We are going to develop a new LiPo battery UPS circuit for our handheld. I have read couple of articles and almost all of them recommend to charge the battery based on the following curve. I have also read articles that authors say that reaching the 4.2V limit is sufficient for most cases. Someone please guide me on this issue.
Recommended charge curve:
should I strictly follow the curve
You don't "follow the curve". The voltage curve itself is a result of interaction of battery chemistry and constant current supply. Just as the second (current) curve is a result of holding the voltage constant.
Having said that, any custom circuit based on discrete components will most likely perform worse than specially designed IC, not to mention half a dozen protection functions you are getting for free.
If you follow the full curve above, you will get close to 100% of the rated mAh on discharge. If you charge to 4.2V and then disconnect, you will get about 80% of the rated mAh.
This may seem like underutilization, but limiting the battery charge to 80% of the maximum will get you around 10X the number of charge cycles (if you don't discharge it below 20% too often). A typical battery has a capacity drop of about 10% after ~500 charge/discharge cycles from 0-100%, but about 5000 cycles if you limit it to 20-80%.
You described a UPS, so I'm guessing it'll be topping off for most of its life. In this case, you'll probably want to limit the max charge, because keeping a LiPo 100% charged full time is really just roasting it to death. That's why a lot of laptop batteries are useless after a year or so. Some BIOSes have a setting you can use to limit the battery charge to 80% for this reason.
