Hi all, Basically I have been having this problem since about 6 months ago. I've been trying to solve it on my own, but as I still haven't had any success, I've decided to admit defeat and ask for help. Yea big dent in my ego but this machine just HAS to work!! The problem goes like this, and I am not sure what is causing it. - When I shutdown the machine, often it boots itself up immediately - When it boots, sometimes it won't be successful - fans spin and LEDs would come on for a second of two, then stop, then after about 5 seconds or so they would try again, then fail again, after a few goes later, eventually it would boot itself up properly - It can go to S3 sleep correctly, but when I try to wake it up (by remote or power button or keyboard), it would do the above ^^^ and then when it eventually boots, it would be a cold boot instead of an S3 resume - BIOS won't remember some of the settings, particularly the date and the setting to enable/disable onboard LAN controller. If I unplug the power cord, it would loss these two (sometimes more) settings. So basically I have a machine that I cannot shutdown and has to be on 24/7, which is wasting a lot of power especially I only use this machine (its an HTPC) about 2 hours a day. And the spec of the machine: - Vista Ultimate 32-bit - Intel Core Duo T2300 (1.83GHz, Yonah core) - Aopen i945GMm-HL MicroATX board - Version 1.05 BIOS (used to be 1.07TR, but I could only find 1.05 online and I don't know where this 1.07TR is from, so I flashed it to 1.05 just to be safe) - Antec Phantom 500W Semi-Passive PSU - Corsair Value DDR2 SO-DIMM, 2x1GB, 333MHz - XFX GeForce 8500GT Passive 256MB - Two of Hauppauge Nova-T TV card (one 90002 revision, one 90003 revision) - Silverstone LC-10M case (with iMon remote and VFD disabled and disconnected) - SwitchTech s478 heatsink with 80mm fan - 3 other case fans - Western Dgital WD2500KS SATA hard drive - IDE DVD-RW Rom - Onboard Reaktek AC97 sound - Microsoft MCE Remote+Receiver - Gyration MCE Remote+Receiver - Some cheapo keyboard+mouse Originally I thought its got to be the iMon playing up (as it was connected to the power button), but I disconnected it and it wasn't. So may be the physical power button? Tested with a multimeter and it wasn't. The PSU? Bought a new one same model (v. expensive!) and it wasn't either. Hey may be the mobo is dead? Replaced the whole thing including CPU and ram... and no it wasn't. How about the gfx card? Replaced that as well. Its got to be the TV cards? Tried disconnecting them, and its still acting the same. By the way I've bought a new battery for the motherboard already to rule it out. Its killed half of my total brain cells!! Any advice/suggestion/comments wecome! Please help!!! Thanks.
Have you trued running it with out a case to make sure it's not some screw / short / anything related to that?
BIOS not saving date? Could it be that your motherboard battery needs replacement.... Try that. I think in most motherboard it's a watch battery CR3232. The battery model should be said on the package.
Hi there wywy, I’ve go a near identical setup to yours and I to cant get the S3 standby/sleep function working on this motherboard. I also suspect I have a cold boot issue also were it keeps restarting when you power it up and then finally loads windows. Did you have any luck in figuring it out buddy??? Mine does this to... But on occasions it will do that but also come up with a DOS prompt saying Vista failed to shut down correctly the last time, do you want to start windows normally, yada yada... From experience this is never a good sign, I was going to flash to an official BIOS (what’s this 1.07TR bios all about?!) but as you have done the same I can’t see it helping me I wouldn’t be all that bothered if my main rig or surfing pc did this, but with HTPC’s a well functioning S3 sleep mode is a must! Especially with vista and its ridicules load up times.
