On Wed, Oct 17, 2012 at 10:16 PM, Sean Breheny wrote: > OK, but my point is this: what aspects of the sum do you care about? > Do you care if the result is a*x+b*y where x and y are your two > signals and a,b are constants that are not exactly 1? If they are not > exactly one, do you care more about the ratio a/b or simply that the > result has roughly some particular total power and both signals are > present? > > For example: if you have a point-to-point RF link and there are two > transmitters (one at 100MHz and the other at 120MHz) and they share > the same antenna, you don't really care if the two signals experience > exactly the same gain and phase-shift, as long as neither of them is > greatly attenuated or distorted. On the other hand, if you are doing > an experiment where you are feeding two signals into a non-linear > device and you expect to see sum and difference frequencies as well as > higher-order products from the output, then you may care greatly about > the _accuracy_ of the sum and the difference in the phase shift seen > by the two signals. > > Sean > I hit send without realizing I didn't answer the question, sorry. A small fixed phase shift is probably inevitable and that's fine. The input frequencies are fixed and not modulated (essentially) The phase response shouldn't drift though. I can't imagine why it would, can you? The amplitudes should be matched to within 10% or less. Fixed offsets aren't a big deal because it can be accounted for in the test results. The point of adding the signals in time domain is so the load can see two frequencies. There shouldn't be significant difference and higher order components. That would be mixing, right? - Martin K --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .