I have worked with the LECO GCxGC systems for several years - and the configuration shown by LECO is with the press fit in the primary oven. And, this is what I have used with good success. You can put the pressfit between columns into the secondary oven. I've known several people who have done this and believe it gives them some advantave. It is a tight fit and you have to be careful of which manufacturer of press-fit you use. Some press-fits are longer than others. And, the longer press-fits become a bit more of a problem to get into the oven without bending the column too much. (The tighter the bend in the secondary oven, the sooner the column will break!)
I have found that some columns do not seal well to the press-fits and have to be ocassionally reconnected (new cuts on the ends of the columns and a new press-fit) ocassionally to keep going. You can use pollyimide to try to seal the columns into the press fit - but I've even that come apart once or twice. It is a lot more convienient to do the reassembly without having to move things in and out of the secondary oven.
As cold as the cold jets get, I would say that you are mostly freezing things out in the modulator. When the hot jet hits the column, it not all that hot - not more than 30 degrees above the secondary column - and heat transfer is driven by the gradient. So, the question is does the stationary phase film "thaw" out quickly. I've been betting on the thinner stationary phase in the second dimension column. And given that each of the two cooled zones covers about half a centimeter of the column, it may not make much of a difference in the end.
See what gives you good chromatography and push it to the limit. GCxGC is still a new enough field that you can try different things and be no more wrong than the rest of us