Chances are the problem is related to the BIOS. I think this link is to a BIOS for your board. Unfortunatly, I don't see any release notes with the downloads so I have no idea what else they fixed beyond the half sentence blurb on their support page. Next, I'd disconnect the LAN to rule out any wake on lan problems, and all USB devices to eliminate any possible wake signals from the USB HCs. If that fixes the issue, the you will need to change power management for those devices. If that fails, try disabling everything except KB, mouse, and video. If that fixes the problem, it could be a driver issue. I'd guess one of the devices is not entering the proper sleep state. The last move of desperation is to get a PCI diagnostic card/post code reader/port 80 card and try to follow the checkpoints. I think Award has a set of checkpoints specific to the S3 resume path, so that may help trace the source of the problem. My guess from your description is something is sending a wake signal before the system finishes shutting down. Maybe the code is in the middle of a CMOS write to indicate going to the S3 state, and information to to properly resume the system. If the write is interrupted, the CMOS checksum could be bad and the BIOS reloads defaults. Your description sounds like the BIOS tried to hand off to the OS, but failed for some reason. It is still possible the fault is with Vista, but I know MS and Intel do very rigorous testing. And if all this fails, the final thing to do is to dump the ACPI tables, and whatever ASL you can get from Windows, attach it to an e-mail to AOpen and whine until they have an engineer look at it. I would assume AOpen has qualified S3 on this platform as Windows thinks it can do S3 which means it is in the ACPI tables. Good luck!
Thanks for the PM wywy! Thanks for the informative post Splynncryth! good few pointers, I will start looking into them when I get home
Hesky, back the BIOS up before you do anything. And if you can't solve the problem no matter how hard you've (or we've ) tried, using S4 hibernation is not a bad option. PM replied to.
Hi buddy, your right S4 hibernation is pretty good, it takes a little longer to load up as expected. I have only one issue with it though, although my MS MCE remote can put it into hibernation it will not bring it back out. When my pc is in S3 sleep mode the remote can wake it up, i know the system actually falses to get back into windows when in S3, but the remote does in fact wake it up. The annoying thing is, Ive been playing around with my new Asus X38 board with my MCE remote and the damn thing needed no setting up at all to work with S3 standby, goes in and out of standby with the remote, same with my dell laptop.
If you use iMon, it can actually bring the system out of S4, because the iMon needs a physical connection to the power switch. But anyway, try the things in the PM first. And oh, do you mind if I post the content of the PMs here? Just so that the future generation at least has something to refer to. And finally, my theory - a faulty component (in my case, the TV tuner) has an intermitten short, which obviously draws too much current and the board tries to protect itself by shutting down. It is definitely not a setup issue, but a hardware issue.
Sure buddy, if you think it will help peeps go right ahead I tried a few of your suggestions and I don’t want to speak to soon but I'm almost 100% sure I have sorted it now The problem seemed to be caused by my new FSP Bluestorm PSU, I’m not sure why and neither can I explain the reason, but it simply would not let it come out of S3 sleep mode. I’m now running my old Antec NeoPower PSU, this PSU lets the MB come out of S3 sleep mode every time oddly enough the FSP PSU works fine on my old HTPC, with no S3 sleep issues at all put it back on my aopen and its no go again. Maybe the board is just a bit picky when it comes to certain components but I’m not to bothered. I can now put my new HTPC into S3 sleep mode via the remote and also wake it up via the remote Im really liking this MoDT stuff, passively cooled pc’s are just bliss Thanks again for the help chaps
This may be of little value to you, but it sounds like the PSU couldn't supply the standby current needed. Since the RAM is still active in S3, it needs power to keep it's contents. If it doesn't get enough juice, it could become corrupted. I have seen a lot of S3 problems at work, but I think this is the first time I've heard of problems caused by a PSU, I'll have to remember this. Is there a spec on the board for a minimum required standby current?
Hi, I have first hand experience with PSU+S3 problems on my old Pentium-M setup, hence the suggestion Hope it goes well for you. MODT is great isn't it! Nowadays all my PCs are MODT... one Celeron-M Yonah (car PC), one Dothan (hackintosh), one Yonah (HTPC), one Merom (desktop), and two VIA C7 (servers)... don't you just love the power consumption and the coolness!!! Now I just need to get my hands on a socket-P board for some Penryn mobile action...!
If I knew what pins on the atx plug it kept powered up I could probably isolate the issue. I'd like to send the psu back as its brand new, but like I say it works fine on other machines. If its dropping below spec and this is what is making the aopen board unhappy, if I could prove it then Im sure I could get it replaced. I just wish these modt boards were a little cheaper though, Aopens i965gmt costs 170+ pounds!! that a good £20 more than my Asus Maximus Formula X38 board cost. I guess the niche market there situated in probably forces their high cost.
I wouldn't worry too much about the price at the moment. Our 915/945 boards were pretty expensive when they first came out too. The MSI 945GT Speedster (my desktop board) was like £170 too a year ago. And if you look at the US prices, they aren't actually that expensive - just a case of rip-off Britain I think. Give it till summer, the price of the boards and processors should become more reasonable. But yea, an ITX board would be nice...
