Announcement

**GSKILL TECH** · 06-03-2021, 09:09 AM

tRAS should be 36 not 34. Try each module individually to see if each can be stable with XMP specifications.

It may also be worth it to try the latest BIOS.

If the memory is indeed suspect you can send them in to try a new kit.

**GrimReaper** · 06-03-2021, 10:51 AM

I knew you would say about tRFC (it's tRAS actually), though I never had problems with tCL+tRCD+2. That's why I was testing 4000 16-19-19-39 right now.
4000 16-19-19-39 1.39V - error after 2min and 1.40V 8h+ stable
4000 20-16-16-40 1.36V - stable
I already know from previous experience only tCL is bottleneck on 4000 16-16-16.
But it's true that I didn't actually test it on XMP. Plus I lowered SA/IO to 1.15/1.15V, though I don't think that matters.
I will do XMP runs with different stress tests, even RunMemtestPro 4.0 which G.Skill uses.

I'm not unfamiliar with testing each module on its own, but it just takes time and I will use both DIMMs anyway. I'm quite sure one is at least 0.01V better though.

I'm happy with the kit and I don't intend to send it back, it's better than 3600 14-15-15 after all. It's just that it's bugging me a little why XMP isn't stable.
Some guy said to me that 52°C causes me to run higher voltage than would be required if they were at less than 40°C. I was quite satisfied with that answer.
It's just an extremely tightly binned kit, which would simply be solved with XMP 1.45V. I'm also sure it isn't XMP stable in at least 50% systems. I doubt my DIMMs are the worst of this bin and I doubt my system is the hottest as I have all Noctuas, 200mm front, 2x120mm top, one 120mm fan pointed towards RAM, so it could be a lot worse.

Now I also remembered I had to up the voltage by +0.01-0.03V going from Z490 Tomahawk to Z590 Gaming Carbon. Exactly on 4000 16-16-16 I had to up from 1.37V to 1.40V. This yet again proves my statement that this XMP should be 1.45V and it would work in 100% systems.
Same goes for 4000 16-19-19-39 1.40V. I'm quite sure it doesn't differ much anyway, even guy on this forum said it works stable 1.38V both on XMP and on 16-16-16.

**GrimReaper** · 06-04-2021, 09:41 PM

It's fine now actually, Karhu 9h+ at XMP 1.40V. I even left SA/IO at Auto 1.25/1.20V. And that's at 53°C.
For now I don't know why it errored before, could be just temperature problem, could be tRAS problem, could be higher SA/IO makes them more stable. I need to do at least 2 more runs to figure that out.
I also noticed board sets 1.392-1.396V for 1.40V and 1.384-1.388V for 1.39V. So I'm not actually running 1.40V on XMP, but 0.005V less.
On 1.39V it errored after 3min.

I would never run RAM at 1.40V if it errors at 1.39V, I would set it manually at 1.41V.
Even me just playing with RAM for 6 months for a hobby, know that with voltages you need at least 0.05V leeway. VCCIO also in two cases needed one day 1.11V, 2 days later 1.16V and one day 1.15V, one month later 1.11V. So minus points for G.Skill here, they need to do better.
I doubt in factory they only got them stable at 1.40V like me, I guess they had to have better board and better cooling and they made it stable at around 1.38V. It's all just guesses of course, I will never know what was going on here. But basically they screwed up, testing at let's say 1.38V and setting 1.40V XMP.
And it is known no B-Die makes 4000 16-16-16-36 1.35V which would be needed to safely claim XMP 1.40V.

**GrimReaper** · 06-06-2021, 07:07 PM

To conclude, it really was the tRAS, SA/IO can be 1.15/1.15V, I already tested my IMC anyway. 20 hours stable now.

I think all these G.Skill bins are tight and don't work with 0.05V less like other kits, maybe if you are very lucky with 0.01 or 0.02V less:
3600 15-15-15-35 1.35V (though only 2x8GB) - my kit needs 1.35V
4000 16-16-16-36 1.40V - my kit needs 1.40V
4000 16-19-19-39 1.40V - my kit needs 1.40V
4266 16-19-19-39 1.50V (2x8GB is 1.45V) - my kit needs 1.50V
4400 17-18-18-38 1.50V - not really the tightest bin, my 3600 kit needed 1.51V and current 4000 kit 1.47V

So basically most chances to get a good kit are with these XMPs.

**Kiriakos-GR** · 06-21-2021, 02:21 AM

Originally posted by GrimReaper View Post

I already know from previous experience only tCL is bottleneck on 4000 16-16-16.

I have no experience with specific product, but you do misuse the word bottleneck.
Best practices for OC this is to find or discover highest system clocks, and RAM settings, and to step back 5% at the CPU MHz, so the PC to stay stable all times.

**kvarq** · 07-06-2021, 05:57 AM

Same family, I received these suckers today
F4-4000C16D-32GTZRA
Don't know why some retailers or on some forums it is stated that these are 10 layer stuff, last Thphoon build 16.6.0.1 reports that are 8 layer though.

Anyways, can't get anywhere with these, left everything on auto, XMP, just played a bit with VRAM voltage - if left default 1.40V, it's actually read as 1.39-ish, small bump to 1.41 still does not help.
IA and SA left on auto

and GSAT error within minutes... =))

Had another kit before, F4-4000C19D-32GTZKK, did also not passed any GSAT, no matter if the voltages were left on auto or tuned, on Z490 had this kit working flawless on XMP and tuned setting, lower IO and SA etc... BUT passed TM5 with anta's profile.

