Jump to content

uwackme

Members
  • Content count

    762
  • Joined

  • Last visited

Everything posted by uwackme

  1. I went looking for a new Infinity for a friend, and they are nowhere to be found. Anyone know of a place still stocking Infinity or LanpartyB NForce2 boards?
  2. Thanks guys. Im always lurking, just having no problems at all, both systems run flawlessly, so I dont come looking for help. OldGuy I'll PM you if I strike out, many thanks.
  3. Got a link, I searched ZZF and found nada in stock. LP UB not even listed.
  4. I use this, fits perfect, compatible with raid and SLI with no issues. I mount it in the 11 o'clock position, so the fins are BEHIND the upper video card. You do need to plugin the 4pin molex BEFORE screwing down the hsink. I remove the fan, it works flawlessly in passive mode, my CPU hsf has plenty of air blowing through it onto the JingTing.
  5. Just scrape off all the old black silicon glue, and use just a tiny dab of new silicon glue at the 4 corners. Clamp the IHS/CPU/HEATSINK inplace on a motherboard lying flat on the table, let the glue dry/harden in the compressed position. This will give you excellent core/ihs contact inside, and should yeild resyults somewhere between NEKED CORE and STOCK. All the glue is for is to keep the IHS from shifting position under use, beyond that nothing else is needed.
  6. An AC unit costs TEN TIMES!!! what it does in the USA?
  7. I use that case alot, I always snip out the grill behind the intake fan. On serious systems, I cut a 120mm fan into the door to blow over the mosfet/dimm areas, works great. Combine with a 120mm fan PSU and you have 2 slow 120mm intake, and 2 slow quiet 120mm exhaust, great combo.
  8. Easy fix. Best if you have a old NF2 system to use...one where you can use a single stick of ram. Get hold of a copy of the Taiphoon SPD burner program. Put the Mushkin in, copy the SPD, then put the Twinmos stick in its place, and burn the Mushkin SPD into the Twinmos stick. Do the same for the other Twinmos. Now your system will think you have 4 sticks of Mushkin, and the SPD's will be the same. You issues come from 2 different SPD's and timing parameter freakout by the bios, clear that up, use ONE set of timings (these are the "load default" settings you end up with, some various timings the SPD sets that the bios doesnt have manual settings for, and so trouble. This solved my 2X and 2Y sticks of UTT-BH that had issues with using all 4 together.
  9. I use the JingTing heatpipe cooler, mounted in the 10 O'clock position, without its fan, as a passive cooler. It max's at 48C on the NF4 in total silence. It sits just behind my X1900 in the top slot.
  10. The brass standoff is not tight into the case, so it is spinning. So just keep turning the screw....the stand off will un-screw from the case eventually. Next time, tighten the "yellow things" really really TIGHT so they wont work loose.
  11. uwackme

    Making MBM work on the new board

    Anyone have thoughts on how to convince/tweak MotherBoardMonitor to recognize the RDX3200? Is the monitoring chip similar enough to the one ued on the NF4 board to make some adjustments to the data/config files and get it to work?
  12. Find the JingTing chipset heatpipe cooler, install it at the 10'Oclock position BEHIND the video card, (remove the fan you dont need it) and it works great. I peak at high 40'sC now WITHOUT the fan.
  13. Jet, there is a trick to make it work, same as with any motherboard transition. Get a Sil3112 PCI sata controller, install it in your old system, boot and load the sil3112 PCi drivers. Now that windows install on your raid KNOWS about the sil3112 drivers. Move the XP raid setup over... sata1 to sata1 channel, sata2 to sata2 channel. Now reboot and the old system should boot off the sil3112 controller. Uninstall the old systems motherboard chipset drivers, etc. Now move the sil3112 and raid array over to the new system. Powerup, you should be able to boot up off the Sil3112 based raid array in the new motherboard, and then load drivers for it. Then once you have loaded the new systems sata drivers, you should be able to once again move the raid array over to the new motherboard based sata controller and all will be recognized and boot fine. Not a garauntee but Ive done that a couple times now with no problem.
  14. Thats great RGONE. On 01', I'd be #12... 35154... with my 2.75Ghz 165, and single X1900XT Wish someone would crack the CFire driver to allow use on SLI boards :nod: I know what you mean Angry, but I see alot of the issues as software not hardware, so I still look forward to getting a CSX3200, so long as Oskar, etc keep working on solving the problems. I also firmly believe that it is ATI's software for CFire/drivers that is at fault. There is no reason 2 X1900's can work as smoothly as 2 7900's other than NVidia's SLI driver software is well perfected, and gaining application support, while ATI has slipped back into thier old ways of piss poor software development efforts.... remember tha days before the 9700pro when ATI had the most gawdaweful drivers on earth? I do, they sucked azz, but they overcame their faults once, I hope they will again..... fingers crossed. I do appreciate your honest opinion, thanks.
  15. uwackme

