Sign in with
Sign up | Sign in
Your question

Screen goes black

Last response: in Memory
Share
May 14, 2011 12:18:35 AM

could it be my memory? I only have 4 GB and I have Windows 7

More about : screen black

a b } Memory
May 14, 2011 12:49:47 AM

Perhaps you could be a little more specific...
When does the screen go black? Have you tried the troubleshooting checklist that address the most common issues for "no display" or "no boot" concerns? If not, click on the link in my signature.

Some common areas for display issues are:

1. RAM - faulty, not seated properly, or incompatible
2. GPU - faulty, not installed properly, or no auxillary power connected (if required)
3. CPU - faulty, bent pins, thermal paste on pins
4. PSU - may not be sending power adequately, P4 plug not connected to mobo
5. mobo - possibly DOA or shorted

m
0
l
May 14, 2011 2:08:34 AM

I overclocked my CPU a couple of months back and it started to give me the BSOD subsequently, after using the computer for 20-40 minutes. It stopped doing that for a while until now. I just put in a new mobo and video card today. I can use the computer with the display running fine until about 20-25 minutes into my session. Could it be, A-my CPU is damaged, or B-my RAM voltage is not high enough or I am running too little RAM? I read your link. Regarding the thermal grease; you are saying that using too much can be a problem or too little or both, correct? One more strange thing about my computer is that when the screen goes black, I can't turn off the computer by holding in the front panel button. When it is first booting, I can do it then. Yet another weird thing is that I seem to have the most trouble when I have a lot of apps running at the same time.
m
0
l
Related resources
a b } Memory
May 16, 2011 10:03:11 PM

BSOD are more commonly the result of a RAM-related issue; however, a bad PSU is also a reason, as it may not be supplying constant power to the mobo. Run MemTest86+ v4.2 on your RAM, and test the PSU voltages. Compare your voltages to the values listed in the troubleshooting guide (link in my signature).

Generally, if the screen goes black, I'd attribute that to a GPU/VGA problem. If you're using a discrete GPU, remove it from the mobo and use the on-board (if you got one) display adapter. Also, download and install the following software to monitor the temps of your hardware:

1. GPU-Z
2. Prime95
3. CPU-Z
4. HW Monitor
5. FurMark

First run HW Monitor and CPU-Z. Jot down the temps, or take a screenshot and post it in this thread. Next, click on the Memory tab in CPU-Z. Double check that your RAM is being detected correctly; if it isn't, this could be the reason for your problem(s).

GPU-Z is similar to CPU-Z, except it will monitor your GPU's performance. Check the temps and jot them down or screenshot them. High temps in the GPU could be causing it to shut off. Ideally, you should try to keep your GPU temp below 75 C, while under normal load. Use FurMark to test the performance of your GPU. Note, it isn't uncommon to see your GPU temps exceed 80 C while under the load of FurMark.

Prime95 is a CPU stress testing software. At this point, running this software will likely crash your system; however, run this test and read the error log it creates if/when your system crashes.

In regards to your questions:

1. As of right now, it is too hard to determine if the problem is isolated to any one (or more) part. I'll know more once I read the results of the tests I recommended.

2. Thermal paste is, of course, an integral part of computer building. The job of the thermal paste is to add a layer of protection between the CPU and the heatsink. The CPU produces a lot of heat, and w/o the thermal paste there to protect it, the metal-to-metal contact would cause the CPU to burn. The problem is that if there is too much paste, it will act as an insulator, trapping the heat on the CPU, instead of transfering the heat to the heatsink.

On the other hand, if there is too little thermal paste, the heat will quickly burn through it and that can cause the paste to solidify and become cake-like. This is a problem because now the paste isn't doing its job of allowing the heat to transfer. In this respect, it is almost no different than metal-to-metal contact between CPU and heatsink.

