Shuttle’s XPC SN25P mini-barebones system

Manufacturer Shuttle
Model XPC SN25P
Price (MSRP) $420
Availability Soon
WITH EXPANSIVE INTERNALS that can accommodate up to three hard drives, a trio of temperature-controlled cooling zones, tool-free drive rails, and a robust 350W power supply, the P series is undoubtedly the pinnacle of Shuttle’s XPC chassis line. Enthusiasts have been eagerly awaiting a P-series XPC for the Athlon 64, but in the seven months since the chassis’ introduction, the P series has been a Pentium 4-only affair. That all changes today with the arrival of Shuttle’s XPC SN25P.

Sporting a 939-pin socket for the Athlon 64, an nForce4 chipset, PCI Express, and an integrated Envy24PT audio controller, the XPC SN25P could be the best small form factor barebones system to date. Does it deliver on its considerable potential? Was it worth the wait? Read on to find out.

The specs
We’ll kick things off with a look at the SN25P’s spec sheet. Right off the bat, notice that the system is based on NVIDIA’s vanilla nForce4 core logic chip, not the Ultra or SLI versions.

CPU support Socket 939-based Athlon 64 processors
Chipset NVIDIA nForce4
Interconnect NA
Expansion slots 1 PCI Express x16
1 PCI Express x1
Memory 2 184-pin DIMM sockets
Maximum of 2GB of DDR266/333/400 SDRAM
Storage I/O Floppy disk
1 channels ATA/133 with RAID 0, 1, 0+1 support
4 channels Serial ATA II with RAID 0, 1, 0+1 support
Audio 8-channel audio via VIA Envy24PT PCI audio controller, VIA VT1617A codec, and Wolfson WM8728 DAC
Ports 1 PS/2 keyboard
1 PS/2 mouse
1 serial
USB 2.0 (rear)
2 USB 2.0 (front)
1 Firewire via VIA VT6307 (rear)
1 Firewire (front)
1 RJ45 10/100/1000 Gigabit Ethernet

1 analog front out
1 analog bass/center out
1 analog surround out
1 analog rear out
1 analog line in
1 analog mic in (front)
1 analog headphone out (front)
2 digital S/PDIF outputs (RCA and TOS-Link)
1 digital S/PDIF input (TOS-Link)

BIOS Phoenix AwardBIOS
Bus speeds CPU: 200-250MHz in 1MHz increments
DRAM: 100, 133, 166, 200MHz
Bus dividers HT: auto, 1x, 2x, 3x, 4x, 5x
Voltages CPU: auto, 0.8-1.7V in 0.05V increments
DDR: auto, 2.7-2.9V in 0.1V increments
Chipset: auto, 1.6-1.7V in 0.05V increments
Monitoring Voltage, fan status, and temperature monitoring
Fan speed control CPU, system

What differentiates the nForce4 from its Ultra cousin? Surprisingly little. When employed in Socket 754 motherboards, the vanilla nForce4’s HyperTransport link is supposed to be limited to 800MHz. That’s not an issue for the Socket 939 SN25P, though. The system’s nForce4 HyperTransport link runs at 1GHz, just like the Ultra’s.

On the storage front, the vanilla nForce4 lacks the Ultra’s support for 300MB/sec Serial ATA II transfer rates, but that’s hardly worth shedding a tear over. Serial ATA II drives aren’t available yet, and while short burst transfers may be able to take advantage of some of the extra bandwidth, even today’s fastest 15K-RPM SCSI hard drives can’t sustain transfer rates that would saturate Serial ATA’s 150MB/s pipe. Apart from the lack of 300MB/s transfer rates, the nForce4’s storage and RAID capabilities are identical to the Ultra’s, so you’re not missing much.

The only other feature that the nForce4 lacks is ActiveArmor acceleration for the chipset’s integrated Gigabit Ethernet controller and firewall. The GigE controller and firewall are still there, just not the hardware acceleration, which could result in higher CPU utilization during network transfers.

The nForce4 doesn’t differ from its Ultra brethren at all when it comes to integrated audio. Both offer generic AC’97 audio without hardware acceleration or support for high definition resolutions and sampling rates. Fortunately, Shuttle completely bypasses the nForce4’s integrated AC’97 sound in favor of VIA’s Envy24PT audio controller. The Envy24PT doesn’t have hardware acceleration, but it does support higher sampling rates and resolutions. Considering the SN25P’s lack of a PCI slot and the scarcity of PCI Express sound cards, it’s nice to see Shuttle making an effort to compensate for the nForce4’s anemic integrated audio.

Outside the box
As you might expect, the SN25P’s exterior is similar to other P-series chassis. Shuttle mates the cube’s deep charcoal body with a blue face plate and silver trim to lighten the mood, making for an attractive aesthetic that’s still reserved enough for the office.

