A lot of the problem I have found was with cables having to run in the same area. As my HDD now sits the other way round with the cables at the motherboard side, it's much easier to get the brace into place. I have no problem with having a brace in general as it does stiffen the thing up a lot but trying to put something in to the front bay is a great feat of spacial perception. I still can't help but feeling that the blanking plate would be a lot better off clipping into the front panel rather than attaching to the brace.

I've sent Silverstone an email regarding some aspects of the design including the 3.5" bracket and the hard drive mounting. Other members here have had replies from them before which is better than you'll get from most manufacturers.

One way to improve the hard drive mounting would be to replace the trays entirely. There's only a big rectagular hole in the underside to fill and 4 screw holes to mount to so a large aluminium plate could fill the gap and then an all new mounting system for 2 disks built on top of the plate.

The separate trays aren't that useful it has to be said. You've got to have the case on it's side or upside down to get at them so it's not as if you can run the system in normal conditions with only one of them in.

Once the new plate has been built it can be attached at the four screw holes with these screws also be the point where a new set of feet would be attached. The current feet are not good at absorbing vibration, aren't in good positions for stability as they're too far from the corners.

This concept has been done once before:
http://www.hardforum.com/showthread.php?p=1032297051

In that mod it was only done so that he could shoe horn more hardware in and it's not going to be very quiet either but a drive mounting structure on top of that would be quite easy.

a month ago I threw in an old pentium 4 lying around and during the hard drive removal process my power wires would get caught on part of bottom of the case. As for the cable management, I tried looping an IDE cable thru the underside of the motherboard tray to the 5.25" bay but the ide cable wouldn't function properly with my side panels on.

Did you notice that the screws that came with the silverstone fan were garbage? I simply threw them out when replacing my fans.

For kicks, I installed my friend's 10.5" 9800GTX. It was incredible it barely fit. There was at most a millimetre of space between the bottom fan and the video card.

If one thing is to be said about this case it is that the sheet metal is really mallable in this case. I've noticed several deformations from changing and installing hardware. especially where you slide the hdd plates into the bottom of the case.

Cable management is very important in this case. I would not say that it's impossible at all as mine is very tidy now but initially it was a complete mess. Going for a modular PSU makes things so much easier. Getting rid of anything unnecessary is also important. My loose end annihilation process is not complete as I still have the second 4-pin 12V connector which is hard wired into the PSU. I shouldn't need this ever in this system and can always adapt back to 8-pin so I may open up the PSU and desolder the wires for this extra lead. The front panel audio cable also has both AC97 and HD connectors so I may remove the AC97 set to make things easier. Coming from a Shuttle this kind of work is not new to me.

Some cables I have hidden behind the motherboard but not behind the tray itself, that seemed just a little too tight for the side panel to have to be installed again. A set of fan wires could be fed round that way if there was a need.


Not something I specifically had trouble with but I don't like those kind of screws in general. I missed off a set of fan mounts when I ordered which I probably should have got to put the front fans in with.

One thing I would have preferred is more thumbscrews. If they were on the HDD trays it would be easier and also if they were on the expansion slot covers and outer cover it would make installing cards quicker.


With or without the PCI-E power connector? I found that area very tight originally and had several start ups followed by a grinding noise as one of the useless extra PCI-E power connectors got pulled into the fan. You might have to replace the PCI-E connector with a right angled unit to make things easier.


I have had a bit of that too. Mostly if a cable gets jammed somewhere and you keep pushing without noticing.

I'm currently struggling with a minor annoyance that's got me quite confused. At some point in the last month or so the PSU fan RPM is no longer reported. Initially I thought I might have nudged the connector loose but it doesn't seem to be that.

There are two possible culprits. The motherboard or the PSU fan. Fan RPM sensors don't tend to break and there's no external sign of anything being wrong. I have done a couple of BIOS updates and one of those may have changed the fan detection algorithms. I have tested the PSU fan header with the RPM sensor from a 5 volted fan spinning at a similar 480rpm and the RPM sensor fluctuated between N/A and ~10kRPM, so perhaps Asus have changed the RPM sensing on this mobo in a BIOS update. I have also tried hooking the PSU fan sensor wire to another header, still no result. I've also slowed down the other fans and stressed the system until the fan steps up and still nothing.

Asus' update utility does not allow you to install an older BIOS so that's one of the things I'll need the command line version to do... without a floppy disk drive. It wouldn't surprise me if they have made some kind of fan sensing change as most users don't use fans this slow and it may be that it was to fix some other reporting problem.

I'm hoping it's not the fault of the PSU, I really can't see that happening or not being fixable but I still have 2.5 years of warranty on the thing so I don't want to set about modding the thing to fix it if it later needs to be RMA'd for a non-trivial matter.

I've also started compiling a list of CPU heatsinks that can fit using the same semi-passive setup as the original NT06. This basically requires a top down cooler with a height without fan of 82mm. Bending either way is a possible mod.

I have two ideas for alternative PSU placements that would make room for tower heatsinks, but I don't have a SG03 so I can't try it out.
What do you think?

Alternative 1:
This will make the most room for a bigger heatsink and fan. You have to make a hole in the case.
This will only work with standard length (140 mm) PSU's with 120 mm fans I guess, if possible at all.
- The PSU is turned 90° and moved as close to the front as possible, with the fan facing the front.
- A hole must be made for the exhaust in one of the side panels.
- It's possible that some changes must be done for the power cable.

