TennMax was not thrilled with the test results, so I retested the Lasagna and Stock heatsinks using a different method: For this test, I milled a small groove in the base of each heatsink; the concern I had is the base of the Lasagna is thin and cutting a groove might go completely through it – I was able to do it OK.
I epoxied a thermocouple in each heatsink and lapped it flat. This method puts the thermocouple right on top of the GPU. I could not drill a hole in the center as the fans would not allow clearance for the thermocouple wire.
This test shows the Lasagna a tad better, 0.8 C, than the stock heatsink, while the previous test (thermocouple at the back of the card) showed it a tad worse (0.9 C). The conclusion is still valid:
“On cooling alone, the Lasagna does not offer substantially better cooling than the stock heatsink on the Leadtek GeForce 2. However, I can’t possibly test every GPU stock cooler out there, so if your card’s heatsink is puny (give it the “eyeball” test – is it as big as the Leadtek?), then the Lasagna is worth it.”
SUMMARY 4/9: The Lasagna GPU Cooler is a good alternative to large, stock video card fans if you want to cool the RAM chips and a good replacement for puny video fans.
TennMax’s Lasagna coolers are one of my favorites for active Northbridge chip cooling – efficient, compact design and quiet. They have recently fed steroids to them and now they have a large unit for GPU cooling, the Lasagna GF Cooler with Type B push pins (spring loaded). For most of the GeForce 3 and 2 boards, the Lasagna measures 50 mm x 50 mm x 11 mm. The bearing hole at the bottom of the cooler is gone, so thermal grease is OK. The fan also has a third wire for RPM monitoring.
I have been using an Intel Socket 370 retail heatsink on my LeadTek GeForce 2; a better cooler but it wipes out one PCI slot. The Lasagna’s compact design avoids this issue, as this pic shows:
The stock cooler hangs over the RAM chips, so placing RAM sinks on them is difficult. Here’s how the Lasagna looks installed:
To test out how the Lasagna stacks up, I tested three heatsinks: Intel’s retail Socket 370 unit, the stock LeadTek and the Lasagna. I tested each one by placing a thermocouple on the back of the videocard at the center of the GPU. I used watercooling for the CPU so there was no air blowing on the back of the card. I found that running 3DMark’s high detail helicopter benchmark gives me the highest GPU temps, so I looped it to stress the GPU for about 30 minutes before taking temps. I measured ambient air at one inch from the fan’s intake to get a delta.
I also bolted each fan onto the GPU; there are two mounting holes for this, and using 2 small bolts holds the heatsink to the GPU much better than pushpins. For the Intel heatsink, I marked the hole pattern, then drilled and tapped for direct bolting. I used fresh thermal grease on each test and ran three test sessions on three different days.
For sheer cooling, Intel’s retail Socket 370 heatsink is hard to beat, although you lose a PCI slot. The stock heatsink is not bad at all, but it does interfere with mounting RAMsinks. The Lasagna is much more compact, although cooling performance is a tad less than stock. It does, however, allow you to mount RAMsinks.
On cooling alone, the Lasagna does not offer substantially better cooling than the stock heatsink on the Leadtek GeForce 2. However, I can’t possibly test every GPU stock cooler out there, so if your card’s heatsink is puny (give it the “eyeball” test – is it as big as the Leadtek?), then the Lasagna is worth it. I have Lasagnas on two systems cooling BX chipsets and they have performed flawlessly for close to two years now.
If you’re looking for a substantial Northbridge chipset cooler, this model is beefier than the smaller Lasagnas but may need a little tweaking to fit. The hole pattern is just undersized enough that you have to be creative to mount it, but worth the effort.
Thanks to Vincent at TennMax for sending a sample for us to test. I am also testing another Lasagna CPU cooler and will report on that shortly.