IMD Spacing for a Two-Tone test (+/- 1.8MHz ?)

My input signal is a multi-channel/carrier OFDM. Each channel have 5MHz bandwitdh.
Now, I would like to simulate a 2-tone test for its IMD. I plan to measure the 3rd order at say +/-0.2 MHz offsets from the edge (4.8MHz and 5.2MHz).
My computation is a spacing of Fcarrier +/-1.8GHz, in order to create the 3rd order tones.

Is this correct?

I thought the third order products were 2F1 +- F2 and 2F2 +- F1. The nearest inband ones are 2F1-F2 and 2F2 - F1. I can't quite see how you can get 5.2 MHZ from 2 X 1.8 MHz
Frank

Hello Chuckey,

here is my sample scenario. my channel is 5MHz operational BW. normally the specified measurement for shoulder distance (i think i was incorrect when i'm referring to as 3rd order, sorry for that) is 0.2MHz offset from the edges of the BW. Example my carrier is 2GHz centered in a 5MHz BW. I would like then to measure the distortion level at 2GHz-2.5MHz-0.2MHz at the low side and 2GHz+2.5MHz+0.2MHz for the high side. Since in MWO, they only have nth order IMD testings, I have to adjust my input tones such that their 3rd order IMD's are the low side and high side. my overall band inclusive of the side is 5+0.2+0.2=5.4MHz...divide it by 3 gives me 1.8MHz seperation...