[SOLVED] How high is safe DRAM voltage for ryzem.

Status
Not open for further replies.
I'm kind of second-guessing myself. How high is too high for dram voltage on 1st generation ryzen?

My ram is stable at 3333mhz with 1.4v and its been stable for months, but im greedy. I tried 1.43v and 3466mhz, and it boots fine, but fails benchmarks.

Specs are in top spoiler in sig.
 
Last edited:
Solution
well u should go with memtest, that way ull see how many errors u get, if changes are improving it or making it worse

as far as your timings go...
tWR should be 10 or 12
tRTP should be from half tWR up to 12 (if twr its 12 then min is 6)
your twr is 26 atm (which is slow for your refresh)

another inconsistency is tRFC 443 while tRC is 80
tRFC is multiple of your tRC (from 6 up to 8)
so if we say tRFC is 443, then tRC can be within 55 ~ 74 range

tCWL should be same as tCL
Well, even OC'ed ram generally doesn't go above 1.35. I suspect you will shorten the life span of those DIMMS at 1.4 or higher.

Of course, as has been proven in a few articles on Tom's recently, the difference between 3200mhz and 3466 or higher is just a percent or two. In some cases you will see no measurable gains, and any gains there are will be mostly in synthetic benchmarks and a marginal FPS increase in games. I'd keep it at lower speeds with min voltage as possible., and timings as tight as possible.
 
I'm kind of second-guessing myself. How high is too high for dram voltage on 1st generation ryzen?

My ram is stable at 3333mhz with 1.4v and its been stable for months, but im greedy. I tried 1.43v and 3466mhz, and it boots fine, but fails benchmarks.

Specs are in top spoiler in sig.
It depends very much on what kind of DRAM chips are on your modules. Some DRAM's are not only tolerant of higher voltage but actually need it to overclock higher... Samsung b-die is famous for overclocking really high at up to 1.5V, for instance. But other DRAM can actually get less stable.

In general, for most chips stay to no more than 1.4, 1.45 for b-die. That seems the good rule to follow.

HWUnboxed provides a great info on how to determine what chips you have and optimal settings with some easily obtained utilities:

View: https://www.youtube.com/watch?v=KOqhyVNPhaM&t=12s
 

scout_03

Titan
Ambassador
before raisng the voltage up did you try to activate xmp on them also i would not go over 1.4 on them this could be drawing more power board could handle and affect others parts in system your always have to be carefull on overclock with voltages
 
My memory modules are Samsung E-die.
3200mhz C16 1.35v is what they are rated for.

I care less about my ram (lifetime warranty) and more bout not burning my CPU's IMC out.

A lot of folk are out there running b-die at 1.45+ volts on Ryzen 1000 chips. I'd not worry too much staying in the 1.35 - 1.40 volt range if that works for yours. Just keep VSoC in a safe range too as that's probably more likely to burn out the IMC.

E-die scales less well as b-die, so you might go take a read through this:

https://github.com/integralfx/MemTestHelper/blob/master/DDR4 OC Guide.md#voltage-scaling

to get some information on memory voltage scaling.
 
Last edited:
Do you think 1.415v or 1.42v at 3400mhz C16 is safe?
It may be a bit high but your memory should still get you 10 years of good service that way. I'd not worry about it, but have you tested for stability? that's a great OC for a 1st gen ryzen!

Get HCI Memtest and run each memory test block to at least 200% before you declare victory. Using something like HWInfo you could also monitor the DIMM temperature to see if they're getting hot during the test.
 
Last edited:
It may be a bit high but your memory should still get you 10 years of good service that way. I'd not worry about it, but have you tested for stability? that's a great OC for a 1st gen ryzen!

Get HCI Memtest and run each memory test block to at least 200% before you declare victory. Using something like HWInfo you could also monitor the DIMM temperature to see if they're getting hot during the test.
Whats too hot for ram?

My ram has nice heat spreaders and runs pretty cool.
 
Whats too hot for ram?

My ram has nice heat spreaders and runs pretty cool.
About all I've ever heard is 'anything less than 85C'. I did not, and would not suggest, taking that literally. I think what is meant is you're probably gonna be unstable and a stupidly insane voltage before you get it too hot. But you want to make sure you have a case with proper air-flow across the DIMM's to cool them so if it ever IS getting up there you should do something about it.

For reference: my 2x8Gb b-die sticks @1.46V, 3600Mtps run in the 30-35C 35-40C range when thrashing them hard with a MemTest.
 
There are 2 fans within an inch of the ram, so it should be fine. My case is loaded full of fans.

I just got 3333mhz at 1.35v @36c stable, but yet 3400 isn't stable regardless of voltage. This leads me to believe my SoC voltage may need to be bumped for 3400mhz.

What do you suggest I do to the SoC voltage? Whats a safe range there?
 
There are 2 fans within an inch of the ram, so it should be fine. My case is loaded full of fans.

I just got 3333mhz at 1.35v @36c stable, but yet 3400 isn't stable regardless of voltage. This leads me to believe my SoC voltage may need to be bumped for 3400mhz.

What do you suggest I do to the SoC voltage? Whats a safe range there?
A few quotes may be in order:

"SOC voltage is the voltage to the IMC and like with Intel, it's not recommended to leave it on auto. You typically want 1.0 - 1.1v as above 1.1v doesn't help much if at all.
On Ryzen 2000 (possibly 1000 and 3000 as well), above 1.15v can negatively impact overclocking."

"There are clear differences in how the memory controller behaves on the different CPU specimens. The majority of the CPUs will do 3466MHz or higher at 1.050V SoC voltage, however the difference lies in how the different specimens react to the voltage. Some of the specimens seem scale with the increased SoC voltage, while the others simply refuse to scale at all or in some cases even illustrate negative scaling. All of the tested samples illustrated negative scaling (i.e. more errors or failures to train) when higher than 1.150V SoC was used. In all cases the maximum memory frequency was achieved at =< 1.100V SoC voltage.
~ The Stilt "

So...there you have it. I never had any success using VSoC to get it more stable for my 1700.
 
SOC voltage was indeed on auto, but it was automatically giving it 1.2v according to the bios.

I manually set it to 1.2v according to this TweakTown page, but I will try 1.1 or 1.15 later.
https://www.tweaktown.com/guides/8506/ryzen-1000-series-overclocking-guide/index.html
Something to keep in mind: when fiddling with voltages you should be monitoring from within the OS with an appropriate tool...like HWInfo64. The voltage as read out in the BIOS isn't going to be the same voltage reading you see when booted into the OS. This is because in BIOS the CPU is only operating on 1 CPU core and none of the OS's power management features working. It's not representative of being in the OS.
 
hi, on ryzen if u want to go over 3400mhz (ryzen 1/2) u will need to change proc ODT and RZT to stabilise link, while changing those things, ddr voltage may need to be readjusted

u shouldnt go over 1.4V for ddr unless u have active cooler
Increasing tRCDRD and tRP by 1 or 2 can improve stability and reduce memory voltage requirements
CLDO_VDDP voltage will improve stability and it will also move mem hole (so go with small steps)
 
  • Like
Reactions: NightHawkRMX
u should first find out about your mainboard capabilities first
reduce clock to 3200mhz
then change procODT (60ohm should be default for this clock, dont go over 60!)
go as low as you can
then post your lowest result which could POST
Will do!

With tRCDRD and tRP increased by 2, SoC at 1.1v, and DRAM at 3400mhz (1.4v) it is not stable, but it is closer to being stable than before. It takes longer to fail test i mean,
 
Status
Not open for further replies.