Like other P-series systems, the SN25P integrates a memory card reader right up front. The card reader consumes two of the chipset’s USB ports, but given the popularity of digital cameras, those are ports well spent. Shuttle hides the rest of the system’s front panel goodies under a series of drop-down doors.

With the touch of a finger, the doors swing down to reveal an external 3.5″ drive bay and a front port cluster that includes audio, USB, and Firewire plugs.

Shuttle also hides the SN25P’s 5.25″ drive bay behind a stealthy door, keeping beige optical drives from scarring the system’s otherwise attractive face. The spring-loaded door opens and closes automatically under the force of the optical drive tray. The chassis is also equipped with an adjustable external eject button that’s compatible with a wide range of optical drives.

Moving to the side of the system, we catch a glimpse of the P-series’ Prescott roots. This chassis was originally designed for toasty Prescott Pentium 4 processors, so the chassis has plenty of venting to provide adequate airflow through the system.

Holes even riddle the bottom of the case, allowing cool air to waft up under the motherboard.

Oddly, the SN25P’s rear panel isn’t as heavily-vented as the rest of the cube. In fact, the grills on the system’s two 70mm exhaust fans look restrictive enough to impede airflow. The 80mm power supply exhaust fan’s grill is more open, but it’s tempting to break out a Dremel to trim some of the extra metal.

Below the fan grills, the SN25P bristles with peripheral ports. The system’s extensive array of audio ports, which includes both RCA and TOS-Link digital S/PDIF outputs, is particularly impressive. Shuttle scores points for integrating a CMOS reset button right into the port cluster, saving overclockers and tweakers from having to crack the system open to clear the BIOS.

Opening ‘er up
Enough with the outside, let’s slide off the SN25P’s aluminum shell and have a peek at what lies under the hood.

A loaded SN25P from the left…

And the right

With support for external 5.25″ and 3.5″ drives and two internal 3.5″ hard drives, the SN25P’s internals are surprisingly expansive for a small form factor system. Hard drives reside side by side along the top of the system and slide into place on a pair of very slick tool-free, snap-on rails.

The snap-on rails make mounting drives in the SN25P easier than in most full ATX systems, which is pretty remarkable. Shuttle also includes a handful of rubber bumpers to dampen vibrations between hard drives and the case’s aluminum shell. It’s a good idea in theory, but the stiff rubber bumpers are too tall for the limited clearance between hard drives and the system’s outer skin. You can force the shell back on with the bumpers installed, but that flexes the hard drive rails to a point that I’m not entirely comfortable with. Thankfully, it’s easy to trim the bumpers down to a more reasonable height.

Like the hard drive rails, the SN25P’s external drive cage is a tool-free affair. External 5.25″ and 3.5″ drives get rails of their own, and the entire drive snaps into place with nary a screwdriver in sight. With the drive cage removed, we can take a closer look at the rest of the system’s internals.

The SN25P’s PCI Express x16 slot is mounted on the outside edge of the system, leaving just enough room for double-wide graphics cards like ATI’s Radeon X850 XT and NVIDIA’s GeForce 6800 Ultra. Such cards will block the system’s PCI-E x1 slot, but until PCI Express peripherals become more prevalent, it won’t be missed.

Around the other side of the system, users have easy access to the SN25P’s DIMM slots.

The FN25 motherboard’s layout is as tight as one would expect from a small form factor system. Notice that the DIMM slots run down the middle of the board rather than along its top edge, presumably to maintain closer proximity to the Athlon 64’s on-die memory controller. Also note the use of an active chipset cooler for the nForce4, which is developing a reputation for running a little warmer than most chipsets.

Shuttle hides the FN25 motherboard’s most interesting feature up in the top corner of the board. Here, we find VIA’s Envy24PT audio controller, VT1617A codec, and a Wolfson WM8728 DAC. With support for 24-bit audio at sampling rates up to 96kHz, the Envy24PT hints at high-definition audio, but it’s not quite that easy. With analog output, codec and DAC sampling rates and resolutions also come into play, and that’s where the SN25P’s implementation falls a little short. The Envy24’s first six output channels are routed through the VT1617A codec, which only supports resolutions up to 20 bits and sampling rates up to 96kHz. The Wolfson DAC supports 24-bit/192kHz audio, but in an eight-channel configuration, it only handles channels seven and eight.

It would be wasteful just to use the high-end Wolfson DAC for channels seven and eight, but VIA’s Envy24PT drivers actually allow users to route two-channel stereo audio through the WM8728. This pipes stereo output through the SN25P’s rear 7/8 audio output, potentially offering superior fidelity to the VT1617A. At the very least, the driver switch allows users to enjoy two-channel 24-bit/96kHz audio output through analog speakers or headphones without having to worry about downsampling sapping fidelity.

