I have recently reviewed the Seasonic Super Tornado 350w and 400w version (SS-350FB and SS-400FB).

It looks to me like these PSU are not designed for silence. As the packaging says, they are "for gamers and overclockers", and one thing is for sure : while reviewing them I didn't feel like I was the "target audience" for this product.

Test : Seasonic Super Tornado (in french).
Google translated review.

To sum up :
- great efficiency, Dr Cable, steady rails, good overall quality, tiny heatsinks
- the fan is a Yate Loon D12SM-12, same fan as in the Fortron PSUs. Not the same fan as in the 300 Super tornado reviewed here at SPCR. The fan starts very reliably
- fan controller is the strangest fan controller I have ever encountered. The PSU is very quiet until the case reaches a normal operating temperature (~32°c), and then its as loud as the 120mm Fortron. From then, the noise changes with the load (which is what many have experienced here). It takes less than five seconds for the fan to change its speed when the load changes, and its pretty annoying to have the noise level change all the time.

The setup I used for my test had little airflow, but the quiet PSUs I have tested with this setup where able to keep their fan at its lowest speed. So it's not a case cooling issue, and it looks to me like this PSU is designed for high airflow, not silence. That's what it says on the box anyway.

I wonder if the unit Mike has reviewed is representative of the units that are actually available on the market. I didn't follow the "Tornado threads" very closely, as I have had much to do these last days, but I also wonder how many people have bought this PSU and were pleased with it. Maybe the Super Tornados sold in europe are different (at least the fan is different, so there could very well be something else, like the fan controller). Or maybe the 300w unit is very different from the 350w and 400w units.

Mike, what do you think about the availability of the PSU you tested?

Could you please state what are the "quiet" PSU's you have tested?

Seasonic SS-300FS, modded Seasonic SS-300FS (rubber mounted L1A and a few airflow tweaks), Seasonic Super Silencer 460w, Tagan TG380-E00, Tagan TG380-U01, and others that don't strictly qualify as quiet : Fortron FSP300-60PN (PF), Nexus NX-3500, L1A modded AOpen FSP250-60GTV (PF) (those couldn't keep their fan at it's lowest speed even in a fridge )

I have also been running higher power setups with low airflow PSUs. The Tagans, Seasonic Super Silencer and SS-xxxFS all do a good job at staying quiet under load. In my tests, the Super Tornado didn't.

Well was I p***ed! Just got home with a Super Tornado 350W in hand. Was waiting for a suitable moment to shutdown, and read Kostik's post & review which did NOT bode well. So a quick change of plan: popped the cover (voided the warranty - you can keep the card, Supastar ) and sure enough, 1 Yate Loon fan inside.

With all my elation evaporated, proceeded to instal (which is a complete pain in my case because the psu mounting screws are 'upside down' on the Seasonic). The good news is, so far(6 hours Folding), I haven't heard the psu fan change speed. The psu fan is quieter than my cpu heatsink that has a Papst 3412NGL @ 9V. The 3412 @ 6V is about equal to the psu fan noise - the psu fan has more air whoosh and the CPU has more growl/motor noise (all sound measurements taken by uncalibrated ear about 30cm from rear of case.)
Probably the major difference between my setup and you: my case inlet temp is 18.5C as I write, about 3cm from the cpu fan intake it is 28.7C (using an indoor/outdoor thermometer converted to 2 * outdoor!). At the moment I have unplugged my front intake fan - the Seasonic is my only case fannage. I may experiment with an 80mm 'flo extracting at the rear later, more likely put the 4412FGL back at 5V. Also my overall W is lower than you: a while back I measured 120W: my config hasn't changed that much since then and with a higher efficiency psu I guess I am somewhere 100-110W AC now.
Well so far I have cheered up a bit. But am well cheesed Seasonic seems so unpredictable: they thoroughly deserve the de-rating Mike gave them in the recommended PSU list. I really wanted the 300W Super Tornado but that isn't available in Europe (Yet/this week/for some marketing reason/Sadam Hussein ...): the 300FT I already have 1, but has shorter DC cables. The fan rpm cable on the Tornado is a waste: so far the fan hasn't registered with MBM.

BTW Kostik - the molex count in your review is (presumably) for the 400W? My 350W has only 7 for hdd - 2 * (2 + floppy) and 1 * 3 (and (English ) manual says 7 on 350W, 8 on 400W )

Yes, I have noticed that this PSU is really of the on/off kind. If you don't go over a certain treshold of temperature/power, it will be *very* quiet, but once you've reached this treshold, it's just too loud. And it looks to me like this treshold is too low for the Super Tornado to be quiet in most systems.

And I agree with you, the seventh position, just behind the Fortron Aurora in the recommended list, is much more logical that the first position.



You are right. I counted and measured the cables for the 400w version and falsely assumed that the cable count was the same for the 350w version. Thanks for pointing this out ! I will add a note to the review.

Interesting thread here, some good insights & more twists in the ST saga.

I've had 350 and 400W ST samples for maybe a month -- before Comdex anyway, and hardly any chance to much with them. I just tried firing up both of those again, and found the default fan speed/noise a bit higher than the first 300s I reviewed. I opened both up and can report that it is the same brand / model of fan as before: SuperRed, .19A, CHB12012AS-A.

My conclusions with these 350 and 400W models I have is that the default/min fan voltage must have been raised a bit -- maybe to 5.5-6V instead of 4.5-5V. It was literally inaudble before at moree than a foot, now it can be heard much more readily -- still quieter than the 120mm Fortrons tho.

The use of those 0.3A Yate Loons in the samples Kostik tested raises very interesting questions:

1) Are the 2 versions (with different) fans representative of production runs at different time periods or are they 2 different vesions that will coexist -- albeit in different markets?

2) Was the voltage controller for Kostik's YL fan models kept at the original 4.5V default or raised as mine seem to be? If the former, this suggests maybe they went back to the SuperRed after the default voltage was raised?

It's all quite confusing, really.

The positive thing is that dukla2000's post shows even the version with the higher speed YL fan can be made to run pretty quietly in an appropriate system. But it can no longer be regarded as the silent champ that my original samples (and the entire first production run, it seems, save the misbehaving ones) were.

BTW, dukla2000, if the Tornados are not available there, where did you get yours?

Have you tried the Nexus 3000 or Seasonic Super Silencer 400?

Any comments about the SuperRed fan?

Kool 'n Quiet (aka SupaStar hereabouts) has the Super Tornado 350W, Super Tornado 400W, some Super Silencer and the 'OEM' Tornado 300FT. Only the 300W (Tornado - which I really wanted!) is not available from KnQ - considering I bought today he will probably get those in on Monday

I must also note, I did a very quick and nasty sound check of the Yate Loon before installing (figuring I would swap in my NMB or Papst if necessary). Because of the speed sense cable 'daisy-chain' I could only plug the Yate Loon into a Fanmate (at max so I presume it got 11V) without cutting. The noise was pretty similar to the NMB 4710NL-04W-B10-P00 (@12V) so for the moment I am sticking to the Yate Loon.
The high (circa 80%) AC/DC efficiency and ATX 1.3 compliance (mainly high +12V loads after efficiency) potentially puts them well clear at the top of the list. This saga to get a predictable 120mm version is soul destroying - I would love to see them clean up but ...

SS 460 was reviewed by pcsilencieux.com

Google English translation

JVM, I have not tested the NX-3000. The Super Silencer 460w I have is, AFAIK, the only Super Silencer sample in France, as this PSU is not yet available here. So I can't tell you about the 400w version.

The Superred fan is the only bad point I have found about this PSU. It's not loud, but it's far from Papst/Panaflo quality, and this PSU has so many qualities appart from that, that it would probably be perfect once the fan is swapped for a better one. By the way the full review (in french) is available here.

Edit : well, like Trip said .

Speaking of Papst, is there any click sound with their fans when undervolted? Only happens on 80mm fans?

Do you recommend swapping the Superred with a Papst? Is the problem then that the Superred spins to aggressively and makes more of the swoosh sound whereas the Papst spins slower and therefore quieter?

Lastly, I will have to search out a translation for your SS 460 review.

