Windows System mysteriously shutting down

Ok, my machine must have some really advanced AI as it's started shutting itself down since last night (in protest?), just a day after I made an inquiry here about what motherboard I might consider replacing the current one with.

This is definitely going to accelerate my plan to buy a new mobo, CPU and RAM, unfortunately at a very inconvenient time. Busy and of course tight on budget during the holidays.

Anyway, I looked in Event viewer and found nothing at all around the time the machine shut itself down. It's done this at least twice in the past 24 hours. When i got up this morning it was off and it just shut down again 10 minutes ago.

This machine is a PVR and music server, so it needs to be running all the time. Something it's been able to do without an issue in the past 12 months. The only two changes I've made are to update to a new nightly of Slim Center and I've changed the UI layout presentation within SageTV.

I'd have an easier time believing SageTV has something to do with it than my minor UI change in SageTV, but neither of these really strike me as the type of change that would bring upon this kind of symptom.

I'd like to try to make sure this doesn't happen anymore until I can swap out the components. Since I'm changing the mobo I'll no doubt have to completely reinstall XP which also means reinstalling SageTV, so that's something that's going to take the better part of a day which I'll have to schedule.

Any insights?

I had a PC doing this last week and the problem was caused by an IDE cable fouling the CPU fan. I'd guess the thermistor on the mobo was detecting an overheat and signalling the OS (XP) to shut down. I have to admit I'm making an assumption about the graceful shutdown, it may have been more abrupt than that.

That's a good suggestion. Also, what about an intermittent short on the power button?

Is the PC actually fully shut down or did it get placed into standby mode? Did it shut down gracefully or did it just lose power?

Some operating systems have an option as to how to behave when the power button is pressed. See if the power button is programmed to "Shut Down" or to "Standby". If the former, set it to the latter, and observe the behavior if you leave it running overnight.

I haven't touched the case in months - it lives on a rack in my basement. But I'll check for obstructions and the like if the problem happens again.



It was fully shut down. Pressing the power button ran it through regular post and boot sequence as far as I can tell.



There was no message when starting Windows saying that it wasn't properly shut down "boot from safe," etc. when I started it up. Other than that, is there any other way for me to know? I have no idea how/why it shut down. I don't even know the exact time of the shutdown.



The power button is set to shut down and the sleep button (which i don't have) is set to stand by. I'll try this if the issue happens again.

I've reinstalled the previous copy of Slim Center I was running to see if the issue happens again. So I'd like to confirm this and then look into trouble-shooting the hardware. In the end, I just need it to behave for the next couple of weeks if I'm to replace the hardware. If I notice it's back to solid, then I can post-pone the new hardware until the new year.

You can set the power button to do "nothing", so if there's a short there, the server should keep up and running. You will have to shut down via software since that moment on, which is actually better, in my view, if your machine is supposed to stay up all times.

Also, shut down events must be triggered by something, whether that is a power button push or a software initiating the shut down process. That "something" ends up in the event log. You say that you did not find anything around the time of the shut down. Do you mean you did not find anything "unusual", or do you mean the shut down event itself was not logged? If the latter, that's not ok and makes me think of a non-graceful shutdown, regardless of XP asking you to boot in safe mode or not uopn restart.

Can you post here the events that preceed and that follow the down time? Maybe we can help with those.

From the description, it really sounds like it must be a software thing. Hardware faults don't result in perfectly clean shutdowns.

Note that PVRs don't have to be powered up all of the time to do their thing -- our MythTV box is only powered up when actually recording or when we're viewing recordings. The rest of the time it completely shuts down, using the real-time clock "alarm" feature to power on again automatically for the next recording.

It is possible, I suppose, that SageTV might also have that feature, and somehow it got activated on your box. Or more likely, some WinUpdate has enabled a different power profile than what was in use before, and the system now shuts down after an extended period of "inactivity".

Cheers

That's not entirely true. A bad fan could cause overheating which could be detected by the system, causing it to perform a shutdown. Sure, software is still intimately involved, but the root problem is a hardware problem.

If it was really that, then just check the system logs to find out more about it. Doesn't MSWin have a "event viewer" or some such syslog interface?

Yeah, but it sucks. It's obtuse and always seems to log the wrong things.

As evidenced by the fact that this seems like a software-related issue of some nature, yet he can find no information in the Event Viewer.

There's absolutely nothing in the event log related to the shut downs. The previous entries in the Event log were hours before the shut down.

After reinstalling the older version of SlimCenter the machine has not shut down. What I need to do now is see if I can find some type of crash note in a SlimCenter log file. I'm pretty confident it's what took the machine down since it was the only change I had made to any software in months.

Windows Update is disabled on this machine because I can't trust MS not to break something. And if it ain't broke, don't fix it.

I would love to run the system in this fashion, but I don't think my hardware setup will gracefully deal with all the power management requirements for this, even if SageTV did support putting the machine to sleep and waking it again (which I'm not sure it does completely yet).

I should have mentioned in the other thread, this is another requirement of a new mobo, power management layers properly and completely implemented to allow this type of scheduling to be done by the OS services.

As long as the new SageTV extender product can wake the server I'll be set. That's something I'm going to look into as well because I'm not currently using extenders. I'm looking forward to disconnecting the computer's video from the TV though since it will open up a lot of possibilities. Running the TV at HD resolutions conveniently, ability to decode HD streams without bogging down the record processes, ability to log into the computer remotely without affecting the video output to the TV and of course being able to get the PVR content onto more than one TV.

So long as the hardware supports wake-on-LAN (WOL, for use with the media extenders), the rest is just about all up to the software.

EDIT: Mmm.. I do keep forgetting that, on MSWin, functionality of hardware is frequently restricted by software and drivers, and the only way to gain features is often to buy newer hardware so as to gain access to the improved drivers/software. Ugh, great way to make money, but disrespectful of one's user base.

Pity that MythTV is still a major pain to set up initially, because it handles all of that stuff with aplomb.

SageTV claims to support shutdown and autowake, in conjunction with Windows power management. But I never did get a straight answer from anyone about the same stuff with SageTV on Linux.

-ml

Mark, to tell you the truth, the only reason I even run Windows on the PVR is for the expanded support of hardware by SageTV. Once the system has been set up it's (the configuration) not something I really want to go back and play with.

I have some future plans for the PVR as far as capture devices go, so I'm going to look into Sage's Linux support on that end as well as power management (search forums and put in a question with their developers).

On Linux, all that's really needed is a configurable call-out from SageTV: the ability for SageTV to call out to a user/vendor supplied script to schedule a shutdown/wakeup sequence.

With that, anything is possible, and not at all difficult.

On the Windows side, the impression I got from the SageTV site/forums, was that normal windows power management could be used to shut the machine down when idle. SageTV prevents the maching from appearing "idle" when it is busy recording/playing. There were no details about when/how it sets the wakeup alarm, though.

Cheers

Odd. Linux supports *way* more capture hardware than Windows for SageTV. But not necessarily "officially" supported by SageTV.

The best reason to use SageTV is that it is MUCH easier to set up and use for the most common situations, choosing from a limited selection of hardware tuners and remote controls. And the menus probably are much better configured by default than in MythTV.

MythTV is more confusing to set up (though not difficult with MythBuntu), and has a chaotic default menu structure. It is also confusing to configure certain common remote controls (such as the older ones that were included with the Hauppauge PVR-250 cards).

And I don't know about "media extenders", but MythTV is natively designed for external "viewing clients" (aka. "frontends").

But MythTV does have very good commercial flagging out-of-the box, works with a zillion tuner/capture devices, and can (with WORK) be customized for any scenario. Such as the multi-antenna setup here with the Frankenswitch V2, and the multiple tuners devoted to specific groups of channels in our config.

I (still) really like the look of SageTV, but it could never work for the situation here.

Cheers

Oh.. MythTV is also available for OSX.

Mac OS has virtually no support for capture devices. There's no capture driver standard to speak of and every vendor has their own proprietary driver architecture, none of them tied into the OS features for video acceleration and high quality display.

Of course display wouldn't be a concern with my future setup of having the PVR doing only record and stream duty. But the hardware support is still down to one or two products, none of which I care for.

From some quick research it doesn't appear that there's stable/good/final Linux support for the Hauppauge HD-PVR. This may be the capture solution I move to in the future. It's a choice between that or modified Satellite STB's that allow pulling the digital stream over USB (R5000-HD mod).

Though I may very well switch the OS on which the PVR is built, it will likely be a long time before I give up SageTV, if ever. I'm not 100% happy with some of its UI, but I have no desire to spend time making mods and redesigning that I just live with it and wait around for newer versions to make improvements. The Sage guys thankfully have a better idea of what a PVR and Media player is than a lot of others out there. Obviously not as focused or consistent as TiVo used to be, but still by far the best PC-based setup I've seen.

Their media extender hardware products pretty much seal the deal since no one else has anything like it except Microsoft.

And MythTV. Just about any little mini/nano ITX box, plus a CF card (or network fileserver), can run mythfrontend, and thereby becomes a "media extender". Ditto for notebooks/netbooks.

But MythTV is a DIY solution, not a retail product.

Cheers

On a similar topic, I recently got a Hauppauge HVR-950Q USB2 stick.

This is a *tiny* ATSC tuner (plus an analog NTSC tuner that I completely ignore), which uses the state-of-the-art XC5000 RF Tuner frontend (a DSP chip connected to the coax input), plus a current generation ATSC demodulator chip.

Performance in MythTV is spectacular -- it easily pulls in the distant WNPI-DT station that I've been targetting for over a year now. The HDHomeRun units (old and new) are hopeless there.

The exact same hardware in the 950Q is also available under other brand names (with different USB IDs), including an EyeTV product (the 2008 version only) for OSX.

We had a side-by-side comparison here the other day, with a MacBook+EyeTV connected to the same antenna feed as the MythTV+950Q at the same time.

The 950Q showed a perfect picture, the EyeTV showed some pixelation (this is a *really* weak station here).

So we swapped the EyeTV into the MythTV box, and suddenly it performed as well as the 950Q. Different "microcode" (downloadable DSP code) installed on the two computers, I expect.

So, if anyone in USA/Canada is looking for a *really good* OTA (or CATV) digital TV tuner, this one is a keeper, and doesn't take up any PCI(e) slots.

Cheers

Those products however, the netbooks included, have relatively weak CPUs and no dedicated video decoder chips. That means no decoding 1080 or likely even 720 h.264 and similarly encoded video.

The beauty of the Sage Theater (HD200 or the older HD100) products is they feature a decently capable Sigma integrated chip which handles AV decoding and things like HDMI output. The products also come in under $200 whereas any DIY solution I could put together would surely cost at least two or three times that (maybe even 4 or 5 times that) and still have the limitations I mentioned up top.

Of course one could use an off-the-shelf retail streamer like the Popcorn Hour, but then you lose the UI and integration with the PVR for things like Live TV and scheduling of recordings.

I should have one of the new Sage Theater HD200 boxes by early next year. I've just unexpectedly come into borrowing a 51" HDTV for at least 3 or 4 months, so it makes sense to try out the HD200 soonish. I won't have a live HD source until later in the spring though (I'll subscribe to Bell sat and install the dishes when it's not so snowy/cold).

I do have the HVR-1600 in my system now which has one ATSC tuner I can probably test out. Unfortunately the snow and cold outside also don't make the best environment to be setting up outdoor antennas. The funny thing about the 1600 is that currently I'm using it for a single channel, Food Network (tuning the clear QAM signal on my cable feed)

There are 15-20 DTV stations within reach of a small antenna from your home. Most of those have 1080i (full HD, and better quality than Rogers digital cable). The computer shops on College St. in Toronto sell 4-bay bowtie antennas for about $40, which are good enough to pull in most of those stations, even if just mounted on a fencepost outside your home.

Alternatively, some cardboard, aluminum foil, and about 10' of copper wire, can be combined into an even better antenna (SBGH) for attic use, with about an hour of effort.

Cheers

I'll try the antenna route of course. Right now there's no fence to mount anything on (new house, won't have a fence until next spring). The attic installation sounds ideal if a decent amount of signal still gets through. The SBGH (googled) sounds interesting and I'll have to give the plans a closer inspection. If you have a link to plans with specific building materials I'd appreciate you passing it along.

For the cardboard and foil version (works VERY well in attics), try post #83 (about halfway down) from this thread.

It can be constructed quite rapidly with pliers, a hot-melt glue gun, an old cardboard box, and some foil and wire.

All of the critical (not very) dimensions are available here.

There are even better versions on the drawing boards now, and some simple modifications can give interesting improvements in many situations (eg. use a 95mm gap between the left/right front elements).

If you were intested in meeting over the holidays (pre-xmas), I could drop by and we could build/trial one for fun some evening or whenever.

I expect to be in Toronto a week or so before Christmas, through to after Boxing Day. Just suggest a date, and supply the wine. I would bring a balun, some coax, and F-connectors / tools to make things easy.

Or not. Whatever.

Cheers

I've always found windows system log too rich, actually. No too poor. Maybe largely redundant and therefore largely useless, but never inaccurate. Usually, you need to filter a lot before you find what you need.

And, while I believe there's no evidence of the shutdown as you say so, I've never heard of a windows box that does not log shut downs unless a system crash occurs. What I would not trust, actually, is XP telling you windows did not shud down properly at next reboot, as that is often unreliable.

Did I understand correctly that you do not see any record of the following reboot either?!?

Out of curiosity, "disabled" Windows Updates means it does not update automatically and you do so manually, or simply it does not update at all?

It was fully updated when I reinstalled Windows sometime earlier last year. Since then no updates have been applied and that's the way it's going to stay. This machine is not used for email, web browsing or any other active internet connectivity. Firewall is enabled and no ports are forwarded to it from my router. The only software that gets installed on it are manual updates for SageTV and SlimCenter.

We might be able to do that. Send me a PM and I'll check what my holiday schedule is like. I'm in Milton right now which is a bit west of Toronto (directly west of Mississauga actually). Lots of wine here though.

Globe & Mail (Toronto) OTA story

Hmm- that story reminded me of this one. Specifically, the last lines of the article:

Seems that digital broadcast may revitalize OTA program transmission.

It should do that. But nearly everyone I mention it to is mesmerized by their existing Cable TV "service", and really just don't get it that paying hard earned after-tax cash, every month, is not necessary to watch really high quality television.

EDIT: Mmm.. just like software, I suppose, except even less of a difference for HDTV.

Maybe if I offered to let them pay me for the service, then it wouldn't be "free" or therefore "obviously" inferior?

Cheers

A slight bit of misdirection there, too, as nearly all of the major TV stations in the USA are ALREADY offering digital high-def content OTA for free. Now. No need to wait and panic on the final day.

Here in Canada, the deadline is sometime in August 2011.

Cheers

I dear say I never had issues with MS Updates in years on a various number of systems and various windows flavours. Rather, I had several issues due to updates not being installed. So, while I would try to reassure you on Win Updates, I am not going to question your fear that MS updates may break some of the system components, obviously our personal experiences may vary and generate different opinions.

Having said that: do SageTV or SlimCenter (I don't know them at all) use DirecX in some way? Or .NET? Or do they make any call to the OS DLLs or other OS components? I assume so. Then, should SageTV fail again, I suggest that you actually consider to run SageTV after another full (manual) system update.
On the same token: did you update video drivers and audio drivers? Drivers are the source of most windows system crashes.

SageTV is written in Java and some C I believe. It uses DirectX for video as well as BDA drivers for capture devices. SlimCenter is written in Perl and only touches minimal parts of the system.

I don't update drivers either. Everything was configured when the system was set up and that's it. Nothing should break when nothing is changed. If a new requirement is published for either Sage or Slim, then I evaluate and install the required component and any dependencies. None of the updates in the past year have required any system updates.

For me, this system is essentially an appliance, not a general purpose computer. Like running a DVD player or VCR. Something I don't want to update unless I absolutely have to.

Back a few years ago I can trace specifically my system troubles to the automatic Windows updates. The system became worse and worse until finally very unstable, strictly because of the updates that were put up by MS. I have no clue what part of them caused the issues as it was probably a combination and the way they were applied. But as I said, 99% of what's updated by MS has nothing to do with the requirements for the two programs I need to run 24/7 on that machine, so I don't risk installing them.

When I convert the system to a pure server (no video output) I will scale back some of the services as well, in addition to some of the eye candy features of the explorer UI. Unless of course I switch it all over to Linux when final/stable drivers are available for the new hardware I want to get.

Hmmm.. Charging a subscription ("service") to keep someone's OTA connection up and running. Interesting...

I will gladly get rid of my analogue cable, but I think I'll still need to hop on satellite since there's a lot of content that just isn't broadcast OTA. Depending on how the OTA experiments turn out however, it can hopefully affect which sat programming package I get. At least in theory. It totally depends on how Bell bundles/groups their channel offering.

Looks like I'm setting up a new system sooner rather than later. This afternoon my ASUS mobo kicked the bucket. As far as I can tell anyway.

My wife plugged a USB keyboard into the system which brought it down - display stayed up but it was definitely hung. No power to the USB ports as the keyboard would light up nor would the mouse.

I turned the system off at the power supply, waited a bit, turned it back on and then pressed the power button to boot it up. No display, just a voice error message (some ASUS boards have this) telling me that "System fail CPU test"

Tried unplugging all cards, memory and CPU and reseating everything. No luck. Tried a different CPU and same error. Tried a different video card. Same same.