Cooling and power
The SN25P ditches Shuttle’s venerable ICE CPU/system cooler in favor of dedicated cooling zones for the system’s processor, hard drives, and power supply and graphics card. Since the CPU zone is the most important one, we’ll start there.

Considering that the P-series chassis was originally designed to encase near-molten Prescott Pentium 4 processors, it’s no surprise that the chassis pays special attention to CPU cooling. The processor cooling zone consists of a shrouded wind tunnel that’s completely isolated from the rest of the system. The tunnel sucks cool air in from the right side of the system with a 70mm intake fan and expels it out the left with an 80mm exhaust fan.

A modified version of Shuttle’s ICE cooler sits between the two fans, channeling heat from the CPU up through a quartet of heat pipes into an array of tightly-packed cooling fins.

The cooler sits on a smooth copper base and screws directly into the motherboard. It’s actually the only system component that you’ll need a screwdriver to remove.

With processor cooling shrouded from the rest of the system, Shuttle can concentrate on dedicated cooling solutions for the system’s other components. The hard drive zone is cooled by a pair of 70mm exhaust fans that draw air out the rear of the system, while an 80mm PSU exhaust fan handles the graphics card and power supply zone. All three of the system’s cooling zones use temperature-controlled fans with linear speed control, which should keep noise levels to a minimum.

Given the Athlon 64’s more agreeable thermal profile, the SN25P should be able to get away with relatively low fan speeds most of the time. The Athlon 64 family’s comparatively low power consumption probably won’t tax the system’s 350W power supply, either. That should leave plenty of juice for high-end graphics cards.

Shuttle’s XPC BIOSes are known for giving users plenty of fan control options to balance adequate cooling with low noise levels and the SN25P’s BIOS is no exception.

Users can select from nine fan profiles, including a smart fan option that uses linear fan speed control to incrementally ramp up RPMs as processor temperatures increase. Unfortunately, the BIOS doesn’t let users set the temperature trigger point for the Smart Fan feature.

A better alternative to adding a temperature trigger to the BIOS would be for Shuttle to release a version of its XPC Tools software that supports the SN25P. However, given the fact that Shuttle’s official XPC Tools release only supports one XPC model, the SB83G, we’re not holding our breath.

As far as tweaking and overclocking options go, the SN25P’s BIOS needs some work. HyperTransport link speeds are available between 200 and 250MHz in 1MHz increments, but there’s no way to lock down the system’s PCI or PCI Express clocks, making serious overclocking a dubious prospect.

Overclockers will at least appreciate the BIOS’s CPU multiplier control, although multipliers are only available in 1x rather than 0.5x steps. The BIOS’s voltage options are a little more robust, with CPU voltages available up to 1.7V in 0.05V increments.

At first glance, the SN25P BIOS’s memory tweaking options appear to be reasonably complete. However, there’s no mention of the DRAM command rate. For that, we have to fire up A64 Tweaker in Windows.

According to A64 Tweaker, the SN25P is running with a relatively slow 2T command rate. As we’ll see in a moment, that 2T command rate can have quite an impact on performance. A64 Tweaker can actually force the SN25P’s command rate to 1T, but that has to be done manually after every reboot, so it’s hardly a practical solution. Shuttle is aware of the command rate issue and is working on a BIOS update to address it, but we have yet to receive a BIOS that can set the SN25P’s command rate to 1T.

For those who are a little shy about poking around in the BIOS, the SN25P is compatible with NVIDIA’s nTune tweaking and monitoring utility. Launching nTune produces a rather alarming warning, though.

It’s a little fuzzy, but the message warns that the system’s PCI clock is set to track the HyperTransport link. In other words, the SN25P lacks a PCI lock, at least according to nTune.

Once you get past the warning, nTune offers control over a number of system memory timings and bus speeds.

nTune can also monitor system variables, including temperatures, clock speeds, and voltages. If only it could adjust the SN25P’s fan speed profiles.

Our testing methods
Today we’ll be comparing the SN25P’s performance to that of Foxconn’s NF4UK8AA and DFI’s LANParty NF4 Ultra-D. I wouldn’t expect to see much of a performance difference between the SN25P and its nForce4 Ultra-equipped competition based on the chipset alone, but the XPC system’s lack of a 1T command rate could slow it down. Since it’s not really practical to force a 1T command rate with A64 Tweaker after every reboot, we’ve tested the SN25P with the BIOS’s default 2T command rate.

All tests were run three times, and their results were averaged, using the following test systems.

