to that application.
On the other hand, my noise measurement was made with Quant Asylum 403, but AES17 filter was not applied. Since PFFB was not applied in the circuit, the output noise was obtained at these levels. Because the datasheet also says that noise should be expected around 60uV. I think mine was a little higher.
In my new project, I will apply PFFB. However, I will not feed the opamps symmetrically, but I will feed them with single 12v supply like in TPA3251EVM. Therefore, there will be a +6v offset on the opamp outputs.