I'ld like to comment on the rad testing discussion a few pages back (sorry, been a bit distracted at home and just catching up)
Heat dissipation that a rad is capable of can be a rather tricky thing to determine, and as I have discovered requires quite a specific testing environment.
Using a running system for the heat/power load just isn't doable because so much heat is lost into the air from the components that the actual load on the rad is unknown.
- a custom test environment where the heat source can be controlled - an aquarium heater is the tool for this. (safety must be bypassed though for a constant load)
- everything must be fully insulated to avoid heat loss into the air, so that a known power/heat load is entering the rad.
- multiple temp sensors on rad ports and air inlet (ambient) to get good averages to base the data on.
Rad testing is definitely not for everyone as it can require a lot of patience, redoing tests is not uncommon to get truly accurate data.
Usually the hardest variable to control is a steady ambient temp over the test period - I've settled for 0.5°C variance over a 15 minute test period (after equalization)
Coolant temp equalization can be frustratingly difficult to achieve, with multiple temp sensors that detect small changes, either from power input, or ambient air.
Another weird one to overcome is the mains power variance - even when I control the aquarium heater with a Variac, the mains supply often throws my test out with different Power levels at different temperatures and times of the day.
So have to keep an eye on this for constant/ish load to the coolant from the heater.
3 Watts (1%) power difference will change the coolant temps - blowing averages for the test, and having to restart.
Likewise the ambient air temps greatly affect the coolant temp readings (obviously) - even walking in front of the air intake of the test chamber gives small spikes in air temp.
It is not very tester friendly - rad thermal testing !!!
A formula for heat dissipation can be used to calculate the power being removed from the coolant by the rad (in watts)
This info in itself is still a specific figure for that particular test set environment for the given test parameters (ambient temp, flow rate, fan rpm, CFM ect)
I find it useful myself so will be including it in my results when I post them.
But a watts dissipated by itself without the parameters is useless.
For example you can have 2 tests with different fan speeds - high and low (high & low CFM) that have very similar Watts Dissipated results - because the power/heat input is the same,
But of course the DeltaT number is hugely different with the 2 tests, as the low speed fans won't be able to get the coolant anywhere near as low as the high speed fans - up to 10°C difference I've seen recently.
Solution is of course that we combine the DeltaT with the known and controlled power/heat input gives us a thermal test result which is much more comparable result.
So the piece of data that matters the most IMHO is the W/10C - being the Watts Dissipated per 10°C DeltaT
or even a W/1C - Watts Dissipated per 1°C DeltaT.