Why is it that ACMA only worry about odd-order IMD when measuring? 3rd & 5th in particular? Why not the evens?
2nd order IMD products (eg. f1+f2) will occur well away from your desired signal, and are easily knocked out by your bandpass filter. However, 3rd order IMD products (eg. 2*f1-f2) will fall very close to your desired signal and may fall within your bandpass filter in your amplifier, spewing rubbish out onto the band you are operating on. CW contesters in particular are very fussy about 3rd order IMD specs in their receiver, because this can produce “ghost signals” that can make it hard to copy a weak DX station.
Before computers became passé, it was necessary to workout IMs with a slide rule – and this was a whole lot of work! (eg to calculate high order IMs on a multi-transmitter site could easily require 10,000s of calculations). So the ACMA (I guess to save work) have historically restricted the necessary IM calculations to the 3rd and 5th Order, because:
- As I previously mentioned, these are the ones that create the majority of “issues” as the higher order are at much lower power level.
- Most (and it once was mandated that *all* transmitters) must be sitting behind some form of cavity filter/combiner. These offer huge attenuation to even-order harmonic products, but much less for odd-order
- As Matt said, the even-order IMs fall into a different band. The ACMA appear to have arbitrarily decided to assume that there wouldn’t be users in “that other band” (without even requiring that to be verified or not).
So, if you look at the ACMA’s RALI LM8 (this is ACMA’s “bible” for this licencing process for Land Mobile), you will see that the ACMA have cull limits to frequency constraints associated with intermodulation analysis. Ie They are limiting intermodulation analysis to frequencies within the same band.
So I suggest that the “answer” is due to historical ACMA reasons to limit their work to a reasonable effort.
So (in real life) are there ever problems on a site with second-order products? The answer is yes (occasionally), but that is left to the users to sort it out!
And even harmonics are cancelled in push-pull amps. Not sure if that has anything to do with the decision to ignore 🙂