New System and ram, G.Skill still hates me; F3-2133C9D-16GXH issues

[Arrow0309]

Hi

After having different issues in the past with the Gigabyte Z77X-UP5 TH and the TridentX F3-2400C9D-8GTXD:

http://www.gskill.us/forum/showthread.php?t=12487

http://www.gskill.us/forum/showthread.php?t=12720

Issues not 100% solved I decided to change my whole system except for the cpu and a couple of ssd / hdd so I went for the AsRock Z77-OC Formula and the 2x8Gb RipjawsX F3-2133C9D-16GXH hoping that, at least this time the fate will be sweeter to me

It seems I was wrong, and I can't tell you how much I like the G.Skill products, used to have the Pi series (ddr2) in the past and even the famous Perfect Storm 2000

Returning to my new system it's happening it refuses to work well at the memory's Xmp settings (i.e. 2133 cl 9-11-11-31 @1.60v) and I'm testing for more than three days at default cpu speed and (of course) default vccio.

It give me the "Not Exact in Round" with the Hyper Pi (32mb) or it stops upon an error from the very first run of the Linx 0.6.5

I'm gonna explain it better:
It will work well with only one ddr module if ONLY inserted on the A2 (yellow) dimm slot but will refuse to work at the above specs when inserted to the last B2 (yellow) dimm and they both would run even worse (with the boot loops and bios reset) when inserted on the two A1 / B1 (black) slots (not recommended from the MoBo's manual)

The RipjawsX modules are working just fine on my friend's pc with an oc'ed 3770K and the Maximus V Extreme

My system too is working fine when tested (at default cpu speed) with a pair of Crucial Ballistix Tactical Tracer (2x4Gb) at their Xmp of 1866 cl 9-9-9-27

Here are some proofs:

One memory module only on the A2 slot, working fine:



The Tactical Tracers at Xmp 1866:




I mention that I was trying with three different bioses without improvement and right now I'm using the latest 2.30 version
I even tried to downclock the ram to 1866 cl9 @1.5v and the same "Not Exact in Round" error with Hyper Pi, tried first with 9-9-9-27 timings and therefore with the 9-10-9-28:



At this point I simply cannot believe my new Mainboard is faulty, it only remains the cpu (strangely working otherwise) ore this weird incompatibility between the OCF and the F3-2133C9D-16GXH

Would you great teachers and masters tell me what is wrong with me or my system?


P.S.:

I have noticed on the G.Skill's product page the confirmed compatibility wit the Z77 OC Formula

[Arrow0309]

It seems is working downclocked at 1866 @1.5v with the 9-10-10-28-1T timings:



Would that be good enough?

Gonna try to see how's running with the cpu in oc at 4.6Ghz


EDIT:

The system is running perfectly even at 4.6Ghz "green" this way, 1.5v dram, +0.030 offset cpu, default vccio and the rest:

http://s27.postimg.org/et9bqs1cx/Fir...LC_Level_3.jpg

Maybe I should keep it this way, for a while ...

[GSKILL TECH]

Post some pictures of BIOS. You should be able to achieve even better timings @ DDR3-1866.

Thank you
GSKILL TECH

[Arrow0309]

Hi and sorry for the delay, it was the cpu's Imc responsible for that, the memory is working nice and smooth, tested on some other systems and it was 100% ok, tested even on a friend's pc with the same motherboard Z77 OC Formula and with both the 3770K and the 2600K and it simply worked well in the Xmp mode, we only had to manually raise the dram voltage to 1.60v (as required)

Yesterday I took a 2600K cpu and tested with my mainboard and the RipjawsX in Xmp, everything is working good at default cpu:

http://s11.postimg.org/67fzw7vsh/Cpu...rident_Xmp.jpg

At 4.5Ghz cpu:

http://s27.postimg.org/l1ahjm1mp/Cpu...mp_Hyperpi.jpg
http://s29.postimg.org/b1cdc3fth/Cpu...6_load_Xmp.jpg

And at 4.6Ghz cpu:

http://s3.postimg.org/reo3igspd/Cpu_4_6_offset_Xmp.jpg

All done with the both vccio and vccsa kept at their default voltages

Now I'm going to ask for a RMA procedure for my cpu

[emissary42]

It happens, even if it is kinda rare. Good luck with your Intel RMA.

[Arrow0309]

Thanks!
I already started a support chat discussion with an Intel guy and he just gave me a RMA number
Keep you informed, hoping the new cpu would be with a better IMC!

[Arrow0309]

Here I am again guys, my bran new 3770K arrived directly from the Intel and I previously bought a new cooler, the new Cryorig R1 Ultimate to "properly" cool down the new cpu

http://cdn.overclock.net/d/d7/d7500e...154113714.jpeg
http://cdn.overclock.net/0/0d/0de4b3...145414099.jpeg

And with that, my dream ended once again.

The "perfect" combination RipjawsX F3-2133C9D-16GXH, Z77 OC Formula (the single Asrock Z77 mainboard certified on the RipjawsX's site) and 3770K is still not working again when I set the Xmp at 2133 and manually set the voltage at 1.60v as required.

