Let us check that. PV cells (even silicon ones, you are ignoring the much lighter organic ones that I mentioned) are heavy only because of the necessary structurally stiff carrier.
Regardless, the mass of PVs is far greater than the mass of cold fusion generators, for the reasons I gave above. Even if the PVs weighed nothing, the mass (and cost) of the per capita portion of the power distribution network alone would exceed that of cold fusion. That, and other reasons, are why cold fusion will be hundreds of times cheaper than any other source. It is not just that cold fusion energy costs nothing. Wind and solar energy also cost nothing, but cold fusion will be far cheaper than they are, because of the equipment costs, as I explained in my book.
They can be added to roofing tiles and in that form do not greatly increase the mass. Performance does not degrade a great deal after 10-15 years. Here is the data: http://energyinformative.org/lifespan-solar-panels/.
Ah, they have improved since I last read about them. 25 years is on par with many consumer appliances. But, as I said, combustion thermoelectric devices last 50 years, and so will cold fusion devices.
(Commercial HVAC equipment lasts a long time. But it has its limits. My office HVAC dates from the 1970s and it is falling to pieces. The landlord is paying a fortune to replace it. He had to hire giant cranes to lift the new units to the roof. Twice! They are so heavy, and they were installed so badly, the roof is now leaking all over the place, although not on me, thank goodness.)
Currently PV embodied energy is < 10% of payback.
Your source says 2.5 years in the U.K., with a product life of 25 years. That's not very good compared to other conventional sources, such as wind turbines or gas turbines.
That is the embodied energy from manufacturing. I wonder how much energy is needed to recycle them. It probably depends on the type.
Rating energy production devices by payback is silly. That would be equivalent to saying that an LENR device that generated 10X the electricity out it required to run it (or, say 40X the heat out) was somehow much worse than one with a higher out/in ratio.
Rating cold fusion devices by payback is silly, because they will reduce the cost of energy to zero, and because their embodied energy will be hundreds of times smaller than conventional generators. For other devices it is not silly. If the payback for PV is 10% of lifetime production, and for wind it is 1% (approximately true), this should be taken into account when comparing lifetime performance. You have to reduce the PV estimate by 9% to make a fair comparison to wind. 9% is significant.