Until some eMergy analysis gets down,

Odum liked tacking the 'how much knowledge is embedded in this product' and as such, PV to electrical power will always be expensive eMergy-wise.

If one applies a 'time value' to electric power, the conversion of a photon to electrical power via PV is the shortest.   Wind/Hydro is shorter photon->power conversion than plant->oil/alcohol->power and that is shorter than wood->power or coal/oil-> power.   eMergy has an ability to track time-value of a photon, but I've not seen a model of it.  

And last I checked, nobody had invented electric-powered bulldozers, trenchers, trucks, scoops, etc.

Check again.  In Milwaukee WI, the earthmoving machines made there are electric.

PV makes low voltage DC, usefull only for direct use (or battery storage).  A significant extra step is required to make this power generally useful, power electronics to convert it.

Wind makes medium voltage and hydro high voltage AC electricity, easily useful doing most everything electricity is used for (convert to medium voltage DC for aluminum smelting, Urban Rail (freight rail can use up to 50 kV AC) and limited other applications.

Strung together PV panels produce high DC voltage. On my system five panels in each string give 350V. The inverter cost only 5% of the system cost. On the other hand batteries sufficient to smooth out even a week's supply cost nearly as much as the rest of the supply and last only a few years.
As long as total PV power is only a few percent of grid power and you can get your supplier to allow you to fit inport and export meters and don't put too great a differential on the two prices grid connection is a far better idea.
PV makes low voltage DC, usefull only for direct use (or battery storage).

Do you have a point here?  

A significant extra step is required to make this power generally useful, power electronics to convert it.

Oh, well to apply your own logic, the power in the home should be 12VDC or even 1.5 VDC.   Because some of man's most usful creations is TTL and CMOS logic.   If you plug a IC into 120 VAC, it goes poof!   A significant extra step is needed to convert 120VAC down to 12VDC, 5VDC, 3.3 VDC and other very low  DC voltages.

A cheap and easy step because efficiency is not an issue.

I can feel the heat coming from the converter on my feet right now.  My guess is <$5 wholesale.

OTOH, when efficiency matters, it costs.  Such as PV > Medium voltage AC.

A cheap and easy step because efficiency is not an issue.  I can feel the heat coming from the converter on my feet right now.

What about the heat from the step down transformer on the pole/neighborhood substation?    What about the 20-50% line loss (depending on who you want to quote)?

If you are willing to accept cheap lossy transformers at your feet, why oppose solar PV that MIGHT need a conversion step?

20% line loss ?

WAY over the US average.  The most often quote is 10% transmission & transforming loss.

Since computers do not run on the EXACT DC voltage that PV modules produce and because DC voltage cannot be transformed (i.e changed) without going first to AC (i.e. 12 V DC > 120 V AC > 5 V DC), any PV energy calculation has to include the costs (energy & capital) to convert to useable AC power.

The only exception that I can think of is direct feeding of series wired PV to Urban Rail (600 V DC to 1000 V DC depending upon system).  PV could be used as supplemental power in that way.  Also, wire house for 12 V DC, use 12 V DC Sunfrost frig, 12 V DC light bulbs (including LEDS), 12 V DC small travel TV, car recharging for cell phone and 12 V DC lead acid batteries.  This would be in addition to regular 120 V AC power for other uses.

You have a point about line mounted transformers.  MUCH less efficient than larger units.  More efficient, more expensive, and somewhat larger residential transformers might reduce US electricity consumption by fairly close to 1%.

From Howard T. and Elisabeth Odum's book "Energy Basis for Man and Nature" 2nd edition 1981:
Capture of sunlight by solar collectors, solar voltaic cells, and concave mirrors that converge energy ... use so much embodied energy in their feedback compared to the light they capture that they are not net energy: most of their energy comes from the main economy and thus indirectly from fossil fuels.  As fuel prices rise the solar technological costs rise just as fast.  They will not become relatively more important as fuels run out.
Whether this was true on not in 1981 I don't know but it certainly is not true now as many studies have shown.
This
 is a study of one particular system that shows energy payback in 8 to 10 months.

Despite many very detailed studies that refuted this claim it  still circulates on the Internet. It would be good if another
trail does not originate from TOD  

Yeah, they try the same lies on nuclear power, too. Don't know why. It makes more sense to say that coal has negative energy payback when you consider global warming is going to flood the coastal plains and reduce agricultural, pastoral, and silvicultural production.
Then again, that much high fish productivity coastal waters will dramatically increase piscatorial production;)
And when Odum speaks of costs, this includes the brain-power that was used to make the PV panel.

Thusly under the Odum eMergy cost Model, the lowest cost tools are chipped rocks.

Feel free to show that your quote does NOT include 'brain costs'.