In order to ensure the perfect conditions to the ram tests I first of all do a proper clear cmos (removing the battery), flashed the latest Asrock official bios for my OC Formula, the P2.30 and did another clear cmos once again.

So I tested before with one module only (at the default cpu), inserted into the yellow slot (A2) closest to the cpu and it finished well the 32mb of the hyperpi:

http://www.xtremeshack.com/photos/20...7075205532.jpg

Tested the other one (dimm) too, on the same slot and it's also working:

http://www.xtremeshack.com/photos/20...7079128539.jpg

But the both of them properly inserted into their yellow slots (A2 and B2) will make the hyperpi fail:

http://www.xtremeshack.com/photos/20...7077309389.jpg

And here again, reversed:

http://www.xtremeshack.com/photos/20...7081201230.jpg

The reason is the very ddr3 slot B2, the yellow furthest to the cpu, here one single dimm inserted gave me instant errors:

http://www.xtremeshack.com/photos/20...7083131101.jpg

Or even a bsod ...xc4 with the other dimm

Lastly I even tried to bump the vdram (1.62v), vccio (1.15v) and vccsa (0.95v) and it's still the same:

http://s17.postimg.org/496xy1oql/Cli...ccsa_error.jpg

Now I simply can not believe that my new 3770K Box that I personally unsealed, received from the Intel should have a crappy Imc as well (and maybe the other one was also good) and not to think of the faulty motherboard.
Consider that the memory is still working well set at 1866 c9-10-10-28 @1.5v
And with the SB 2600k it even worked well at 2133 at default cpu and 4,6ghz

I'm looking forward for your answer and starting to consider even a MoBo Rma
Thanks!

[GSKILL TECH]

Have you tried A1 B1?

Thank you
GSKILL TECH

[Arrow0309]

Not this time, I only had tried them once and it was even worse, with boot loops and bsod's

But I did try another thing yesterday:
I installed the 2600K cpu once again and tried the ram in Xmp at default cpu first, OK and then at 4.0Ghz and 1.22v:













Then I replaced the 3770K without any clear cmos, keep the same bios settings and ran the Hyperpi again.
It crashed

http://www.xtremeshack.com/photos/20...5430166407.jpg
http://www.xtremeshack.com/photos/20...5431946332.jpg
http://www.xtremeshack.com/photos/20...5431376830.jpg
http://www.xtremeshack.com/photos/20...5432444278.jpg
http://www.xtremeshack.com/photos/20...5434474258.jpg


Right now I sent the ram kit back to my friend who tested them successfully before on it's Z77 OC Formula and 3770K and I sent him the new 3770K also this time.
He is also working with the hardware as full time job for the city of Imola (Italy).
I'm quite sure it'll work nice, I just need one more testimony of the three new hw components fully compatible and that they're working correctly altogether.
Then it is clear it's a motherboard failure

I'll keep you informed

One last question:

How come, on the RipjawsX F3-2133C9D-16GXH site's QVL there's the single Z77 OC Formula only as a certified Z77/B75 Asrock motherboard?

And yet, this should be another "proof" that they're compatible and must work as specified.

[GSKILL TECH]

That is the board we had on hand during release, but we test and will update afterward as well.

Thank you
GSKILL TECH

[Arrow0309]

Bad news, my friend has finished a first session of tests with my RipjawsX and my new 3770K at 2133 (Xmp) and using his Z77 OC Formula with all the latest bioses (P2.0, P2.20, P2.30) and every test ended up in Hyperpi errors just like my tests.
It's hard to believe that a bran new 3770K coming directly from the Intel (Nederlands), capable of 4.6Ghz with only 1.17v is having such a crappy Imc.
At least my OC Formula is ok, no need to rma this one as well (and maybe my previous 3770K was ok too)

Now we're gonna make some "crossfired" tests:
1- The RipjawsX @2133 Xmp, the new 3770K and other Z77 mainboards (Gigabyte UP5 TH, UP4 TH and Asrock Extreme 6)
2- The new 3770K, another ram kit (2x4gb Geil Evo Corsa 2133 cl 10-11-11 @1.5v) and the Z77 OC Formula

Depending of these last results will decide to either Rma (for the second time) the cpu, sale the RipjawsX and replace them with other ram or keep them at 1866 cl 9-10-10-28 @1.5v

I'll come back with the new results

[Arrow0309]

The new 3770K is OK, no issues with a Geil Kit 2x4Gb Evo Corsa 2133 cl 10-11-11 @1.5v and the Z77 OC Formula
It handle the Geil nicely even at 2400 @1.5v (with proper timings)

I'm really disappointing for the G.Skill ram choice I've made this time

Now I'm gonna sell them and return to a 2x4Gb kit (which is more than enough for mi right now) one thing is for sure, I don't want any Samsung IC anymore, not with this motherboard

I'll probably choose the Geil or the Corsair 2133 @1.5v