Sapphire HD3870 512MB TOXIC Edition Review

Does Vapor-X Really Work?

Throughout our testing we kept wondering one thing: with Vapor-X technology behind it, does the Sapphire HD3870 Toxic really need to be as loud as it is? This is supposed to be an efficient cooling solution and in theory the fan only has to push enough air for a constant flow over the aluminum fins since the heat is supposed to be dispersed so well. So, we had to put it to the test.

What we did here is took a stock HD3870 with the reference heatsink along with the Toxic and remounted both heatsinks with Arctic Cooling MX-2 thermal compound. We then used Rivatuner to manually adjust the speed of the fans on both coolers so their fans would operate at a constant 25%, 35% and 50%. The 3DMark Batch Render Test was then run for 30 minutes while temperatures were logged with Rivatuner. The stock card was overclocked to the same speeds as the Toxic.


What can we really say? We were so gob smacked by these results that we didn’t run the tests once or twice; we ran it four times with four separate mounts all with nearly identical results. Bear with me while I try to explain what I THINK is happening here.

First of all it must be said that while the Vapor Chamber is obviously a very good cooling solution, the stock HD3870 heatsink is an absolute dog. The issue comes with the fact that all the stock heatsink contains is about two dozen small copper fins that don’t take up the whole height or width of the internal air duct. Thus, as the fan spins, air is allowed to pass beside or over the fin assembly with very little of it working its way between the fins where it is really needed to help disperse the heat they hold. Add to that the fact that the fan is so far away from the copper fins and we have a perfect recipe for disgustingly high core temperatures. However, as fan speed increases, the large fan is able to push more air with greater pressure so more and more moving air makes its way THROUGH the fins above the core. We can see this illustrated quite clearly in the 50% fan speed test where the stock cooler suddenly increases its cooling capability by leaps and bounds due to a mere 15% increase in speed from the pervious test.

Now, when we have a look at the cooling the Vapor-X cooler provides; its performance is leaps and bounds above that of the stock heatsink because of a number of factors. First of all, if you remember our diagrams of the vapor chamber’s inner workings we showed that it disperses the core’s heat over a much broader area than the stock heatsink does. To make use of this, Sapphire has covered the entire top of the vapor chamber in aluminum fins to take advantage of the even heat distribution on the upper cover. This much larger fin surface area in conjunction with its close proximity to the fan means that the fan can work at a lower RPM while maintaining a proper core temperature. Simple, no?

Here we have shown that with the right utility we are not only able to curtail the noise output of the Vapor-X cooler but also keep more than optimal core temperatures. With these lower settings combined with its limited height this card becomes the perfect companion for a HTPC environment.

This all really boils down to one question you all must have on your minds right now: if this heatsink is so good at moving heat away from the core, why in the world did Sapphire figure they needed the Toxic’s fan to spin so fast to keep the core at an optimal temperature? Unfortunately, we have no idea.

Please check the Sapphire HD3870 TOXIC Comment Thread where we have attached the temperature log files for these tests.