That took maybe 30 to 45 minutes and was all the time I was going to spend.

The first unfortunate part is now I have to spend about CAD$400 on a CPU, mobo, ram and a $50 video card (need to get a plain PCI card as all I have right now are AGP boards). The biggest unfortunate part however is having to reinstall Windows. which is probably going to take the rest of the night if not half the day or longer tomorrow.

This is what I'm about to pick up in an hour or two:

Palit 128MB NVIDIA FX5200 video card (need the S-VIDEO right now) ($50)
Intel Core2Duo 2.66GHz (part E7300 socket LGA775, 1066Mhz FSB) ($155)
Asrock P43Twins1600 (Intel P43/ICH10) ($98)
Crucial Ballistix 2x 1GB DDR2 1066Mhz memory ($66)

No you don't need to spend so much time on this, follow the directions on this site and everything should just work. I used this method of installing starting with a P3 motherboard and upgrading the motherboard 4 times.

http://www.michaelstevenstech.com/XPrepairinstall.htm#RI

It only failed because I decided I was going to try and get my network adapters to be numbered 1 and 2 instead of 13 and 14.

I probably should have waited until the morning to start putting this together.

Right now I've got the board in the case, the CPU and RAM installed and the PCI video card plugged in. PS connectors are also plugged in. System won't turn on at all. The CPU fan spun about 20 degrees at one point but that's it

Now I'll need to remove the mobo to make sure there isn't a short somewhere. Apart from that I can't think of anything else at the moment. Have tried two different ATX power supplies, both known to be working.

I'm impatient... Tested some thing a minute ago.

Removed mobo from case and sat it on anti-stat bag. Without any cards in, one of my PS will allow the CPU fan to spin up and stay spun up. The PS I was just using in this same case however will not spin up the fan.

If I put the PCI gfx card in, then neither PS will make the CPU fan spin. I've tried every PCI slot (there are only 3).

The issue right now is definitely the PCI gfx card - the mobo installed in the case wasn't shorted and it spins up fine with my two PCI TV tuner cards installed.

I tested the gfx card in my Shuttle system and it worked fine - system booted while outputting video.

What a hassle. I don't have a car during the day so I can't even go back to the store until the evening - at which point I'm stuck working on this at night again.

You shouldn't do that. Most antistatic bags are conductive.

It's not actually on a bag, but rather the anti-stat foam that came on the back of the mobo.

I just checked the power supply that wasn't doing anything with this mobo by shorting pins 14 and 15 (with nothing else connected to it) and it powers up.

So now it looks like I have a working mobo, 2 working PS (only one of which will power up the mobo) and a working video card that caused the mobo not to work. Arrrgh.

Same again. Some types of antistatic foam are conductive. The device has to be completely enclosed in a sealed bag or have all its connections shorted together to protect it. Putting the board on the foam/bag isn't doing anything helpful.

You might have a bios support issue with the motherboard. Check the cpu support list on the website for the motherboard maker. They usually say in a newer version of a bios is needed to make a cpu work. Also try doing a CMOS clear as the system could be set to boot video from something other than PCI.

Oh you could also have a video card voltage issue. Newer motherboards only work with 3.3 volt PCI devices.

PCI slots are keyed to prevent physical insertion of an incompatible voltage card. Non-issue.

Thanks for the feedback. tman, the bag (which I'm not using under the mobo) and foam are non-conductive.

The CPU support list claims that all BIOS revs of this mobo support my CPU (E7300 Wolfdale, 2.66 "Core 2 Duo")

I have so much stuff to do today that I'm starting to get overwhelmed. I need to leave this until I can visit the store again and have them test it all on their bench until they find a setup that works and I can leave with a functional set of parts.

Hmm. Yeah. Take it back to the store. Thats a very strange combination of problems.

Sigh, living in a big country has many advantages, one of them the number of TV stations. Here in Zagreb only stations with national coverage transmit OTA digital - a grand total of four...

So I took everything back and waited there while some tech tested this and that. To make a rather long story shorter, here are the highlights:

When then first connected it up, it worked fine and he was telling me there was no problem. He used a PS he had behind the counter for testing. He told me I needed a new power supply because even though the mobo claims it works with 20-pin ATX, I really needed a 24-pin.

Ok, I bought a new PS on the spot (430W Thermaltake) and told him I'd like to test it with that PS before leaving the store.

It didn't work. Same problem as my own PS. Fan on CPU won't spin up and neither will PS.

tested with his PS again and it was still working that way. His PS was a 460W older Thermaltake.

Next he tells me perhaps the components in the one I'd just bought aren't high enough quality. Goes and gets a 400W Crucial. Same story, won't work. he tried different combos of ram, etc... Finally tried a different mobo. But a completely different make/model. Doh.

In the end I took an ASUS P5-something-or-other mobo instead which has built-in video (less one PCI and one PCIe slot), and returned the graphics card. The new mobo was almost $30 cheaper than the previous one and the eVGA gfx card was about $69. That offset the $50 PS I had just bought and put a bit more money back in my pocket.

The only drawback to this setup is that until my SageTV extender comes in, I can't connect the computer to my TV. It has VGA out but the TV I'm using right now only has DVI, Component, S-Video and Composite. If the extender ends up with a long shipping time I'm going to have to borrow a PCIe card from one of my friends at AMD or something. For now I have a 20" monitor sitting on the coffee table in front of the TV.

The tip of doing a Repair Install for Windows worked beautifully. I simply had to delete a few ATI drivers and software after it was completed and then install all the drivers for the new components on the Mobo (chipset, VGA, LAN, audio).

I still have one unknown device listed but so far regardless of what I've found by Googling I can't for the life of me get it recognized. I don't even know what it is. I'll post about it later because I trust the info in here to a much greater degree.

That biz with the PSUs is bizarre.. what brand of motherboard was the original (failing) one again -- ASRock ?

For DVI-D out from the ASUS mobo -- if the new mobo has integrated Intel video, then one can get/use a tiny little PCIe card that has a single onboard chip and a DVI-D output. Uses practically no power -- just provides a DVI-D connector for the integrated Intel video.

But it does take up a PCIe slot.

EDIT: the search terms for this kind of adapter are ADD2 and SVDO.

Cheers

Here is an example of one of these cards. The one I have here is half that size (nearly all of the card is dead space regardless).

The repair install is really nice.

As for your unknown device try downloading Unknown Devices from http://www.halfdone.com/Development/UnknownDevices/ it should tell you what the device is.

The program didn't find the device, but did list a string to help me find out what it is.

ATK0110 ACPI Utility

I put the CD that came with the mobo into the CD drive and ran the Intel INF Update which apparently updated something so that device is no longer unrecognized. When I first set up the system I had misplaced the CD and thought it had stayed at the store, so I downloaded all the software updates from ASUS' site. Evidently, whatever they have up on the site is missing something to cover this one item.

Its some Asus specific chip that does the monitoring. Install whatever software they use to do hardware monitoring. It should be in there.

ATK = Asus Tech I guess.

I had a similar device on my Abit board and installing the monitoring software also installed the driver.

Bruno (and everyone else),

I stumbled across a keyboard recommendation for HTPC systems, and now that I've filled my own Christmas stocking, here's the scoop!

CAD$23 keyboard w/integrated touchpad mouse, no drivers required.
Wireless, of couse, with a tiny USB dongle and 30'+ range.

Walk-in and pick-up in South-Eastern Ontario and the lower BC mainland; otherwise order online with reasonable shipping costs ($7 within Canada).

-ml

That looks like a great HTPC keyboard. When I get mine back up and running, though, I'm probably going the separate keyboard and mouse route. Granted, it's not very convenient on a couch, and it's two things to keep track of, but I just hate trackpads with a passion.

I'm looking at this one for a keyboard. Again, not very practical for touch-typing, but for typing out a few words or a couple sentences at a time, it's okay with me.

I came across this one the other day. Also quite small, and it includes a trackpad (sorry, Matt, but your dislike of trackpads is freakish) and a protective cover.

Okay, yeah, it's like six times as expensive....

Trackpad > TrackPoint (tm) style pointer

I found this one was under my tree last year. The circular track pad seems odd, until you learn that continuous scrolling, horizontal and vertical, is done by circling the perimeter of the pad. Wither it goes horiz. of vert. depends on the start point of the circling motion.

I would have been happy with that playstation kbd but didn't find any confirmation, of it working problem free, when used with a PC.

Circuit City is much more expensive.

Ya, that still sounds bizarre as hell if you're talking about using a track pad to move a mouse cursor. If you're scrolling a list or page, I can see how that interface would be fine though.

I'm waiting now for my Sage HD Theater to be delivered, which I'll use connected to my TV and let my media server do what it does best, record and serve.

