I have been struggling with stability problems on my Soyo Dragon Plus! motherboard for months. While there were some DMA issues (copying from master to slave was particularly unstable), after weeks of experimentation (changing one BIOS setting at a time, adding and removing one hardware component at a time), I determined that the system was fairly stable with the DDR clocked to 100 (200) MHz and very unstable at 133 (266) MHz, independent of all other settings.
Soyo tech support was unable to help me and in some cases quite rude ("why don't you just run at 100 MHz", "we don't support Linux, please install Win98 and test the system", "you have bad memory", etc.)
I've tried two power supplies and two different memory modules. I even resorted to removing all the dust inside my case (there was a lot around the Northbridge due to turbulence from the fan . . . touching the Northbridge is significant here even if I did not know it yet). The system was stable for about 8 hours afterward before becoming unstable again. Days of fiddling with the computer allowed me to notice the following qualitative things:
1. The instability was worse when I first came home from work when the air conditioning had been off all day (I live in Arizona) and most stable late at night when it is very cool.
2. When I took the computer out to mess with it, and tested it on its side (it is a tower case) it often ran stable only to be unstable when stood upright again.
On a hunch, I removed the two nylon push pins with expanding anchors that hold the small heatsink to the Northbridge, noting that it was a rather loose fit. I found no thermal compound at all on the back side and put some high quality thermal compound from Rawn (rated to 500F and non-drying) on the back of the heatsink and the top of the Northbridge.
My system in now in it's second day of rock solid stability in all ambient temperatures and while standing upright.
This raises teh following questions:
1. Was there supposed to be thermal compound on the Northbridge heatsink that was omitted on my board?
2. Do you think I got a bad Northbridge (I'm in the industry and know that some customers get the edge-die ;-) )?
Soyo tech support was unable to help me and in some cases quite rude ("why don't you just run at 100 MHz", "we don't support Linux, please install Win98 and test the system", "you have bad memory", etc.)
I've tried two power supplies and two different memory modules. I even resorted to removing all the dust inside my case (there was a lot around the Northbridge due to turbulence from the fan . . . touching the Northbridge is significant here even if I did not know it yet). The system was stable for about 8 hours afterward before becoming unstable again. Days of fiddling with the computer allowed me to notice the following qualitative things:
1. The instability was worse when I first came home from work when the air conditioning had been off all day (I live in Arizona) and most stable late at night when it is very cool.
2. When I took the computer out to mess with it, and tested it on its side (it is a tower case) it often ran stable only to be unstable when stood upright again.
On a hunch, I removed the two nylon push pins with expanding anchors that hold the small heatsink to the Northbridge, noting that it was a rather loose fit. I found no thermal compound at all on the back side and put some high quality thermal compound from Rawn (rated to 500F and non-drying) on the back of the heatsink and the top of the Northbridge.
My system in now in it's second day of rock solid stability in all ambient temperatures and while standing upright.
This raises teh following questions:
1. Was there supposed to be thermal compound on the Northbridge heatsink that was omitted on my board?
2. Do you think I got a bad Northbridge (I'm in the industry and know that some customers get the edge-die ;-) )?
Facebook
Twitter
Instagram