By virtue of visiting the hallowed realm of OC3D you have already identified yourself as a connoisseur and as such appreciative of all things "high end", something of an aficionado when it comes to selecting the very best kit your hard earned wedge can buy. It comes as no surprise then that you will already be aware that Scythe produce some of the very best fans on the market. The Gentle Typhoon 1850rpm can be thought of pretty much the Wayne Rooney of the fan world, like Mr Rooney, it's no looker, but it can be guaranteed to get the job done.
Some of you will also know that the Scythe Gentle typhoons actually originate from a company called Nidec Servo, in fact their logo can be clearly seen on the box.
All this is very interesting of course, but you're probably beginning to wonder where it's going. That said as astute devotees of all things great and good you've probably already worked out exactly what we're going to be looking at in this review. And if you haven't, then the title of the review should have given you an idea at least.
So what's all this about then? High RPM? How high? Well simply put, should you so desire you can actually get the quite wonderful Gentle Typhoon fans in some quite ridiculous RPMs. How ridiculous? Well how does 3000rpm sound? not ridiculous enough? Try 4250RPM. Still not mad enough for you? Then take a bite out of the totally insane 5400RPM model, and while you're at it, be careful it doesn't take a bite out of you!
Lets have a look at the numbers then shall we.
Some pretty impressive rpms and CFMs, but look at those noise levels!
|Operating voltage range||10.2-13.8V||10.2-13.8V||10.2-13.8A|
|Air Flow||83 CFM||116.5 CFM||150.1 CFM|
|MTBF||100000 hours at 35c||100000 hours at 35c||100000 hours at 35c|
Up Close: Packaging and contents
The Gentle Typhoons (GTs) come simply packed in plastic boxes with shelf hangers. A clear front allows us to see the fan inside. The black sticker towards the bottom of the pack carries two very important messages
"Caution, do not injure yourself. Fan speed is very high"
"Only connect to peripheral 4pin from PSU to draw power".
Can't help but think it's probably best to pay attention at this point, as manufacturers aren't generally in the habit of warning that their product can hurt.
A card insert supplies all the info you need with regards to the specification of the fan within. The upper part of the insert gives details of voltages, RPMs, CFM, dBA etc, whilst the bottom section describes fan blade and bearing characteristics. Aside from the fan itself the only other think in the box is 4xfan mounting grub screws.
Up Close: The Gentle Typhoon
In a world of coloured fans with bright coloured LEDs and fancy shaped cowlings, the Gentle typhoons do perhaps stand out simply because they are none of these things. The lower RPM series that you may be familiar with has a black cowling and grey fan blades, and to be honest has always looked to me as if it's been made from recycled wheelie bins. The high RPM series we're looking at today however is an all black affair, with the plastic looking a bit more glossy and refined than the lower RPM models. It is strongly recommended to connect the fans directly to the PSU supply by the 4 pin connection only. Each of the fans comes with a molex pass through as well as a tachometer connection for attachment to the Motherboard
Probably the most striking feature of the high RPM series is the circumferential stabilising ring. This ring is designed to aid the passage of air rearwards in a more narrow linear fashion. I can't help thinking that they also serve to add rigidity to the blades. If you've ever stood near to one of these when it's running you'll be as appreciative as I am of anything that adds strength to the blades. The picture below right also shows that the cowling is of the semi closed corner design, which may cause issues if you wish to mount these units on a CPU heat sink.
The sticker on the rear of the fan gives info with regards to operating voltages and currents as well as informing as giving the model number of the fan. For those not already "in the know", the AP 29 is the 3000rpm model (pictured below), the AP30 the 4250, and the AP31 the 5400rpm model.
Performance and testing
It's important to remember from the outset that these fans were originally intended to be used in the IT industry, specifically to keep large volumes of air flowing over rack-mount systems, or in other situations where airflow and CFM is more of a concern than noise levels. Why do I mention noise? Well take a little look back at the dBA figures and you'll see why. These fans are loud, and I mean ear bleedingly loud.