The only times I ever need a keyboard right now are after having done a remote login to the media machine. Getting it to output to the TV requires clicking the login ID and typing my password again. Now I won't even ever have to do that. The video output of the media computer can stay fixed to the remote connection when I need it.

Heh, yeah.. they don't exactly advertise it!

The combo board I'm recommending works perfectly without any special drivers or anything. The only quirk (for others, not me) is that it has no "windows keys". Rather, those two positions in the keyboard matrix have duplicates of the two mouse buttons.

Cheers

I knew I shouldn't have gotten an ASUS mobo. Every single one I've ever had has been a total pile of crap.

Last night, for the first time since I set up this new system, SageTV crashed on me. It didn't bring down the system. WHen I restarted however, Windows went to a blue screen (I don't know if it did it while booting or while shutting down because I wasn't at the display).

Restarting Windows (XP SP2) in SAFE mode brought the system up. Then I shut down, turned off the power and rebooted. System came up as it should and there were no errors in any of the system event logs.

Today I noticed the system wasn't on the network. I go downstairs and see that it's in a restart loop that always stops saying Windows didn't boot properly. No matter what option I pick, including safe mode without networking, the system just blanks the screen and restarts.

Booting from the Windows XP (SP2) DVD produces a "STOP 0x000000A5" error as detailed here: http://support.microsoft.com/kb/314830

The error says that my BIOS is not ACPI compliant and that I can boot and press F7 to omit installing support for ACPI or that I should update the BIOS from the manufacturer.

The first parameter on the error is "11" and this is what that page says about it:



This is why people are willing to spend extra on a Mac. Shit like this just doesn't happen.

So now I'm about to go look for a BIOS update that for some reason was not required for every boot of the system since last December. No new hardware and no new software has been installed recently.

Macs never have a hardware fault or get corrupted? I'm impressed.

Go check your memory.

Windows updates?

I have Windows update turned OFF, so that's not the culprit.

And Trevor, I didn't say that Macs never have hardware or other faults. But the situation I'm describing is quite unique to a Windows-based system. Mac OS doesn't have a craptastic installer with cryptic messages. Let alone messages that tell you that one thing is wrong when in fact it might be something completely different.

Comparing the quality of the hardware in any Mac to what ASUS produces for their own branded MOBOs is like comparing the quality of a current BMW to a 1982 Hyundai.

Updating the BIOS didn't help with the ACPI error. It flashed/updated successfully but still the same error when booting the Windows installer.

I finally had to press F7 at bootup to omit the "complete ACPI" - right now it's progressing through the initial stages of a repair installation (the same type recommended earlier and that I used to reinstall Windows last time).

If anyone can think of a way for me to check my memory without booting into Windows and without a floppy disk I'll give it a shot. Something that can boot from a USB key would be ideal.

I believe that Trinity Rescue Kit has memtest on it. Actually, both Memtest86 and Memtest86+ have bootable downloads.

Linux, of whatever flavour you want?

Couldn't proceed with installation because Windows installer for some reason disabled my USB mouse and keyboard. Unplug/replug into any of the 10 ports on my system did nothing to change this.

Now I've plugged in an old IR keyboard that uses the PS2 ports (and doesn't require drivers). Waiting to get back to the same point in setup.

The next reboot will be to run some of the tests you guys linked, which I'm about to download now.

Yeah, definitely you need MemtestX86.

Let it run for many many loops. Sometimes subtle RAM errors don't show up until it gets warmed up.

MemTest 86+ started reporting errors immediately. In fact, after letting it go through 4 complete tests it was reporting that only 8% of results were good. AT least that's how I interpreted it - the top said PASS: 8%.

The bottom of the screen was perpetually RED (error/fail).

I'm bringing the system upstairs now to start pulling components. I don't understand how the memory could suddenly go bad in a system that runs cool all the time. Not only that, but if the memory is truly as faulty as reported, how is it that the Windows installer loads at all?

I would have a much easier time believing it's the motherboard and memory controller that have gone to crap. More finger pointing at ASUS.

Prior to running MemTest, the Windows install failed again. This time after confirming some drivers (that were already on the system and have been there since day 1). The blue screen error this time was "irql_not_less_or_equal"

As mentioned, I brought the machine upstairs to work on, next to my Mac so I can hit the net at the same time.

I've removed the only two cards in it, both Hauppauge capture cards, removed the memory (Crucial dimms) and then re-installed the memory and booted the system to the MemTest.

Right now we're on the 4th test with no reported failures yet. The only other difference at this time is that I don't have the USB nor PS2 keyboards or mouse plugged in.

This might be power supply or a heat issue. Try adding the cards back in, does it pass memtest? Close the case back up does it pass memtest?

Didn't the x86 motherboard architecture run one of the PS2 keyboard lines and the memory lines together? Maybe it's a keyboard problem or a jack problem.

Also, with the way DMA works for things like PCI video cards, it could be a problem with one of the cards.

That meant that it was 8% of the way through that pass.

Ok, done two complete passes now with not a single error. Now I'll add the cards back in.

The issue is not likely a heat issue since the system has run pretty cool since it was set up. It's using on-board video, the cards are for video capture only (PVR). It's also rare that one of the cards (the QAM/ATSC) is used at all.

Normally only a USB keyboard and mouse are plugged in - if at all. This did not change to cause the problem I first reported today. The PS2 thing was connected only to try and get past a point in the Windows install during the repair. It could have been what was causing the memtest errors which I'll know shortly.

BUt if Memtest doesn't fail, I'm back to square one, not knowing what the problem is at all.

Think of everything that changed including subtle/weird things.

For instance, simply reseating the RAM or cards could have fixed an intermittent pin connection.

Opening the case to remove the cards might have been a factor.

Once I had a problem with a whole series of identical systems, they were all reporting "out of memory" dialog boxes in Windows when our company's software was run on those systems. When I took the CRT off the top of the computer's "pizza box" style case, the errors went away. It turned out that all the systems had been built with the IDE cables for the CD-ROM drive folded across the top of the drives and the cables were being crimped between the case top and the drives. (I'm not saying that's your situation, I'm just saying it could be something as simple as moving the case around could make a hardware problem go away).

The system has been untouched physically since December and has been running without an issue.

Only updates to SageTV have been applied.

Right now I'm back with the Windows setup... And again I notice that I can't move the mouse pointer, so I know I'm going to be screwed when it asks me to click on something...

EDIT: OK, at least the PS2 keyboard/mouse work.

This time no IRQL error when installing devices. I'm now at the Regional/Language prompt. I'm going to see if I can finish this and then I'll do a repair again and try WITH ACPI...

So far so good. I had one problem device (PCI-e root hub) that Windows said it didn't have the resources for, but that seems to be OK after doing a new repair without crippling ACPI.

In the above situation I also had Windows showing that my floppy device wasn't installed properly (it was disabled in BIOS). Now it doesn't show up at all, which is what I expect (and how it used to be before).

I'm about to install AVAST since I've tested it on another system successfully. This will be the first time I've had anti-virus software running full time on a PC.

Currently in the process of applying SP3 and if all goes well I hope to have the system connected back to my TV sources and recording by 8pm.

I do wish I knew what the hell happened to cause this headache.

Also, check for Bad caps on the mobo -- Mac's can suffer from this too!

Yeah, but all the ones I've seen on Macs have been on known-bad motherboard revisions and they've all produced tell-tale smoke.

Don't you just get a sad mac when something goes wrong?


You have enough good memory that the installer can load and run. It doesn't have all the drivers for your hardware installed so lots of things are dormant. The installer itself doesn't use much memory since it has to be able to fit within the minimum system requirements. If the extra ram is used at all then it will be for caching.


Memory issues can/will cause bizarre problems or errors. If it is using that part of bad memor for ACPI then it will think something is wrong with ACPI.

Nope. The sad mac was part of the ROM on machines older then the CRT iMac. They also did away with the death chimes that would accompany the sad mac.

I think Mac folk just don't admit they have problems.

I was doing a publishing project one using Pagemaker on a PC, which would crash several times a day, sometimes taking the data file with it. I bought a Mac based on this idea that they don't fail when doing such things and found that Pagemaker crashes just as much and still thrashes the files. I've since learned to hate all things Adobe.

We used to have a Jaguar XJ6, a car which, in the US at least, has a mostly undeserved reputation for unreliability (they were never as unreliable as, say, a Chevy). One day my wife had it and it wouldn't start after work, so she went back in to call me. When her cow-orkers asked if she was having problems with the Jag, she said "No, I just forgot something".

Well, something is definitely bad in the system.

This morning Windows was stuck in a restart loop. After running last night for a good 6 hours without any visible issues.

When I did the memtest today at 9am it was coming up solid failures. We're talking thousands of failures by the time it got even part way through test #4.

