|
发表于 2015-7-17 20:29:08
|
显示全部楼层
It is possible to explain why the unloaded buffer is stable. But let's put it aside as another story.
If the buffer is abruptly loaded with a R-C network, the case is different. First is slewing transition, during this phase the buffer works very hard in order to recover the output, which is rising edge in your figure.
After this phase, the buffer goes into class-A (or linear ) region, in this period if you look back into buffer's output, it will exhibit inductive impedance other than resistive. Then the whole circuit is simplified by voltage source, inductive output impedance and r-c network, if r is too small, then L-C oscillation will appear.
Looking back to slew rate tuning, if SR is chosen to be slow, then the buffer requires more time to return class-a opeation. If such a span crosses the inductive-impedance region, then a smaller ringing can be observed. But it never disappears.
Such an issue it common in SAR ADC buffer design. You can refer to one TI's application note for more detail (title: I don't remember. Sorry,,,).
Good luck! |
|