Processor Athlon 64 3500+ 2.2GHz
System bus HT 16-bit/1GHz downstream
HT 16-bit/1GHz upstream
Motherboard Foxconn NF4UK8AA DFI LANParty NF4 Ultra-D Shuttle XPC SN25P
BIOS revision 1.1 NF4LD209 FN255901
North bridge NVIDIA nForce4 Ultra NVIDIA nForce4
South bridge
Chipset drivers ForceWare 6.39
Memory size 1GB (2 DIMMs)
Memory type OCZ PC3200 EL Platinum Rev 2 DDR SDRAM at 400MHz
CAS latency (CL) 2
RAS to CAS delay (tRCD) 2
RAS precharge (tRP) 2
Cycle time (tRAS) 5
Hard drives Western Digital Raptor WD360GD 37GB SATA
Maxtor DiamondMax Plus D740X 40GB ATA/133
Audio nForce4/ALC850
Graphics NVIDIA GeForce 6600 GT with ForceWare 66.93 drivers
OS Microsoft Windows XP Professional
OS updates Service Pack 2, DirectX 9.0c

Thanks to OCZ for providing us with memory for our testing. If you’re looking to tweak out your system to the max and maybe overclock it a little, OCZ’s RAM is definitely worth considering.

With the exception of the XPC SN25P, our test systems were powered by OCZ PowerStream power supply units. The PowerStream was one of our Editor’s Choice winners in our latest PSU round-up.

We used the following versions of our test applications:

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at a 75Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests. Most of the 3D gaming tests used the Medium detail image quality settings, with the exception that the resolution was set to 640×480 in 32-bit color.

All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

Memory performance
For our memory performance tests, we benchmarked the SN25P with both its default 2T command rate and a forced 1T command rate. This will illustrate the command rate’s impact on both memory bandwidth and latency, which should help us to understand subsequent benchmark results.

With its default 2T command rate, the SN25P lags considerably in our memory bandwidth and latency tests. Forcing a 1T command rate improves the performance considerably, although Shuttle may need to tighten additional memory timings to maximize the performance of the Athlon 64’s on-die memory controller.


The SN25P is only a point off the pace in WorldBench, which isn’t too bad. None of WorldBench’s individual tests stand out as being particularly problematic for the SN25P, which tends to lag by a few seconds across the board.


With the exception of 3DMark05’s overall score, the SN25P falls behind in our gaming tests. The performance differences aren’t huge, but they’re consistent.

Cinebench rendering

The SN25P is barely slower in Cinebench…

Sphinx speech recognition

But Sphinx is much more sensitive to memory performance, and the XPC’s 2T DRAM command rate can’t keep up.

Audio performance

The SN25P’s integrated Envy25PT audio controller doesn’t offer much in terms of improved CPU utilization with positional audio, but since the Envy24 lacks hardware acceleration, that’s to be expected.

Audio quality
For RightMark’s audio quality tests, I used an M-Audio Revolution 7.1 for recording. Analog output ports were used on all systems. To keep things simple, I’ve translated RightMark’s word-based quality scale to numbers. Higher scores reflect better audio quality, and the scale tops out at 6, which corresponds to an “Excellent” rating in RightMark.

Despite its lack of hardware acceleration, the Envy24PT should shine in RightMark Audio Analyzer’s quality tests, but it’s not that simple. We tested the SN25P’s output through the front, rear, and Wolfson-powered outputs, and performance was mixed. With the exception of the stereo crosstalk test, the Wolfson DAC doesn’t appear to provide improved audio quality over the VT1617A codec.

Not content to let RMAA evaluate audio quality alone, I donned a pair of headphones and ran the SN25P through a series of back-to-back listening tests with uncompressed WAV audio files. To my ears, the Wolfson DAC’s output is a little richer and more detailed than that of the VT1617A. The two are remarkably close, but the Wolfson DAC presents subtle background sounds with slightly more clarity.

While I had the headphones out, I also tested for audio interference while performing disk-intensive tasks. The front headphone output on Shuttle’s Pentium 4-based P-series XPCs is particularly prone to interference during hard drive defrags, but as far as I can tell, no such issue plagues the SN25P. No amount of window-dragging, disk access, or other mischief could reproduce the SB95P’s noticeable hum.

ATA performance
ATA performance was tested with a Maxtor 740X-6L ATA/133 hard drive using HD Tach 3.01’s 8MB zone setting.

Scores are pretty even when we look at ATA performance, but the SN25P is a little sluggish in the write speed test.

Serial ATA performance
Moving to Serial ATA, we tested performance with a Western Digital Raptor WD360GD SATA hard drive. Again, we used HD Tach 3.01’s 8MB zone test.

For the most part, the SN25P’s Serial ATA performance is spot on. As you can see, our 10K-RPM Western Digital Raptor WD360GD can’t even burst fast enough to saturate Serial ATA’s 150MB/sec of available bandwidth. Serial ATA II who?

USB performance
Our USB transfer speed tests were conducted with a USB 2.0/Firewire external hard drive enclosure connected to a 7200RPM Maxtor 740X-6L hard drive. We tested with HD Tach 3.01’s 8MB zone setting.

