Interesting and good review. SPCR's reviews are always "good", but not always "interesting", but this review shows a new leader (maybe).

However would it be possible to add an extra page on the review which compares Ninja PASSIVE vs Ultra-120 PASSIVE performance?

It's not conclusive which is best for passive use. However it is clear that the Ultra-120 is slightly better than Ninja with airflow.

That's not really possible. If you have read our review of the Ninja, you'll know that the Ninja was not capable of cooling our test processor passively on an open bench. I would not expect the Ultra-120 to be any different.

Passive heatsinks rely on system airflow for their effectiveness, so the "passive" performance of the two heatsinks will depend on how much system airflow there is. As a general rule, system airflow across the heatsink will tend to be lower than any amount of forced air, so if you want to get an idea of how they will perform passively, look at how the difference in performance changes as airflow drops. It's not hard to extrapolate that the Ninja would have a slight advantage if there was still another test point with even less airflow than a Nexus generates at 5V. For this reason, the Ninja would be my choice for a passive system.

However, because passive performance depends on system airflow, it is not possible to definitively state whether the Ninja or the Ultra-120 is better in a passive situation — it's system dependent.

Can the fan be mounted on the other side? If so, if the heatsink were also turned around, would the bend in the fins be enough so the fan air would blow onto the VRM's? Would this cool the VRM's enough?

This won't work. If you rotate the heatsink 180°, then you are back to the same shape. This is because the fins are not mirror symmetrical along their axis, but point symmetrical over their centre point.

If this sounds like hocus pocus, then consider the following: If you look at the bent fin tips through the fan, then the tips are bent up on the left and down on the right. This is on the side where the fan is. When the air travels through the radiator to the other side, then it encounters bent down fin tips on the left and bent up tips on the right.

So half of the air that is deflected from the straight path, is always deflected downward. But I don't think the angle is great enough for that air to reach your VRM's.
_____
"Typo" patrol: I don't think "K8 systems restricted to one orientation" should be listed as a Pro in the table at the conclusion.

Unfortunately, the S-type heatsink clip is only for HR-01, Ultra-90(A), SI-97A, and SI-128. The Ultra-120 is probably too heavy.

How is the weight distribution? I remember it was mentioned that the Ninja was bottom heavy, thus better for the motherboard. Not the case for this heatsink? I may have to stick with the Ninja if it will make moving safer.

From the article:

why not test inside a case and get some real-world results:)
with a single 120mm fan @ 5v in the rear, it should be doable to cool passive

yeah, testing simularly to how you tested the ninja in a P180?

http://www.silentpcreview.com/article251-page5.html

I prefer the Ninja for a tower heatsink design because of the square design...which means equal cooling ability no matter how the CPU socket is orientated on the board. With this Thermalright you must first determine which way you want the air to flow, then buy the correct MB.Big hassle.....

I dont normaly spot spelling mistakes but i found one, yay to me



I only read conclusion, in work, looks like a very good performer.

Jozi

EDIT: 2nd page, below 3rd picture

i agree with DanW (and all before), since most people who have ninjas seem to run it semi-passively. judging from the results, the ninja is likely to perform better, but thats not conclusive, and by how much is likely to be important to some people; as the thermalright is clearly better at high air flow (ie, better for heavy OCing).. so weather the difference between semi-passive is significant or not may be a decision factor.

Devon, under the most critical listening conditions, is there any audible difference between the Ninja and Ultra-120 with the Nexus fan? Between the fin density and the winglets, there might be a very subtle change in the signature. While this would not be important in any real-world system, I'm most curious to hear if you noticed anything.

I'm not speaking for Devon, as he may have heard something, but I can't say I did. For me, the Nexus 120 is still a bit too loud at 12V, but it really doesn't spin that fast, and at this airflow level, there did not seem to be any appreciable difference in noise mounted on the Ninja or the U-120.

Thanks, Mike. I'm sure Devon would have included a note in the article if he'd heard something, but I just wanted to be clear. While my experience with fan grilles implies that heatsinks can affect the fan noise, it's probably too minor an effect for us to worry about.

As the reviewer noted in the comparison between the ninja and the U-120, the difference between the two was nonexistant at 5V. However at lower airflow (something that seems possible in the SPCR crowd, if using a duct where it only receives a portion of the airflow from a 5V nexus, or a hypothetical fan that blows even less) there isn't much to go by.

Maybe a possible idea to look into for future reviews is a column with lower airflow or just passivly cooled (if so you'd need a seperate test system, possibly a resistor coupled to a heatspreader to simulate the CPU, that way you don't need to worry about overheating the CPU), and with that data one would be able to extrapolate probable data for lower airflow situations.

Its nice to see thermalright stepping up to work on better products.

As good as the performance is, the thing that would make me still choose a ninja over it is the physical deminsions and weight. Its alot bigger in every aspect, and i personally dont feel all that add'd "Stuff" just doesnt compensate for the slight cooling performance increase at higher airflow. The biggest drawback for me would be its hight, and then weight. People were complaining about barely able to fit a ninja inside their case, and this thing is another 11mm taller.

What i would be more interested in would be a mini-ninja. Something around 50-100mm tall with all the other measurements the same, with the same fin spacing. Sure it wouldnt perform as well, but it would be very viable for passivly cooling processors in SFF low airflow systems.

I'm not really sure that you can say that the ninja is better at passively cooling, as this would assume that the cooling performance is strictly linear in relation to fan speed, which may not hold true.

Good review, someone has finally made heatsink that beats Ninja, I thought this day would never come

In any case, like others said I would still choose Ninja over U120 for three reasons, one, Ninja plus is $10 cheaper than U120 here in US, two, Ninja is smaller (it's 100 grams lighter, 1cm lower and 12-22mm narrower depending if you count in the fan or not), and three, at extra low aiflow Ninja is just as good as U120. Depending on the ambient temperature I can run my nexus fans as low as 5-6V, which is very very low airflow (although in the summer heat I turn them to 12V). And fourth, unofficial reason, I just like my Ninja, it's a macho heatsink that looks and performs right

I'd like to second Aris's desire for a mini-Ninja. From my cursory researching, it looks like many of the SFF cube cases have a power supply just where it is worst for CPU heatsink vertical clearance.

Perhaps something with a horizontal fan (XP120-like) but widely spaced fins (for low airflow), with a fan clip. You can attach a fan, or leave it off and put a bottom-fanned PSU right above it.

I had e-mailed Hank at Thermalright last week regarding using the S-type clip (to allow the heatsink to rotate 90 degrees) with the Ultra-120. Here is what he wrote:

The S-type HS clip does indeed work on the Ultra-120 as well. They should be available in most major on-line stores by now.

If you check the Thermalright website they now show this:
S-type heatsink clip- allowing heatsink to rotate 90 degrees on the CPU socket.(For HR-01, Ultra-90, Ultra-120, SI-97A, and SI-128)

Just thought I'd mention that clicking on this link in the recommended heatsinks list takes me to the old Ninja review, not the Ultra...

Thanks for the heads up. Fixed it -- and a few other trifles... on the Aug 21/06 revision of this reference piece.

MikeC, would it be possible for you to correct the Thermalright Ultra-120 review, as well as the synopsis on the Recommended Heatsinks page As I posted on August 13, the Ultra-120 heatsink can be rotated 90 degrees on K8 systems by using the optional S-type heatsink clip. This does require you to mount the heatsink using the stock retention bracket on the mb, rather than through the motherboard.

Thanks for your concern 1111111;

We have been aware of this for some time, and have already updated the article once to mention that the clip is available. However, because the clip is not included, we stand by our original comment; you have to go out of your way and spend more money for compatibility.