Should be a good experiment provided you can get the cooling just on the lnbf. Ie I'm not sure where the switches are on these things.
The symptoms to me could be either LNBF or SWITCH, particularly since the missing EVEN channels are the channels that are band-stacked and shifted to another freq after they leave the LNBF.
I don't know anything about the layout of these contraptions, but for a little more info about what is done inside, you might be interested in:
http://forums.satforums.com/SatForum...003.16#6003.16
{you might have to register with satforums to read this}
It explains how the even channels are moved up to higher freqs.
But I guess that downstream of the Dishpro circuitry, it's hard to tell if the problem is the LNBF or the dishpro switches and converters. I'm also not sure if running the cable directly from your receiver to the lnbf in question helps, because according to the above, the receivers apparently only supply one voltage, so the Dishpro circuitry must be responsible for providing the 13/18 polarity voltages, not the receiver (I'm not sure of this, but this is the way I read the above URL). I think that if you have access to an FTA receiver, and can bypass the Dishpro circuitry, and go direct to the LNBF with that, that you could verify that it is indeed the LNBF that is at fault.
Another thought I had, is that I wonder if rather than a temperature effect, that it might possibly be a brownout situation. Ie some places, in the summer with all the air conditioners running, the power line voltage goes down during the day, and comes up again at night. Just a thought. The temperature thing is more likely though. I once had an LNB switch freq by 20 MHz on me after a particularly hot day.