Intermitent shutdown and alternating tones - HELP!!!

B

Beliskner

Weaksauce
Joined
May 1, 2003
Messages
72
Hi everyone,

Recently I purchaced an eMachines W3050 after accepting the reality that I just don't have the time and money to build the machine I want, plus my poor laptop of 5 years is really starting to show it's age. So I figured I'll get this for now, build my PC over the next year and then give this one to the kids.

Now I'm running into a big annoyance with it, I am experiencing intermitent shutdowns followed by an alternating tone (Sounds like european police sirens). I have to unplug the PC just to get it to stop so I can reboot the PC. (Who was the f*n genious who decided to remove the power switch on the power supply?) Anyway I need to find out what this tone means. eMachines support sucks ass and are impossible to get ahold of.

Currently I have not added/removed any hardware and the only software related stuff I've done is added 2 user accounts, uninstalled McAfee and Norton and installed PC-Cillin's Internet security and a couple of games. I also ran Windows updater and updated the system to SP2 and updated DirectX to 9c. I also tried updating the video drivers and thought that was the cause however it still happes without the new updates. (I rolled it back to the day I first fired it up.) I have read that some other eMachine models have experience shutting down due to heat and one of the fixes was adding more thermo grease and loosening the CPU heatsink. Apparently it was on to tight and causing a "hot spot". I haven't checked this out yet but is on my list of things to try. Right now I'm trying to find out what that tone means. If anyone know what it may be or can direct me to a site that describes these tones I would really appriciate it.
 
Well overheating was my first guess. I'm going to load MBM tonight and let it run and see what happens.
 
it might be a faulty PSU.

HP has the same thing in some of their systems where when the PSU is going wonky it makes an odd alarm sound.
 
If it's a heat problem, you can try to lap the heat sink. Just pull it off and use fine grit sand paper to smooth it, then use some artic silver 5 to reinstall the heatsink.
 
Rumrunner said:
If it's a heat problem, you can try to lap the heat sink. Just pull it off and use fine grit sand paper to smooth it, then use some artic silver 5 to reinstall the heatsink.
Click to expand...

Depending on how big the stock HS is, I may end up just getting a new one. I'm thinking about just getting a new HS and a duct system this weekend and give that a try. I loaded up MBM last night but I don't think I have the right board selected because it was telling me that the CPU was at 30 degrees.
 
I agree about the heat issue, but there's one other possibility that hasn't been mentioned yet. I know that Abit's AV8 board will make the exact same siren sound if the fan fails, or fails to be sensed.

Perhaps the fan's RPM is not meeting the minimum expected level. I'm not sure about the BIOS that eMachines has, or if you can even see temperatures/RPM settings, but if there is some kind of FOC (Fan-Off Control) you might want to try disabling it and see if it continues.
 
Ok, according to the online documentation I can find, it looks like the board is a K7MNF-64/1.1A

As for the manufacturer I haven't been able to locate that info yet. It looks like an ASUS number but I'm not an expert. :D So if anyone know who makes that board and what setting I should use in MBM that would help a lot!
 
_Durandal_ said:
I agree about the heat issue, but there's one other possibility that hasn't been mentioned yet. I know that Abit's AV8 board will make the exact same siren sound if the fan fails, or fails to be sensed.

Perhaps the fan's RPM is not meeting the minimum expected level. I'm not sure about the BIOS that eMachines has, or if you can even see temperatures/RPM settings, but if there is some kind of FOC (Fan-Off Control) you might want to try disabling it and see if it continues.
Click to expand...

You might be on to something there Durandal, The PC is really quiet, to quiet I think, and I have often wondered if the fan is even working. I think I need to pop open the side and just watch the inner workings for an evening.
 
Saw your mention in another thread and took a look. I'm inclined to agree. Sounds like a clogged up CPU fan to me alright. What I'd suggest you do is run the heck out of Prime 95, then, reboot and look in the system monitor in the BIOS itself. CPU fan failure would explain your problem EXACTLY and I've heard a similar sound when I nerfed two CPUs in mine.