Despite sharing the same nForce4 USB controller as the DFI and Foxconn motherboards, the SN25P manages quicker USB transfer rates.

Firewire performance
Our Firewire transfer speed tests were conducted with the same external enclosure and hard drive as our USB transfer speed tests.

The XPC’s Firewire performance doesn’t distance the system from the rest of the pack.

Ethernet performance
We evaluated Ethernet performance using the NTttcp tool from Microsoft’s Windows DDK. The docs say this program “provides the customer with a multi-threaded, asynchronous performance benchmark for measuring achievable data transfer rate”.

We used the following command line options on the server machine:

ntttcps -m 4,0, -a

..and the same basic thing on each of our test systems acting as clients:

ntttcpr -m 4,0, -a

Our server was a Windows XP Pro system based on Chaintech’s Zenith 9CJS motherboard with a Pentium 4 2.4GHz (800MHz front-side bus, Hyper-Threading enabled) and CSA-attached Gigabit Ethernet. A crossover CAT6 cable was used to connect the server to each system. The nForce4 boards were tested with the NVIDIA Firewall and Jumbo Frames disabled.

Despite the nForce4’s lack of ActiveArmor hardware acceleration for Gigabit Ethernet, the SN25P manages higher throughput than the nForce4 Ultra’s GigE controller. Update 6/13/2005 — We recently discovered that the ntttcp CPU utilization results included in this review were incorrect. The CPU utilization results have been removed, but they didn’t factor prominently into our overall conclusion, so that remains unchanged. A full explanation can be found here.

For our overclocking tests, we swapped our low-latency OCZ PC3200 memory out of the SN25P in favor of some of the OCZ’s PC4400 sticks, which are rated for higher clock speeds at more relaxed latencies. PC4400 memory is designed to operate at speeds of up to 550MHz, so it shouldn’t bottleneck our overclocking efforts. However, running the PC4400 memory at more relaxed 2.5-3-3-8 timings could result in lower overall performance if we can’t crank the clock speed high enough to compensate.

In testing, we were able to get the SN25P stable with a 230MHz HyperTransport link and a 10x CPU multiplier that kept the processor reasonably close to its default 2.2GHz clock speed. Unfortunately, we had to drop the HT link multiplier down to 3x to get the system stable with a 230MHz HyperTransport link. The system would boot with a 4x HT multiplier, but only up to HyperTransport link speeds of 225MHz.

I’ve presented scores for the overclocked SN25P at both 10x225MHz with a 4x HT multiplier and 10x230MHz with a 3x HT multiplier.

With a 230MHz HT link speed, a 3x multiplier clocks the HyperTransport processor link at only 690MHz—much slower than our 10x225MHz configuration, whose HT link is running at 900MHz. However, neither Unreal Tournament 2004 nor Sphinx seem to mind the slower HT link speed. Both tests prefer the 10x230MHz configuration’s slightly higher memory bus and processor clock speed.

Noise levels
We measured noise levels 1″ from the SN25P’s front, side, and rear using an Extech Model 407727 Digital Sound Level Meter. Measurements were taken after 10 minutes at idle, and then after another 10 minutes of a Folding@home CPU load. We also measured noise levels with Shuttle’s XPC SB95P V2 and SB86i for comparison. Those systems were equipped with a Pentium 4 520 2.8GHz processor and Radeon X600 XT graphics card, but otherwise had similar hardware to our SN25P. The SN25P and SB95P V2 used their default Smart Fan setting, but the SB86i requires its Mid fan speed setting to maintain stability under load.

With an Athlon 64 3500+, our SN25P system is much quieter than either Pentium 4 520-equipped XPC. That isn’t a particularly surprising result. However, notice that the SN25P barely gets louder under load, and that’s not even with Cool’n’Quiet enabled.

It looks like the XPC SN25P was worth the wait, but it will take a BIOS update for the system to realize its full potential. At the very least, Shuttle needs to implement more aggressive memory timings, including a 1T command rate. While they’re at it, a PCI clock lock and 0.5x CPU multipliers should also be added to give overclocking enthusiasts a little more flexibility. Finally, it would be nice if the smart fan’s temperature trigger were exposed, either through a BIOS update or an official XPC Tools release that works with the SN25P.

All those issues could, ideally, be resolved with a quick BIOS or software update. Hopefully that update will come, and sooner rather than later, because the rest of the system borders on sublime. The more time we spend with the P-series chassis, the more we love its spacious interior, tool-free design, and robust 350W power supply. The SN25P’s three cooling zones might be overkill for an Athlon 64, but with temperature-controlled fans, the system doesn’t make any more noise than it needs to.

