I can make a couple of comments, but may not solve your problem directly... that will take a bit of research on your part. First, it generally isn't a good idea to tie off unused opamp inputs to Vdd. A preferred method is to configure the unused part as a voltage follower (output tied to - input) and tie the + input to ground. Second, though the 662 has rail-to-rail *output* capability, you need to check the data sheet and see if the input range (sometimes called common-mode range) can handle all the way up to Vdd. The behavior you are describing suggests that it can't. Sorry I don't have a databook here to look it up, or I could give a more concise answer. Finally, do you have a bypass capacitor right at the opamp power pins? If this doesn't help, let us know. Brian Aase > Hi all, > > I'd appreciate it if anyone could give me some hints on a strange op-amp > problem I'm having: > > I have an LMC662 dual CMOS rail-to-rail op-amp hooked up so that one of > the two amps is in a non-inverting buffer config and then other has both > inputs tied to Vdd (5v in this case). Vss is 0 volts. > > When I apply a voltage to the non-inverting input,the output follows it > perfectly until I get to 3.7v,at which the output suddently swings up to > 5v. If I decrease the non-inv input,the output stays at 5v until I get > below 3.7v at the input at which time it begins to follow the input again. > I also have a scope hooked up to the output and I detect no oscillation or > other strange behavior throughout the entire input range from 0 to 5v. > > I have also tried two of these chips(they have a different date code),they > behave the same way! > > Am I going crazy,or do I probably have bad chips?! They were shipped in > anti-static foam and I have stored them that way,too. > > Thanks, > > Sean