RAM speed oddity/question

I've had some ongoing adventures with my new Q9450 system, and now that I thought I had all of my hardware issues resolved, I decided to try some OCing.

C2Q Q9450
2GB Crucial Balistix RAM x2 (4GB total 2 sticks)

Well I ran into a rather odd occurance (at least to me) that I didn't expect.

I did extensive testing with my RAM previously with both Memtest86 and Prime95 to where I was pretty sure it was fine stable running at the CPU default FSB of 333Mhz, with the RAM set to a ratio that made it 400Mhz. So basically the RAM was running at 800Mhz 4-4-4-12 2T and all seemed to be good.

When I got around to OCing my CPU I pushed it up to 400 FSB, but I changed the RAM ratio to 1:1 to keep the RAM running at 800 Mhz, with the same timings. Well after playing with the CPU voltages and trying to troubleshoot the OC, I noticed that my RAM was failing tests like crazy.

Now my question is, why would it fail on the OC, but not the original settings, when in theory, both are running at the exact same speed (400Mhz)?

Is this just pure coincidence that maybe my RAM decided to flake out or go bad now? or is this something people run into commonly?
12 answers Last reply
More about speed oddity question
  1. No, that is not a common problem.
    Double check your timings and voltage settings.
    If everything is fine, test each stick individually.
    If one stick fails, try it in another slot.
    It is probably just a faulty stick that needs to be warranted out.
  2. Before you do that, bump the voltage to your Front Side Buss by one or two notches and see does that help.
  3. I agree with both Outlaw6669 and Scotteq. Running your FSB up like that could have weeded out a faulty DIMM. Could weed out your video card too if you don't lock PCIE frequency in BIOS. I would increase the voltage to the RAM, PC6400 = 2.0-2.1v or slightly higher for an overclock like that. I have had Crucial ballistic DIMMs fail and have to be RMA's under warranty doing exactly what you are doing. Anyway.
  4. I bumped up the voltage on the FSB to make it stable at 400Mhz (I was using Prime95 to test it). I just thought it was odd that the RAM would suddenly start getting errors, even though its running at the same speed. The only difference is the CPU/FSB is faster, but I lowered the Ratio to keep the RAM running at 400 Mhz still.

    Currently i'm playing with RAM voltages to see if that helps. I just don't see why I should have to Raise the Voltage to run at the same speed I was running at previously. I suppose I should also bring the FSB back down to 333 and run the RAM at 400 Mhz to verify that it still runs clean at those settings (what I was using before).

    I should note that I do have the Vid card locked at 100 in the Bios, as I don't want to OC that at all. Also my original 333Mhz FSB (RAM at 4-4-4-12 2T 400Mhz) was done at 1.8v on the DRAM.

    I'm just trying to sort it out, I have had heard lots of stories about Failing Crucial Balistix RAM lately and its not filling me with confidence.
  5. I recently had some Crucial Ballistic on an older machine run for a long while clocked from 3.2 GB/sec. to 4.0 GB/sec. data transfer rate with a voltage increase of 2.8-2.85v. A DIMM failed, but I got a BS. RMA'd that kit. So, the ballistic is vulnerable too.

    If an increase in voltage to the RAM doesn't solve the error problem, you have a bad DIMM (no BS?).
  6. Actually, once you get your OC situated, you can increase the PCIE setting from 100 to 115-119 or so and get an even better OC, provided the card can take it.
  7. Maybe do as outlaw6669 suggested, test the DIMMs individually and remove the 'bad' one. Then see if you get errors like you are having with the good DIMM. Set the RAM voltage for PC6400 to Mfg. Specs., not default 1.8v.
  8. Tech - One thing you may want to consider is UNDERclocking yoru RAM for the time being. Once the Proccy is stable, then you can go back. Just simplifies troubleshooting since you're only tweaking one subsystem at a time.
  9. Quote:
    2GB Crucial Balistix RAM x2 (4GB total 2 sticks)
    Also my original 333Mhz FSB (RAM at 4-4-4-12 2T 400Mhz) was done at 1.8v on the DRAM.

    According to the manufactures product page, you need to run your RAM at 2.0V to ensure stability with your 4-4-4-12 timings. Give it a run with this voltage and post back the results.

    Fixed the link
  10. You can RMA, if needed, from the Crucial support link Outlaw6669 provided. Crucial also has an online chat tech support if needed. They also have a 'prepaid' immediate replacement shipment option. If you bought from Newegg, they may let your RMA also.
  11. I'm running some tests right now to see what I get. Currently I'm Running it back down on 333 MHz FSB, with the RAM set to 400MHz, 4-4-4-12 2T at 1.8v and it seems to be running clean so far.

    I'm going to definitely try the higher voltages in my other test as well, but I guess my original question/thought was isn't 400Mhz the same regardless of which Ratio i'm using to achieve it for the RAM? Why should the RAM care about my FSB, other then the MHz i'm trying to run The RAM at?

    In short why should 5:6 333:400 act any diff then 1:1 400:400? Just seems strange to me.
  12. Although you can control the inner working (speed, timings and voltage) of the RAM individually, the RAM is subject to the increased frequency you are running the system BUS at, as is all the hardware running on your system BUS. So, apparently setting the divider alone is not sufficient to compensate for the increased frequency of the system BUS. Weak or bad DIMM? Try 2.0-2.1v with the RAM. Like I said, I had a Crucial Ballistic DIMM give out and fail to be RMA'd under the same circumstances. Same as OCZ, LOL.
Ask a new question

Read More

Hardware RAM CPUs Overclocking