Very recently, we posted a trilogy featuring memory from Crucial’s recently refreshed Ballistix series. Before Crucial have massively switched to Micron D9PFJ, they had to resort to other IC makers in order to produce their fast memory. Tonight, we will take a second look at some of their older stuff to judge if going Micron made things any better.


Last winter, we reviewed a Hynix-based set of DDR3-2000 Ballistix Elite and saw that using LGA1156 and loose enough timings, it cleared the rated clockspeed with 200MHz to spare. Such feats are impossible with Crucial’s newer 4GB modules on our LGA1155 platform, which makes their predecessors a bit more interesting to test.

Motherboard	Gigabyte Z77X-UD3H (BIOS F9)
CPU	Intel Core i7-3770K (ES, E0)
Graphic card	XFX 8600 GT
Memory	Crucial Ballistix Elite BLE4G3D2001CE1TX0CEU
HDD	Samsung 40 GB
PSU	Silverstone OP1000
OS	Windows 7, 64 bit SP1

As our method of stability testing, we will use 150% pass of eight 750MB instances of HCI MemTest as it’s a very tough, convenient and familiar for us way of testing things.
To make sure we got frequency range fully covered and our scaling chart made perfect sense, we employed a secondary platform to verify things we were not fully sure about.

Motherboard	ASUS M5A99X (BIOS 0902)
CPU	AMD FX-8150
Graphic card	XFX 8600 GT
Memory	Crucial Ballistix Elite BLE4G3D2001CE1TX0CEU
HDD	Samsung 40 GB
PSU	Silverstone OP1000
OS	Windows 7, 64 bit SP1

What we didn’t show last time were the ICs used on these Ballistix sticks. Now that we know how to safely get rid of the heatspreaders, we will eliminate the information gap we left unfilled.

Memory kit used:

Manufacturer	Crucial
Series	Ballistix Elite
Part Number	BLE4G3D2001CE1TX0CEU
Type	DDR3
Capacity	16 GB (4 x 4GB)
Frequency	2'000 MHz
Timings	9-11-9-27
VDIMM	1.65 Volt
Cooling	Passive Heatspreader
Review:
Photo Gallery

Results:

Seeing Hynix 2Gbit BFR on memory would make anyone feel optimistic as, together with revision D Samsung, they are the ICs hunted for the most by the overclocking community.
As this is a “revisit” type of article, we will try to keep things simple. Having thoroughly investigated Hynix BFR on a set of Ripjaws-Z, we discovered that these ICs show the best compromise between performance and frequency, when primary timing formula is CAS + 2 = tRCD = tRP + 1.



While results don’t look bad in general, by Hynix BFR standards they are quite disappointing as our Ballistix fail on three different accounts.
First, while voltage scaling is nice and linear, our memories need a lot more juice to reach same frequencies in comparison to their GeIL and G.Skill counterparts equipped with inferior (but well binned) Hynix CFR.
Secondly, while it can’t be easily seen from the diagram, our memories have relatively low voltage-independent tRCD and tRP walls. For example, x-12-11-x timings are generally expected to sail past 1200MHz on BFR and good samples of CFR, but on ours they barely clear 1100, which makes them practically useless even for not voltage-shy benchers.
And finally, we spot the Achilles’ heel of Ballistix memory on LGA1155. No matter how hard we tried, we couldn’t repeat our previous results and get the memory stable at 1200MHz. Since we didn’t even touch the kit since February, there was no degradation going on so the only possible culprit is be Ivy Bridge compatibility at high frequencies, which seems to be causing Crucial some problems.

Photo Gallery