ASRock Z170M OCF funhouse

[GtiJason]

Thanks l0ud. Gave it a shot...no go Still same codes.

 

I've had similar things happen to me, usually happens to me from switching mounts. I think it's because of all the changes in pressure and temps can cause issues. My pot is heavy (kpe f1ee dark), my SS is hart to tell how tight it is, then I also use water occasionally and air for quick tests using 3 different backplates. I just remove cpu hold down and start from scratch rebuilding. I've noticed if just the 3 standard torx screws are too tight my board will curve, add a heavy pot or a lot of pressure and the pcb is leterally being pulled up... away from the cpu socket. SKL pcb is so weak that it just needs to be slightly askew to cause problems

[Splave]

Recover mei in tools section of bios will fix

[Guest Digg_de]

Looks like i got my first B-Die Kit, that can handle Pi32m with Waza @1933 12-12.

 

 

Last B-Die Kit runs great Win7/8/10 (2040CL12), but failed @Pi32m without waza 3600CL12.

 

Recieved 2-3 Hours ago.. needs more testing.

[nachtfalke]

Hi folks...having a bit of a problem so wanted to check in here for thoughts/ideas.

 

z170m ocf has been running great with a 6320 and, recently, a 6700k (which I'm slowly figuring out). All was well during an ln2 session a couple days ago. Tore the setup down, put it back together and put a waterblock on the 6700k so I can set up an os, load some drivers, etc, etc.

 

Problem is that when I start the board, it flashes 19 then 00 (cycles those 1-3 times) then goes to 23. Per Splave's OP, 23 is "23 RTL/Mem failure -Try different RTL / timing or different vdimm or VCCSA". Board can't post atm so can't do anything in bios.

 

Pulled everything apart and sprayed the socket, cpu and dimm slots real good with contact cleaner. Also checked for bent pins and any other signs of damange...everything looks fine. Put the setup back together then tried again. Same problem Tried single sticks in each slot (used 6 sticks total for tests).

 

Any ideas?? Is something dead?

 

same here but was cpu issue , from mounting pressure the pcb was bent , just check on your cpu if the pcb is convex ( with something perfectly straight- but NOT metall , in lenght) from mountig pressure that was my issue, code 00 and 99 (boot _device_led -stuckt) , now i use in the socket center were the square hole are and the MLCC´s cap are a small pice of armaflex , help me alot now to avoid this issue!!

Edited August 9, 2016 by nachtfalke

[funsoul]

Tried with a different cpu (6320k) and the board booted right up

 

The 6700k looks perfectly flat, pcb doesn't seem warped. Figure I'll reinstall mei from the bios menu. Hoping the cpu's not damaged Any ideas?

 

Not able to follow/understand nachtfalke..."use in the socket center were the square hole are and the MLCC´s cap are a small pice of armaflex , help me alot now to avoid this issue"?

 

Update: Should I be using the 121B (which is already on the board) or 131 bios?

Edited August 9, 2016 by funsoul

[Guest Digg_de]

Finally i've got my B-Dies, what i'm looked for... i guess.

It is easy to handle... after one Failrun with waza, i put 10mV more Vdimm, and it is done.

My very first 2000 12-11 run:

1.89v/0.94v Vdimm, 1.16v/1.16v IO/SA.

 

Now it's time for tweaking! :nana:

[RULE]

That io and sa is very low man! Sick, I usualy go straight to 1.25-1.35

 

Inviato dal mio SM-A300FU utilizzando Tapatalk

[Guest Digg_de]

Yea, i am glad, the CPU don't need that much. And it is all @air.

[RULE]

Realy impressive man. Keep pushing it.

 

Inviato dal mio SM-A300FU utilizzando Tapatalk

[GtiJason]

Yea, i am glad, the CPU don't need that much. And it is all @air.

 

1.25 & 1.30 are the sweet spot for me, but sometime as high as 1.35v on both can help. Really nice that it's all air and rare to find ones that are easy to handle

[Guest Digg_de]

Maybe i need more voltages, when i tweak the Mem? I'll test it next days for sure.