PS. According to google, the K7MNF-64 is a FIC (brand?) nForce 2.

A power supply could have a similar effect mind, just usually not to such an extent. Still, you can test this by opening the side of the case and plugging in another PSU (eg, you don't have to actually try to squeeze it into that proprietary shaped case, just set it somewhere safe and plug it in directly.)
 
Nazo said:
Saw your mention in another thread and took a look. I'm inclined to agree. Sounds like a clogged up CPU fan to me alright. What I'd suggest you do is run the heck out of Prime 95, then, reboot and look in the system monitor in the BIOS itself. CPU fan failure would explain your problem EXACTLY and I've heard a similar sound when I nerfed two CPUs in mine.

PS. According to google, the K7MNF-64 is a FIC (brand?) nForce 2.

A power supply could have a similar effect mind, just usually not to such an extent. Still, you can test this by opening the side of the case and plugging in another PSU (eg, you don't have to actually try to squeeze it into that proprietary shaped case, just set it somewhere safe and plug it in directly.)
Click to expand...
I'm gonna check that stuff tonight and thanks for the info on Prime 95. I'll run that tonight.

Great, FIC doesn't even list my model on their site! LOL i get the board that the "interns" designed! :D

Don't have access to another PSU so i'll have to wait on that, I wondered about that as well.
 
Did some more thorough googling and found this thread: http://www.rhcf.com/sisubb/ultimatebb.php/topic/14/171.html

According to one person there, that board is most likely an OEM version of this: http://www.fic.com.tw/product/motherboard/amd/k7m-nf18g.aspx Try that in MBM if it lists it. If not that, maybe see if there is some generic nForce 2 chipset listed.

The Prime 95 trick -- beyond the obvious stability test (probably will report errors when the cpu gets too hot before it actually starts rebooting and such -- is that if you run it for 10 minutes or so to get the CPU nice and hot, then quickly reboot and jump into the system monitor part of the bios as quickly as you can, you might get just a VAGUE idea of what your CPU was running at. Unlike using an uncertain chipset choice in MBM5 or similar programs, the BIOS utility should be semi-accurate (not perfect, but, better than an incorrectly configured MBM.) Don't really mean for you to run it overnight or something as that won't really test anything. In fact, under the circumstances, I'm recommending you keep the system off when not in use, just in case it may actually be the CPU. If it runs too hot too long, permanent damage could occur. Once the system reboots and freezes, the CPU will be minimally used, but, without the fan running -- if it's not running -- it could still get hot enough.

On that subject, have you tried just leaving the case opened to keep an eye on things? If the CPU fan completly stops, it won't exactly be hard to recognize.
 
I found that thread as well and posted a message there. Thanks!

I'm going to email the company tonight and ask them if those two boards are the same.

My plans for this evening are as follows...

1. Set the box on my desk and open the side panel.
2. drink a beer.
3. See what the innards are doing and determine if it's a fan.
4. drink another beer
5. If not a fan then load Prime 95 and do as you suggested.
6. have another beer.

I'm not to worried about the system going belly up since it is brand new and under warranty. I haven't put to much on there yet just for this reason. this all started the second day I had it out of the box.
 
Well I had left the PC running all day just to see if it would crash and it was still on when I got home! So I removed the side panel and began to observe the fans (CPU and PSU) Both were operating just fine and there didn't seem to be any excessive heat coming from inside the box once I opened it. After running through a few things like downloading mail and launching Firefox I decided to adjust the wallpaper and then it crashed. Both fans were working just fine when it crashed however there was a red light on the MB with a D21+ underneath it. (If that means anything to anyone) So, bieng the tester that I am I (foolishly and knowing exactly what could happen!) touch the PSU and it was cool, I then touch the CPU heat sink and it was cool to the touch. So I think heat is not my problem. I'm now leaning torward either a PSU, graphics, or memory issue. I have noticed that it will usually crash after I have been doing something relating to graphics like gaming, changing my desktop, ect.

Thats what I have so far.
 
