For the next stage of the review, I disassembled the RBX and replaced the factory installed number-1 accelerator nozzle with the number-5 accelerator nozzle, in the photo below on the bottom right. It's ideally suited to higher pressure, and it's this nozzle which most closely resembles the cascade design. Will this nozzle's increase on pressure by virtue of its dispersion pattern, and that pattern itself have improved the RBX's cooling ability? This is the question we seek to answer.

The number-1 factory installed nozzle has the widest orifice, and does not isolate individual streams into corresponding heat-voids. Comparatively the number-5's multiple stream, or "spray" design; is more conducive to the array of "heat-voids" situated beneath it, and above the CPU-core. Below I've disassembled the RBX, and placed the nozzles side by side. Their inverted to show where the water will exit (the paper on number-5 protects adhesive);




I foolishly placed the number-5 accelerator nozzle into the block upside down, which left the heat-void impressions, (seen below) on the gasket. I cannot begin to describe my embarrassment when I learned purely by accident in Xtremesystems Forums I'd done this. I'd like to thank Rokk1972 for pointing this out. Were it not for him, an entire review may have slipped through the number-5 nozzle juxtaposed.

Perhaps some day I'll print it regardless, it was here for a short time. The good news is the temps are better with the nozzle seated as intended, shown below;



Prior to the next series of tests, there is one issue which I feel deserves mention. Intel's Thermal Clock Throttling, or CPU-Throttling. Besides the Abit IS7-E and Abit IC7 BIOS', I've yet to see a Canterwood/Springdale BIOS which provides the option to disable this feature indigenous to the Pentium-4. In Asus P4C800E-Deluxe BIOS-1014, there are two primary adjustments pertaining to Thermal Clock Throttling. First; Throttle Clock Slow Ratio, and just beneath there's a feature labelled; System Thermal. When System Thermal is Enabled, your able to set the temp, and Ratio at which Thermal Clock Throttling will be activated, the lowest being 45C/113F. The screenshot below exemplifies the feature's options if Enabled;





Since owning my Asus P4C800E-Deluxe I've kept System Thermal Disabled, however; this is deceptive, as the setting above it; Throttle Slow Clock Ratio, infer's one hasn't the option to totally disable Clock Throttling, but only adjust the Ratio % at which it occurs, as indicated below;





This portion of the review was performed on a much cooler morning, so the ambient temps (Temp-1 or motherboard) will differ from the day prior. As I explained above, the relevant data exists in the differential between Temp-1 (motherboard/case ambient) and Temp-2 (CPU) under Idle, Load, default speed, and overclocked speed. Extrapolating from this data shall emphasis the difference between nozzle's and the cooling prowess innate to the RBX block itself! Being the TTGI case remains open throughout the review, diode readings are not simply limited an isolated ambient case temp, as if the case were enclosed thereby creating it's own environment. When the case is wide open, as it always is in my tests I experience much lower temps overall and much more consistent, such that one doesn't necessarily need to employ the Delta-T principle.

For Load conditions I ran SETI@Home ver.3.8, and Pifast ver.4.1 until the system levelled at its highest temps. I did this either at default speed 3.0GHz for the 3.0C, or overclocked at 3.7GH, 250FSB 1:1 Ratio. Seti@Home can be very CPU/cycle intensive, as I experienced several re-boots when running overclocked, under Load. All screenshots from this point represent the system running with the number-5 accelerator nozzle. I made no changes, leaving the Z-Chipset NB-block as primary feed from the Black Ice Xtreme radiator, using the Hydor L30, and 1/2" tubing throughout.

The first screenshot below exemplifies the number-5 nozzle's cooling prowess, whilst running Idle, at default speeds. Or 3.0C @ 3012MHz (Vcore 1.550V) 200FSB (1:1 Ratio, VDIMM 2.75);





These next measurements are stressing the system running Seti@Home and Pifast, until a temp plateau was reached. Once again the CPU remains at default frequency or 3012MHz (1.550Vcore) 200FSB (1:1 Ratio);




I'd like to take a moment to point out, albeit under full LOAD conditions, or merely overclocked, the temps achieved with the number-1 accelerator nozzle installed, remained above that of the motherboard-thermistor/case-ambeint temp. The first set of benchmarks where the system ran at default speeds; there was a significant difference between the motherboard-thermistor/case-ambient temp, and CPU. The latter being lower by almost 10F, and several centigrade, however; with the number-5 accelerator nozzle installed a consistently lower temp is maintained then motherboard-thermistor/case-ambient.

The next set of measurements represent the CPU overclocked to 3.7GHz (Vcore 1.550V) 250FSB, 1:1 ratio, with no other programs running. The screenshot below represents Idle, overclocked with the number-5 accelerator nozzle;





For the final series of tests, running the RBX with nozzle nuber-5, I had to raise the Vcore to 1.70V, to ensure there were no system crashes. I stressed the system running Seti@Home, and Pifast, until a high temp plateau was reached and maintained. Since CPUCool is slightly inaccurate where CPU-frequency is concerned I ran WCPUID's Real Time Clock Checker to verify the "actual" frequency. Therefore, there are two screenshot's below, the first emphasising this frequency differential, and the second taken a short time later for the sake of thoroughness. Ergo; these final screenshots represent 3.0C overclocked to 3.7GHz, via a 250FSB 1:1 ratio, Vcore 1.70V, VDIMM 2.85, under full Load running Seti and Pifast concurrently;