[nachtfalke]

Tried with a different cpu (6320k) and the board booted right up

 

The 6700k looks perfectly flat, pcb doesn't seem warped. Figure I'll reinstall mei from the bios menu. Hoping the cpu's not damaged Any ideas?

 

Not able to follow/understand nachtfalke..."use in the socket center were the square hole are and the MLCC´s cap are a small pice of armaflex , help me alot now to avoid this issue"?

 

Update: Should I be using the 121B (which is already on the board) or 131 bios?

 

we use different board , but the issue seems to be the same , code 00 all time

 

don´t need to understand in word what i´m try to say , just use your imagination!

 

the cpu are NOT dead!! he don´t make a good contact in socket with some pin´s

 

now my platform work again

 

some pic for you, and for others with this issue!!

 

[GtiJason]

Tried with a different cpu (6320k) and the board booted right up

 

The 6700k looks perfectly flat, pcb doesn't seem warped. Figure I'll reinstall mei from the bios menu. Hoping the cpu's not damaged Any ideas?

 

Not able to follow/understand nachtfalke..."use in the socket center were the square hole are and the MLCC´s cap are a small pice of armaflex , help me alot now to avoid this issue"?

 

Update: Should I be using the 121B (which is already on the board) or 131 bios?

 

He was talking about putting piece of Armaflex in middle of cpu socket. Here you can see I poured in let for same effect.

 

[funsoul]

hmmm....thanks guys.

 

Reset mei without issue but when I put the 6700k in again, still same 00>19>23 and no post issue. The chip doesn't look bent but maybe it is?? If so...how can I straighten it out and/or otherwise get it working again? Don't see how putting a little armaflex in there would help?

 

Good to hear it's likely not a dead cpu! Now just to figure out how the heck to get it working again

[Guest Digg_de]

Today i've tweaked my Kit a little bit @2000 12-11

 

This Setting was running with 1.90v/0.95v Vdimm and still 1.16v/1.16v IO/SA. But this time CPU cooled with SS.

Next step is ~2040MHz with 102.x FSB.

 

edit:

Tested some more Settings, but not every timing lowering give me a gain... like tCKE 6->4..

Edited August 11, 2016 by Digg_de

[websmile]

Good luck, you might run out of voltage though, 1,9v BIOS on MOCF is 1,98v real for me, so this can be a close call. IOLs look marvelous with IMC cold, definitely a good move

[Guest Digg_de]

Yea, the more Vdimm you give, more Overvoltage from the Board. With 1.35v i've got 1.357v With 1.90v its around 1.97v like you wrote.

 

When i start with 2040, i'll begin with 12-12 and looser timings.

 

edit:

4133 12-12 Win8.1 64 boot works fine with 1.95v

But Geekbench (thats all i've tested) don't pass.

Edited August 11, 2016 by Digg_de

[Splave]

Good luck, you might run out of voltage though, 1,9v BIOS on MOCF is 1,98v real for me, so this can be a close call. IOLs look marvelous with IMC cold, definitely a good move

 

I like to pretend it doesnt, makes me feel better lol

[websmile]

No problem at all, it is well known issue, I had this on big OCF already and everyone should know this by now, also Asus or MSI guys. I often also refer to BIOS voltage therefore, it is reliable landmark for me.

[RULE]

I'm close to sell my Geil Dragon memory so i've try one last run.

 

 

I've found some interesting stuff:

-I can run with proper RTL/IOL this time, but the price to pay is high, 1,30Vio/1,40Vsa

-For correct RTL retrain may be necesary overvoltage memory. In my case Geil not fear High voltage, can pass waza @ 1,95Vdim(BIOS), but for a correct retrain need to apply 2,15Vdim(BIOS). After some strage error codes i can bootup go straight in bios and lower the Vdim, reboot and go in windows easy.

-TCKE 5 is lower in performance compared to 6, but with XTU close to max efficiency can help.

-TRFC 240 is fastest than 220, but 180 is fastest than 240, not all mem can handle.

 

 

 

 

[GtiJason]

I'm close to sell my Geil Dragon memory so i've try one last run.

