On 24/12/16 16:00, T i m wrote:
On Sat, 24 Dec 2016 15:42:08 +0200, The Natural Philosopher
wrote:

snip

I think the typical installation orientation of nearly vertical would
not help with any 'finning' that would in turn help the conduction and
the convection that supports it.

If the (typically orientated) element had some longitudinal 'finning',
maybe in the form of a cross, that may increase the surface area
sufficiently and not interfere with the convection too much?


Why bother?

It was a hypothetical question. You may not have understood such? ;-(

Its all expense to make something 100% efficient...100% efficient!

I bet you don't know why people climb to the top of mountains when
there is a perfectly good cable car either.

See, it's not all about efficiency directly but questioning *if* there
were any *other* improvements that could be made that *could* have
other benefits ... like faster recovery-time or reduced scaling.

Cheers, T i m

well the OP was about efficiency.