I turned off the power supply, waited about 2 seconds and turned it back on. Still lots of failures, but noticeably less.

Turned off the supply again and unseated and reseated the RAM modules. This time only a handful (10) failures by the time it finished test #4.

WTF? Neither of the shutdowns waited long enough for anything to cool down. The only thing I can think of is that with a quick shut down there's still data in the memory that isn't getting cleared. But I still don't know whether to blame the RAM or the motherboard.

And here I find myself with one of the pitfalls of building my own system. Which item do I return? If this were a Dell or Apple, I'd just swap the whole system with the vendor, probably the same day.

Should I just go ahead and RMA the memory and mobo each on their own? I wonder how long this is going to take - the memory is Crucial and the mobo, as mentioned before, ASUS.

I'm past the 90 days of warranty offered by the store for the mobo, so now I have to deal with ASUS directly. For the memory I think the store offers service for 1 year - Crucial's warranty is lifetime I believe.

That phenomenon was in the news last year -- it can take *minutes* for the DRAM capacitors to drain. Apparently governments use this "feature" to extract data (such as hard drive crypto keys) from the RAM of recently powered-off notebooks.

-ml

Except that Dell would probably want the hard drive(s) back as well, which is a major PITA.

Cheers

Yeah, I used to do the same thing on the Commodore 64 to liberate code from memory.

Mmm... bad PSU ?

Can you try it with first one DIMM, then the other? Or is it one of those motherboards that requires paired DIMMs to work?

Peter

Reading through all of this again, I feel for you, man.

You really could use a second system right now, so that components can be swapped to isolate the failure. I have a couple of them here, but like that does you any good 5.5hrs down the highway from me.

We're supposed to be trying to narrow down the bad parts, not adding more questions.

The PS is as new as everything else, from December. A big Thermaltake. Any suggestions on how I might isolate that?

I can run a single module at a time, but it will run single channel instead of dual channel. The mobo can also use either memory slot when using a single DIMM. I'll go unplug one right now. But I suppose to be thorough I've got to then later swap the DIMM to the other slot to see if it's a slot problem...

I have a Shuttle system which I was just about to throw SageTV onto and then let it run my recordings for the week while I'm away, and while waiting for replacement parts. Not ideal, but I think I can get it set up quickly, since it's mostly just installing the PVR software and the drivers for one capture card.

Most of my media is on a RAID0 box, which does get me thinking that I should likely have some redundancy in the hardware as well. I'm tempted to have a spare identical motherboard and RAM for this such issue.

Here, I have a backup box (mobo, CPU, PSU, etc..). Any of the components of the Mythtv system can be swapped for those of the backup box.

Including just moving the hard drives and tuner cards from the Mythtv system directly into the backup box.

Linux/Mythtv automatically reconfigure themselves for changes like that without user intervention, driver reinstalls, or telephone calls to Microsoft, so it's an easy swap.

A buddy of mine here recently turfed his pirated SageTV setup in favour of Mythtv, despite me suggesting he was probably better off with SageTV. But he begs to differ on that point, and is really happy he made the switch. But of course, he's a non-technical user, so he doesn't know about that kind of thing.

It was especially fun to watch his expression when he later moved the hard drive and tuner card from the original old-P4 system into a faster AMD64 box, and didn't have to tweak anything for it to boot up and work perfectly. Useful feature, that.

Cheers

My only slowdown right now in setting up the Shuttle for Sage is that the system has only a singe PCI slot - which was occupied by a SCSI card to run the boot drive and CDROM. So, to clear out that slot I had to swap in a new boot drive and optical drive, which means I have to also install Windows onto that new drive.

Unfortunately I my car is awaiting some repairs, otherwise I would have gone out already and picked up another spare mobo and memory which I'd use in the original system while sending the current parts off for repair/replacement.

I ran the memtest on a single dimm for an hour and have just now put back the previously removed module and started another test round. I'll run it the whole day if necessary to see at which point in time it starts to show failures.

Everything always comes at just the right moments in time too. I leave for Vegas at 7am and should be using this time to get together stuff I need to pack as well as practicing some flash and umbrella photo techniques in preparation for some photos I need to take on the weekend. Not to mention putting up an out of office notice on my site, doing some last minute shipments and preparing some auto-responders.

I can't find it now, but I remember reading in the last few weeks about someone having issues with their Asus motherboard, dual channel ram and ACPI mode. They fixed the issue by using an older bios.

If I had just updated to a new BIOS out of the blue I would strongly be leaning toward that type of problem as well. But the system hadn't changed in that regard since I put it together in December (using the original BIOS it came with).

You can't blame a poorly written application on the OS. The difference between the Mac and the Windows (although Windows has admittedly gotten better in recent years) is that when a program crashes with the Mac, it doesn't take the rest of the system with it.

...

Um, clearly the symptom is that accessing the RAM doesn't work properly. But the question is whether that's because a RAM module is faulty, or the motherboard, or maybe even the PSU.

Peter

The biggest issue here is that it's not something that happens instantly nor can be verified right away from a cold state.

The MEMTEST has been going for some hours now (new run) and no errors have been reported yet. The memory modules are quite a bit hotter than when they're running normally in the system - just noting this by touch.

My PVR software is happily running on a different system right now, but I don't have access to all my previously recorded content because that's still siting on SATA drives in the original system. The spare machine has only PATA, so it's going to be handling only the record duties while I'm away and until I can get this memory stuff figured out.

Now I'm in a scramble to finish up all the other crap I need to do before leaving in the morning - which it turns out is more than I was anticipating. Scramble time.

Oh, well there's your problem. The universe knows when one is going away, so breaks as many things as possible to make one's life a living hell.

Tru fax.

Update: Memtest has gone through 16 complete passes and not a single error yet. That's with both modules in place and over 5 hours of run time.

WTF.

I'm planning on leaving it until at least morning, when I have to leave. Should I leave it running all week while I'm away?

Do you have a blow dryer, put it on low heat and aim it at the PSU does it start to fail?

To help reveal any intermittent solder joints on the mobo, give it a few moderate taps with the plastic handle of a screwdriver while running Memtest86 or better still XP.

Sounds like the PSU.

I'll try both the hair dryer on PSU and tapping.

What circumstances are pointing to the PSU right now? I suspect there's more energy required when running Windows, but how about the fact that it was failing memtest this AM and Windows had been failing to boot for at least 6 hours...

PSU failure cause the strangest issues. I had a PC that would run 100% fine but would fail to burn a DVD 1 out of 10 times a few months when buy and it wouldn't burn DVD's anymore. I replaced the burner and it wouldn't burn. I tested both in a different PC and they worked. A friend had a PSU Tester and it showed that my PSU was bad.

The PSU is under very little load just sitting there running memtest.

But with SageTV up and running, recording a program or two, and transcoding or commflagging them, things start to suck a little more power.

I'm betting that eventually something gets hotter than it wants to get (a failing component or whatever), and then it misbehaves until it gets to cool down substantially again.

PSU is the most likely candidate, but it's hard to tell without being able to swap in a known good PSU.

Cheers

Ok, I'm back from Vegas... Unfortunately I came down with a cold the night before leaving and it's still with me now. The trip was still cool (lots of hiking in the "mountains"), but I was only able to actually taste food 3 times during the whole week. The rest of the time I might as well have been eating salted cardboard.

I ran the memtest all afternoon yesterday and hit the PSU with a hairdryer for quite some time on/off. The PSU itself got quite hot and the air blowing out the back of it, even after turning off the dryer, was very warm for some time. No errors reported.

While heating the PSU I also knocked on it and the rest of the case, something that never naturally happens.

The test has continued to run all night and right now it's at over 20 hours without a single error.

If I didn't know any better I would say there don't seem to be any hardware faults.

What's next?

I really hope I'm not going to regret writing this later, but right now everything seems to be okay.

After 30-something hours or running MemTest and multiple sessions with the hair dryer and knocking on the case, PSU and mobo, I decided to reinstall Windows.

I formatted the boot partition, wiping all traces of the past WIndows install. So far I've installed only SageTV, TweakUI and Service Pack 3 along with the drivers for the mobo and capture cards. The system has been up since last night without any signs of issues, recording and playing back for pretty much the entire time.

Fingers crossed.

I knew I'd get hammered for saying things were ok.

Not more than 30 minutes after posting, SageTV quit out. I went to the system and it was still up and running. Trying to start Event Viewer produced an error. Starting a few other apps produced similar errors. I restarted the system via Windows shut down menu and immediately booted into Memtest.

Continuous errors reported. It was climbing into the thousands and the bottom of the screen was solid red and scrolling non-stop.

Turned off the system at the PSU for 10 seconds and restarted. Now it ran for a couple of minutes with only two errors reported. Shut it down the same way for 10 minutes and now there are no errors reported.


