Jump to content

PsyKo-Billy

Members
  • Content Count

    14
  • Joined

  • Last visited

About PsyKo-Billy

  • Rank
    New Member

Contact Methods

  • ICQ
    29413825
  1. I now declare the system stable @ stock. The fun part comes in where it wont do the same overclock it used to. In fact it won't even boot with 280HTT, 6x Multi, 2X LDT, 100Memclock. I might revisit the latest BIOS. Anyone know of some supercool modded BIOS for these boards or does noone care about NF3 'cause they're all playing wth NF4?
  2. It is not a weight issue. It's the tension of the clip bending the board. The bottom of the HSF tries to push the CPU through the board therefore bending it. Hacking up the clips seems to have done the job. System still stable so far.
  3. I don't think this is a standard thing. I know of 4 machines which all got the same cooler on pretty much the same time and mine is the only one with an issue. The other boards are DFI NF4 x 2 and one Epox NF3 S754. Edit: I seem to have the machine running stable again with the stock cooler so I'll try the freezer again.
  4. Thanks for the tip. I've done that but the problem still exists. I'll take a pic next time I have the board out but the RAM doesn't seat fully in the middle of the slot. Mind you I can't see any pins it's just the RAM isn't in as much in the middle. The system won't boot fully into windows with both sticks in the first two (oragne) slots and will only boot in the 2nd pair with some extra jiggling. Can somone have a looksee at their boards and let me know if the RAM appears to be a little more out of the slot in the middle? Maybe I'm chasing a falut which isn't a fault...
  5. It's not the weight of the HSF that's bending it it's the force of the clip. Which is bloody tight to get on.
  6. Here's a good one... I recently put an Arctic Freezer 64 on my system. Even tho the temps have dropped I've been having stability problems (seemingly RAM related) pretty much ever since. One curous thing was the system was more stable with the RAM in the yellow, outer, slots. So tonight I pulled the board out and built it on the bench. It was then I noticed the board was a little bent and the RAM wasn't seated properly. So with a little nudge in the middle the RAM seated properly and the system proceeded to run for a while after. Now I'm running with one single 512mb stick in DIMM4 in the hope of some stability while I work out what I'm going to do. So far it seems all good but no doubt the RAM will work its way lose eventually... So who's got some ideas for me to stop this board bending? I haven't heard of this causing issues before. I must just be lucky. I've considered trying to bend the mount for the HSF but I'm worried I'll bend it unevenly and it won't seat properly. The HSF is def on properly I've reseated it 3 times...
  7. FWIW It makes absolutely no difference to me which slots I use. Granted I'm still on NF3 and it seems everyone else in the world is using an NF4 board but... I still get 2 errors in test 5 of memtest if I'm anything over 225MHz RAM clock...
  8. Well that sounds rather poo. Shouldn't I be getting higher clocks at loose timings tho even with lower VDIMM? I always have in the past.
  9. I haven't tried one stick. I have swapped slots and the order of the chips and it's always right at the end of test #5.
  10. Possibly. Tho it happens first loop of test 5 and repeats at the exact same location (right at the end of test #5) each loop after. Same number of errors. There is no fan blowing on the DIMMS but they don't get too hot to the touch and there's some airflow in the area. It's something to try I guess.
  11. The most the board can dish out is 3.2V and that's what it's been on when I've been overclocking. I still would have expected it to have gone higher with loose timings Vs tight timings regardless of VDIMM.
  12. Oh yeah, shoulda mentioned that I guess. The SATA drive is connected to Port 3 so it's locked. Wouldn't that be conencted to HTT speed anyways? I can ramp up the HTT to pretty much what I want (tho I havent tried much over 300) but the minute I go over 225 memclock it's error city.
  13. Here's an interesting one. As the topic says I get the same max mem clock of 225MHz no matter what the RAM timings are. I've tried everything from 11-3-3-3 to 6-2-2-2 and I still get 225. Same with CPC. 1T or 2T doesn't matter, still 225MHz before errors in memtest. "So what?" I hear you say. Well the little dilemma I have is my max CPU clock is 2500MHz with a max multi of 9 which gives me a HTT of 280 or so. Now with a HTT of 280 I can run 166 mem divider and get a mem clock of 233, which is bad, or 133 and get something like 187, which is also bad. If I drop down to 8x then I need a HTT of 312 which my poor little DFI board might not well do and even if it does then using that 133 divider I still only get around 207MHz. I've tried both pairs of RAM slots and that made no difference. I'm also pumping 3.2V into the RAM. The thing I think it may be is one of the myriad of other options in the BIOS for RAM timings which I'm not changing is holding the RAM back. Tho I've been thru the RAM setup guide and relaxed the timings as much as possible and even tried having them all on AUTO. About the only thing I haven't tried is dropping the VDIMM back from 3.2V to something a bit lower but the stories I heard is BH-5 loves VDIMM Anyone have any bright idea's? CPU is a 3000+ Venice (no idea what week) Board is a DFI Lanparty NF3 Socket 939 RAM is 2 x 512mb sticks of OCz Value VX with BH-5 modules. The only thing I do with this machine I want the performance for is CounterStrike:Source if that makes a difference to anyone's suggestion... Edit: It seems my sig is missing so here it is. Edit2: No it's not.
×
×
  • Create New...