Diode Ring VFO Part II: How Much LO into a Diode Ring?

Last week we were trying to determine how much LO injection we really need in the SolderSmoke Direct-Conversion receiver.  The answer seemed to be "enough to turn the diodes in the diode ring on and off."  Ok, but this brought us to the question of how far we should go with this.  Does it make sense to go for more LO signal? If so, why? And how much more?   Todd VE7BPO offered a very thoughtful comment.  He pointed out that for a simple receiver like this, turning the diodes on and off would probably be sufficient.  Sometimes we hear 7 dbm, others say 10 dbm, or even 0 dbm.  But what is the logic that underpins these figures?   Solid State Design for the Radio Amateur (SSDRA) provides the answer on page 120.  See above.  

With a diode ring (or other switching mixer) you want the LO (VFO or PTO) to be the signal that is switching the diodes  You do not want the incoming RF signal to also be strong enough to switch the diodes.  Having the RF do this would result in something of a mess at the output.  

If you have a weak LO signal going into the mixer, it might on peaks reach the level of turning the diodes on.  You will get some mixing action.  But as the SSDRA paragraph indicates, during much of the LO cycle the diodes will not be switched on.  And they won't be firmly turned off either.   A strong RF signal could come in, add to the LO voltage, and switch the diodes.  That would not be good. 

So if you put a strong LO signal in there, on half the cycle that signal will be turning two of the diodes on.  But on the other half of the signal, that same LO signal will bereversed in polarity,  turning those same diodes off.  Hard off.  Definitively off.  It would take one very strong RF signal to overcome the reverse bias signal put on those two diodes by that LO voltage.  That is the advantage of a stronger LO signal.