Oh, and don’t forget that there’s an nForce4-based motherboard lurking under the hood with PCI Express, Gigabit Ethernet, plenty of Serial ATA RAID, and Envy24PT audio. The vanilla nForce4 may lack a couple of the Ultra’s features, but few users will miss them. In a PCI-less small form factor system like the SN25P, I’d rather have an nForce4/Envy24PT combo than an Ultra, anyway.

Shuttle XPC SN25P
February 2005

In the end, the SN25P is a BIOS update away from being the ultimate enthusiast-oriented small form factor system. The XPC’s $420 suggested retail price isn’t cheap, but it’s the first Athlon 64-based small form factor system with PCI Express support, and it could be worth every penny.

Comments closed
    • runkelstrunk
    • 15 years ago

    Excellent review, as usually at TRs.

    One note: AFAIK the “standard” NForce4 SATA-Implementation also does not support NCQ (Native Command Qeueing) opposed to the NForce4 Ultra and SLI.

    That does matter IMHO.

    Unfortunately, that was not mentioned in the review when analysing the SATA-Differences between the different NForce4 versions.

      • Dissonance
      • 15 years ago

      Incorrect. The nForce4 supports both NCQ and TCQ, just not 300MB/sec SATA-II transfer rates. 300MB/sec transfer rates are exclusive to the nForce4 Ultra and SLI.

    • sjankech
    • 15 years ago

    wow shuttle does it again by managing to skimp on the chipset. for $400 it should come with the ultra version not the plain. they seem to skimp on all the AMD stuff that i’ve seen. last time it was not using the gigabit ethernet on the nf3. i think i’ve given up on shuttle and gonna build a matx comp instead.

    • SezaGeoff
    • 15 years ago

    What in the world does a case review tell you? How loud and how sharp the edges are.
    A SFF review does the motherboard and the case, and gives you an overview of how the whole thing hangs together. SFFs are basically systems with your choice of CPU, HDD, Optical and Video, so these reviews are of much more interest than a simple case.

    • Rousterfar
    • 15 years ago

    I actually really like the SFF reviews. Keep up the good work!

    That being said, can we see some normal case reviews mixed in more often?

      • Dissonance
      • 15 years ago

      Stay tuned. We have a lot of interesting things coming at you in the next little while.

    • dosbox
    • 15 years ago

    Nice review of a nice looking unit. The P chassis is a good compromise in size.

    In the future, can you guys comment on noise a little more? 50db of high frequency whine is different from 50db of lower frequency hum.

    • Freon
    • 15 years ago

    If it had a G chassis and a standard PCI slot instead of the PCI 1x slot it would be about perfect.

    Oh well, I’m out of the market. I couldn’t wait and sprung for a midtower AMD64 system.

    • adisor19
    • 15 years ago

    OK, how the heck can Firewire be slower then USB 2.0, especially when writing ?

    Can you guys post the chipset used in your Firewire enclosure for the USB 2.0 and Firewire tests ? As well, what is the firmware version of the chipset in that enclosure ?

    Something smells fishy here..

    Currently there are 2 main chipsets used for Firewire enclosures on the market :

    Oxford 911 and PL-3507 Hi-Speed USB & IEEE 1394 Combo to IDE Bridge Controller.

    The Oxford 911 is a Firewire ONLY controller so it must have an ccompanying USB 2.0 chipset if your enclosure uses it. This is a very high quality implementation allowing up to 60MB/s transfer rates assuming you’re using the latest firmware v.4.0. It is also very expensive..

    The Prolific PL-3507 is the cheap alternative used by many cheap enclosure manufacturers because it combines USB2.0 and Firewire in 1 chip. It’s inexpensive and very problematic. The problem is that the Firewire part of it is a verry crappy and buggy implementation. A max transfer rate of around 30MB/s sustained can be achieved with this POS using Firewire mode. It also needs the latest firmware if you don’t want to loose your data randomly.. USB 2.0 on the other hand is better implemented according to what i’ve seen on forums all over the net..

    So Geoff can you sched some light on this ?

    My beef is that if indeed, your enclosure uses the prolific chipset, you’re spreading the wrong idea to the readers that Firewire is slower then USB 2.0, even though that is not the case.


      • Hattig
      • 15 years ago

      I didn’t know that. I was looking to get a Firewire enclosure in the long term, so I’ll bear that in mind when looking for one. The motherboard uses a VIA Firewire controller, and lots of motherboards use this controller on VIA, Intel and nVidia based motherboards.

      How does $420 compare with other SFF barebone systems?

      • Dissonance
      • 15 years ago

      The enclosure uses the PL-3507. We’re really not comparing USB to Firewire, though. Just each platform’s Firewire or USB implementation.

        • Hattig
        • 15 years ago

        How can you test a Firewire implementation with a dodgy peripheral that doesn’t actually work that well in Firewire mode?

