This is a continuation of my previous "Building a Stable Athlon" article. If you haven't read that yet, you probably want to before reading this one. Basically, by trying a variety of RAM sticks I was able to get my system stable with a non-AMD-approved power supply unit (PSU). I wanted to see if an approved power supply would allow me to use the other RAM, which was extremely unstable with the current PSU.

Once again, if you're thinking of mailing me for help/suggestions with your system, please use the Forums instead so everyone can learn/contribute..

I received my 400 watt Athlon-approved power supply the other day (full review of the unit here). I wanted to know if it would make a difference in stability when using inferior RAM, so I thought I would test my old Siemens stick of RAM with it. I had already sold my NEC RAM (which I knew didn't work properly with the Macase supply), but I hadn't tried the Siemens stick in this machine before. It was the worse of the two sticks in the FIC and MSI boards, so I never bothered trying it in the K7M. Much to my chagrin, the Siemens stick was working perfectly after several hours of torture testing and DVD watching with my old Macase power supply. This made it impossible to compare for stability with the new power supply. Then I remembered that Agg, during his review of the K7M, had discovered similar RAM-related instability with a stick of Compaq RAM. The board I have is the same one he used for his review (I bought the review unit) AND, Agg used a Macase K10 case for his review, the same as I am using for my system. Presuming the PSU's in the 2 K10 cases are identical compatibility-wise, this gave me the opportunity to build an almost identical machine except for the PSU - which should prove once and for all if the PSU was the issue. I got the Compaq stick from Agg and whacked it into the machine.

The Compaq RAM is marked "PC100 3-2-2" and as you'd expect the CAS was detected in bios as 3, so it was definitely slower than either of my old sticks of RAM (both the Siemens and NEC were CAS 2). I left the machine running overnight with Prime95 torture test and several other applications (just to use up all free RAM), until tonight when I got home from work. Upon switching the monitor on, I found that everything was still ok after more than 24 hours of intensive use. I can only conclude that the power supply has made all the difference.

However, there can still be no denying that better quality RAM enabled me to use my old inferior power supply stably for weeks with the Athlon. If you have an Athlon and are in need of more RAM, try to find some that will work on a non-approved power supply, whether or not you are going to get one in the future or even already have one. This way you will be able to use any old supply in the interim period if your approved one blows up - it may take a little while to track down another approved PSU. Also, since my "worse" quality stick of RAM was working fine while the "better" one got constant blue screens and spontaneous reboots, I'm inclined to believe it's not so much the quality of the RAM but just general hardware compatibility. The K7M didn't mind the Siemens stick while the FIC and MSI boards detested it. For the time being, I'm using both my new PC133 RAM and my old PC100 RAM at the same time, and have had no problems so far - with either power supply.