High output, low noise – Brian
SUMMARY: High output, low noise
Note: While this link above takes you to a similar 500W unit, I actually received a 550W specimen when ordering this item using this link.
“Switching” refers to facility to run the fan in this at three different capacities, which I’ll cover in detail later.
This particular PSU comes from Directron’s stock of “Quiet Computing”
There’s extremes in computing in regards to noise. One end of the spectrum takes a
“Damn the torpedoes” approach, and any noise required to achieve that highest overclock possible is acceptable.
The other extreme is to make the PC as quiet as possible, while still trying to achieve a high overclock, or just run as silent as possible at default settings.
Most of us “live” somewhere in between. Generally, some noise is tolerable, provided the end result meets our needs.
This PSU attempts to satisfy all the “extremes”. High output, low noise, appearance, and versatility. We’ll cover all of those bases here.
Let’s start off by a close look at the PSU, and the wire harnesses.
- The PSU
- Standard power cable
- Small bag w/ 4 mount screws
- Data sheet of “Recommended AMD MP” PSU’s (with the TTGI/SF units highlighted)
- Owner’s/User Manual
An “Owner’s Manual”?? For a PSU? Yup. This isn’t your usual PSU.
The manual is bi-lingual, printed in English and Chinese only. The bulk of the information contained within it is PSU specifications for the entire TTGI/SF product line of PSUs,
and pinouts for all the connectors. There’s the standard (more or less required) paragraph on “Do NOT open the PSU!”, and little else. Hidden in the text is a very brief description
of what the three switching “settings” are. Again, I’ll cover that later.
The PSU itself is a very attractive unit, featuring a single 120mm fan that has four blue LEDs built into it. They cast a nice glow when powered up, illuminating a good portion of your case.
Fit and finish is good, with no sharp edges that I could find.
Signore? * =)
The harnesses feature a great selection of connectors, and are almost long enough to reach your next door neighbor’s PC, if needed. The main ATX power lead is sleeved in black plastic “flex”
material, a nice touch.
- 1x 20 pin ATX power
- 1x 4 pin ATX12 power
- 1x 3 pin fan monitor lead
- 2x floppy drive connectors
- 2x Serial ATA (SATA) connectors
- 9x (NINE!!) 4 pin Molex connectors
It’s good to see manufacturers incorporating SATA into the harnesses now, eliminating the need for power adapters. As more SATA drives are released, and
they become a “mainstream item”, these will be very useful indeed.
The nine 4 pin Molex plugs is amazing. But it also raises the question of; If you need that many power connectors, how quiet can the rest
of the computer be?
(From left to right as pictured above)
- 20 pin ATX power: 21″ long
- 4 pin ATX12 power: 21″ long
- 1st accessory lead: 21″ to first 4 Pin Molex, 7.5″ to 2nd 4 PM, 7.5″ to 3rd 4 PM, 7.5″ to floppy connector
- 2nd accessory lead: 21″ to first 4 Pin Molex, 7.5″ to 2nd 4 PM, 7.5″ to floppy connector
- 3rd accessory lead: 21.5″ to first 4 Pin Molex, 7.5″ to 2nd 4 PM, 7.5″ to SATA connector
- 4th accessory lead: (identical to 3rd lead)
- 3 pin fan lead: 24″
Here’s a good visual reference of just how long those leads really are;
The PSU is sitting on top of a mid-tower case (Raidmax Scorpio 868W), and in front of a Chieftec full-tower case (the one with SIX cd-rom bays). The “height”
it’s sitting at is about 1/4″ lower than the PSU would actually mount in the big Chieftec.
Suffice to say, you should have no problems reaching anything inside any case that needs power. The Chieftec is one of the biggest towers
commercially available. When I set this PSU on the file cabinet (two drawer, 29″ tall) below the two PC’s pictured, the accessory leads reached the floor. Easily.
If you plan on installing this PSU in a mid tower case, your problem is likely going to be how to hide all of that harness, rather than not having enough connectors,
or the leads being long enough.
The two systems pictured above are also the two systems I tested this PSU in. When we come to the testing part of this review, I’ll further detail the systems themselves.
Pictured is the sticker from the side of the PSU, showing the minimum and maximum output levels for the PSU. Some quite impressive numbers, for a PSU that caters to the “quiet
Note the upper right corner…AMD K8 support. It’s good to go for AMD 64′s, and Opterons.
* Italian to English translation: “Would you like anything else with your huge plate of spaghetti, Sir?”
Next, I bench tested the TT-550SS, with a number of other older PSU’s, to try to gauge just how quiet it is in relation to other PSU’s available.