WTF. It seems that these memory errors only happen after running Windows (and some software) for a day or so.

What can I do here other than just buying a completely new system (and putting a fresh copy of Windows on it)? I don't even know how I'm going to file warranty claims on any of the hardware I have now because I can't prove definitively that any single piece fails at any specific time.

PSU.

Need to swap the existing PSU with another known-good one and see if that cures it or not. Got another PC in the house somewhere?

I have a crappy spare PSU here that could be used as a freebie loaner, and I have better ones that unfortunately cost money.

But I do plan to be in Toronto next weekend..

Cheers

I agree with Mark and think you should try a differnt PSU. If it isn't the PSU it has to be something overheating. Do you have large fan you could aim into the case when your geting the continuous errors in memtest?

With the capture card running (perhaps even just with its driver loaded), not to mention the other peripherals and their drivers, the system will be using more power than it does with a bare boot into Memtest. I agree with the others, it sounds like your PSU is flaking out under load.

Peter

I can see the load issue, but it's not a heat issue. The system was absolutely cool when I saw the issue and saw he errors in memtest. Much cooler than when I had run the memtest for 30 hours. And certainly much cooler than when I blasted the PSU and memory with the blow dryer.

I have another PSU, I just need to check to see if it will work with this mobo/setup at all.

TUrns out I don't in fact have an extra PSU as I had thought. I'll have to pick up a new one this week sometime.

I've also filled out an RMA a Thermaltake's site - without the advance replacement option.

I'm thinking of buying an OCZ StealthXStream 500w which can be had for $72 minus a $30 mail-in rebate.

For the TT RMA I'll have to ship the PSU to California. I'm wondering if it's even worth it, given the high price of shipping things. The RMA has a 90 day window anyway, so I have plenty of time to decide. And plenty of time to test the new PSU.

Now I'm also wondering if I should buy new RAM and/or mobo while I'm at the store for the PSU - just to save another possible drive out there later.

If you still don't have it sorted by the weekend, then I could pack a spare system in the Scoobie before we head down that way. I'm free on Sunday..

Thanks again for the offer Mark.

The replacement PSU was put into action last night. So far the system is still running. If it doesn't fail, I'll restart it myself in a couple of days and immediately run memtest. Which a few days ago would have shown at least a few errors.

I just did the restart now as I noticed that the SageTV service had quit at around 6pm.

Sure enough, by the time memory test#3 had completed there were already over 60 errors reported.

Seems like it's not the PSU.

Now I'm going to look into returning the mobo and memory at the very least, but I think I should try swapping out the processor as well. Unfortunately I just can't get any conclusive test that proves the faulty component (so I might forward it along to the manufacturer)

I wonder if any of these components can be cross-shipped. At least they're a lot lighter and easier to ship than a PSU.

I just can't wrap my head around the memory failures even if the contents of memory were still intact/corrupt upon restart. The test should be inserting new data, so the previous contents should have no bearing on the testing. This is puzzling because the longer I leave the system off, the less errors will show up (off for even close to 1 minute removes all traces of memory errors and they never come back regardless of how long I run memtest or blast with a hair dryer).

It seems like only running Windows with the other apps I run causes the memory to test bad.

If you have multiple DIMMs, run them one at a time and see if it's a particular one that's having problems.

Have you ruled out the TV Tuner card? I don't recall how it's figured into your other diagnostic tests.

Thanks for the reminder. I just realized that I only did the DIMM isolation testing when using MemTest and not while running Windows for an extended period.

I've now popped out one of the 1GB modules and we'll see how it goes in the next couple of days.

I also already come up with a plan to run Sage along with my drives on my other system if I should need to. I did this while I was away in a more cut-down fashion.

Tony, I haven't ruled anything out, it just didn't cross my mind that the TV tuner card would be the cause of the memory errors though. Especially after rebooting the machine. I ran my primary capture card in my alternate system for the whole week I was away without any issues.

I can conceivably do a card shuffle using that other system to isolate the cards. I'll run through this memory isolation first and then move onto that, likely setting up my other system for primary recording and continuing the testing on my primary system to see if I can get it to a stable state and identify any suspect parts.

So far so good. We had a crash/reboot this morning as I had hoped.

Situation so far:

Ran the machine with only DIMM#1 in memory slot #1 for a few days with no issues. Did a few manual reboots and memtests with no errors reported.

Switched over to running with DIMM#2 in slot #2 for a couple of days. At some point last night the system rebooted itself which then loaded memtest which I had left in the optical drive (on purpose). 25 errors tested within the first pass, then the errors stopped happening. It was at 26 passes by the time I checked the system this AM. I suppose I could have checked the memtest "uptime" to tell when the system rebooted.

Now I'm on to the third test where I've swapped DIMM#2 into slot #1 - this will hopefully let me know if it's a memory module or memory controller/mobo issue.

I restarted a little while ago into memtest. DIMM#2 in Slot #1 showed a lot of errors. Over 500 before test 4 was complete and it looked like the number was still rising.

I've now removed that DIMM and reinstalled DIMM#1 into slot #1 as a sanity check. I will run this module longer than I did the first time and under greater load. Once I've done that then I'll run it in slot #2 to complete the whole test pass and help make sure I'm not actually seeing two failure points (mobo and memory).

I hope you're intending to give DIMM#2 a good kicking for all the trouble it's caused you. Um, unless it's still in warranty I suppose, in which case that might prejudice your return process...

Peter

The modules are a matched set from Crucial and are definitely under warranty. Otherwise I'd likely just snap it in half.

I have to take a look at their replacement policy, hopefully they can cross-ship a new pair so I don't have to shuffle my PVR parts to my other system again.

Blah. After running seemingly without issue for quite a few days, the machine restarted tonight and showed a lot of errors on the module I thought was likely ok.

Turned the system off briefly (obviously not long enough) and it ran for about 2 hours before crapping out again.

I've since shut it down for about 5 minutes and verified that a new run of memtest isn't producing errors.

I think I'm going to RMA the memory and the mobo now.

Before heading down the RMA route I did a bit more searching... Came across a few people that also had errors only after crashes. And some other people with occasional errors.

Lots of people in these search results seem to be using Crucial Ballistix memory. Made me a little worried. More than a few also with ASUS mobos.

Anyway, based on some info I decided to check the timings and voltage of the memory. FIrst I checked to see if the options existed in my BIOS. They did, I had to turn off a couple of AUTO settings to be able to set mem voltage to 2.0 and timings to 5,5,5,15.

I then set it back to AUTO because I realized I hadn't checked what the system was reporting it had been automatically set to. Memtest reported that the memory timings were 5,5,5,18 but I didn't look elsewhere on how to check what the voltage was.

So now I'm running with the specs of the actual modules manually set for the mobo. 2.0v and 5,5,5,15. CPU settings are still set to AUTO.

I'll give this a few days. This brings me back to one of my original points. What kind of POS mobo requires you to make manual memory settings to maintain stability? ASUS=POS.

Oh, I'm also currently running Prime95 trying to stress the system with a pair of blended tests

The auto settings just uses the SPD settings from the memory module, so i think you have to blame the memory module for containing the wrong information, and not asus.
It does happen that spd info is programmed wrong (or even faked, to pretend to be faster memory)
You could try to read your spd with: http://www.techpowerup.com/spdtool/

Did you buy EPP or EMP memory by any chance? Unless your motherboard can read the extra data in the DIMM SPD EEPROM, it won't be able to actually run the memory at its rated settings without manual help.

The SPD tool says the modules are "EPP Supported" in the EPP identifier property. There's a ton of information in the report, which is (and should be) irrelevant to me.

When all is said and done, there's still the fact that this setup ran since December until last month without so much as a single hiccup. No system changes were made in that time.

If could afford it, I'd have snapped the memory module and motherboard in half already and chalked it up to experience (again). Every ASUS mobo I've ever had eventually goes bad.

My system uses EPP memory as well and it was mostly stable with the defaults. It'd work okay for months but then I'd get a random crash or two. I had to manually set it to 2.1V as per the memory datasheet for it to actually be 100% stable however.

After 5 hours the Prime95 test halted with an error in one of its test processes (it runs two at the same time, one for each core I suppose).

Neither the CPU nor the memory modules are even remotely hot. I wouldn't even call them warm.

As mentioned, I'm currently running with settings manually input as per the label on the memory modules.

Restart the system and Memtest shows a significant number of errors. It was up to 80 while on test #4 before I stopped it.

I think I'm done testing, so it's time to check out how to get the mobo and memory replaced. Though I can't see why this won't happen again a few months from now.

It is threads like this that make me glad I gave up building my own PCs years ago. Sure, the up front savings look tempting (not that the savings are as great as they used to be anyway), but the occasional complete screw up like this more that wipes it out with hassle and wasted time.

