Hello Sean, Re #2) I am relatively new to the concept, googling shows that the second and third order distortion products are mixed results from a "two tone" test. Would I be correct is assuming the in real life the second tone would typically be harmonic in applied signal? Re: #3) I suppose I presented things in a rather backwards fashion. Now I see I don't have to assume anything about gain to make these calculation. The reason I did there was to make the distortion a function of input power. Like you mentioned, it probably is more conventional to work directly in terms of output power. Re: #5) Thank you, I'll make sure to put things more clearly next time around. Thanks Sean, I appreciate all the advice. On Tue, Mar 22, 2016 at 8:41 AM, Sean Breheny wrote: > Hi Jason, > > A few things: > > 1) The article you reference defines P1dB as the input level which causes > 1dB gain compression. I've never seen it specified that way - it's usuall= y > the output power where 1dB gain compression happens. > 2) No, you can't assume that 2nd order products would be negligible when > 3rd order products are low. It is very likely that 4th, 5th, 6th, etc. > products would be negligible when 3rd order are negligible (in a > well-designed linear amplifier), but 2nd order can still be significant. = I > think the reason why 2nd order isn't spoken about as much as 3rd order is > that 3rd order products include the 2*f1-f2 and 2*f2-f1 products, which c= an > be within the desired frequency band and cannot be filtered out (See the > article you linked to). This only applies, however, when you have a > non-sinusoidal input (so that there are multiple frequencies in the input= ). > 3) Why do you have to assume that the gain is 20dB? Can't you tell what t= he > gain will be if you are designing or choosing the amplifier? > 4) IP3 can be specified as output (OIP3) or input (IIP3) power. > 5) Your formulas seem right (assuming that IP3 is specified as OIP3). The > distortion power one would be easier to understand if you showed the step= s > to derive it. I wasn't sure where 2dBm came from until I re-derived it. > > Sean > > > On Sat, Mar 19, 2016 at 3:57 PM, Jason White < > whitewaterssoftwareinfo@gmail.com> wrote: > > > Hello, I am looking to build a RF amplifier (~900MHz), its output will = go > > directly to a 50ohm antenna. I would like guidance on how to determine > the > > maximum usable (clean, low distortion and harmonic content) output powe= r > > for an RF amp like the RF3374. I don't need to know the exact figure, b= ut > > rather I'd like to make sure I know how to go about finding that value. > > > > In this case, I am assuming the worst case performance of the amplifier > > with no output filtering whatsoever. The RF3374's P1db compression poin= t > is > > specified at around 17dBm. Its worst case IP3 point is specified at > 29dBm. > > > > Would I be correct in assuming that if I reduce the IP3 distortion to > > negligible levels then I could assume that all of the other harmonic > > produced by the amplifier will also be negligible? > > > > I have read [1] which describes what P1DB and IP3 are, and I have > > formulated a guess, but I am not confident that I have got things right= .. > > > > Guess: > > I am going to assume that the amplifier has a gain of 20dB. I also woul= d > > guess that 3db down from the 1dB compression point that things are pret= ty > > linear. To start I am going to choose that point and calculate the IP3 > > distortion at that power. > > > > Below I have put together formulas for output power and IP3 distortion = in > > dBm: > > > > Gain=3D20dBm > > Input power at IP3: 29dBm - 20dBm =3D 9dBm > > Fundamental power in dBm: Out=3D20+input > > Distortion power in dBm: Out_IP3=3D(Input*3)+2dBm > > > > So at 14dBm output power, assuming a 20db gain, the input power is -6dB= m. > > Using the formulas above that places the amplifier's IP3 distortion at > > about -16dBm. If this is right then the distortion would be 30dB down > > relative the signal, and thus is negligible. Is this calculation correc= t? > > > > >From the above, I would think that anywhere 3db (or more) down from th= e > > P1db compression point would be a safe starting point for most amplifie= r > > designs. > > > > Thanks, > > Jason White > > > > References: > > [1] > > > > > http://electronicdesign.com/what-s-difference-between/what-s-difference-b= etween-third-order-intercept-and-1-db-compression-point > > > > -- > > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > > View/change your membership options at > > http://mailman.mit.edu/mailman/listinfo/piclist > > > > > -- > http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist > --=20 Jason White --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .