On 11-12-08 03:16 PM, Jed Rothwell wrote:
Putting a heat source ~4" away on a copper pipe would bring it much
closer than Rossi's arrangement, because the heat exchanger design
would not be good if the heat conducted to the cold end on the outside
of the pipes. The fact that heat exchangers work well -- they exchange
heat efficiently -- means there is not much heat conducted by the
metal surfaces of the pipes from the hot end to the cold end. If there
was significant amount of heat conducted by that path, it would not be
"exchanged" (that is, it would not heat up the cold fluid). It would
be lost to the surroundings.
You may have missed the point.
It's a counter flow heat exchanger (as they typically are) which means
the EFFLUENT from the secondary circuit in the heat exchanger (which is
the secondary "hot" side) is immediately adjacent to the INLET for the
primary circuit (which is the primary "hot" side). In fact, the *goal*
of the heat exchanger is to conduct heat from the primary to the
secondary pipes, as rapidly and completely as possible. Consequently,
the primary inlet and the secondary outlet are placed in extremely
intimate contact as soon as they enter the heat exchanger. (When most
normal people imagine a heat exchanger they think of a device where the
two flows are going in the same direction, but that's actually a far
less effective design than the counterflow scheme which is used in
practice.)
The issue is that, assuming the exchange of heat isn't perfect, the
secondary outlet may actually have been substantially cooler than the
primary inlet, in which case heat traveling through the surfaces of the
pipes (and, possibly, other parasitic paths) may have caused the
thermocouple to read some temperature between the value for the
secondary effluent and the primary inlet, which would give an inflated
value for the secondary effluent reading. This can happen, once again,
because the two flows are necessarily adjacent at that point, due to the
design of the heat exchanger.
In fact, heat leaking to the *cold* side, as you suggested, would tend
to produce a lower overall power measurement, because the temperature
increase across the exchanger would be reduced. It's also far less
likely, because the cold and hot sides are typically separated by the
full length of the exchanger.
(These comments relate to any use of a counter-flow heat exchanger with
thermocouples used to determine the heat gain across the secondary
circuit. How the power output of the E-cat was actually calculated is
something else again; from first principles, it seems like it should be
possible to do that more accurately by looking at the heat gain across
the E-cat itself and ignoring the heat exchanger.)