I have never noticed any clicking with Papst fans, be they 80mm or 120mm, running them between 5v and 12v. But many forumers here have had this problem, so maybe the quality is different in the US. I don't know. Anyway if you have access to Panaflos, you would better use an L1A or M1A, they vibrate less and are cheaper.



The biggest problem I see with the fan is the quality of the noise (quality of the bearing). It doesn't push that much air, and the "wooshing" is very acceptable.

For a high price psu you should not have to

Actually, I was already figuring I should sacrifice my Folding output tomorrow and do a serious comparison of my 120mm fans: I have the SuperRed & Yate Loon from Seasonic, Papst 4412FGL & NMB as above plus 2 others that should have been retired with Concorde but survive on a shelf somewhere.

You are absolutely right. It boggles the mind that a PSU around $100 dollars should need any kind of fan swapping.

Is there a PSU which is fairly quiet with no annoying sound and not in need of fan swapping?

The plot is thickening. As I write I have my 300FT and Super Tornado 350 in kit form on the floor.

1) The picture in MikeC's test of the Super Tornado 300W fan connector shows a 3 pin header on the psu mobo (if my eyes dont deceive?). Both my FT (admitedly no speed sense anyway) and my 350FB only have 2 pin headers and neither mobo even has a 3rd pin location. (The Yate Loon from my 350FB has the speed sense wired direct to the external connector.)
2) The S2FC voltage supplied to the fan is dependant on the fan! Details to follow later, but in short between my Papst, NMB, SuperRed & Yate Loon I have a low of 3.75V and a high of 6.95 under the same conditions! Based on Mike's comment in another thread that even fans in the same batch can vary significantly, this could be the cause of all the different experiences folks are having.
3) More trivial: just double checking the pics in Mike's review: the DC outlet (back) side of his sampla has a single row of (sparse) vents. Both my psus have double, dense (fully populated) rows!

Somehow I missed this earlier. Great review and thanks for the link!

Dukla2000, same thing goes here. As you can see from this pic, the samples I have reviewed had a two pins connector (and RPM monitoring cable wired directly to the PCB), and more rear vents than the one in Mike's review.

4) Another thing I have noticed - both my psus & Kostik's appear the same. But Mike's review pic hereshows a blue wire connected to the immediate right of the fan pin header - we have no wire there at all. Not even a hole in the pcb.
5) And also to the immediate left of the pin header we have a transistor where Mike appears to have a yellow & black wire. It may be Mike had a transistor that is obscured?

Differences 1, 2, 4 & 5 could just be the result of Seasonic reworking the fan control circuit? Would be interesting to know what Mike's 350 & 400W samples are like.

Coming back to the fans and Mike's comments above, I figured I would measure the lowest voltage being fed to my fans. Room temps were about 18C, only load on the psu was a hdd, psu case was open on the floor and the fan was aimed away from the psu. After starting the psu a thermistor balanced on the low voltage heatsink generally rose about 10C which is when I took the 'stable' readings below.

First off I measured my tornado 350W with Yate Loon. Start 4.2V, stabilised at 4.46V after 30 minutes.

Then measured 300FT with SuperRed: the multimeter was all over the place with no reading constant for more than 1/2 second - start fluctuating 4.2-4.5, after 20 mins fluctuating 4.63-4.74.

So figured the FT fan controller could be very different to the 350FB and put the Yate Loon on the FT: start 3.63V, 3.75V after the temps rose. Rock solid (on the multimeter) readings, so concluded the fluctuations with the SuperRed were a feature of the fan itself!

But figured while I was at it would try some other 120mm fans I have on the FT controller (more details on these fans coming in the fan forum later):

NMB - start 6.86V, 6.95V when psu warmed up (5-10 mins)
Papst - start 5.2V, in 10 secs rose to 6.1V and eventually to 6.2V
Evercool - start 3.74V, quickly to 3.8V and stabilsed at 3.83V

Which lead to my conclusion 2 above. But the real gotcha is even swapping fans on these psus is a nightmare as you cannot predict which part of the S2FC curve a particular fan will land on!

