Instead of the Cross CHOP, try the Compositing CHOP with the first input being your original channel, the second input your new value channel and the third input should be a channel with values between 0 and 1 which acts as something like a mask to layer the first 2 inputs together.
It sounds though that you might be trying to do something else - can you describe a little bit ore about your approach and maybe we can find a solution where you don’t need to do that.
As well, if you are looking for completely custom CHOP channels, you can use the Animation COMP and set it to Output Full Range. Use the Range page of parameters to set the length, ie. how many samples/frames.