Alpha Intel S478 Heatsink - S-PAL8952 Review

Cooling/CPU Cooling by jmke @ 2004-03-16

Alpha has delivered a solid solution for the P4 with their new updated release of the PAL heatsink. A few minor adjustments to the pins (height/form) has brought the performance on par with other high end products while the mounting system could have benefited from an overhaul as the installation is still quite a time consuming process.

  • prev
  • next

Test Setup and Results

Test Setup:

JMke's Test Setup
CPU Intel P4 2.6 "C"
Mainboard Chaintech 9PJL2 Silver
Cooling * ALPHA S-PAL8952
* Zalman CNPS7000-CU
Memory 2 * 256Mb PC3700 OCZ Rev.2
Video nVidia Geforce 1 DDR


I used Arctic Silver 5 during testing. I remounted every heatsink 3 times and took the average results. To achieve maximum "Load" temperatures I launching the application K7burn twice (our P4 has HT enabled, launching the program only once gives 50% load)

Fan / heatsink combo’s used are the following:
  • M81 = Delta AFB0812SH-F00 (80x80x25.4mm) - 4000RPM / 40db(A)
  • M82 = Sanyo Denki 109P0812M701 (80x80x15mm) - 2000RPM / 21db(A)
  • Coolink Cool-VS1 Low = 80x80x25.4mm - 1500RPM / 19db(A)
  • Coolink Cool-VS1 High = 80x80x25.4mm - 4600RPM / 44db(A)
  • Zalman Fan High = custom size - 2400RPM / 25db(A)
  • Zalman Fan Low = custom size - 1350RPM / 20db(A)

    Environment temperature is mentioned where necessary, the test setup was not installed in a closed case, but as you can see, the environment temp is pretty high, as the test setup is located in a warm area of the room.

    Madshrimps (c)


    All idle temps are in the same region but when we stress the CPU we see our winners step forward, the Coolink at its highest settings is able to take the lead at the cost of noise. The real competition is between the different “silent” solutions. The Zalman @ low is bested by the M82 fan Alpha includes!

    To give you a better idea of the noise / heat ratio I’ve made this graph, lower score = better

    Madshrimps (c)


    Combining the fan noise and load temperatures clearly shows you that the extra noise created by the high CFM fans is not really worth it in the end. The Zalman @ High is able to keep a very nice ratio between generated noise and performance, although the difference is clearly audible between 25 and 21db(a)

    Now let’s turn up the heat, overclocking the system by raising the FSB to 240 and the vcore to 1.63v netting us an average heat output of 100 Watt.

    Madshrimps (c)


    Again the Zalman and Alpha give it their best to provide the most silent system without letting the CPU temp go into dangerous regions. The M82+S-PAL8952 seems to be preferred solution for those who can’t stand noise.

    Madshrimps (c)


    The graph doesn’t change much from our results at 68Watt heat output. The lead of the high CFM fans has diminished, the Coolink running at its lowest settings falls back while the M82 is cornered between the Zalman running high/low.

    Overall very good results from the new Alpha, in our earlier P4 heatsink comparison we’ve put the Zalman CNPS7000-Cu against the Swiftech MCX478-V and Thermalright SP-94. From our comparative results obtained here we can conclude that the S-PAL8952 is able to keep up with the competition nicely!

    As we mentioned in the introduction, Alpha has their fans sucking hot air away from the heatsink instead of blowing down through the pins. I’ve flipped the M82 fan to find out if it makes a significant difference

    Madshrimps (c)


    Sticking to Alpha’s guidelines gives you 2°C improvement, this difference will increase when using higher output fans as the efficiently of the “sucking”-method ;-) also increases. Another benefit from this fan setup I noticed is the decrease in temperatures when it comes to motherboard components in the vicinity of the socket. Instead of blowing hot air on the capacitors and chips you are removing it, a welcome bonus.

    Onto the final words ->
    • prev
    • next