One last observation to report - the high voltage heatsink is NOT live - this was tested using a variable speed finger flick until I gained confidence Anyway, time to get my FB back together and get rid of my 'silent' FSP that was doing temporary duty during testing!

Did you measure the current through the fans? Just a thought but maybe the circuit is adjusting fan current in proportion with the output? That would certainly explain a few things...

How about a nice PSU that doesn't rev up and down like a yo-yo and has a quiet fan and costs less than $100 dollars?

Dear Santa...

That's what I thought/hoped too. So I was planning to hold a 120mm fan tournament anyway which is now posted here including some current measurements. Certainly the basic current use at 12V can't explain things - or not as far as I can see. I didn't measure the current at the start voltage, mainly because I didn't think of it at the time but also because I only have 1 multimeter. (In theory though each fan is fixed resistance, so the current at the start voltages could be calculated?)

I don't know if you can treat a fan as a simple resistive load but, assuming you can and that I didn't make any mistakes, a quick calculation based on the power ratings of each fan at 12V gives the following:

Not quite conclusive but it seems that the power figures are closer to each other than either the voltages or the currents. Pity we don't have the heat sink temperatures for each case. That might show whether the differences are in line with termperatures.

I am too dumb to check the numbers quickly, but feel the big gotcha is the Yate Loon which has extraordinarily high current use, compared to the SuperRed which has a more 'normal' current use: except they get similar start voltages. [Edit] Just double checked and realised this para is complete rambling! The double check is to get the Evercool current which I can do quickly. [/Edit]
I did log the room temp to try keep things fair, as well as the (low voltage) heatsink temps but figure they are meaningless: first the thermistor was losely 'balanced' between folds of the heatsink and probably moved between tests. It was certainly replaced more than once. But second the fans were aimed so they were NOT cooling the insides of the psu - the sirflow was parallel to the side of the psu case. So the temps I measured are independent of the fan cfm - at best an indication of the load the fan was putting on the psu: the only psu cooling was air convection (in a cool room, and the only load 1 hdd and a fan!) Last, the (heatsink) temps I recorded (at stability) just dont have enough discrimination/variation: min 29, max 31.

The reason I did record temps (which is/was valid) was to see when the psu had got to an equilibrium and the fan voltage would (presumably) stop rising. That equilibrium scenario did occur on both the temp and the fan voltage. One corollary I tried was to aim the fan at the psu (after 'stability' peak) to see the voltage drop. These (unthorough and incomplete) results are messy: in some cases the fan voltage did drop but never back to the start voltage. In 1 case (with the SuperRed) there was no obvious voltage drop: arguably it even increased but as first reported with the SuperRed the voltage fluctuations were so large (as observed on my multimeter) at all times it is hard to assert anything.

Kostik did a review on the SS 460 with SuperRed fan and did not note any wild revving up and down. So how could the problem be the SuperRed fan?

Yup he did (earlier) review the SS460, but I was talking about the Super Tornado review he has done most recently and title of this thread Some/many/early 120mm Seasonic's have SuperReds (my 300FT and the Tornado MikeC reviewed), others (Kostik's 350 & 400 Tornados and my 350 Tornado) have Yate Loons.

When I say I had voltage fluctuations on my (120mm) SuperRed, they were not resulting in audible fan speed differences or speed revving that I noticed. The (digital) multimeter would flash e.g. 4.1, 4.3, 4.4, 4.5, 4.3, 4.1 in the space of about 2 seconds. I guess the possible effect on the fan speed was more like PWM.

Okay, but is there a conclusion that the 120mm SuperRed is operating differently than the 80mm SuperRed? I thought you were coming to a conclusion that the revving up and down with the Super Tornado had something to do with the SuperRed fan.

here's an idea. why dont you just put your own 120mm papst fan in it and 7volt it.

end of discussion. now its just as good as any other PSU with a 120mm fan in it and just as quiet if not more so.

1. It costs more
2. It voids the warranty
3. Messier cable arrangement
4. Can't adapt to varying room temperatures and system loads. Sometimes it might not be cool enough, other times it won't be as silent as it could be.

By not using the built in fan controller, you're negating one of the reasons for buying this PSU in the first place.