Announcement

Collapse
No announcement yet.

TridentX 32GB 2400 MHz - odd problems on Gigabyte Z77X-UP5

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • TridentX 32GB 2400 MHz - odd problems on Gigabyte Z77X-UP5

    (Wow, you should remove that silly "moderation queue" for new accounts; I already proved I'm human and not a spammer ... why waste everybody's time with the moderation queue? Paying customers shouldn't have to wait unnecesarily)

    In any case, can someone from G.Skill confirm whether you tested the TridentX F3-2400C10Q-32GTX kit on the Gigabyte Z77X-UP5 motherboard?

    I'm using a 3770K and the integrated Intel HD 4000 graphics. I've tried many different bioses, including the latest beta F11 (03/09/2012).

    I can see the Z77X-UD5H and Z77X-UD3H in the compatibility list for this TridentX kit.

    The same kit worked fine on an Asus Maximus V Gene and P8Z77-V Pro. On the Gigabyte UP5 I have boot loop and power on issues.

    Sometimes it powers on but does not post, other times I can't power it on at all (I have to unplug and clear the CMOS), other times it gets into a long boot loop (resets and powers itself off up to 10 times in a row), and when I'm lucky it posts and boots into Windows but the network card is not always working until I reboot once again.

    i tried with both XMP enabled and disabled. All default values in bios as well as overclocked. Issue still exists.

    If you have tested this kit on the Z77X-UP5 motherboard and worked fine for you, can you tell me which brand and model of PSU you used?

    I'm using a Corsair CX 400W ... could it be that this motherboard is more picky about supplied power? (3 other mobos worked fine with the same CPU and RAM kit). I noticed that disconnecting all fans and the water pump (Laing D5, about 1A sinked) results in less boot loops and power on issues, but they are still there.

    Cheers!
    Last edited by mastabog; 09-17-2012, 07:24 PM.

  • #2
    The queue is there to keep Spammers out, appr 8 out of 10 folks that sign up and go to the moderation area are confirmed Spammers, I check each signup individually (as a result of which, we stay rather SPAM free where many other forums I visit have upwards of 10-100 SPAMs a day).

    As far as GSkills testing results, I don't have access to that. My initial guess is the BIOS is the problem - especially if dated 3/9...Check back in about 10 minutes and I'll dig up my settings, am running them on my Z77


    Pls offer comments on support I provide, HERE, in order to help me do a better job here:

    Tman

    Comment


    • #3
      Originally posted by Tradesman View Post
      The queue is there to keep Spammers out, appr 8 out of 10 folks that sign up and go to the moderation area are confirmed Spammers, I check each signup individually (as a result of which, we stay rather SPAM free where many other forums I visit have upwards of 10-100 SPAMs a day).
      That's because you have a very easy to crack "anti-spam" system, i.e. "Type 'ares' to prove you are a real person." ... seriously? That's not anti-spam at all. You don't change the question either! Of course it's dead easy to automate a bot to fill in "ares" so no wonder you have spammers. You should install a real image CAPTCHA and forget about spammers. Approve accounts manually is not quite 21st century technology to put it mildly . Other forums using real CAPTCHA have no spam bot problems.

      Originally posted by Tradesman View Post
      As far as GSkills testing results, I don't have access to that. My initial guess is the BIOS is the problem - especially if dated 3/9...Check back in about 10 minutes and I'll dig up my settings, am running them on my Z77
      I did say I tested many different bioses (4 actually), including the latest beta dated 2012-09-03. Tested bioses were F4, F8, F9, F11beta.

      When you say "on my Z77", are you talking about the same motherboard, i.e. Gigabyte Z77X-UP5 or some other Z77 motherboard? (the kit works fine on three other motherboards, see my post above).
      Last edited by mastabog; 09-17-2012, 08:57 PM.

      Comment


      • #4
        Sorry, up top you show it as 3/9 just figured Mar 9th, regardless, many of the BIOSs out there aren't ready for 32GB/2400/CL10. They generally set the base right but many are off on the advanced/secondary timings. My Z is an ASRock and after working with them on their BIOS (originally tRFC was limited to 255) which I could run, but it took forever to get the rest of the timings to gel, I've found the following work pretty well with most Z77 mobos

        tRC 43
        tWR 16
        tRFC 324
        tRRD 7
        tWTR 9
        tRTP 9
        tFAW 26
        tCWL 7

        if still a little unstable might loosen tWTR and tRTP to 10 along with tRFC to around 340 - if stable can always try to tighten down...let me know if you need any help, either way


        Pls offer comments on support I provide, HERE, in order to help me do a better job here:

        Tman

        Comment


        • #5
          Thanks, I'll try those timings and report back later.

          Comment


          • #6
            Before I try them, am I right to assume those timings for the kit I mentioned, i.e. TridentX F3-2400C10Q-32GTX and for XMP @ 2400 MHz?

            Comment


            • #7
              That's the set, would have to dig up the original numbers, believe the actual XMP is a lot looser on the tRFC (like 348), and I run it tighter than what I posted, which I've used on a few different mobos. It's also fairly similar to my 32GB 2133/CL9 Ripjaws X set, but they don't tighten up quite as well


              Pls offer comments on support I provide, HERE, in order to help me do a better job here:

              Tman

              Comment


              • #8
                Actually, the XMP is tighter on tRFC (313).