Totally agreed. I'm sure young me would call current me an old man.

No kidding.

To tell you the truth, there really aren't any significant savings to speak of. If I did't already have something and needed a system, I'd likely just buy one pre-built. The simple fact is that this time around I only needed a few parts (mobo, memory, processor and PS). I suppose those few parts are the majority of what constitutes a system though. And at some point you just have to stop using parts you already have around. I could have given up the case as it's a home theater style case, but the system is now being used exclusively as a server, hidden in the basement.

It occurs to me that back in the day you used to have to add all sorts of expansion cards to computers to get what you wanted, whereas now, most people, I think, run with very few cards at all. Chances are now that a desktop system has cards that override devices that are built into the motherboard: video, audio, network, etc.

In the days of yore, if you ordered a preconfigured system, you probably weren't going to get the expansion cards you wanted, which meant that you were going to have to discard a good portion of the system you got.

Now, though, even if you build it all yourself, your motherboard is going to have all those baseline devices built in anyway, which means that you're not overspending by buying the base preconfigured system. Plus you get the fact that it was tested as a unit and you get some support on top of that. Also, Dell, et al., are doing a really good job with cases these days.

</rationalization>

For non-notebooks, I still assemble from scratch. It's the only way to avoid having to pay for a copy of an operating system that I really REALLY don't want.

And since I purchase wholesale, the prices are quite reasonable, and I get to customize with the exact specific brands/models of components desired. Ditto for ongoing upgrades and the like.

This is particularly handy for the MythTV boxes -- an off the shelf Dell or something would simply lack the required expansion room for drives, tuners, VFD front panel display, etc..

For notebooks, I buy new/used according to what's available at the time, and obviously don't assemble those myself. The useless copy of Windows that I'm forced to pay for ends up sitting in the shrink wrap until the unit is resold years later.

Cheers

Oooh. Have you found a decent one of those?

Peter

The fancy Antec Fusion case I used for my own Myth box included a very readable 16x2 VFD as stock gear. I've programmed it to show useful status info, and it's just peachy (stock Myth can also drive it, but my own software does a much nicer job of it).

Unfortunately, Antec has since switched to using unreadable backlit LCDs in the current versions of the Fusion case. Ugh.

Cheers

Well, my argument was mostly for normal desktop computers, not special purpose stuff. You have a point.

I was going to say that you couldn't get a big discount buying parts yourself, even with the MS tax. But I just went to Newegg and spec'd out the cheapest computer I could. Less than $200. Given, it's not exactly a power house, but that's a complete, functional computer.

Damn.

Looks like you can buy a VFD from Antec that just pops into a 3.5" drive bay. Is it the same one that came with yours?

...and if you spend a whole $75 more with Dell you'll get a PC with 4 times the disk space, 4 times the memory, a RW DVD drive and a faster CPU. The time I'd spend putting it together (let alone any time I might spend troubleshooting if it didn't work) is worth more than $75 to me.

If you buy from the Dell outlet store you can pretty much match the newegg price, within $5. I had never considered the outlet store before, but used it a couple of times recently with great success.

That one does look rather similar, yes. On the Fusion case, the knob is separate, way over to the right. But otherwise probably the same unit.

Speaking of the knob -- what were they thinking?
I mean, who in their right mind wants to get up off the chesterfield to adjust playback volume ??? That's what remote controls are for!

My software simply uses the knob as a way to scroll the text display, useful for checking on the status of the machine when not actually sitting down to watch something.

Cheers

I've been doing the same recently, it's really a great way to go. Ironically, it's an even better for businesses, as XP Pro/Vista Business don't incur the $100 markup.

My latest find was an Optiplex 760 minitower with a 3Ghz C2D for $320 out the door. Small HD/Memory, but those weren't a priority for this one. Optiplexes have much better cases than inspirons and come with a 3 year warranty. I'm adding in a Intel NIC and using it as a home ESXi server.

I just now had a look at the Dell Canada "outlet store". Lots of "Sorry there are currently no products in this category" messages there.

Looks like perhaps a place to hunt for discounted notebooks, but most of them are subspec for what we require here. And he desktops there are definitely more expensive than what I can buy/build from parts.

The warranties are also inferior, typically just a few months, compared with 1-2 years on purchased parts elsewhere.

I suppose that might all vary from week to week. And country to country.

Cheers

It does, a lot. Luckily in the UK with our relatively strong consumer protection laws we are get the same 12 month warranty on the outlet stuff as we would picking the lowest warranty option on the non-outlet stuff.

You have to be careful as well, sometimes the price can be a little "random", on the UK store at least.

Incidentally, the problem I'm having with my system is getting worse. Even using the DIMM that I thought earlier might be OK, the system will no longer stay up for a full day. I suppose since the issues are memory errors, this could be totally random depending on how the memory is used at any particular moment. But so far three reboots in 24 hours without doing anything out of the ordinary.

Hm. I so expected the parts computer to be pricier, and then so surprised that it was so cheap, that I failed to finish my comparison.

Getting the specs a little closer brings me to within about $60 of the Dell. Regardless of putting it all together, the fact that Dell makes sure their stuff works together is worth that amount.

Weird, I submitted a reply to this thread and it's not showing up. Might not have finished posting and I may have closed the window...

Anyway, on Dell's Canadian site I can't see any detailed specs for any desktops. Can't see images or details about the back panel nor the inside of the machine. Don't know what ports nor slots any of the machines have.

With Dell you generally have no clue what you're buying and you can't even see or touch the systems in any stores.

I'm now waiting for replies to my RMA requests from both Crucial and ASUS. I wonder if I should RMA the processor while I'm at it? The machine will be down anyway, and I can't really be sure the processor isn't a contributor...

Dell Canada Vostro 200 product page. Click on the "Gallery" link to see a lot of pictures of the machine, including the back panel. Click on the "Tech Specs" tab to see innards details. Seems pretty through and precise to me.

Hmmm.. WTF. I could swear that Tech Specs tab wasn't on the pages I was looking at. I even remember scanning the tabs to see if there was a useful one.

I'm pretty sure the text gallery link was always there and I missed it.

Bottom line is I knew if it did exist, someone would let me know.

Still can't necessarily configure the system you want though. For instance, if you want the big Vostro case, then you're also taking a 256MB video card. If you want a Core 2 Duo instead of a Core Quad and only 2GB instead of 3GB of memory, then you're paying only $50 less ($600) than for their featured Quad system ($649).

So there are plenty of reasons and plenty of dollar savings to be had by building your own apparently. Especially if you're just wanting to build-out a case you already have in addition to some other parts. The bad part is the integration issue crap-shoot as has been mentioned. And having to warranty one item at a tine to different vendors if something does go wrong.

Ok, Crucial's memory RMA came in over a week ago. Two new modules.

The ASUS mobo came in today. Same board I sent in. No visible signs of any rework. No description of what was tested/done included with shipment nor available on their RMA status web interface. The board was in "repairing" status for over a week before it was shipped back to me.

Time to put this all back together and then burn it in again with Prime95. Fingers crossed.

Heh.. I got an ASUS board back from RMA here last week. This board was extremely fussy about RAM -- the modules would work only in two of the four slots, and then only with one particular model of RAM. It was even weirder than that, but very difficult to describe.

The board that came back may or may not be the exact same unit. It certainly has the same MAC addr for the onboard ethernet, so it's probably the same board.

RAM slots seem to work now, though. No notes or anything about what was done to it.

Cheers

I certainly couldn't tell simply by looking at the board itself whether it's the same one or not. But what gave it away was the serial number sticker.

It's all put together now the same as it was when I last ran the Prime95 tests. I'll let it run all night if I have to - last time it failed in a single afternoon.

EDIT: Prime95 running 11 hours so far. Knock on wood.

The torture test has now run for over 24 hours. I'll restart into memtest in a little while and then hopefully move the recordings and cable connection back to this machine and let it resume its PVR life.

Now I'm going to run the Prime95 for another 24 hours.

The ASUS mobo was running the RAM under-spec. It was using a clock of 399MHz (DDR2-800) instead of 533 (DDR2-1066) and the latency was 5-5-5-18 instead of 5-5-5-15.

If I set the mobo with Memory Frequency to AUTO, it will run 399MHz even though it's also set to use SPD. So I turned off SPD to set the latencies manually, set the clock manually and even set the voltage manually.

I'd really love to know if the SPD data is incorrect or if the mobo is simply not using it properly. Is anyone able to understand the output of the SPD tool sufficiently if I post it?

Is this the same EPP RAM from before? If so then unless your board specifically supports the EPP extensions, it will run it at conservative settings. Blame the tweakers for EPP. It is mainly for overclocking RAM.

What board and RAM do you have?


533MHz is DDR2-1066.