    Has AS5 met its match?????

    Not gonna do it. I can see it now, a tiny drip gets on my case, the next morning I have a pile of parts on the floor. The IHS is solid copper, with a nickel plating. No mention of how that stuff handles nickel.
  16. I took a pic, need to download it from the camera. I just installed an Accelero2, took temps down from 42C/79C stock unmod'd cooler to 40C/69C now. That's at 701/802 1.52/2.151/2.151/1.50V fans @ 100%on each test. Plus its very very quiet. So 9C drop at load and silence vs the leaf blower :nod: With the AC-2 and the passive NB cooler in, the system is sooooo quiet and still 2.75Ghz 165, 700/800 X1900XT 2x36G raptor raid. Simply amazing. I have a 120mm Nexus intake, exhaust and internal over DIMM/mosfet area. An 80mm yateloon sleeve 2200 rpm on the Alpha A64 cooler, and a Powerstream 520W with 80mm fan inside. Remarkably quiet for the performance level.
  17. uwackme

    PCI-E Mhz

    I have read over on XSystems that X1900 overclocks can be improved by upping the PCIe clock. Not sure why, just emperical testing. The PCIe wont BREAK anything by being turned up. The bus is very robust, just like HTT but way way slower than HTT. At 100Mhz you should actually be able to hook up a 20ft ribbon cable and plug your video card in over at the other side of your computer room, and it will work just fine. That's how robust the design of PCIe is. Ive been running 120Mhz ever since I got the X1900, its been fine. The only NEED to ever bother going past 100Mhz is if ATI releases drivers that allow CFire without dongle on the X1800/X1900 boards. At which point you would NOTICE the difference between 100Mhz PCIe and 150Mhz. Doing dongle-less CFire is about the only thing that would tax the current bus to the point a speed increase would be meaningful. But the worst case result would be the NF4 or RD580 getting a bit warmer when pushing 150Mhz instead of 100Mhz on the PCIe. But just like your 1800Mhz Opty 165 doesn't explode at 2600Mhz, neither will your PCIe bus. But 99% of the time it's simply a pointless act, and as AG says, dont bother. Trust me, Im an engineer with intimate knowledge of the PCIe spec.
  18. Im very puzzled by the AC vs VF900. Lots of missed results. Lots of "guerilla" marketing of the Zalman going on, fake reviews, etc. I also think both coolers suffer from a concern with all ATI cards, The SHIM. The stock cooler has a bump out, which stick down to assure solid contact with the core. The AC and Zal both are FLAT, so if the core is BELOW the level of the SHIM you wont get good contact, and you may end up with lousy results from an otherwise good cooler. Always TEST whether the core is above/level/below the level of the SHIM before installing a non-standard cooler. If your core is BELOW the shim, you will need to POP the shim in order to assure good contact. Ive had to remove the shim on a dozen+ 9700-9800-X800-X850 cards over the last 3 years because of this. So maybe the reason for the odd disparity in results comes from differences in shim height from card to card. Just a warning and thought.
  19. Get a maxtor drive of some kind, use MaxBlast4 which now knows all about sata and raid. Copy your raptor's image to the maxtor doing a "new boot drive" operation. Save the two MSoft files involving activation if ultimately you want to avoid activation in the end. Now, take your raptor(s) and build whatever raid you want. Then use maxblast to once again do a "new boot disk" install proceedure, this time from the maxtor drive to the raid array. It will make a mirror image on the array of what you had on the single raptor. ONLY ISSUE: Maxblast defaults to implementing 4K cluster size, and I haven't figured out how to get around that. But otherwise, it is a very handy way of doing what you want, and for free.