(The PSU from the Raidmax case (a 420W unit (Model No. KY-520ATX)) shown above)
Now, I don’t have any “Decibel Meter” to truly accurately measure sound volumes (it’s on the “shopping list”), just a fairly good ear. One of the good things I retained from my
“roadie/guitar tech” days. =) I figure if I used to be able to take a guitar with a broken string just offstage, replace it, stretch it, and have the
axe retuned by ear and back in it’s stand before they finished the song being played, I ought to be able to pick up fairly subtle volume
levels between these PSUs.
I will be investing in the same Radio Shack unit the Joe Citarella uses for testing purposes in future.
Looking at the back of the TT-550SS, we find the usual suspects, plus one additional button. This small red button (just to the right of the main switch) controls the
three switching “levels” this PSU can use. As they do effect noise levels this unit produces, I’ll elaborate on those now. They are (top to bottom):
- “A”, “Auto”. The fan is thermostatically controlled on this setting.
- “S”, “Silent”. Disengages the thermostat (I believe…it’s not documented however), and runs the fan on a low voltage, for quiet operation.
- “T”, “Turbo”. Full out boogie! The fan is given full voltage and spins up faster, and the four blue LEDs get noticeably brighter, as well.
The three settings have a corresponding LED next to them. “A” uses a blue LED indicator, “S” a green one, and “T” uses a bright red LED to indicate
Obviously “T” is going to be the loudest of the three settings. It is. But the question is, how loud?
While I had the power supplies out of the two machines I tested with, I bench tested those, plus a number of other older PSU’s I had handy. Most of these had
been culled from dead PC’s, or replaced when upgrading my machines over time. They vary from 250 to 350W units (with the exception of the 420W Raidmax unit).
I tested seven other PSU’s (eight, including the TT-550SS) on the bench as pictured at the top of this page. Well, six, really. In the process of testing I found one
unit I’ve been hanging on to was dead…that unit was the “quietest” of all of them. =)
Bench testing these involves jumping two pins in the 20 pin ATX power connector, to allow the PSU to operate without connecting it to a motherboard. I shut all of my systems down,
so there was no background noise while I tested them.
All of the “other” PSUs use one or two 80mm fans for cooling. Usually, these are very low spinning (thus low flowing) fans. A few of the older units (and the 350W
Antec PSU from the blue Chieftec) were audible at a short distance. A few of them (including the 420W Raidmax unit pictured above) were very quiet. The TT-550SS set
on “A” and “S” fell in with the quiet ones. The Raidmax and TT-550SS were the quietest of the two, only being audible with your ear close enough to the fans to get
your hair sucked into it.
With the TT-550SS set to “Turbo”, it produced a lot more noise. Not an unreasonable or undesirable amount, but definitely audible at two or three feet away. When
testing the TT-550SS mounted in the two systems on “Turbo”, the effect was noticeably lessened being enclosed in the case itself.
However, as I mentioned above, those PSU’s use low flowing 80mm fans to achieve the (lack of) noise they produce. The TT-550SS uses a 120mm fan that is
capable of pushing a lot more air through the PSU. On “Turbo”, it does move a tremendous amount of air from inside of the case to outside.
- “A” (Auto) 1475 RPM (average)
- “S” (Silent) 1300 RPM (average)
- “T” (Turbo) 1950 RPM (average)
So we see that the fan does indeed spin up much faster in “Turbo” mode, accounting for the noise level increase. But this still moves a lot more air at just under 2000 RPM
than a pair of 80mm fans would at the same speed. To achieve the same airflow, those 80mm fans would have to spin up much higher (and therefore be much louder).
In researching this article, I could find nowhere that listed the CFM rating of the fan used in the TT-550SS. Even a “Google” search of the numbers on the fan itself turned up
nothing. The bulk of the 120 x 25mm 4 LED fans I did find push just under 70 CFM @ 2400 RPM. Assuming that this fan fall in line with those (which was about 90+% of the fans I
found of identical description), that might place this one at around 50CFM or so on the “Turbo” setting.
Still, that’s a lot of air, for a very little noise.
Chances are, as well, if you’re concerned about the noise level from a PSU, you’re not putting this unit into a high end gaming rig, but rather a multimedia system or server.
In something like this, you probably wouldn’t need (or use) the “Turbo” setting anyway.
Like I mentioned, with the setting on “Turbo”, the LED’s in the fan do shed some light. Believe it or not, all that light is coming from a pair of fans; the 120mm fan in the
TT-550SS (blue), and the 80mm fan in the side panel (red/orange/yellow). The bright spot on the front of the case is reflected light from the monitor on my desk.
One other item I tested these power supplies with is a plastic “gasket” that can be used to eliminate noise further by isolating the PSU from the chassis.
This works by cutting off the transfer of vibration from the PSU to the chassis, which can lead to other parts of the case to vibrate (notably side panels).