Alternative 2:
No holes have to be made in the side panels, but I'm not sure that this alternative will give enough room for big heatsinks.
- The PSU is moved as close to the front as possible, with the fan facing the motherboard.
- A duct must be made, which directs the exhaust air to the side vent and prevents the case from heating up.
The easiest way to solve this is to use a Silverstone PP02, but I think it takes too much space.

#0,
What heatsink do you have on your e8200? And is the way you placed the PSU the standard way to do it, and if not: could it have fitted in the top of the case?
cheers, and nice rig.

Silverstone NT-06 Lite (the fanless one)

This is the standard configuration I have used. There is no space for an ATX power supply at the top of the case however with some modification I'm sure that a PicoPSU or other small form factor unit could be placed there.

As a continuation of my spare cable annihilation programme, I have done some work on my PSU today. Here's the original setup:


The cable layout was:
20+4 pin ATX connector at 475mm
ATX12V connector also 475mm long with a spare 110mm for the second 12V connector which is not needed in this system
470mm long fan speed sensor cable.

This is what it's now like:


I opened the PSU up and cut off the cables for the second ATX12V connector so that it could be removed. I then fed the fan sensor cable down the same braiding as the remaining ATX12V connector as they connect in a similar place at the back of the motherboard. As part of this work I also took the time to remove the fan and clean it. I've left the fan guard off as it is unnecessary in this setup. I also removed the annoying labels on the cables which could get caught in fans.

The cable layout is now:
475mm 20+4 pin ATX connector
475mm ATX12V + 100mm for fan header.

This has meant I've removed 450mm of cables in this mod which has cleared things up a little further.

In total I have now removed:
700mm by disconnecting Firewire, something I have never used
175mm + 280mm in modding the optical disk power cabling to be a single SATA rather than a string of molexes plus a molex-SATA adaptor
Around 250mm from removing the extra PCI-E power connectors
250mm in changing modding the HDD power cabling to have a single SATA connector rather than 3
300mm by using a short SATA cable for the HDD rather than a massively long one
450mm by PSU cable culling/rebraiding

That's a total of 2.4m! In a small case like this that's a massive change.

The only real cable modding left would be to shorten power and fan cables.

As a nice aside to the tidyup which may have improved airflow, the missing PSU fan RPM sensing has returned! Two guesses: either the sensor cable had become dislodged from the plug previously or if the PSU fan connector itself had been dislodged somehow.

I've now done some exploration of the undervolting abilities of the 9600GT. The card runs at stock at 650MHz, 1.1V. There is also now a 'Green Edition' which runs at 600MHz, 1.0V. I thought it was a bit limp wristed so I had a go at testing overclocking capability at different voltages. The BIOS was modded each time with NiBiTor using the 1.1V, 1.05V and 1.0V settings it has, these voltages were not actually measured on the card so no idea how accurate this is. Stability testing was done with Furmark 1.6.5. I only tested at 10MHz increments as I'm not looking for the max overclock, just some idea of the zone of stability.

This card can actually crack 800MHz but with artifacts so lock ups aren't a problem.

1.1V


1.05V


1.0V


Only a 30MHz margin separates 1.1V from 1.0V. On a max overclock to base power basis, assuming P is proportional to f and proportional to V squared:



And on a frequency comparable basis:



A 10C drop in temp isn't bad and there's still a fairly big amount of overclocking headroom. The power draw will probably have fallen over 10W as well. I'm going to stick with this undervolt as I'm really not fussed about the extra 30MHz headroom I've lost. It does also make me wonder why the Green Edition has only dropped to 1.0V when the speed has also dropped to 600MHz. Why didn't they go for 0.8V? Especially when they're running the better 55nm process, unlike this earlier 65nm card.

An update:
The NT06 is clearly not the best topdown cooler available but does fit in the space available. Having looked at the specs at lots of coolers, the Scythe Zipang 2 (running fanless) seems ideal as it fits in the space under the PSU tighter than the NT06. It is larger, has more heatpipes and many reviews have shown it to be one of the better topdown coolers. Here it is sat in my system sans PSU, graphics card and hard disk trays:



It should fit with most motherboard/CPU combinations in any of the Sugo series (SG01, SG02... etc.) as well as some of Silverstone's media cases. It does sit a little further forward than the NT06 in this system. The fin arrangement is also beneficial as it allows air from the front fans to run through the fins. Other than removing the fan the only other change I made was swapping the pushpins for the bolt-through kit used by the NT06. This makes it easier to install I think although you still need the motherboard out of the case as space is so limited.

Performance wise I think it's dropped around 6C from the CPU idle temp but I don't have the time nowadays for detailed testing. As it now allows positive pressure cooling from the case fans rather than relying on negative pressure from the PSU, ducting may now be a lot more beneficial.

That SG03 is a nice case. As for cable annihilation, man that is an extreme way to say it. I guess I don't worry as long as the cables don't interfere with anything and the temperatures are acceptable.

After my Modu82+ blew up recently I ordered a Seasonic X-400 - it took 2 and a half weeks to come, dealforu.co.uk are hopeless. This now means that the system is fully positive pressure only with large heatsinks on the CPU and GPU just taking the air that is passed through them from the 2 fans at the front and exhausting out of the back. If anything this works better for the CPU cooler than before as the air is forced all of the way through it rather than being exhausted through the PSU part way along.

The X-400 is fantastic but there are a few things which for this case are a problem:
- The cables are quite long.
- The SATA connectors themselves stick out some way from the drive, this is a problem with the homebuilt HDD suspension system that I have in place as it makes it very hard to install drives and leaves pressure against the case, allowing vibrations to be transmitted.

Both of these issues could be resolved with cable modding although it would be nice if Seasonic made some shorter cable kits available as an extra.

I will try to get some pictures up at some point.