This page is dedicated
to providing updates and development of my radiator estimator
is it? How does it work?
It is an educational excel file that will help you understand the
importance of heat and fan performance relative to radiator sizing.
As I get more radiators tested, I'll update these sheets to
further help you in selecting the appropriate size radiator.
It's a common misconception that a radiator has a "Capacity" and
that performance is good up to that capacity as if you have a fuel tank
of gas you can expend. Unfortunately to define any sort of
capacity, you have to select a temperature delta. It's common to
think of heat dissipated according to a 10C delta, but it's not always
used 5C or 15C are also used for those numbers. To make matter
even more complicated the performance of the radiator is highly
dependant on the fans selected, it's not uncommon to see 3-5X the
radiator performance difference simply by using different fans.
This pretty much makes trying to understand what size radiator to
use extremely difficult, and why I've developed these helps.
The delta I previously mentioned is more specifically the Water
Out (water coming out of
the radiator) temperature - the (Air In) temperature. This is an
very good measure of how your radiator and fan combination is handling
the heat. There is also no set in stone number for what is
required, I would simply suggest that an "Extreme Performance" level
system would seek to attain 5C or less, and an "Average Performance"
level system would seek 10C or less. I havn't pinpointed when
a system meets air cooling standards, but I would expect a 15C delta
and have assumed that to be "Poor Performance". In the end, this
number simply can be used to estimate your water temperature, and gains
or reductions in water temperature will correlate to the same amount of
gain in processor temperature. For example if you have a 10C
water delta and your processor temperature is 55C under load. If
you make changes to your radiator or fan setup to get down to a 5C
delta, your new processor temperature would be 50C (assuming your
ambient was constant).
The spreadsheet works by using the data I've collected in my radiator
testing and using that data to plot a C/W vs RPM power curve equation
trendline. It's not perfect because I do use two different
fans, but it gives you one curve to pull data from. It then
uses the combination of the entered heat load and RPM to produce a
Water Out - Air In delta number. In addition, I've added a
shroud performance pull down where you can bump performance for the
added use of shrouds. I left this up to the user to decide
because I don't belive there is enough data collected on shroud
performance on different radiators to do any better with that.
I know that I gained around 7% shroud gain using the TFC
shroud in pull with 25mm fans 1350RPM and nearly 15% gain with the 38mm
fans at 2000RPM. In addition I gained nearly 28% for using
two fans in push pull with shrouds. I expect those results
will vary depending on the radiator and fan combinations, but I'll
leave that to you to decide.
This is for educational purposes only! I would not recommend
using these spreadsheets to compare radiators as there are far too many
variables in radiators to know if performance is the same with your
particular fan setup and pumping power. The curves included
in these estimators are based around the Yate Loon D12SM12 (25mm)
"Curved blade" fan for low speed fans, and the Scythe "Ultra Kaze" 3000
(38mm) for the 2000-3000RPM fans. These fans have shown to
have decent static pressure, but any fan that varies from these fans
will utimately perform a little differently per RPM depending on the
particular radiator used and how much pressure that fan has.
Generally 38mm fans will produce more pressure per RPM than a
25mm fan, etc.. but there is no hard and fast rule. But
rather than have no tool, I would accept that using different fans may
yield different results. In addition all of the radiators
were based around a fix 1.5GPM flow rate. Flow rate
differences are generally fairly small between radiators, but they do
exist so there is likely to be some differences. If you are
getting greater than 1.5GPM , you should be getting better performance,
and less than 1.5GPM...less performance, etc. These are
fairly new, so I have very little information on checking accuracy
other than the data I collected to create the curves. With
that said, I'm just going to assume, the accuracy of the results is no
better than 30%.
By clicking the link below you understand:
Excel 97-2000 spreadsheets zipped
- Educational spreadsheet for educational purposes only
- Curve database is built from a fixed 1.5GPM, two
specific fans in pull, no shroud.
- Estimator is possibly in error by as much as 30% or
- Author is not responsible for accuracy or error.
- Use this information at your own risk!
FURTHER DEVELOPMENT! or say THANKS!
If you would like to show your appreciation for this work and
contribute to further developement. Just click the link below
to send me any amount
through paypal. Thanks!!! Martin