I have searched for a while and doesn't seem I'm able to find some discussion on this.
I have a few class D boards such as the Sure TK-2050, YJ Tpa3116d2, original sonic TA-2024 and the Tpa3118d2 PBTL. I have also read the datasheet and noticed most of them have differential input. I noticed typically the negative arm of the diff input is connected to ground by a cap of same value as the positive arm, which will take input audio signal.
When I modded the input caps, I usually change both the positive and negative caps. My questions:- Is it necessary to change the negative input to ground cap also (for a better matching) ? How about if I increase the capacitance of the positive side cap but not the negative to ground cap ? Will there be a problem since they are not equal in value and will give asymmetrical corner frequency for bass roll off ?
I have a few class D boards such as the Sure TK-2050, YJ Tpa3116d2, original sonic TA-2024 and the Tpa3118d2 PBTL. I have also read the datasheet and noticed most of them have differential input. I noticed typically the negative arm of the diff input is connected to ground by a cap of same value as the positive arm, which will take input audio signal.
When I modded the input caps, I usually change both the positive and negative caps. My questions:- Is it necessary to change the negative input to ground cap also (for a better matching) ? How about if I increase the capacitance of the positive side cap but not the negative to ground cap ? Will there be a problem since they are not equal in value and will give asymmetrical corner frequency for bass roll off ?
After having walked the dog (or the other way around) I dare make a fool of myself and present an argumentation.
The majority of monolithic class D chips are designed with balanced input lines. When you do not use the balanced input lines feature, one input is tied to ground without signal. It could be any of the two input lines but normally the one tied to ground is designated the negative input. The negative input needs to be tied to ground through a capacitor that allows for the bias voltage at the chip input pin.
The negative input pin needs to be decoupled to ground in order to reject noise coupled onto the pin but unless the chip input bias circuit allows a signal to propagate via the bias circuit (like on the TDA8932), no real signal but noise is present.
I assume the rationality behind using two similar capacitors at the two input lines, even with one input grounded, is that the unused input line is decoupled in the same frequency range as the frequency range used at the active input line. But, even if the input capacitor at the active "+" line is changed (moderately), it won't make a difference at the "-" input because there is substantially no signal anyway.
Therefore, I see no need to adapt the value of the capacitor in the "-" input line.
However, if the input capacitor in the "+" line is changed for a better quality (ceramic->foil)), I would do similar for the capacitor in the "-" line. This, because foil capacitors have less noise than ceramic capacitors which is important also for the unused "-" input.
The majority of monolithic class D chips are designed with balanced input lines. When you do not use the balanced input lines feature, one input is tied to ground without signal. It could be any of the two input lines but normally the one tied to ground is designated the negative input. The negative input needs to be tied to ground through a capacitor that allows for the bias voltage at the chip input pin.
The negative input pin needs to be decoupled to ground in order to reject noise coupled onto the pin but unless the chip input bias circuit allows a signal to propagate via the bias circuit (like on the TDA8932), no real signal but noise is present.
I assume the rationality behind using two similar capacitors at the two input lines, even with one input grounded, is that the unused input line is decoupled in the same frequency range as the frequency range used at the active input line. But, even if the input capacitor at the active "+" line is changed (moderately), it won't make a difference at the "-" input because there is substantially no signal anyway.
Therefore, I see no need to adapt the value of the capacitor in the "-" input line.
However, if the input capacitor in the "+" line is changed for a better quality (ceramic->foil)), I would do similar for the capacitor in the "-" line. This, because foil capacitors have less noise than ceramic capacitors which is important also for the unused "-" input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.