 

I've found some interesting stuff:

-I can run with proper RTL/IOL this time, but the price to pay is high, 1,30Vio/1,40Vsa

-For correct RTL retrain may be necesary overvoltage memory. In my case Geil not fear High voltage, can pass waza @ 1,95Vdim(BIOS), but for a correct retrain need to apply 2,15Vdim(BIOS). After some strage error codes i can bootup go straight in bios and lower the Vdim, reboot and go in windows easy.

-TCKE 5 is lower in performance compared to 6, but with XTU close to max efficiency can help.

-TRFC 240 is fastest than 220, but 180 is fastest than 240, not all mem can handle.

 

Thanks for sharing your experiences with them. 2.15v bios seems crazy high, I know mine would set real volts to 2.20 + and some may set it closer to 2.25v real. Farthest I've set mine was 2.05v real and they were fine, but only ran 32m once. I've noticed similar results as you with tCKE and tRFC but never tried below 190 yet since I run 4133 most of the time and 200 can give them trouble with some benchmarks.

[RULE]

Yes is a lot of voltage (2.239v real) but this is releted to mem. This method works also on g.skill but with much lower volt.

 

Inviato dal mio SM-A300FU utilizzando Tapatalk

[RULE]

If you miss my FB post, i'll resume here my early test on tCKE combo.

Setup with LiL Yellow, strong retail IMC, strong retail G.Skill mem.

Tamb 20°c (my b-die realy like this one) - Cpu under SS - VCCio 1,25v/VCCsa 1,35v

 

 

So i was sceptical about the real gain on lower TCKE, but look at this:

 

tCKE 5 - tRDRD_sg 6 - tRDRD_dg 4 - tRDRD_dr 6 - tRDRD_dd 6 = 1792 Best out 5 rerun

 

tCKE 4 - tRDRD_sg 6 - tRDRD_dg 4 - tRDRD_dr 4 - tRDRD_dd 4 = 1778 Best out 5 rerun

 

tCKE 4 - tRDRD_sg 6 - tRDRD_dg 4 - tRDRD_dr 3 - tRDRD_dd 3 = 1787 Best out 5 rerun

 

tCKE 4 - tRDRD_sg 6 - tRDRD_dg 5 - tRDRD_dr 5 - tRDRD_dd 6 = 1792 Best out 5 rerun

 

As you can see tCKE can be efficient, just need to tweak around the other tminigs accordingly, tCKE 4 is always a bit unstable for me, 5 is overall more stable.

tCKE 3 is very hard to bootup, and too much unstable to compare.

 

Not a mem engineer but sounds like:

tRDRD_sg tRDRD_dg can't go any lower for me

tRDRD_dr tRDRD_dd best choice 5-5 (with tCKE 4)/ 6-6 (with tCKE 5)

MRC Fast Boot = Disable - can be your friend

 

Need to retest with SPI32, because you know xtu hit at some point efficiency limit and you can't see if scale in performance anymore.

Edited September 25, 2016 by RULE

[Elkim]

Thank you for sharing Franco.

 

I'm trying to get 50/51/7/7 but I cannot pass the train, I tried 2.15V asi you mentioned, didn't helped.

 

 

 

 

These are my waza stable settings, but cannot do anything else. I tried to play with 3rds, last 7 4 5 5, 7 4 4 4 or 1st ones in differents ways, no go.

Edited September 26, 2016 by Elkim

[RULE]

You are welcome, as far i can see from your screen your RTL are telling you 50-6/51-6 is possible.

Don't force RTL as you like, but try to undestard what mem+board can do.

For example if at first retrain, with RTL all AUTO, you have 51-7/58-13, you might consider iol 6 as you lower limit, so CHB 13-7=6 and 58-7=51, leave CHA still AUTO and try to retrain, also a couple of time, sometimes board can adgjust the other CH, if can't try to insert the value you want, considering always 6 as a lower limit.

 

This is not an empirical way, is just my way

 

Also notice from your screen your tRDRD group looks good and also tWRWR_sg 6 is not bad at all. I can't lower down tWRWR_sg to 6 @ 4133.