CBS240,
Now I see what you were getting at. That is a clever idea. However, beware that the voltage at the joint emitters does not only change because of the common mode voltage, but also changes a bit because of the differential voltage. The reason for the latter being that the Vbe vs. Ic transfer curve of a BJT is not linear (and very much non-linear, since it is exponential). That is, if we have no common mode voltage, but just +Vd/2 on one input and -Vd/2 on the other, Ve (which is the same for both transistors) will vary somewhat with Vd. I haven't tried to figure out how you feedback topology would affect differential mode amplification, though. Anyway, as I said earlier, I think it is much better and simpler to just try achieving a high CMRR in the diff pair, if worrying about the common mode voltage at all.
Now I see what you were getting at. That is a clever idea. However, beware that the voltage at the joint emitters does not only change because of the common mode voltage, but also changes a bit because of the differential voltage. The reason for the latter being that the Vbe vs. Ic transfer curve of a BJT is not linear (and very much non-linear, since it is exponential). That is, if we have no common mode voltage, but just +Vd/2 on one input and -Vd/2 on the other, Ve (which is the same for both transistors) will vary somewhat with Vd. I haven't tried to figure out how you feedback topology would affect differential mode amplification, though. Anyway, as I said earlier, I think it is much better and simpler to just try achieving a high CMRR in the diff pair, if worrying about the common mode voltage at all.
Hi Christer
One of the most important rules when building something to me is KSS or Keep it Simple, Stupid.
Audio is so subjective though
. I agree that CMRR is not likely a problem in a well designed/built circuit. But sometimes I try to vary from the beaten path a bit, it helps me understand the subject better. It's a good point you have about the Vbe vs Ic of the VAS BJT's. I hadn't looked at that variable closely. The amount of current change needed by the output stage would be a factor for delta Ic. But if the Vbe of each transistor is changing, that change in voltage would reference from where the emitters connect to each other, the point where the CM feedback derives. Each base would have it's own voltage fluctuation related to Vd. Since this is driven by the drain(or collector) of the input stage, which is the dependent current source, this voltage should be related to the VAS base and not the current in the VAS transistors or input stage, thus eliminating the effect of Vbe in the circuit. However, it might create an extra current change in the input stage related to delta Vbe/R2(or R5), from the drawing. This should offset each other(if the Vbe's are matched) in terms of CM current in the tail of the input stage, but I think it could create an extra current component in each leg of the input stage
. Even if nothing is achieved here in the realm of CMRR, at least the DC bias seems surprisingly stable.
One of the most important rules when building something to me is KSS or Keep it Simple, Stupid.
Audio is so subjective though. I agree that CMRR is not likely a problem in a well designed/built circuit. But sometimes I try to vary from the beaten path a bit, it helps me understand the subject better. It's a good point you have about the Vbe vs Ic of the VAS BJT's. I hadn't looked at that variable closely. The amount of current change needed by the output stage would be a factor for delta Ic. But if the Vbe of each transistor is changing, that change in voltage would reference from where the emitters connect to each other, the point where the CM feedback derives. Each base would have it's own voltage fluctuation related to Vd. Since this is driven by the drain(or collector) of the input stage, which is the dependent current source, this voltage should be related to the VAS base and not the current in the VAS transistors or input stage, thus eliminating the effect of Vbe in the circuit. However, it might create an extra current change in the input stage related to delta Vbe/R2(or R5), from the drawing. This should offset each other(if the Vbe's are matched) in terms of CM current in the tail of the input stage, but I think it could create an extra current component in each leg of the input stage. Even if nothing is achieved here in the realm of CMRR, at least the DC bias seems surprisingly stable.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.