TR si-120 vs Scythe Ninja

stukovx · Post by **stukovx** » Sun Mar 05, 2006 8:09 pm

I have both these heat sinks and paired them up with a nexus at 7v. To me, the noise levels are about the same for each one. Performance wise, I would say the ninja beats out the SI by about 1-2c max.

The only reason I want to switch back to the SI is because of the size and the ability for it to cool the extremely hot mofsets on my dfi nf4 ultra-d.

I just wanted some opinions from ppl who have had both on a ultra-d.

Felger Carbon · Post by **Felger Carbon** » Mon Mar 06, 2006 3:21 am

stukovx wrote:The only reason I want to switch back to the SI is because of the size and the ability for it to cool the extremely hot mofsets on my dfi nf4 ultra-d.

Just out of curiosity, what CPU are you running? And how do you know that your mosfets are extremely hot?

Post by **Ralf Hutter** » Mon Mar 06, 2006 7:40 am

stukovx wrote:
The only reason I want to switch back to the SI is because of the size and the ability for it to cool the extremely hot mofsets on my dfi nf4 ultra-d.

I just wanted some opinions from ppl who have had both on a ultra-d.

I agree 100%, and I think this is relevant to nearly all motherboards.

qviri · Post by **qviri** » Mon Mar 06, 2006 7:52 am

The authoritative guide to tower vs. blow-down heatsinks:

Tower heatsinks: work better with front-to-back case airflow. May be able to be cooled "passively" with nearby case and/or bottom-intake PSU fans.

Blow-down heatsinks: cool the motherboard better. Do not work with traditional front-to-back case airflow as well as the tower heatsinks, but are able to take advantage of a Chassis Air Guide (which may or may not let fan noise out).

Both: if improperly designed and/or copied without design considerations *cough*, may apply excessive torque to vertically-mounted motherboards. Are able to cool modern CPUs relatively quietly.

Pick your own priorities.

stukovx · Post by **stukovx** » Mon Mar 06, 2006 12:24 pm

thanks for the replies, I think I'll go ahead and use the si-120. And for the question about how I know my mofsets are hot, SmartGuardian reported 50C+ while gaming and caused freeze and reboot. I put my hand around the mofset area and it was like a heatwave.

jmke · Post by **jmke** » Tue Mar 07, 2006 4:09 am

In my tests with DFI NF3 board, the PWM area were about the same with both SI-120 and Ninja.. the Ninja has 2-3°C lead in CPU temps.

the SI-120 is blowing air on the motherboard.. true, but that's not really "cool" air either, that air is above the temp of the rest of case as its warmed up the by the CPU.

Name - PWM / CPU

Ninja: 65/59°C
SI-120: 66/62°C

JJ · Post by JJ » Tue Mar 07, 2006 6:46 pm

Sorry about hijacking this threa, but what, if any are the advantages of the SI-120 over the XP-120? Was the reason for the SI-120 only to give added clearance for objects on the motherboard, or were there other changes/advantages to the design?

jmke · Post by **jmke** » Wed Mar 08, 2006 12:20 am

the XP-120 trails the Ninja -- the SI-120 is practically on par. This is tested inside a case with 1x120mm fan in the rear @5V and a heavily overclocked/overvolted A64.

SI-120 mainly for higher compatibility, some reviews have it on par with XP-120, but most of these tests were conducted outside a case, not really "real life" scenario.

Laz Winterz · Post by **Laz Winterz** » Wed Mar 08, 2006 1:43 am

http://www.madshrimps.be/?action=getart ... rticID=406

Barely in favor of the si-120 in a case. Was also discussed here http://forums.silentpcreview.com/viewto ... highlight=

jmke · Post by **jmke** » Wed Mar 08, 2006 1:49 am

3-5°C difference is no longer "barely" in my book

if you know that all the high end hsf are within 1-2°C of each other

Laz Winterz · Post by **Laz Winterz** » Wed Mar 08, 2006 1:15 pm

It's just 1 degree difference at full load which is what I measure, not idle.

The testing was done on the same system unless I misread it since I was just skimming.

jmke · Post by **jmke** » Wed Mar 08, 2006 1:30 pm

different systems, different outcomes.

Post by **MikeC** » Wed Mar 08, 2006 4:19 pm

jmke wrote:the SI-120 is blowing air on the motherboard.. true, but that's not really "cool" air either, that air is above the temp of the rest of case as its warmed up the by the CPU.

The airflow still helps the VRM cooling, guaranteed. More than no/little airflow from cross-flow HSFs like the Ninja. With lower power CPUs, there's not much gain, but as you go up in CPU power, it becomes significant. I measured >20W increase in AC power draw on a Pentium 670 system at full draw as the VRMs were allowed to heat up. I've been promising an article on this.... and it will come.

patord · Post by **patord** » Wed Mar 08, 2006 7:43 pm

On a Gigabyte nForce4 system I have, I tried both the Ninja and a SI-120 using a 120mm YL spun down to 7v. I ran CPUBurn for 1 hr with each heatsink on a A64 Venice @1.4v. With ambient temp +/- .5c throughout the testing, I used an IR thermometer aimed directly on top of several of the MOSFETs that comprise the VRM on this board. The additional airflow (even if it was just 7v) from the SI-120, made a enough of a difference for me to standardize on the SI-120 for my system deployments and relegated my sole Ninja to my test bench. I don't remember the exact numbers, but let's just say it is worth the extra $ for the SI-120 vs the Ninja (@ jab-tech.com).

The MB was mounted vertically in a test case of mine (Antec super lan boy) with a rear 120mm YL @ 7v and a Seasonic ST400. This was done a while back, and I'm sure this probably is not a truely controlled experiment like MikeC would put things through. So take my findings with a grain of salt. Also, I would have used a thermocoupler on my fluke to do the temp readings... but I blew a fuse on it during that time and it was out of action. So the question would be, can the IR thermometer be trusted? I'd say somewhat, since I calibrated it sometime ago to my fluke and the fact IR measurements are subject to other external factors just like anything else.

diver · Post by **diver** » Wed Mar 08, 2006 8:03 pm

One should also consider the effect of air from a "blow-down" cooler on cooling a video card wih a passive heastsink where the passive heatsink is on the back of the video card.

jmke · Post by **jmke** » Thu Mar 09, 2006 12:11 am

MikeC wrote:
jmke wrote:the SI-120 is blowing air on the motherboard.. true, but that's not really "cool" air either, that air is above the temp of the rest of case as its warmed up the by the CPU.
The airflow still helps the VRM cooling, guaranteed. More than no/little airflow from cross-flow HSFs like the Ninja. With lower power CPUs, there's not much gain, but as you go up in CPU power, it becomes significant. I measured >20W increase in AC power draw on a Pentium 670 system at full draw as the VRMs were allowed to heat up. I've been promising an article on this.... and it will come.

it will largely depend on the motherboard layout; the DFI NF3 has the PWM area near the rear exhaust, having a rear 120mm fan in the case (even at 5v) is enough to keep that area cool, even with Cross-flow heatsinks. The SI-120 does not improve performance in this case.

there are 1001 HW combinations and it's impossible to test them all; if you do decide to test HSF vs PWM cooling, don't do it on an open bench; it's as far away from a real-life situation as can be.

TR si-120 vs Scythe Ninja

TR si-120 vs Scythe Ninja

Re: TR si-120 vs Scythe Ninja

Re: TR si-120 vs Scythe Ninja