The last sentence in your post above suggests that your computer crashes while under heavy load. Although 4 GB is what the average system needs, some memory-intensive apps are happier with more RAM to use. Perhaps you should list your system specs.
m
0
l
May 17, 2011 3:56:04 AM

T_T, thx for the input so far. I have done a lot of research and I am going to clean the CPU and heatsink of thermal paste and reapply thermal paste to the CPU, run Prime 95, and then borrow my Dad's multimeter and check the PSU voltages. Here are the temps:

THRM-35C
TMPIN0-35C
TMPIN1-41C
TMPIN2-38C

all CPU cores run @ 28C
GPU runs @ 48C

these are all idle temps. The RAM is a manufacturer's replacement from bad RAM. I think that the RAM company must have tested the new RAM before they sent it out, so I didn't run the memtest. Let me know if you are not convinced that the RAM is good and I will run the memtest. The RAM is running 9:9:9:20 and CPU-Z says 1.5v. I don't know what you mean by "discrete" GPU, but I had the same problem when I used the integrated, Radeon HD 4200 video. Question: how will I be able to read the error log from the Prime 95 test if the system crashes? Does it save the log and make it available upon reboot? I have posted a question about CPU_Z before about my CPU clock speed. It fluctuates between 800 MHz and 3200 MHz and mostly stays at 800 MHz. I imagine that it could mean that the CPU is damaged. Here are the specs; let me know if you need more info. :

120 mm fans plus another similar sized one on the bottom
Xion 500w PSU
AMD Phenom II X6 1090T CPU
BIOSTAR TA785G3 mobo
OCZ 1333 MHz 4 GB RAM
Sapphire ATI Radeon HD 6850 video card
250 GB HDD
DVD-ROM RW
Realtek PCIe GBE Family Controller
Broadcom 802.11g Network Adapter

The old mobo wasn't displaying the overclocking options properly in the BIOS, so I had them test it for me and they sent out a new one; I received a few days ago. I doubt that the mobo is causing a problem, since they probably did extensive tests on the new one to make sure that I didn't send them back another defective mobo. It now takes longer for the system to go to that black screen. I have run my computer for hours now with no problems under heavy load. Btw, when the screen goes black, the fans stay running. Any idea on what that means? How long should I run the Prime95?
m
0
l
a b } Memory
May 17, 2011 5:54:20 PM

The CPU and GPU temps are normal, so that's good. Personally, I would just test the RAM anyway. Without testing, you're just assuming; however, in theory, the RAM should be OK if was just sent back, but if you're wrong about the manufacturer testing for sending out replacements...

As for your RAM specs, is CPU-Z showing what they should be?

"Discrete GPU" is a term used to imply an add-on video card, and if the problem is consistent while using on-board graphics and add-on graphics, it's likely not a display problem.

Now, to answer your questions...

1. how will I be able to read the error log from the Prime 95 test if the system crashes? Does it save the log and make it available upon reboot?

P95, if you haven't noticed yet, has several windows open, depending on what test you've selected to run. In these windows, you'll see the current test being run and it's status. Should P95 encounter a problem, you will be directed to a text file. In this text file you will see the description of the error message P95 created.

2. Does it save the log and make it available upon reboot?

The file is actually created as part of the installation.

3. Btw, when the screen goes black, the fans stay running. Any idea on what that means?

This suggests a power problem, might want to check your PSU. Another common problem that causes this symptom is faulty RAM. Running the latest version of MT86+ would be the best bet for testing RAM.

4. How long should I run the Prime95?

Generally, one hour is enough to time to stress the CPU and find errors; however, some overclockers have run the test for 24+ hours to really feel comfortable.

In regards to your CPU frequency fluctuating, this is likely because of Cool n Quiet being enabled in the BIOS. CnQ adjusts the CPU speed in an effort to reduce energy consumption. If you plan to overclock, you should disable this feature.


"Here are the specs; let me know if you need more info. "

I would like more info on:

RAM - what's the model number? Use the Memory tab in CPU-Z to find out, if you can't find the model number elsewhere.

PSU - what are the load specs? PSUs should have a sticker on the side that label what the load capacites are. You should see several values in Amps (A), how many +12V values do you see and what are the values?
m
0
l
May 24, 2011 5:09:01 PM

ok, T_T, I am willing to look at the PSU as a possible problem, but I want to rule out the CPU being the problem first if possible as it may be damaged from the overclocking that I wrote about. Sorry about not getting back to you earlier; I have trouble doing this diagnostic on a daily basis. Anyway, I have a problem with Prime 95. I ran Prime95 and the screen goes blank before I can record the results. Is there a way that I can retrieve the log after I power down the computer and start it back up? You gave me a reply to this issue but I don't quite understand your answer. Also, is there a way to save the log onto a file that can be displayed as a picture, like in jpeg format? I want to show the log as a picture so that I can put it on a forum without having to take a picture of it with a camera and post the camera image. When I press the CPU option in the "options" menu, it gives me an option to configure when daylight begins and ends. Does this affect my regular usage of my machine? In other words, is this option dictating when the program will start up on its own without me manually going in and starting the program? I am wondering if the program is starting on its own. Do you have some advice for my predicament?
m
0
l
May 29, 2011 1:37:14 AM

I forgot to run the Furmark test. The GPU temps were the idle temps found on HD Monitor. Furmark doesn't run. That seems to be a clue. I have read up on PSU 12v rails. Is it possible that my only problem is that I am putting too heavy of a load (amps, I guess) on a 12v rail or group of 12v rails? If so, maybe it would help to reconfigure which connectors are being used for what component. Maybe buying a better brand with a higher wattage could help. I read that having too much wattage on a PSU can be a problem, too. I read much of the detail about this on one of the suggested reading websites from inside your signature link: http://www.playtool.com/pages/psumultirail/multirails.h... . I am not sure which kind of power supply I have. I have read a lot of the detail but in the interest of time, I thought that I'd ask you about this. Is it possible that, based on the kind of PSU I have, I could be using too many connectors on the same 12v rail and thus drawing too much ampherage on a single connector? Could this problem be solved simply by buying a PSU with more wattage? Here are the PSU specs:

+3.3v 23 amps
+5v 28 "
+12v1 17 "
+12v2 18 "
- 12v .6 "
+5vSB 2.0 "

I am not sure if the specs show that I have 2 12v rails. I am confused whether the + output means AC or DC and - means AC or DC or if it has nothing to do with it. In any case, if I have multiple rails, that seems to make me wonder which of the 3 kinds of PSU's I have. Maybe that doesn't matter, but only that I may just be using the wrong components on the wrong rails or too many components on a single rail or both. Maybe you could clear this up for me.

Side notes: I have no alternate CPU to swap with. I believe that I was having the same problem when I switched to the integrated graphics. Something that is strange to me is that when I play a components intensive app. such as a new game called The Witcher 2, sometimes I can play it for hours and sometimes I can only play it for a couple of minutes before the screen goes black.

My RAM's model # is OCZ3SR1333LV4GK
(OCZ Signature 4GB (2 x 2GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666))

The RAM passed the memtest, but I only ran it for one test. Do I need to push the F1 "options" button to test it for longer? Do you want me to run it longer?

-12v reading is -6.34v; is that normal? My CPU is designed to run @ 125W, if I am not mistaken. HD Monitor shows it is running @ 139.2W . Is this normal? GPU-Z shows GPU voltage is 1.15v. Is this normal?

Could a 650 Watt PSU be too much wattage?

Can you tell me why a CPU can be damaged or a crash will happen if there is not enough voltage applied to it? The only thing that makes sense to me as to why it would crash is that there is too much voltage being applied to it. I know that you don't don't educate people on everything under the sun about computers, but if its not too long of an answer, I would appreciate it. I just thought I'd ask.



m
0
l
May 29, 2011 8:31:00 PM

I want to buy a Dell supplied PSU because they can finance my purchases. They have a Corsair 650W PSU that I am interested in, but I am clueless about some of the specs. Will someone please look in the tech specs about the power output connectors details on this link and tell me if it will suffice for my build?

http://accessories.us.dell.com/sna/productdetail.aspx?c...

My problem is that I have a 4 pin connector for my power supply going to my mobo and I can't tell whether or not the 4 pin mini power connectors are the ones that go on my mobo, separately from the 24 pin with detachable 4 pin connector. What is each connector for? I want to find out I it has enough of the oval shaped, 4 pin connectors for my build. Ultimately I just want to know if this PSU would work fine with all of my components combined.

My problem is that I have a 4 pin connector for my power supply going to my mobo and I can't tell whether or not the 4 pin mini power connectors are the ones that go on my mobo, separately from the 24 pin with detachable 4 pin connector. What is each connector for? I want to find out I it has enough of the oval shaped, 4 pin connectors for my build. Ultimately I just want to know if this PSU would work fine with all of my components combined.
m
0
l
a b } Memory
June 1, 2011 12:53:52 AM

dragonfly522 said:
I forgot to run the Furmark test. The GPU temps were the idle temps found on HD Monitor. Furmark doesn't run. That seems to be a clue. I have read up on PSU 12v rails. Is it possible that my only problem is that I am putting too heavy of a load (amps, I guess) on a 12v rail or group of 12v rails?


Yes, it is possible to overload a PSU. Most, if not all, reputable brands (OCZ, Thermaltake, Antec, Cooler Master, Corsair, to name a few) have a load balancing feature. This feature allows the load to distribute among the +12V rails.

dragonfly522 said:
If so, maybe it would help to reconfigure which connectors are being used for what component. Maybe buying a better brand with a higher wattage could help. I read that having too much wattage on a PSU can be a problem, too.


As mentioned above, a good quality, name brand PSU would be better, as the PSU will likely have a load balancing feature. With such a feature, you don't have to worry about how you connect devices to the PSU. As for too much wattage, the person who posted that crap is full of... well, you know what I mean. A PSU only distributes what is necessary.


dragonfly522 said:
I read much of the detail about this on one of the suggested reading websites from inside your signature link: http://www.playtool.com/pages/psumultirail/multirails.h... . I am not sure which kind of power supply I have. I have read a lot of the detail but in the interest of time, I thought that I'd ask you about this. Is it possible that, based on the kind of PSU I have, I could be using too many connectors on the same 12v rail and thus drawing too much ampherage on a single connector? Could this problem be solved simply by buying a PSU with more wattage? Here are the PSU specs:

+3.3v 23 amps
+5v 28 "
+12v1 17 "
+12v2 18 "
- 12v .6 "
+5vSB 2.0 "


Yes, it is possible. According to your posted specs, you have a combined total of 35A, but if these +12V rails don't share their loads, it is possible that the PSU isn't good enough to support your GPU. Just so you know, your GPU uses ~12 Amps at reference speeds, but depending on what you will do once your system is stable, you may even requires ~15 Amps to use the card effciently. If you plan on doing Crossfire, you'd be using ~20 Amps.



dragonfly522 said:
I am not sure if the specs show that I have 2 12v rails. I am confused whether the + output means AC or DC and - means AC or DC or if it has nothing to do with it. In any case, if I have multiple rails, that seems to make me wonder which of the 3 kinds of PSU's I have. Maybe that doesn't matter, but only that I may just be using the wrong components on the wrong rails or too many components on a single rail or both. Maybe you could clear this up for me.


Yes, you do have 2x +12V rails. These specs do not reference AC/DC at all.

dragonfly522 said:
Side notes: I have no alternate CPU to swap with. I believe that I was having the same problem when I switched to the integrated graphics. Something that is strange to me is that when I play a components intensive app. such as a new game called The Witcher 2, sometimes I can play it for hours and sometimes I can only play it for a couple of minutes before the screen goes black.


This suggests that the problem is either in the RAM or GPU. Considering that the problem is random, it is hard to determine exactly which. This why running MT86+ on each stick, individually, for 3-4 passes (or more, if you can afford the time) would help.

dragonfly522 said:
My RAM's model # is OCZ3SR1333LV4GK
(OCZ Signature 4GB (2 x 2GB) 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666))

The RAM passed the memtest, but I only ran it for one test. Do I need to push the F1 "options" button to test it for longer? Do you want me to run it longer?


Yes, run the MT86+ tests longer.

dragonfly522 said:
-12v reading is -6.34v; is that normal?


PSU tolerances are ~5-10%, anything beyond that margin is considered defective.

dragonfly522 said:
My CPU is designed to run @ 125W, if I am not mistaken. HD Monitor shows it is running @ 139.2W . Is this normal?


Are you still overclocking?

dragonfly522 said:
GPU-Z shows GPU voltage is 1.15v. Is this normal?

Yes, this is normal for stock (No OC) values. However, like a CPU, the GPU can also be overclocked, which requires voltage increases, too.

dragonfly522 said:
Could a 650 Watt PSU be too much wattage?


The wattage a PSU has isn't as important as the load (amps). So, asking about a PSU and only mentioning the watts isn't going to help any knowledgable person at all. Ideally, posting a link or the sticker specs would be the best way to determine if the 650W PSU in question is enough.

dragonfly522 said:
Can you tell me why a CPU can be damaged or a crash will happen if there is not enough voltage applied to it? The only thing that makes sense to me as to why it would crash is that there is too much voltage being applied to it. I know that you don't don't educate people on everything under the sun about computers, but if its not too long of an answer, I would appreciate it. I just thought I'd ask.


With current technology, I don't see how the CPU can be damaged by undervolting. Simply put, if there isn't enough power going to the CPU, the CPU won't turn on. The result of this would be no display seen on the monitor, with the exception of "no signal" or some other message indicative of no display coming from the computer.
m
0
l
a b } Memory
June 1, 2011 1:02:44 AM

dragonfly522 said:
I want to buy a Dell supplied PSU because they can finance my purchases. They have a Corsair 650W PSU that I am interested in, but I am clueless about some of the specs. Will someone please look in the tech specs about the power output connectors details on this link and tell me if it will suffice for my build?

http://accessories.us.dell.com/sna/productdetail.aspx?c...

My problem is that I have a 4 pin connector for my power supply going to my mobo and I can't tell whether or not the 4 pin mini power connectors are the ones that go on my mobo, separately from the 24 pin with detachable 4 pin connector. What is each connector for? I want to find out I it has enough of the oval shaped, 4 pin connectors for my build. Ultimately I just want to know if this PSU would work fine with all of my components combined.


Last paragraph removed from quote due to redundancy

The 4-pin connector that you would attach to the 20-pin connector is shaped in such a way that it will fit only the main power slot on the mobo. Similarly, the 4-pin (P4) CPU connector is shaped in its own unique way so that it fits into the CPU connector slot. On top of that, if you look closely, the 4-pin connector that is used with the 20-pin connector has a couple of notches on it to help secure it to the 20-pin connector.
m
0
l
June 1, 2011 2:05:55 AM

Thanks very much for helping me out. All of your answers are very informative. I have my CPU at stock settings and they have been that way since months ago when I stopped overclocking because I was having stability issues with my computer and I couldn't seem to get even the smallest overclocks to work. I wasn't using safe practices, either, even though I read up on the subject for over 40 hours. I know what to do now the next time I decide to overclock. I decided to go with the Corsair HX650W PSU because it is high quality for what I need and it isn't terribly expensive. I'm also buying a new CPU, because I think that I must have damaged it since I didn't have any problems before the overclock, if I remember correctly.
m
0
l
!