Better availability and lower prices would be great one thing I really like about this Aopen board is the fact its rammed with I/O features. I guess it’s just part and parcel of using an Intel chipset ITX is great but I always find them a little to basic and you usually end up stuck with the onboard graphics. Definitely going to continue to play about with these low power pc solutions though!
I went looking for the power supply requirements for that board and turned up empty handed. It doesn't look like the desktop division at Intel has made a MODT board. I was hoping they had and I could get the technical product spec which would have the standby power requirements if it existed. I with Intel's embedded platform division would consider some consumer level marketing. Aside from a lack of PCIe x16 slots, I have been impressed by what I have seen from their reference boards. There is a 'Xeon ULV' that is a merom. The northbridge and southbridge were consolidated into one chip and, IIRC, you could set it up to be legacy free. It took standard DIMMs, had 6 SATA ports IIRC, and plenty of USB ports. But it only something like 4 PCIe x4 slots and 1 PCI slot. Then there is the new Tolapai system on a chip. It's pentium M based, and not super fast, but it could be very interesting in something like a low power file server, or other app that isn't a CPU killer. These have some features that are a little more desktop oriented than MODT boards, but they have their quirks too. I'd be nice to see specific "low power" desktop boards designed around mobile CPUs but with a desktop oriented chipset (cut some of the laptop features, and put in some better for desktop). I havent's seen any 965/crestline based MBs out there. I wonder if many manufactures considered the MODT boards based on the 945 an experiment, and are not eager to try again. One problem I did notice when I upgraded my laptop from a Yonah to a Merom was how hard it can be to find mobile CPUs. Again, I with Intel would consider MODT more seriously.
Aopen'ss i965GMt motherboard uses an GM965/ICH8M chipset, but that's the only board I've seen so far that supports 800MHz Merom cpu's.
http://www.msicomputer.com/product/p_spec.asp?model=Fuzzy_GM965&class=ipc http://www.ibasetechnology.net/mi910.html http://www.liantec.com/product/emboard/ITX-6965.htm It seems most of these are 'industrial for factor' boards. They key may be to look for the GME965 chipset, though I need to see a side by side comparison with the gm965 and pm965 chipsets. I'll throw out a few examples of what I've seen from Intel Infrastructure Platform Division. they did a chipset, the 3100, an "MICH", basically the northbridge and southbridge are integrated on the same chip. It can run Pentium Ms, Core Duos, and Core 2 Duos, all the mobile form factors. The C2Ds are limited to the 667FSB parts though. IIRC, it could also handle a DP capable Core Duo part codenamed Sossaman, but I'm not 100% certain of that. Tyan made a board based on the chipset http://www.tyan.com/product_board_detail.aspx?pid=502 The integrated video and lack of a PCIe x16 slots are the worst shortcomings IMHO, and the support for DDR2 is limited to 400mhz according to the spec. Power consumption for the platform is really good though if the above shortcommings, the ATX form factor, and price don't scare you off first And here is Tyan's Sossaman board http://www.tyan.com/product_board_detail.aspx?pid=415. Another low power 'chipset' that looks interesting is Tolapai. It's certainly no performance king, but I think it would be interesting as a media system is couples with the right hardware. Maybe this idea could be taken firther and we see some Core architecture stuff given the SoC treatment. To veer away from MODT, there is the 5100 chipset. IMHO, skulltrail should have been based off of this as it uses DDR2 instead of the much maligned FBDIMMs. This assumes that the memory bus speed could be clocked over the 667mhz max mentioned in the spec. Again, there is a serious lack of PCIe on this, but it looks similar to the limits on the 5400 chipset. Check out the dev kit offerings for the embedded stuff. There are some neat things there that would be very cool if they were tweaked a bit to be more desktop friendly. You can find single board computers based on their mobile stuff, but these are aimed at the embedded market and often have a number of 'odd' features for desktops like compact flash boards for the primary IDE drive, additional PCI slots, a lack of high bandwidth PCIe slots like x16 slots, additional disk controllers, requirements for ECC, registered DIMMs which usally run at lower clock speeds than the non ECC stuff typical on the desktop, etc. Still, I think some of these chipsets would make really solid MODT boards.