These are really only effective if the entire PSU is isolated, meaning that no part of it is touching any part of the chassis. If any part of
the PSU is touching the chassis, the vibrations you’re trying to eliminate will be transferred through that contact point, defeating the purpose of installing the gasket.
Both cases I tested the TT-550SS in had contact points other than the side of the PSU the gasket fits around, so any “noise reduction” using it was negated.
I’ve got a generic mid tower case here that will allow me to set up a PSU in it totally isolated if the gasket is used, and plan on testing the noise reduction properties in that soon.
I’ve got enough bits and pieces to cobble together a system in that case (it certainly won’t be a “powerhouse”, by any means…likely a Duron based system with a
2GB hard drive and PCI graphics), but it will let me test the gasket out fairly.
I took some Motherboard Monitor 5 readings in both systems I ran the TT-550SS in (as well as the PSU that was in the system previously), to try to gauge how the
voltage rails would react to having the extra power (comparatively speaking, over the smaller output units I have been using).
I set the machines up by starting the PC, launching the Folding@Home client, and then let the machine run for a minimum of 90 minutes.
Nothing other than the PSU was changed (IE, no BIOS settings for CPU, DDR, AGP voltage), and I tried to route the harnesses as close to how they were as possible,
to see if the extra exhaust flow this PSU is capable of might effect temperatures. This was real tricky in the Raidmax case, where the TT-550SS PSU has not only twice
as many harness leads as the Raidmax PSU, but the harness leads are almost twice as long, as well. I wound up stuffing a whole bunch of wire down the back, between
the motherboard tray and right side panel. Not pretty, but effective.
The Chieftec case, being as big as it is, didn’t present this problem. However, this case has a horizontal brace built in, that blocks the majority of the airflow
from reaching the top half of the case, limiting any temperature reduction gains from the extra exhaust capabilities the TT-550SS might offer.
Fortunately, the lower half flows enough air that this isn’t a concern. I get extremely good temperatures despite this.
While we’re discussing the Chieftec tower, we’ll start there. This system used to be my gaming server, and after the hard drive in my main system crashed, became
my “workstation”. I lost a ton of data and partially completed articles in the crash, and moving these apps over to the Chieftec case, with two drives already set up in a
RAID array alleviated my concerns of a repeat of this in future. I changed the configuration of the two Maxtor 30GB (ATA133, 7200RPM) drives over from “Striped” to “Mirrored”,
reinstalled Windows, and started rebuilding.
(Yeah, yeah, I know…back up more often, and this wouldn’t be a problem. *sigh* )
The rest of the system is built around an MSI K7D Master-L motherboard, running a pair of modified (to run in SMP mode) AMD XP1600+ Paliminos. These are cooled by a pair
of Vantec Aeroflows. A lone stick of Kingmax ECC 512MB PC2100 sits in the ram slots.
Case cooling includes a trio of 80mm intake fans, a 92mm and two 80mm exhausts, and one 80mm fan in the left side panel (blowing in).
The PSU I have been using in this system is an Antec PP-352X 350W unit. It’s never given me a bit of trouble, even though everyone I’ve mentioned this to is surprised
a PSU of such a low output is running this system.
Well, let me elaborate on that one point actually. One thing I’d noticed after building this machine, was the CPU voltage would fluctuate greatly. I’ve currently set this in the BIOS
to run at 1.7v, and it will occasionally drop as low at 1.55v. In my
back then detailing the initial buildup of this machine (in a different case back then), I said:
“MBM5 has facility to track variations in temperature and voltage, and the CPU voltage fluctuates as far as 1.65v to 1.92v, with it set in the BIOS for 1.85v. Not good.
I’ve gotten into Windows at 144 FSB (1800 MHz), but it locks up fairly quickly… I truly believe this is the reason. Once I get a better PSU, I think 1800 MHz will be within reach.”
Back then, I had the CPU’s set to run at 12.5 x 140, or 1750MHz, at 1.85v. I’ve since backed that off slightly to 12.5 x 133 (1667MHz) , which I can reach at default voltage.
I received an e-mail about this point shortly thereafter from a fellow who mentioned the MSI K7D Master-L has a bit of a problem in this area, and it was likely the motherboard’s
voltage regulator causing this, rather than the “weak” PSU. I never had the opportunity to confirm or deny this until now, as I never upgraded the PSU in the machine. So reviewing the
TT-550SS gave me a chance to see if a stronger PSU was needed, or it indeed is a motherboard issue.
I ran the Antec PP-352X first (being already installed). The label on the side of the PSU lists it’s specs as follows:
- 3.3v = 20 Amps
- 5v = 33A
- 12v = 15A
- -12v = 1A
- -5v = 1A
- +5vSB = 2A
- Combined 3.3 + 5v = 180W
This PSU uses an 80mm fan mounted in the exterior side, and a 92mm fan located in the bottom for exhausts.
