Quote:
|
Originally Posted by imbabawu
may be the best answer is to do it again
so i just did once more
|
To be clear here, the problem with your test is also repeated.
The temperature is still rising.
The graphs are pretty but they are meaningless, simply because it shows that equilibrium has not been reached. Leave the block on for an hour or two under load (CPUburn is good), and then measure the temperatre over a 15min period. That temperature shouldn't vary by much at all, and hence you need not produce a graph by rather just report the average.
Repeat for the other block.
Ensure you follow the mounting directions for any comparison block, and you don't modify it.
Ideally, you'd measure the ambient air temp and the coolant temp and thus "weed out" any variances (someone opening a door etc).
Cathar's comments about temp measurement mean that anything you "measure" will have a wide degree-of-error. Perhaps you can reduce air flow around your sensor by building a "shroud" which stops air flow around the entire socket, or maybe just around the diode?