Hmm, things like desktop are actually more memory than video believe it or not. It's just a 2D image, using none of the abilities of the card itself (other than just raw drawing speed.) I'm more inclined to say take a look at that memory as a possibility then. People get bad chips all the time, and it's possible you might have (especially if they set them too fast in the bios or something.) No harm in testing it anyway.

First up, the obvious thing to try is the famous Memtest86+ program (which just about anyone here will recommend when it comes to testing memory.) You can find it here: http://www.memtest.org/ Just get one of the bootable binaries -- CD ISO or floppy, whichever you prefer, they are both the same. I'm pretty sure this is better than the binary executable since it should have more direct access and such. Run that for a couple of hours or so. If it reports no errors, step up to the CPU+memory+heat test, Prime95 (set it to the blend mode in the benchmark settings so it uses lots of memory.) The thing is that Memtest will tell you more directly if an error is the memory itself or not (still not 100% perfect as it's theretically possible for an unstable CPU to cause an error to report, but, with next to no CPU usage, it would have to be pretty unstable in most cases.) If you have a faulty chip, it will probably report an error at the same address(es) every time, but, if it's timings or something then it will get errors seemingly at random. While you have it up, look at your timings and speeds to make certain they are what that memory should be running at.

Prime95 on the other hand will test the CPU and memory as well as getting things nice and toasty for a very thorough test under extreme conditions. Most people feel that if their system is Prime95 stable overnight, you just can't blame CPU or memory for any problems. You kind of need to let it run at least a couple of hours as a slight instability of a nature that could cause reboots and such may not show up for a good while (such as when I run my memory at 200x2MHz. It takes apx one hour and ten minutes to finally find an error.)

Finally, when you want to test stability of your video card, the best thing I could really come up with is running rthdrble for a good while and then closing it and switching to artifact tester 5 as quickly as I can (idea being to give it as little time as possible to cool down) and start up it's test. The reason I do rthdribl first is because artifact tester 5 is outdated and, IMO, not the best tester in the world. Thing is, you won't always notice an artifact on rthdribl, so it's better to not just assume it's stable but to let AT5 check too. I always set all the options as high as I can in rthdrble and find a resolution just under where it errors out with the warning it's out of video memory so that I can push my video card as hard as possible, hopefully both in memory and GPU.

Good luck. I hope this little regime helps you narrow the problem down some little bit at least.
 
The issue seems to be with the RAM.

I ran Memtest86 like you suggested and it found 48 errors on the first pass. It would also find the same errors on multiple passes. Now to determine if it's the stick itself or the bus. I'll swap the stick over to the second bus and rerun Memtest again tonight.

Thanks again!

Oh and for a little laugh, I have finally recieved a reply back from eMachines. Now before you read this keep in mind that when I originally sent off my support request I had included the specs of the machine, the exact model number (W3050), when the issue started, what changes I have made (hardly any since it started the second day I had it), and that I have done a full restore from the backup on the computer. (No restore disks came with it.) So sit back and enjoy a good laugh!

Dear Steven:

Thank you for contacting eMachines.

I noticed that you are having issues with your T5026 eMachines
computer shutting down unexpectedly.

I know that this has caused you inconvenience and I realize how
frustrating this can be.

This situation could be environmental, hardware or software related.

This routine will make it so that Windows will show an error instead
of just shutting down unexpectedly when a problem is encountered.

You need to be logged into an account that has ‘Administrative’
privileges to do the following routine.

1) Click on Start.
2) Right-click on My Computer.
3) Click on Properties.
4) Click on Advanced tab.
5) In the Startup and recovery area, click the Settings button.
6) Inside System Failure area, take the check out of Auto Restart
then just click on OK.

This will cause your computer to come up with an error. You will
want to search Microsoft's web site for this error or email us back
with the exact error message and we will search for you.

The easiest way to tell if it is software is to restart the computer
in MS-DOS mode.

Put the Recovery CD in the CD-ROM drive and restart the computer. At
Microsoft Windows Startup Menu, choose to boot to the command prompt.

Leave the system on this screen overnight and follow the instructions
below based on one of the two outcomes and solutions listed below.


1) If the system does not restart, then you probably have a software
problem. You may consider restoring your system with the recovery
CD. This will verify that there is no software corruption that is
causing the issue.

To run a full restore:

Disconnect everything from the tower except the monitor, keyboard,
mouse, speakers and power cord. Do not leave any cable or DSL
modems, scanners, cameras, etc connected to the tower.

Backup any important data that you cannot afford to lose and have
added to the computer since taking it out of the box.

Turn on the computer and repeatedly tap the "F11" key until the
System Recovery menu is displayed

Full System Restore (Destructive)

**WARNING** This option will format your hard disk and reinstall all
factory-shipped files. Use this if all other options fail to restore
the machine to satisfactory condition. All your personal data files
and applications will be lost.

1. Select this option and click on Next to begin.
2. The System Recovery process will take a few minutes, depending on
the size of the recovery image.
3. After the restore process completes, remove the System Recovery
Disc from the CD/DVD drive.
4. Click Restart to exit System Recovery.


2) If it reboots during the night, then it is either defective
hardware (most likely the power supply) or there is a problem with
the power outlet to which the system is attached. You may wish to
test it on another outlet.

If you are not confident enough in doing the above steps, you can
also have a one-on-one chat session with us for more interactive
troubleshooting and it is available Monday through Friday, from 8 AM
to 5 PM PST. Just browse our website at www.emachines.com and you
can find more information regarding our support in our Support
options section.

I hope this information is helpful to you in resolving the situation.

Should you require further assistance with your eMachines system,
please feel free to send another message. Your comments and
suggestions are welcome.

Sincerely,
Bernard

eMachines recommends you frequently backup data files off the system
to prevent data loss.

eMachines now offers one and two year Extended Service Plans (ESP).
For more information or to purchase an eMachines ESP**, please call
us at 877-566-3463 or visit us on the web at www.emachines.com/esp.
**Currently not available in Puerto Rico, Arizona or for Canadian Notebooks.
============================================================

eMachines authorized replacement and upgrade components are available
by calling 1-888-368-7898 or visit us on the web at
http://parts.emachines.com/emachines
 
I'm not even a tiny little bit surprised. That is why I try always to find more direct means of contact, such as the "one-to-one chat" they mentioned. I find most support sections of companies to be next to useless. I can top you btw. I once had troubles with a soundcard. It wouldn't play sound on my right speaker (or maybe it was left -- and this was back in the days us normal people had 2.1) so I sent an e-mail to creative's support contact with all the information they requested and some VERY thorough descriptions of the problem and all steps I had taken to try to fix it. I explicitely told them I had tried things such as using it in DOS, carefully checking that the mixer was correct, plugging other devices such as headphones into the card to ensure it wasn't the speakers, changed irqs, etc. They promptly (two months later) sent me back an e-mail that told me to do each of the things I had explicitely told them I did.

I threw the card away and bought a Santa Cruz.

BTW, interesting thing about that e-mail. At first, when I saw the stuff about restarting in DOS mode and all, I presumed the model they mistakenly thought you said you had was an older one running Windows 98 or something. But, then I noticed the stuff about system restore and all and how they explain to you how to turn the BSODs back on (stupid microsoft, if it still has an error, just hiding the error from you fixes nothing...) Lol, there is no reboot to dos mode. Now, maybe their recovery disc (which maybe comes with the other model) is bootable, but, they don't explain that clearly at all, and the average joe won't know what to do.

Anyway, if you get the errors at the exact same addresses, it's definitely the memory. CPU would cause errors to occur somewhat randomly or constantly once they start, not start and stop at the same place each time. Go ahead and try one stick at a time to see what happens, but, I'd have to say that there isn't exactly much room for doubt here. Normally I'd say buy a new stick of memory, but, since your machine is in warantee, I guess you have to get it all packed back into it's original box and have them replace it to save yourself rather a lot of money.
 
You must log in or register to reply here.
Back
Top