It's the same brand/model of memory as before, but new parts. Crucial Ballistix DDR2-1066. I must have written "1600" due to my newly emerging dyslexia.

The mobo is the exact same one as before. As mentioned I have no idea what has been done to it.

The memory before with all settings at "auto" weren't the same (low) as they appeared today.

The mobo supports EPP and the modules do as well. I'll try posting the output of one module below.

Yup. You've got Tweaker RAM. The SPD contains a conservative profile. The motherboard is doing the right thing at least regarding reading the standard SPD data.

Do you have to enable EPP on your board? My Abit board needed you to change the settings for it to enable all this stuff. By default it'd run everything at its listed speeds with no tweaks.

It doesn't make any sense to me that I'd have to manually configure specs or "tweak" to get the RAM running at 533MHz, its advertised speed.

I didn't buy this RAM to overclock or otherwise speed-bump any of my settings. I just want to run at rated speeds on everything. CPU running at default settings to get my the advertised 2.66GHz speed. And I assumed, pop in the memory, have the system read its settings, and it should also run at advertised/labeled speeds.

Why label the memory 5-5-5-15 2.0v and then set up the SPD with different numbers?

EPP ("SPD") is enabled by default on my motherboard.

Basically you're buying RAM that was tested by the manufacturer to be able to run at specified overclocked settings. The standard SPD data specification doesn't allow for some of these settings. Thats why it defaults to a very conservative setting. You need the EPP extensions to be able to automatically set everything. Even Crucial say that you may need 1-2 setting changes to handle EPP.


EPP isn't the same as SPD. It specifically says EPP?

No clue why it doesn't work for you though.

Tweaker RAM? Send it to a home in Arizona for a few months.

I can't help noticing that Crucial also sell plain, non-Bollistix DDR2-1066 memory -- at a higher price. The "plain" ones seem to do that data rate at 1.8V without being overvoltaged to 2V.

Peter

I'm not certain if I have to set the voltage to 2.0v to get the 533. They're separate settings in the BIOS, I haven't tried setting only the clock.

On the Crucial site, out of 23 reviews the average rating is 2/5. Lots of people mentioning errors and BSOD. Great. MOst people claiming they can run relatively error free at 400MHz.

This is starting to seem like under-spec memory being falsely advertised as higher-end parts. Not cool to advertise DDR2-800 as DDR2-1066. The funny thing is that they also have DDR2-800 Ballistix modules. Are those 300MHz parts in reality?

To get the rated speed you have to run it at the specified settings including voltage.


Online ratings/reviews are skewed towards the bad side anyway. The posters aren't running it at the specified voltage but still at the higher speed/timings.


No. You bought RAM which tested by the manufacturer to run reliably at modified settings/voltages. If you don't run the RAM at the specified settings then you won't get the rated speeds.

If you bought a regular stick of DDR2-800 would you complain that it won't run properly at 1.2V and it crashes lots? You bought a stick of DDR2-1066 that required 2.0V for DDR2-1066 operation. Pay more and you get a stick of DDR2-106 that runs at 1.8V for DDR2-1066 operation.

Or just buy el-cheapo generic DDR2-1066, no silly heat spreaders, and it will also work fine.

Pretty much. Thats what I do. You won't notice the tiny improvement you get from overclocking it and reducing the timings unless you're using a benchmark utility anyway.

That's precisely why I bought these modules. I didn't give much thought to branding. I just did a search at Canada Computers for the specs I wanted and the best price at the time. This is what was there.

As reported previously in Matt's Shuttle thread, one of the new replacement Ballistics modules died while running at its recommended (and supposedly tested) settings last month.

Yesterday night the second stick finally bit the bullet while running at even lower (safer) settings of 800Mhz.

Bottom line is that Crucial do test these parts for overclocked performance to achieve a DDR2-1066 rating, but the modules are just poor quality crap and they will all fail eventually. One needs only to read the reports on the net of widespread failure of these Ballistix modules.

Taking Crucial up on their warranty is a waste of time, money for shipping and also potential data loss when the replacements fail again. This stick is going to get snapped in half just like the last one.

They're being replaced with Kensington DDR2-800 (on-spec 1.8v parts).

I'm also thinking of replacing the mobo next month with a Gigabyte P45 that features two GigE LAN ports and a good number of PCIe and PCI slots. It doesn't have built-in video so I'll have to toss a cheap $30 video card on it. I do wish more systems had built-in video. It's a waste of slots and money to put a video card into a system that's going to be a server. Seems like everything with built-in video is MicroATX and lacking in the slot department.

I still think you've got something wrong elsewhere. You're getting an excessive amount of failures with new components.

The only other thing that it can possibly be is the motherboard. Which ASUS claims is in perfect working order. The PS has been replaced and both the new one and old one seem to be working flawlessly.

The CPU won't cause the memory to go bad. And the tuner boards (the only cards in the system) work flawlessly in my Shuttle system and have certainly not ever caused any components in that to go bad.

I'm definitely getting that Gigabyte mobo. I'm now off Crucial shit memory products and I don't plan to come back to ASUS any time soon (as mentioned earlier, every single ASUS mobo I've ever had has eventually gone bad). To show you how crap the modules are, I peeled off one of the aluminum "heat spreader" covers and half the memory chips came up with it. The adhesive didn't seem all that strong as it easily peeled off the other remaining chips.

If anyone wants the mobo, they can have it for free - just pay for however much it costs to ship it.

I'll be dropping my Hauppauge PVR-500 very soon since I'm going OTA-only for add-in cards (HVR-1600 and 2250) with an HD-PVR externally for HD satellite capture.

The new memory has been put into the computer and now I just have to repair the Windows installation which unfortunately was damaged in one of the spontaneous reboots/crashes during the night. It claims my hal.dll is corrupted or missing - luckily MS has a KB article about this very issue, so I'll get to see how well their solution works.

Repairing the Windows install was a PITA. Either the whole System32 folder was missing or it went missing after I ran chkdsk. It was instead replaced by a file with 0 byte size. What I didn't know at the time however was that chkdisk had recovered the folder and put it inside a different folder in the root.

None of the recovery steps outlined by MS would work because they weren't intended to repair the whole folder. What should have worked, using the Repair option when re-installing Windows was also not working...

At first I simply wasn't getting that option as the installer would never find a windows installation. The boot.ini was also toast and MS's previous steps didn't repair nor rebuild it. So I copied the boot.ini from my other machine which has been set up previously with the same disk names and folder structure. I also copied that system's System32 folder which was itself also a PITA which required multiple tools to accomplish.

That finally led me to being able to select the repair option during Windows installation when booting from CD. But I still couldn't get the issue fixed because the installer would disable the mouse and keyboard - both USB and PS2 models.

I decided that perhaps I'd better build a new installer disk by slipstreaming SP3 onto my WinXP SP2 install disks. nLite is what I used and it was a smooth build. The new install disk finally let me use my mouse and keyboard and accomplished the repair. After some more dancing in having to swap drives letter assignments for C and D (not as straight forward as using Disk Management) and after some driver re-installs I was back in business.

The new memory has been running in the system without incident. 4GB of Kingston value RAM. A new Gigabyte mobo has been ordered along with the new HVR-2250 and HDPVR. As mentioned, this mobo also required a gfx card, but at least it's only costing me $30. Next week the ASUS mobo goes bye-bye and I do another Install-Repair with another specialized slipstream disk I made (containing the ATI drivers for the new GFX card, all the Gigabyte mobo drivers and the Hauppauge drivers).

Fingers crossed.

New system up and running.

Gigabyte EP45-UD3P (this is one solid-feeling mobo with a lot of expansion room)
4GB Kingston memory mentioned before
New Hauppauge HVR2250 dual-ATSC tuner
Gigabyte ATI Radeon HD 4350
Plus the drives and cards I had before.

Right now I have 6 active tuners in the system. The NTSC tuner will be removed once I cancel my cable subscription and go OTA-only for a few months.

I'm quite impressed with the mobo in comparison with the ASUS. Seems better built and nicely laid out. It's also got a great amount of connections, including 12 USB, 3 Firewire, 2 GigE LAN, 8 SATA ports (6 Intel and 2 Gigabyte), 5 PCI-e slots (2 x16 and 3 x1) and 2 PCI slots.

The Slipstreamed CD I made failed to install any of the third-party (new/additional) drivers. I'm not sure why. I had to pull them off my other system's download folder to get all the new devices configured.

The only complaint/oddity I have is that the mobo brought up my CPU as 2.26GHz instead of 2.66. It was using a multiplier of 8.5 instead of 10 on a 266MHz clock. Its memory voltage setting at "auto" was also over-clocking my ram a bit at 1.9v. I manually set it to 1.8v.

Going to have to burn it in using Prime95 in the next couple of days.

I've been using that very same motherboard for about a year now. Very happy with it, never had a problem.