That said, as we're always up for a bit of madness we couldn't help wondering how much of an effect the GTs would have on the components within a standard PC. To that end we'll be strapping these babies to an air cooler and a water cooler, as well as looking at some other ways in which they can be used.
Test 1: Air Cooler test
For this test we took the Zero Infinity Free Flow+ cooler. A twin stack cooler with 5x8mm heat pipes. Having removed the 3 stock 140mm 1200rpm fans we attached 2 of each speed of GT to the cooler in turn before running the torture test at various clock speeds, with kit and voltages etc as per our usual CPU heat sink testing The theory being that by attaching high speed high RPM fans we'd be able to significantly lower the CPU temp and squeeze out enough performance to enable a higher overclock than that possible with just stock fans.
|Max Average||Ambient||Delta T|
|3000rpm at 4.2GHz||81.5||18||63.5|
|3000rpm at 4.4GHz||FAIL|
|4250rpm at 4.2GHz||76||18||58|
|4250rpm at 4.4GHz||FAIL|
|5400rpm at 4.2GHz||75.25||18||57.25|
|5400rpm at 4.4GHz||FAIL|
To put these results into perspective we need to know that with it's 3x140mm 1200rpm fans the Free Flow+ returned a Delta T of 64.75 degrees at 4.2 GHz. The figures above show that the 2x3000rpm fans don't manage to bring this down much further. Moving to the 4250 we see that there has been a drop of 6.75 degrees over the stock fans giving a Delta T of 58 degrees. Finally moving to the beasts that are the 5400rpm units we see that they've only managed to drop the Delta T by an additional 0.75 degrees over the 4250s. This would indicate that the ability of a CPU cooler to move heat away from the CPU is as dependant on the heat transfer at the contact plate an transmission up the fin stack as it is on the ability of the combined fins and fans to dissipate that heat. In short, if the cooler can't get the heat away from the contact plate of the CPU and up into the fin stack, it doesn't matter how much air the fans push through the stack.
Test 2: Liquid Cooler test
For this test we used the Zalman CNPS 20LQ. If you've read the review of this cooler you'll already know that the fan it comes with is quite loud at full tatt. Quite loud, but not nearly as loud as the GTs. Seriously chaps, don't think for a second that any of these GT fans can be used for every day cooling. Having removed the stock fan we actually fitted a pair of the GTs either side of the radiator, set to pull cool air in from the outside of the case and as such maximising the cooling capabilities of the 20LQ. having seen the air cooler fail to reach 4.4GHz, and knowing that the Zalman on staock fan can hold 4.4GHz at a Delta T of 63.75 we set about finding out what the GTs could do.
|Max Average||Ambient||Delta T|
|3000rpm at 4.4GHz||81||18||63|
|4250rpm at 4.4GHz||77.75||18||59.75|
|5400rpm at 4.4GHz||76.75||18||58.75|
The results although better follow the same pattern as with the air cooler. There's little difference between the stock fans and the 3000rpm models, with both giving Delta Ts within a degree of each other. There's then a bit of a step down in temps with the 4250 fans giving a Delta T of 59.75, and then only a very small step of 1 degree down to 58.75 when we use the 5400rpm model. So again we're left with the conclusion that fan speed and CFM only impact on the cooling ability up to a point, and that once this factor has exceeded the coolers ability to Moe heat away from the processor the gains are negligible.