                I attached the relevant screenshots from the F8 bios. Could you confirm that these are correct?

                https://www.dropbox.com/sh/7imocx5dm0b8e5d/iJUoqowidW

                Should I increase the IMC and Vtt voltages? (I'm not overclocking the memory, but XMP @ 2400 MHz is effectively overclocking)

                What about tWTP? Did you mean tWL instead of tCWL (although manufacturers don't always have the same acronyms)? What about the other timings and settings, e.g. tREWFI, stability level, etc?

                Also, what is the "Performance enhance" option (possible values "normal", "turbo", "extreme")?

                Many thanks again, and sorry for the multitude of questions.

                Comment


                • #9
                  Hi,

                  we had the GA-Z77X-UP5 TH, http://www.gigabyte.us/products/prod...px?pid=4279#sp
                  this board should be able to support 2400Mhz @32GB by XMP as well, and will test/confirm the result again for you.

                  Comment


                  • #10
                    Originally posted by GSKILL View Post
                    Hi,

                    we had the GA-Z77X-UP5 TH, http://www.gigabyte.us/products/prod...px?pid=4279#sp
                    this board should be able to support 2400Mhz @32GB by XMP as well, and will test/confirm the result again for you.
                    Yes, that's the board I have. That would be great if you can test this particular TridentX F3-2400C10Q-32GTX kit!

                    Please do try to send the machine to sleep from Windows a few times, and then bring it back from sleep. That's mostly when it craps out in my case when I have XMP enabled.

                    I would also appreciate if you could save the bios profile to a file on a fat/fat32 usb stick or partition and then attach it here so that I try the same bios settings.

                    If it works fine for you, can you also please indicate what bios version, CPU, PSU you used as well as what other devices you had connected to the PSU (hard drives, fans, etc)? I have a feeling that this motherboard is picky about the PSU.

                    I have an older Corsair CX 400W and if I don't connect any hard drives or fans (only the ssd and mobo) then it almost never powers on at all. Maybe voltage levels go down too much on very light loads and the mobo doesn't like it.

                    This motherboard is driving me mad. The only reason I switched away from Maximus V Gene was the very poor video outputs on the MVG. The ram kit was working fine on the MVG.

                    I paid quite a premium on this ram so I'm not willing to let it go.

                    Thanks again.
                    Last edited by mastabog; 09-17-2012, 10:29 PM.

                    Comment


                    • #11
                      Might also see if you can try with a heftier PSU, would suggest a min of 500-550 Watt


                      Pls offer comments on support I provide, HERE, in order to help me do a better job here:

                      Tman

                      Comment


                      • #12
                        Hi,

                        got it, i'm just enable XMP profile 1 with the F9 BIOS(8/23), and will try sleep mode later, and i think your PSU is good enough to support the system, don't have to worry about this part.
                        Last edited by G.SKILL; 09-17-2012, 11:15 PM.

                        Comment


                        • #13
                          After many tests in the past two days, it seems the Corsair CX400W psu was the culprit .. or rather, the Gigabyte UP5 mobo + Corsair CX400W psu combination.

                          With an old 350W psu from work I had no issues on bios F8. It seems the UP5 motherboard is quite sensitive to the PSU used since the COrsair CX400W has worked finw on 3 other Z77 motherboards.

                          Even though I figured out the psu problem myself, G.Skill has been in close contact via email. They tried the odd tests I asked them to do, shared the bios profile, responded promptly etc. Very professional, and as expected given I paid a large premium for this 2400 MHz 32GB ram kit.

                          I can't say the same about Gigabyte who still has not replied to my support tickets for more than a week. This motherboard is top of the line, being their most expensive Z77 motherboard after the UP7 if I'm not mistaken. Very disappointed. I wrote with detailed symptoms and I would have expected they might be interested as they could revise the UP5 to make it more robust.

                          Questions about this TridentX 32 GB 2400 MHz ram kit - F3-2400C10Q-32GTX

                          G.Skill has said they managed to get this kit to 2800 MHz on the same motherboard and a 3770K. Would anyone care to share the timings, voltage and LLC levels that were used for that?

                          Did anyone tested to undervolt this kit on XMP settings (2400 MHz, 10-12-12-31)? What's the minimum stable voltage for XMP on this kit? I set 1.60v instead of 1.65v and have been running Prime95 with 16384 MB memory usage for 1.5h and so far so good.
                          Last edited by mastabog; 09-21-2012, 04:47 AM.

                          Comment


                          • #14
                            Like you discovered with the PSU, it will vary depending on the combination of PSU and mobo (even the individual CPU and BIOS can affect the required voltages)....normally I just try and get things stable at lowest possible voltage, then maybe add + 0.01 or so to add another safety measure and run with it, unless/until I have problems


                            Pls offer comments on support I provide, HERE, in order to help me do a better job here:

                            Tman

                            Comment


                            • #15
                              Undervolting to 1.60v on XMP Profile1, 2400 MHz 10-12-12-31-2T seems to work fine for this kit.
                              Prime95 with 16384 MB memory usage, 8K to 2048K FFT size for about 8 hours was fine.

                              Touching the ram heatsinks felt a bit less hot than when they were at 1.65v, but that's a very subjective test.

                              Can anyone confirm that the temperature does indeed go down with this kind of undervolting?

                              Comment

                              Working...
                              X