This one does not pass TM5, default settings, voltages etc...
Bumped a bit the VCore, also VRAM to 1.45, nothing.
Let the SA and IO voltages default (which on Asus are quite high, over 1.47V SA and almost 1.4V IO), still nothing. Reduced them to 1.3-ish, still error.

Asus XIII Apex paired with a 10900K, tried lots of bios versions so far...

Funny thing is that not even the classic memtest is not a pass...

Edit: Wonder if the binning is too tight, or unlucky and I bought a mess...
Temps seem really high compared to other kits I had, 47C in idle =))
Meshify S2, is quite well ventilated

**Kiriakos-GR** · 07-06-2021, 09:20 AM

Originally posted by kvarq View Post

Anyways, can't get anywhere with these, left everything on auto, XMP, just played a bit with VRAM voltage - if left default 1.40V, it's actually read as 1.39-ish, small bump to 1.41 still does not help.
IA and SA left on auto

With out power supply, a branded 850W or better, you will not go anywhere.
Prepare 160$ USD as minimum cost, for your upcoming PSU.

**kvarq** · 07-06-2021, 11:26 AM

Seasonic TX850 already

Asus XIII Apex Crap / 10900K / G.Skill 2x16GB 4000C16 4000C16D-32GTZRA / Samsung 980 Pro 2TB / EVGA 3090 K|NGP|N / SoundBlaster Z / Thermaltake Ultimate 3.0 / Seasonic Titanium 850W / Meshify S2

anyways, I’ll test them on an Apex XII as well, a completely different animal / system :P
if these suckers are crashing there as well clearly I bought some crap... good that I stopped myself when I saw the 4000C14 kit

Binning is too tight already for some time, quite superficial, the good old days are gone, G.Skill is not anymore what it used to be

**kvarq** · 07-08-2021, 01:36 PM

I wonder what kind of tests G.Skill performs for these kits...
Probably I expected a little too much =))
This is the second time in years a G.Skill kit fails me. For some reason I have some kind of Ram addiction, buying few kits a year...
Once it was that dodgy 2x8 3600C15, which ended to be faulty, an now this one.

To be honest, I wanted to buy the 2x16 4000C14 one, but good that I stopped myself :P
Acc to the S/N, these have been manufactured on 2021/06/16...

I am trying to find somehow some kind of voltage combination for my crappy Asus XIII Apex.
Auto voltages set by Asus MB (which are quite high 1.38V IO and 1.47V SA), but last nights I tried IO and SA voltages from 1.20V till 1.50V IO and 1.6V SA, increased the RAM voltage to 1.45V, even bumped the VCORE from 1.27V till 1.35V.
Also lost some time with the skew settings, but got bored in the end...
But I shouldn't touch any settings, in the end, that was the idea with this kit, XMP & that's it

In general, I'm asking for some time GSAT stability, but couldn't believe that it won't pass not even the memtest inlcuded in mobo's bios =))
Activating XMP I or II and changing timings to 17-17-17-37 is rock solid, passes GSAT and so on.
But if I'm using this at C17 then I should have actually bought the cheaper one.
And I'm coming anyways from C16 already, with that KK kit, but wanted to be plug and play this time.

**GSKILL TECH** · 07-09-2021, 11:04 AM

Have you tried each module individually to see if both have the same maximum results? if you feel the memory is not of spec, feel free to exchange them and see if a replacement can work better.

**kvarq** · 07-09-2021, 11:30 AM

I did not test each module by itself, to be honest I kinda lost my patience... last week I lost lots of nights trying to find a stable voltage

but on the other hand... I think everything should work "plug and play", XMP and should be fine and pass all the tests w/o adjusting stuff, despite Asus's high default voltages - to be honest that was the idea purchasing this kit to have it XMP only

this kit is manufactured June 16, someone else tried exactly the very same kit from May on Apex XIII & 11900K and eventually fails after few hours of testing, the only troubleshooting being rising stuff to C17

it could be an idea trying another kit, but right now I'm not willing to pay shipping across EU to Germany/NL wherever G.Skill is located in EU and wait at least 2 weeks the replacement - probably this is the standard procedure...

anyways, I left them overnight on a Z490 computer with a 10900KF, they "almost" passed GSAT...

**GSKILL TECH** · 07-09-2021, 11:42 AM

There is a new BIOS version that may be worth trying.

Otherwise will need to see if a module is not performing as it should.

**kvarq** · 07-09-2021, 01:29 PM

yeah, yesterday they uploaded 2 of the beta versions which have been available since may, but for some reason now these "betas" became "official"
already tried lots of bios versions so far, no solution yet
indeed even after half a year since Z590 became available, they still need to work on bios, now much more that ever...

on the other these failed after a while on XII, maybe I should test the modules separately

**kvarq** · 07-10-2021, 03:08 AM

both sticks failed tested individually
tried to bump processor voltage from 1.27V to 1.33V, also VRAM from 1.40 to 1.42, even 1.45V
nothing helped

could be bios stuff, as bios now for Z590 is pretty much a mess

Announcement

F4-4000C16D-32GTRGA 4000 16-16-16-36 1.40V - not stable on XMP voltage

F4-4000C16D-32GTRGA 4000 16-16-16-36 1.40V - not stable on XMP voltage

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment