If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#21
|
|||
|
|||
Homer J. Simpson wrote:
I would expect that you are going to get a lot of memory errors or complete system freezeup or spontaneous reboot when you run memtest86. You cannot rely on the results from memtest86 when you are not supplying the correct power requirements to the memory modules. You need to test the memory on a known stable system board to determine if your memory modules are defective. Thanks for the warning, makes sense. Details below, but I'm not confident I can consider this board stable yet. One thing puzzles me. Why do the problems seem to manifest mainly at startup, occasionally at shutdown but not when Windows is running normally? Almost always when the two initial Windows screens, the one with the progress bars (or the CHKDSK screen) have just ended. I can run the box on a full load (SETI screensaver) for several minutes, no problem. Burn CDs, no problem. Occasions I might have guessed problems would occur, but not so. Does the point at which the problem occurs tell us anything about the hardware in the machine, or anything about what Windows is doing as it starts? Think about what what I've just written, it's when a screen mode changes - low res to higher. It's late, I'm tired and I readily agree I could be talking rubbish. But aren't memory and an AGP graphic card on the same PCI bus? So might lowering the AGP voltage also have a positive effect? Do they share the same power line? The card is an Nvidia GeForce FX 5200, 128MB of memory. It's just a thought, so be gentle with me if I'm talking b******s. The DDR Voltages in the BIOS, that are above the standard 2.5V, are there for the overclockers so that they can attempt to achieve system stability when running their memory modules faster than its original design specifications. Are significant gains possible? Aside from SETI it's rare I max out my CPU for any length of time. Occasional WAV or JPG editing needs it, but most of the time I'm just ticking over. What are these guys doing that they need the power? And are these systems reliable? Or is it just because they can? No problem with that, by doing so they're helping me as well. Just curious. BSODs more frequent today, various types, half a dozen in total - four just trying to start the box earlier this evening. So I increased the DDR voltage to 2.6, then 2.7 volts before they stopped. Forgot to record the Vtt value, will come back with that another time. Thank you for taking the time to help. -- Simon |
#22
|
|||
|
|||
The AGP bus and the DDR memory bus are both connected to th KT333
Northbridge chip. The PCI bus is connected to the VT8233ACE Southbridge chip. So the answer to your question is no, the PCI bus is not on and not the same as the AGP or memory bus. Notice that the AGP and DDR memory buses are both connected to the same chip. The default AGP Voltage is 1.5V, you cannot lower it, you can only raise it to 1.8V maximum by 0.1V increments. When you say that an increase in video resolution seems to cause the problem, that would point to the video card drawing more current as the video resolution increases (more video memory on the graphics card being used) causing the voltage to drop to both the video card and the memory module(s). This would indicate that you really need to get that capacitor soldered onto your board to solve this voltage regulation problem. This motherboard can run the DDR memory synchronously at 266MHz (preferred) or asynchronously at 333MHz. When overclockers increase the FSB beyond the normal 266MHz this also increases the speed of the AGP bus (above 66MHz), memory bus (above 266MHz), and PCI bus (above 33MHz). When the buses are run beyond their original design specs the digital signals on the buses start to degrade or become deformed from the ideal square wave and the system is unable to tell the difference between a logical 0 and a logical 1. It is the voltage level that determines a logical 0 or a logical 1. If the digital signal on the memory bus transitions below the Vtt voltage then this is a logical zero, if the digital signal's voltage is in the Vtt voltage (1.25V) to DDR voltage (2.5V) range this is considered a logical one. Overclockers will increase the DDR Voltage to "strengthen" the digital signal on the memory bus. This is also the same reason for raising the VCore Voltage for the CPU when overclocking. Overclockers usually overclock because they can. The performance gains are usually noticable, otherwise they wouldn't be doing it. Overclocking success and stability/reliability are dependant upon the CPU model and the quality of RAM and good cooling of the CPU. "Simon Elliott" wrote in message ... Homer J. Simpson wrote: I would expect that you are going to get a lot of memory errors or complete system freezeup or spontaneous reboot when you run memtest86. You cannot rely on the results from memtest86 when you are not supplying the correct power requirements to the memory modules. You need to test the memory on a known stable system board to determine if your memory modules are defective. Thanks for the warning, makes sense. Details below, but I'm not confident I can consider this board stable yet. One thing puzzles me. Why do the problems seem to manifest mainly at startup, occasionally at shutdown but not when Windows is running normally? Almost always when the two initial Windows screens, the one with the progress bars (or the CHKDSK screen) have just ended. I can run the box on a full load (SETI screensaver) for several minutes, no problem. Burn CDs, no problem. Occasions I might have guessed problems would occur, but not so. Does the point at which the problem occurs tell us anything about the hardware in the machine, or anything about what Windows is doing as it starts? Think about what what I've just written, it's when a screen mode changes - low res to higher. It's late, I'm tired and I readily agree I could be talking rubbish. But aren't memory and an AGP graphic card on the same PCI bus? So might lowering the AGP voltage also have a positive effect? Do they share the same power line? The card is an Nvidia GeForce FX 5200, 128MB of memory. It's just a thought, so be gentle with me if I'm talking b******s. The DDR Voltages in the BIOS, that are above the standard 2.5V, are there for the overclockers so that they can attempt to achieve system stability when running their memory modules faster than its original design specifications. Are significant gains possible? Aside from SETI it's rare I max out my CPU for any length of time. Occasional WAV or JPG editing needs it, but most of the time I'm just ticking over. What are these guys doing that they need the power? And are these systems reliable? Or is it just because they can? No problem with that, by doing so they're helping me as well. Just curious. BSODs more frequent today, various types, half a dozen in total - four just trying to start the box earlier this evening. So I increased the DDR voltage to 2.6, then 2.7 volts before they stopped. Forgot to record the Vtt value, will come back with that another time. Thank you for taking the time to help. -- Simon |
#23
|
|||
|
|||
Homer J. Simpson wrote:
The AGP bus and the DDR memory bus are both connected to th KT333 Northbridge chip. The PCI bus is connected to the VT8233ACE Southbridge chip. So the answer to your question is no, the PCI bus is not on and not the same as the AGP or memory bus. Notice that the AGP and DDR memory buses are both connected to the same chip. The default AGP Voltage is 1.5V, you cannot lower it, you can only raise it to 1.8V maximum by 0.1V increments. Please excuse my late night ramblings then. When you say that an increase in video resolution seems to cause the problem, that would point to the video card drawing more current as the video resolution increases (more video memory on the graphics card being used) causing the voltage to drop to both the video card and the memory module(s). I follow you. This would indicate that you really need to get that capacitor soldered onto your board to solve this voltage regulation problem. I would agree. I'm getting BSODs almost every time I start the box now. It's like an old car on a winter's morning. You've got to turn it over several times and eventually it will fire up. That's with DDR voltage maxed. This motherboard can run the DDR memory synchronously at 266MHz (preferred) or asynchronously at 333MHz. When overclockers increase the FSB beyond the normal 266MHz this also increases the speed of the AGP bus (above 66MHz), memory bus (above 266MHz), and PCI bus (above 33MHz). When the buses are run beyond their original design specs the digital signals on the buses start to degrade or become deformed from the ideal square wave and the system is unable to tell the difference between a logical 0 and a logical 1. It is the voltage level that determines a logical 0 or a logical 1. If the digital signal on the memory bus transitions below the Vtt voltage then this is a logical zero, if the digital signal's voltage is in the Vtt voltage (1.25V) to DDR voltage (2.5V) range this is considered a logical one. I follow you. Overclockers will increase the DDR Voltage to "strengthen" the digital signal on the memory bus. This is also the same reason for raising the VCore Voltage for the CPU when overclocking. Overclockers usually overclock because they can. The performance gains are usually noticable, otherwise they wouldn't be doing it. Overclocking success and stability/reliability are dependant upon the CPU model and the quality of RAM and good cooling of the CPU. That bothers me. I bought this board because, at Rev 1.2, I thought "That's got to be fairly stable. They'll have ironed out the bugs from the first version.". I don't want to overclock, all I want is a stable board that delivers what it promises. And at the moment this isn't. Gigabyte boards were ok by me, until recent experience. How do I know the next board I buy is going to be reliable? I guess I'm going to have to research the next purchase, rather than buy on whim. -- Simon |