          • Dissonance
          • 15 years ago

          Dodgy or not, the enclosure is representative of what’s on the market. And we’re testing comparative Firewire performance, not comparing Firewire to anything else. Regardless of a chip’s shortcomings, the playing field is level and we’re not making any claims about the value of Firewire versus USB.

        • adisor19
        • 15 years ago

        Ok Geoff, i agree that you’re not testing to see wich one is faster, but keep in mind that you’re nowhere near close to reaching the full potential of Firewire with this enclosure. The PL-3507 chipset itself is maxing out at around 30MB/s when testing the Firewire interface, so you’re barely scratching the surface of what the VIA Firewire chip can attain.. Who knows if that VIA chip has a bottleneck somwhere around let’s say 40MB/s.. With this enclosure you’d never know.

        It is a well known fact that Texas Instruments makes the best Firewire PCI chipsets out there, folowed by Agere( aslo known as Lucent), and then the others.. There is a reason why manufacturers prefer using brands like : Realtek(the crab) and VIA.. they’re cheap and probably don’t perform as well.

        You have reviewed in the past motherboards with Firewire chips made by Texas Instruments. Trust me, they cost the mobo maker more to use them, but they are worth it. Get an Oxford911 based enclosure (i think ADS makes some good ones) and test it out. You should see the TI chip out perform the VIA one in both write and read performance AND CPU time.

        With your PL-3507 enclosure, all you’re telling us is : Yes, the VIA firewire chip can sustain 30MB/s transfer rate.

        Well we knew that already(well.. ok, i did), but that is nowhere near the max throutput of Firewire400 and i bet nowhere near that of the IDE haddrive itself.

        P.S. Also, if you don’t wanna spend 90$ on an Oxford911 based enclosure, you can just buy the Firewire-to-IDE board off of Ebay. Just search for “Oxford 911” and you’ll find this company that sells them for 45$.(just the board +1.5$ for the power injector) wich is a much better price IMO just for testing purposes.

        Please take this as constructive criticism, as all i want is a more realistic benchmark when dealing with testing Firewire and USB PCI chipsets.

        I enjoy reading all your reviews and think you’re doing a great job 🙂 Keep it up.


          • Thresher
          • 15 years ago

          He’s got a point.

          You can’t really judge where the bottleneck is if the device you are testing it on is known to underperform.

        • adisor19
        • 15 years ago

        Also Geoff, just to emphasise how important it is to have the latest firmware for these enclosures here is what was changed in one of them :

        “09/22/2004 –

        1. ICP function: support new FLASH – EM39LV010. [USB]
        2. Mutiple buttons support (for One Touch Button application) for Windows. [USB]
        3. Fix Windows XP+SP2 with VIA 1394 host. [1394]
        4. Fix IDE identify device on boot. [1394]”

        This is strait from the readme file of their latest firmware update. There are sooo many fixes for the Prolific chipset that it makes you wonder who in their right mind agreed to put it for sale with all those bugs in the first place.. and i’m not even talking about their crap performace.

        Anyways, for your reading pleasure and for those of you unfortunate enought that have this Prolific chipset in their enclosure, here’s the link to thier firmware update page. Make sure in case you deicide to update it, that you follow instructions to the letter.


        • Spotpuff
        • 15 years ago

        But wouldn’t you run into a problem if say you were to compare Nforce 4 Ultra and Nforce 2 benchmarks using DDR 200? You could say memory performance in both were “equal”. The same thing is happening here whereby the device’s performance is being saturated before the motherboard’s firewire performance can be examined.

    • MaceMan
    • 15 years ago

    “Enthusiast” means gamers, in my mind. Does the audio support Qsound, or whatever freaky alternative to what Creative is pushing? Does it support all the gaming EAX standards we expect to have available?

    I love the 6 exposed USB ports, and the space for a double-wide video card is so sweet. Add in the snap-rails for everything and I’m stoked to grab one.

    Did you ever find the thermal pads (better than the ones Dan of Dan’s Data commentedon) per your thermal grease question in a previous post? Care to share the secret of where to get them?

      • Dissonance
      • 15 years ago

      The SN25P uses VIA’s reference Envy drivers, so Qsound is in there. Qsound does EAX and EAX 2.0.

      And the thermal pad thing was Scott, not me.

        • MaceMan
        • 15 years ago

        Oh, sorry Diss. I wasn’t paying attention.

        What WAS the result of the thermal pad research? Any hope?

    • dukerjames
    • 15 years ago

    yikes!! not another!

    • zgirl
    • 15 years ago

    While I have been waiting for this and I really like SFF boxes. I do wish it was a little smaller.

    All need is 1 HD, 1 Optical Drive, and one slot for Media Cards.

    I wanted my SB75G2 with PCI-E and AMD64 support. Is that too much to ask?