Note the CPU voltage fluctuation. 1.57v to 1.81v, with it set to 1.7v in the BIOS. That’s a very wide swing (.24v !) Note also the CPU and
System temps; 33C/26C at full load, average, respectively. This is typically where they run. Not bad for a pair of XP1600+’s masquerading as
Swapping in the TT-550SS, with no other changes, and running it for over five hours got these readings:
My question is finally answered. Even with a much more powerful PSU, the voltage to the CPUs still fluctuates wildly, indicating the issue indeed
lies with the motherboard voltage regulator circuit, not a “weak” PSU.
However, note that the rest of the voltage rails did benefit from the TT-550SS, particularly the 3.3v, 5v, -12v, and -5v rails, which are now much
tighter in variation, and closer to spec.
While upgrading the PSU in this system probably won’t help my overclock as I once thought, having the rest of the rails much more stable can only help the system
be more stable overall.
Also of note here, as I thought might be the case (given the already superb airflow properties of the Chieftec, and the large cross brace mentioned), the CPU and
System temperatures were not effected whatsoever by the additional exhaust capabilities of the TT-550SS.
Was the system quieter? Hard to tell, as the bulk of the case fans installed are high flowing Sunon (80mm, 39CFM) fans, that produce a good amount of noise. In all
honesty, none of the systems I use are truly of the “silent” variety, so I’ll have to build something up specifically to test this out. I need to, to evaluate the PSU
Next, it was time to try it out on my heavily overclocked Intel P4 machine. (pictured on the bottom of Pg 2 here) Again, nothing was changed in any way other than
swapping out the PSU itself.
This system features an Abit IC7 v1.0 motherboard with a P4 1.6a CPU that’s been shoved to 144 x 16 (2350MHz). It takes a good bump in core voltage to get there (1.65v,
up from 1.5v default). Cooling that CPU is a Swiftech MCX4000 HSF, with yet another 39CFM Sunon case fan attached to it.
The numbers on the Raidmax PSU:
- 3.3v = 26 Amps
- 5v = 32A
- 12v = 13A
- -12v = 0.5A
- -5v = 0.8A
- +5vSB = 2A
It’s interesting to note that despite the higher “peak output”, this PSU has lower ratings on four of the six rails, compared to the Antec PP-352X.
With this PSU and motherboard, there’s very little fluctuation in the voltage rails. The CPU voltage is very close to spot on, as are the rest of the numbers.
Swapping in the TT-550SS, with no other changes, got these readings:
Again, very little variance, low to high. A couple of rails moved closer to spec, and one glaring exception. The +12v rail for some reason dropped about .25v
across the board. Not dangerously low, but very curiously low. Looking at the times at the top of the screenshot shows this configuration was let go for almost five hours.
None of my systems are hooked up to a UPS (also on the “shopping list”..LOL). If there was a random “dip” in the numbers, I might attribute it to a “sag” in the electricity
coming in from the wall, but this went on for a full five hours of “low”. Oddly, looking at the second screenshot from the Chieftec/MSI rig above, the TT-550SS hugged the
+12v spec quite closely in that machine.
I still have the TT-550SS installed in the Raidmax/Abit system, and it still shows low across the board. Curious indeed. Stability has not suffered in any way, however, so
while it “looks” bad, it’s not effecting performance in any way.
Also noteworthy is that the CPU and System temperatures again stayed constant. I thought in this configuration, the extra exhaust might help a bit, but there was no change
between the two PSU’s used.
This PSU is designed to cater to the “quiet PC” crowd primarily, with the ability to wring it out and handle the performance crew’s needs as well.
It certainly seems to accomplish both of these tasks quite well. Plus, with the big 120mm quad LED fan throwing a fair amount of light into the PC case, it’s got an angle for
the “modders” among us as well. This might just be the best “triple threat” I’ve seen around here since Larry Bird walked off the parquet at Boston Garden for good.
Is it silent? Absolutely. Even on the “Turbo” setting, its still (maybe) only as loud as the average generic PSU, but moves a ton more air than that generic unit possibly could.
If you put this PSU on your keyboard in front of you on either the “A” or “S” setting, you would not hear it running.
Is it powerful? 50 Amps of +5v goodness, and 30 Amps from the +12v rail should be more than enough to satisfy anyone’s needs. This PSU is strong enough to pull down server duty
with ease. And the power harnesses it’s equipped with can easily cover this application as well.
I’m pretty impressed with this product. Silent as a lamb, strong as an ox, and forward thinking to include SATA power connectors make this an item well worth investing in.
I’d like to thank Directron
for sending this item out.
Tags: Systems & Components