Hybrid Car – More Fun with Less Gas

Re: GFX vs home brew - Page 5

register ::  Login Password  :: Lost Password?
Posted by Solar Flare on April 14, 2006, 10:15 pm
But you also claim you don't need to breathe oxygen for the full 10
minute shower so that you can seal your shower stall.

Posted by daestrom on April 15, 2006, 4:41 pm

Guess again.  If your setup was restricted to just 5 feet long, would its
performance be anywhere near as good as the GFX???  And that's the point.
To get performance on par with GFX, you have to resort to something several
tens of feet long.

Heck, If I had someone build a GFX that was 100 feet tall, I'm sure it's
performance would put your setup to shame.  But who has space for 100' of 4"
pipe (vertical, coiled or otherwise).

<snip ASHRAE calculations for steady-state problem>

Well, *you* can calculate using average flow, but the results are *NOT*
meaningful.  Just because you found a formula in a book, doesn't mean you
can apply it to different situations, like intermittent and 'average' flow
and still get meaningful results.  Those ASHRAE formula are for calculating
the steady-state performance of a heat-exchanger.  Trying to apply them to
'burst' mode is a waste of time.  The results do *not* mean anything.  And
they don't prove anything except that you don't know when to apply them.

But just to humor you, if the 'bursts' are 1.25 gpm, then the steady-state
answer would be Cmin-Cmax=1.25*60*8.33 = 624.75 Btu/h-F.  With an area of
78.4 ft^2 and U Btu/h-f-ft^2, NTUx.5*10/624.75 = 1.26 and EU.7%.

Notice how the answer depends on the flow rate *when water is flowing*??
Not the average amount of water that flows during some arbitrary time

The fact that the two different flow rates give such drasticly different
answers should be a clue that you're missing something.

The formulae you are using from ASHRAE are for steady-state, *flowing*
heat-exchangers.  The NTU and effectiveness assume *steady-state* conditions
(i.e. a constant flow rate).  So the efficiency of your system when water
flows and has reached steady-state is only 0.45.  But since your showers are
less than the time needed to reach steady-state, even that number is

Your other post with a step-wise simulation is probably much closer for this
sort of transient behavior, but it too has some flaws.  You posted the
outlet temperature for the greywater as 72F while the outlet for freshwater
as 94F.  This is with constant 55F inlet freshwater and 100F inlet
greywater.  The fact that your freshwater is picking up more energy
[(94-55)*flowrate] than your greywater is losing [(100-72)*flowrate] is a
clue that something is wrong in your calculation.

Your simulation printed out the numbers after 350 minute stagnation period,
not when there is flowing water.  You should print out the temperatures
*during* the last shower, when there is actual flow.  *That* is when there
is energy flowing down the drain.  Print out the numbers for fresh and grey
water outlet temperatures *during* the last ten minute shower.

Calculate the energy removed from the greywater during those ten minutes and
the energy being picked up by the fresh-water during those same ten minutes.
Since the inlet temperatures are both assumed fixed (100F and 55F), if the
energy picked up by fresh-water does not equal the energy given off by the
greywater during those ten minutes of flow, then something is wrong with
your calculations because energy must be conserved. (we're neglecting any
ambient losses)

To find the true effectiveness for this non-steady-state operation, just
calculate the amount of energy picked by the freshwater during the shower
and divide by the total energy to heat that same water to the greywater
inlet temperature.

*hint*, if the water outlet temperatures change a lot while the shower is
running, you might reduce the time step to less than one minute intervals so
as to get better resolution.  This would make for better integration of the
temperature versus time to get total energy.  Too course a time step could
lead to mismatch between greywater and freshwater energy calculations.


Posted by Robert Gammon on April 15, 2006, 5:08 pm
 daestrom wrote:

The issue with a 100 foot tall GFX is the required size of the potable
water tubing and the pressure required to get a reasonable flow rate
thru the 100ft stack.  

As it is on a 60 inch GFX, manifolding is performed to limit pressure
loss (coil height is about 27 inches each) with the base of each coil
tied to the inlet water, and the top of each coil tied to the outlet.

The engineering drawings on gfxtech's web site clearing indicate an
asymptotic behavior.  Adding additional length brings lower and lower
incremental benefit.  Still with two S4-40s in series, pressure loss is
about 2.5psi on a 2 gal/hr flow rate.  and 80 inches of gfx recovery
will get efficiency up another 5-10% over a 60 inch model and a 40inch
height is easier, in many cases, to find a spot for.

daestrom is doing us a great service by pointing out the issues with the
home brew system.  I too do not believe that the home brew system
proposed will work as well as a 60 inch GFX to recover waste heat from
grey/black water and pump that heat to DHW and cold side showers.

Posted by daestrom on April 15, 2006, 5:45 pm

I wasn't seriously recommending a 100' GFX.  Just pointing out that 100' of
any sort of piping takes considerably more space than a 5' GFX.

Yes, that is exactly how mine is constructed.  The problem with piping the
freshwater side in parallel is that the two coils form a sort of
series-parallel flow heat exchanger.  One heat exchanger cannot heat its
outlet as much because it only receives already-cooled greywater from the
other heat-exchanger.  So when it's cooler freshwater outlet water mixes
with the warmer water from the upper one, there is a reduction in overall
efficiency.  I can detect this when someone is in the shower by touch alone
on the two coil outlets.

This is the same sort of thing that multiple-pass conventional
heat-exchangers suffer from.  Less than ideal, but a compromise of
heat-transfer performance versus hydraulic  performance (pressure drop).

I've toyed with the idea of restricting the flow through the lower coil to
improve on this.  Would increase the pressure drop some, but not as bad as
the full series model.  Some weekend project I may put a throttle valve in
series with the lower coil and play around with different settings.

True, but you would need two 40inch heights, or a pumping arrangement.
Since mine is installed in the main waste line for the entire house, pumping
blackwater did not seem very attractive.

It could perform rather well.  And if I know Nick at all from his postings
over the years, it will cost less than my GFX did, even though I installed
in myself.  I'm just trying to keep Nick 'honest' by not letting him apply
steady-state calculations to a transient system.  But it does require more
space to install, and may have some maintenance issues.

Hopefully when Nick is done building it, he'll post his performance numbers
(good or bad).  Direct measurements and experimentation are always better
than theory.

"In theory, theory and practice are the same.  In practice, they're

Posted by Solar Flare on April 15, 2006, 7:30 pm
               GFX      nick
efficiency     6        7
price          5        3
convenience    9        1
wife likes     6        0
total score   26       11

for hot

ft^2, 60%


would its

the point.


sure it's

100' of


rate thru

100' of

pressure loss

tied to

piping the

from the


touch alone

coil to

valve in

loss is

recovery will



with the





require more


This Thread
Bookmark this thread:
  • Subject
  • Author
  • Date
---> Re: GFX vs home brew AstickfortheMUL...04-14-2006
please rate this thread