      • d_rogue
      • 15 years ago

      that product will be out next month. I think it is called the ST20G5

        • zgirl
        • 15 years ago

        yeah, but it is an ATI chipset not Nvidia. 🙁

          • indeego
          • 15 years ago

          The ATI chipset is good, from what I remember of TR’s preview. You can’t complain of more chipsets on the AMD platformg{<.<}g

    • indeego
    • 15 years ago

    Now you’re reviewing systems based on their ability to overclock? Was that always a criteriag{

      • Dissonance
      • 15 years ago

      Yes. Overclocking potential, particularly the appropriate BIOS features and options, has always been one of many criteria.

    • Thresher
    • 15 years ago

    Excellent review as always.

    That being said, I’m getting tired of SFF. Seriously.

    Time to move on.

      • flip-mode
      • 15 years ago

      What do you suggest we move on to? Seriously, I don’t have a problem with your statement, in fact I somewhat agree. But we need something to move on to.

      On the other hand, SFF is pretty much the most exciting thing the PC has to offer right now. So it seems to me that you either move to mobile, or home electronics, or stay with SFF. I do think there’s some room for a good round up of TV/HDTV cards.

        • Thresher
        • 15 years ago

        I would agree with that, but it seems like there have been a boatload of SFF reviews here lately.

        I’d like to see some case reviews. If no one has noticed, there have been some outstanding HTPC cases put out lately by Silverstone and Ahanix. I’d like to see some reviews of standard sized systems. Some more motherboard reviews (like is it really worth moving to PCI-Express yet or should I just stick with my AGP?).

        How about a PDA review or two? Maybe an MP3 player (although that’s kind of been done to death too). Monitors would be nice, I haven’t seen a fullfledged review of the Dell 2005FPW from anyone yet.

        My point is that there is a lot of hardware out there, some of which isn’t covered very well (like PC cases for grownups, not the ricer style cases). Whenever Shuttle produces a new box, I just yawn and pat my Antec Aria and remember that I can change motherboards at anytime, if only someone would review a Mini or Micro ATX board…

          • Tuanies
          • 15 years ago

          Patience my friend 😉

    • Captain Ned
    • 15 years ago

    If it had 1 PCI slot, there’d be one on my desk next week. I’m just not yet willing to give up my Turtle Beach Santa Cruz.

    • flip-mode
    • 15 years ago

    If I were going to lay down some cash on a SFF, this would be the one as it has everything I’d ask for. However, I’ve just joined the ranks of the mobility endowed and find it to be quite pleasing. In fact, its a good thing this XPC wasn’t released sooner or I might have folded to the temptation – and now that I have a laptop – Athlon 64 mR9700 Uniwill 258KA0 – there’s no going back. I LOVE this lappy.

    • Hattig
    • 15 years ago

    Is that some sort of picoBTX motherboard, with the DIMMs rotated from the standard?

    It is a nice system though, certainly more appealing than the Intel version which seems to be a lot noisier (and noise is one of the reasons to get these machines, and lack of overclocking is much less of an issue). I’m still thinking that this system will have 6 fans operating inside it though, so is it (in an otherwise quiet room) noticable when it is on your desk?

    Oh, and power consumption figures would be nice, compared to an equivalent performance Intel version (2.8GHz is hardly competition for a 3500+!).

    Certainly though if I was in the market to buy a PC, I’d look long and hard at a system like this even if it is more expensive up front.

    • Zenith
    • 15 years ago

    Too expensive to begin to attract my attention.

    • dragmor
    • 15 years ago

    Dear Shuttle.

    We (the shuttle buying community) are a shallow bunch. We buy Shuttle PC’s because they are small and pretty/elegant like our Asian girlfriends. In this vain I would like to request that Shuttle:

    a) Work on the units visual appearance. The SN95G5 is you best looking case, please move more towards this style rather than P Series look.
    b) Size is important. The smaller the better, the new shuttle seem to be getting larger and larger. We don’t like when our girlfriends get larger and we don’t like when shuttles get larger either.

    Shuttle Customer.

      • The H
      • 15 years ago


      Are you sure you have a ‘real’ girlfriend? j/k

      • Buub
      • 15 years ago

      Mmmm… Asian fetish…

      • 5150
      • 15 years ago

      Is this the gleek?

      • d0g_p00p
      • 15 years ago

      now we really piss off!

      • jennyfur
      • 15 years ago

      I totally agree that the G5 is a far better looking and useful case. If they did an nForce 4 version of it instead of this ugly, large P series, I’d buy it without hesitation. Why oh why must you constantly torture me Shuttle?

    • Coran Fixx
    • 15 years ago

    msrp of 420.00, what a bargain 4 U !

    • TO11MTM
    • 15 years ago

    Finally, some Via Envy action on a SFF… It’s just too bad that it is still the PT version of the chip. A step in the right direction, though. Good review!

Pin It on Pinterest

Share This