Find out how Crucials latest does

HERE

Crucial don’t need an introduction so I will skip the really basic stuff.

Crucial’s Ballistix Tracer line is aimed at gamers and case modders who want something a bit fancy. With surface-mount LEDs along the bottom of each stick, as well as flashing LEDs at the top that match RAM activity, the Tracer line is certainly eye-catching. The model up for review here is Crucial’s first for DDR3, but instead of releasing a product at the high-end of DDR3 capabilities, Crucial have picked a fairly pedestrian set of MHz and timings for launch- 666MHz at 6-6-6 latencies. The lower spec allows a more competitive street price (no doubt in part to offset the bling,) and if that is coupled with the undeniably unique look, Crucial have the potential to be onto a winner.

Packaging

I believe I received a very early set of sticks that came in single-stick blister packs, therefore I cannot comment. For a case-modders line, blisters with an interesting background would be a better style of presentation than a plain cardboard box. Heres hoping Crucial are learning to show themselves off a bit more strongly- bling does sell, especially to the casual shopper!

The Kit

The heatspreders are a flat black with chrome edge. The yellow, outlined-in-red lettering stands out very well from the base colour. The PCBs are black.

When installed, the sticks display only their blue “under-lighting” for the first few seconds of boot, then the red and green activity LEDs kick in.

Test setup, notes, drawbacks and assumptions

Motherboard: Asus P5E64 WS Pro, Intel X38 Chipset, BIOS 0702.
CPU: Intel Core 2 Duo E8400, 45nm, C0 Stepping.

Two low-speed 80mm fans were positioned over the RAM at all times.

All voltages were measured directly with a multimeter. I do not trust software readings. Ever.

There is more than one place to measure ground on a motherboard. If a PSU ground line is attached to a mobo hold-down screw and used as a measuring point, it reports a higher voltage than if the casings of the I/O plugs are used. I find the difference to be around 0.1V. For this review, I used the PSU ground line as a reference level. This is the worst case scenario voltage. ie… if I report 1.80V in the review, it would have been closer to 1.70V if I had used the I/O casing.

My motherboard does not allow VDIMM values of less than 1.75V if using the previously mentioned ground spot

A MAXMEM value of 600MB was used to limit the effect of random data scatter in RAM. The results should be more consistent using this method.

vNB values were changed where necessary to allow the RAM MHz to be stable. If values significantly higher than normal are used, they will be stated.

I did not spend hours finding the absolute limits of the RAM, I picked ballpark numbers that should be within reach of most owners. There will obviously be some variation between sets, so posting results obtained from the ragged edge would give a skewed review.

Due to the way these sticks arrived, I do not think I received a matched set. its possible that retail sets will have higher ceilings.

No tweaks were used to speed up SPi32M. The longer the test, the better sign it is for stability anyway.

Testing

I will not insult the in-house test procedure or quality assurance by testing at default, so I will jump straight into the RAMs capabilities for tweakers and overclockers:

Sadly, I had very little luck in testing the kit below its rated voltage. No worthwhile results were obtained.

However, if I kept the voltage at 1.80V I was able to tighten up the timings:

680MHz 6-5-5 1T, 1.80V:

Nothing to be ashamed of! I could not improve on the MHz at that voltage, but by increasing the juice by 0.1V I obtained another 40MHz:

720MHz 6-5-5 1T, 1.90V:

If the timings are loosened to 7-6-6 1T at default voltage (remember that 6-6-6 2T is the default) the ceiling is lifted up to 780MHz!

780MHz 7-6-6 1T, 1.80V

Again, adding 0.1V to the supply offered another 40MHz:

820MHz 7-6-6 1T, 1.90V

I see a pattern emerging… will it continue further?

880MHz 8-7-7 1T, 1.80V

adding another 0.1V to supply 1.90V….

920MHz 8-7-7 1T, 1.90V

Conclusion

For a low-spec RAM kit, this RAM offers a real range of MHz and timings options. Irrespective of FSB, there will be a good RAM performance option awaiting. Crucial did not have to use ICs of this quality given the rated spec, I am pleasantly surprised!

The CrucialUK site lists this at £108, which is £33 higher than the plain Ballistix equivalent. I believe this is in part due to it being a new line. The primary appeal of this RAM lies in its appearance- the look is still unrivalled. If it appeals, I would say this RAM is a very good buy- its range of working settings sees to that and act as a very nice bonus to the appearance. If the look doesnt float your boat…buy the plain Ballistix- it will probably have the same quality.

The downside is that Nehalem is near launch. Much has been made of the recent warnings about VDIMM values over 1.65V, and this kit will probably not be suitable for that limit. All major RAM manufacturers are working on the new JEDEC DDR3 spec, and Crucial will have revised tri-channel kits by the time it matters. Given that Nehalem maturity (and lower prices!) are still several months away, I beleive this RAM still has a place in enthusiast-level rigs.

My thanks to Crucial UK for the review sample.

Best place to buy is direct from source: http://www.crucial.com/uk/store/part…KIT12864TA1336

Update!

I was asked about tRAS performance, so heres two runs with tighter settings: tRAS of 12 at 666MHz and 14 at 880MHz. Setting it to 10 at 666MHz provided a no-boot, and attempting 12 at 880MHz resulted in the same. No voltages increases were necessary- 1.80VDIMM was used for both.

860MHz 7-6-6 1T, 2.00V

900MHz 7-6-6 1T, 2.10V

I appreciate that these results need higher VDIMM than some higher-rated kits, but remember the baseline of 666MHz 6-6-6 please.