Test 3: GPU cooling
For this test we used the GTs essentially as case fans blowing air directly into the intake of the GPU. For the test I used my old faithful ASUS EAH5850. I've overclocked this card on more than a few occasions from it's stock configuration of 7.25MHz engine, 4000MHz memory and a Vcore of 1.097. I know that it will handle a Vcore of 1.3volts with it's internal fan set at a rather whiny 100%, but have always wondered how much higher it will go. I started of the test today at the highest i've taken the card to previously, 950MHz/4700/1.3V, with recorded max DeltaT temps as 63 for the TMPINO and 80 for the GPU (ambient room temp is 20 degrees). At this setting it can be a bit flaky for benching and either returns a score or crashes.
|Eng/MHz||Mem/MHz||Vcore||TMPINO DeltaT||GPU temp DeltaT||Ambient|
We can see that positioning the Gentle Typhoon fans adjacent to the air intake of a GPU does have an effect on the cooling of the card, with the 5400rpm model dropping the cards temperature by a full 10 Degrees. This drop in temperature enabled a higher Engine and memory frequency to be used on this occasion enabling the card to complete the bench test at 1000/4800/1.3V with temps not exceeding those used at the lower clocks with the slower fans.
Before we go any further, and just in case you've skipped to the conclusion without reading the rest of the review (Yes you!), it's important to stress that these are not everyday case fans. These fans are designed to be used in servers and in other industrial IT applications. They are loud, very loud. To put the dBA ratings into context let me tell you that while I was testing with the 5400rpm model, my good lady decided it was the perfect time to insist I hoover out what I call my office and what my wife calls my play room. Now having been playing with these fans all day I know they're loud, what I wasn't expecting was for the 5400rpm model to be louder than a hoover in the same room. Seriously if you want to know how loud these are, sit next to a domestic hoover and you'll get a pretty good idea. Please don't think i'm exaggerating here, i'm not.
With that out of the way we can move on to make sense of some of the numbers. In the two heat sink tests run we see pretty much the same pattern of results The 3000rpm model makes little difference from the stock configuration, then there's a bit of a jump in cooling performance with the 4250rpm model, with the 5400rpm model failing to make much more of an impact. To me this suggests that the as we increase in fan speed, beyond a certain point although the fans are disipating the heat from the cooler, the cooler itself is failing to get the heat away from the contact plate and up into the fin stack. In simple terms the fans will make a great cooler better still, but invariably won't make and average cooler great.
When we look at the GPU test results we see a more linear set of results to begin with, but again the invariable drop off as the amount of air pushed into the GPU intake fails to garner any more reductions in temperature. Again we can surmise that the heat sink in the GPU has simply reached the limit with regards to it's capabilities.
So if we're not going to use them as everyday fans, and we're not in need of some fans for a rack mount server what are we going to use them for? Well the reduction in temperatures we achieved should have you all thinking about overclocking possibilities. Not everyday Overclocking, oh no, we're talking Death or Glory suicide runs where every little degree of cooling helps. Or perhaps your part of the dedicated folding community and are lucky enough to be able to have a rig set up in an area where noise pollution isn't going to be an issue. If that's the case then having the peace of mind that your hardware has max airflow passing over it will be worth the £20 per fan that the Gentle typhoons will set you back.
Finally I'd like to thank Tom for the best bit of advice i've had in quite a while, which was to fit fan guards to these fans before turning them on. I have a rule of thumb when i'm modding in that any modding sessions that ends with me having the same number of fingers as I did when it started is automatically classed as a success. I never thought i'd have to apply that rule to reviewing. This is reviewing for Gads sake, it shouldn't come with a health warning! (But I'm glad it did).
So are they any good? Should you buy them? Well it all depends on what you're looking for. If you're looking for rack mount server fan then yes. If you're looking for an everyday case fan then no, definitely not. And if we were appraising them on that set of criteria they would bomb. If however you're looking for something a bit nuts to help you with those extra few degrees then yes they're good, and yes you should buy them. Imagine a test bench with these strapped to the CPU cooler, whether it be air or water, additional fans pointed at the GPU, and even more aimed at the chipset. What tantalising bench scores await. I think every hard core system builder and overclocker should have a selection of these in their cupboard "just in case". You might never need them, but they're there if you do.
Thanks to Scythe for the fans on test today, you can discuss your thoughts in the OC3D Forums.