  20. Sorry AG. I thought by posting here I was talking to the DFI engineers....at least through thier domestic surrogates. Could you, as he wouldn't take my phone call no doubt (and I dont have his number) ask Oskar or someone else in enginnering for me: Is the Vtt circuit free of the droop that plagued the SLI-DR/SLI-D/Ultra-D design? Thanks.
  21. Worst case, you could move the video card to the lower slot, run PCIe at x8/x8 and the board would be along way from the NF4 chip. I assume you monitor the temp on your chipset, and have an alarm on it, so you should never have a serious issue of overheating.
  22. Answer it straight up, it's a simple question. You and I both know what Vtt issue Im asking about.... the huge DROOP in Vtt at high speed/load on the DFI NF4 SLI-DR. Did it get fixed on the Expert? Is it of any concern on the CFX3200? How to prove it, have oskar slap a storage scope on the Vtt and Vdimm signlas and run OCCT, just like with the SLI-DR. Do I need to link in the topic on XtremeSystems where this was all discussed? It is so pronounced a droop you dont even need a scope to see it, a good DVM can see the fluctuation. Why is it an issue, because it takes a good 20Mhz off the ram top speed, thats why. It came from the 9713A being pushed beyond spec for what it can produce for max current. The same chip is being used on the Expert, and it "looks" like it is also used on the CFX3200 too. Two solutions where most likely, use higher value termination resistors on the DIMMs... 47ohm on the two closest, 51ohm on the two farthest from the CPU are the values used on the SLI-DR. Bump the values up to the next available off ths shelf value, and the total current required by the Vtt termination will be lowered enough to be within spec. OR change to use a seperate Mosfet as a 2nd stage for the Vtt circuit to eliminate reliance on the 9713A's current limit. I realize it's a technical question, but it is not something un-answerable. If the same Vtt droop exists on the CFX3200, then why would I bother "upgrading" to end up with the same result? Im really excited that this product is coming out, it should be the ultimate socket 939 board, I just would like to know if the principle limitation on the SLI-DR was addressed (ie: did we learn from our mistakes and fix it?).
  23. I think there is a SMD fuse on each of the fan headers, as well as the USB headers. It's a little greenish thing, looks like a wider resistor. You could always get a replacement from DigiKey if you are handy at soldering. Or just short a jumper across the broken fuse.... just avoid shorts in the future.
  24. The only thing about this that can get me to switch from the SLI-DR board is.... ...is the Vtt issue of the NF4 boards FIXED on this ATI board? Im sitting here with my BH5 and VX sticks held back by the Vtt glitch and will only bother to change if the Vtt is clean and I can get the most out of my ram finally. For that, Ill pony up $200 max.
  25. uwackme

    OCZ PC3200 Gold trouble

    Just set Vdimm to 3.2V and go from there. They are BH5, 3.2V is like falling off a log to them. I get 250Mhz 4x512M same sticks 7,2,2,2.0, @ 3.2V, and remember BYPASS MAX should be set to x7, I think the SPD sets it to x4 and that will limit the top Mhz. You wont hurt those rams, they are very tough stuff.
×