#24
|
|||
|
|||
It's a good idea to check the news groups before buying. You will get an
idea of what problems people are having with particular models. Or you can post your questions or requests for an opinion on a particular board model before you purchase, especially if the board is obsolete or f it is very new. "Simon Elliott" wrote in message ... Homer J. Simpson wrote: The AGP bus and the DDR memory bus are both connected to th KT333 Northbridge chip. The PCI bus is connected to the VT8233ACE Southbridge chip. So the answer to your question is no, the PCI bus is not on and not the same as the AGP or memory bus. Notice that the AGP and DDR memory buses are both connected to the same chip. The default AGP Voltage is 1.5V, you cannot lower it, you can only raise it to 1.8V maximum by 0.1V increments. Please excuse my late night ramblings then. When you say that an increase in video resolution seems to cause the problem, that would point to the video card drawing more current as the video resolution increases (more video memory on the graphics card being used) causing the voltage to drop to both the video card and the memory module(s). I follow you. This would indicate that you really need to get that capacitor soldered onto your board to solve this voltage regulation problem. I would agree. I'm getting BSODs almost every time I start the box now. It's like an old car on a winter's morning. You've got to turn it over several times and eventually it will fire up. That's with DDR voltage maxed. This motherboard can run the DDR memory synchronously at 266MHz (preferred) or asynchronously at 333MHz. When overclockers increase the FSB beyond the normal 266MHz this also increases the speed of the AGP bus (above 66MHz), memory bus (above 266MHz), and PCI bus (above 33MHz). When the buses are run beyond their original design specs the digital signals on the buses start to degrade or become deformed from the ideal square wave and the system is unable to tell the difference between a logical 0 and a logical 1. It is the voltage level that determines a logical 0 or a logical 1. If the digital signal on the memory bus transitions below the Vtt voltage then this is a logical zero, if the digital signal's voltage is in the Vtt voltage (1.25V) to DDR voltage (2.5V) range this is considered a logical one. I follow you. Overclockers will increase the DDR Voltage to "strengthen" the digital signal on the memory bus. This is also the same reason for raising the VCore Voltage for the CPU when overclocking. Overclockers usually overclock because they can. The performance gains are usually noticable, otherwise they wouldn't be doing it. Overclocking success and stability/reliability are dependant upon the CPU model and the quality of RAM and good cooling of the CPU. That bothers me. I bought this board because, at Rev 1.2, I thought "That's got to be fairly stable. They'll have ironed out the bugs from the first version.". I don't want to overclock, all I want is a stable board that delivers what it promises. And at the moment this isn't. Gigabyte boards were ok by me, until recent experience. How do I know the next board I buy is going to be reliable? I guess I'm going to have to research the next purchase, rather than buy on whim. -- Simon |
#25
|
|||
|
|||
Homer J. Simpson wrote:
It's a good idea to check the news groups before buying. You will get an idea of what problems people are having with particular models. Or you can post your questions or requests for an opinion on a particular board model before you purchase, especially if the board is obsolete or f it is very new. Yeah, sorry, another late night rant on my part. Apologies You're quite right, shop around and ask advice before buying. BTW, two boots today and no BSODs. -- Simon |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
GA-K8VT800 memory problems | Mitch Newman | AMD x86-64 Processors | 1 | October 9th 04 06:12 AM |
A7A266: Problems with 1GB DDR Memory? | Paul Moloney | Asus Motherboards | 4 | August 10th 04 09:55 AM |
msi 648 max memory problems - what worked for me | Roch Gadsdon | General | 0 | June 7th 04 12:07 PM |
SATA or memory problems on GA-7N400 PRO2 Rev 2.0 | Yiannis | Gigabyte Motherboards | 12 | May 6th 04 05:21 AM |
What RAm for A7V8X-X? | Devast8or | Asus Motherboards | 9 | August 2nd 03 02:05 PM |