• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

Are there different bins of B-Die? how much voltage is too much?

D

DoLlyBirD

Joined
Feb 28, 2022
Messages
263 (0.22/day)
I have some TeamGroup UD4 DDR4 3000 8GBx2 (K4A8G085W[B/C/D]-BCPB) that I believe may be B-Die though could well be C-Die, I have ran them with a 1600AF for 2+ years at 3400 CL14/1.45v with no issues, now I can push them to 4000, though 1.45v doesn't seem to be cutting it (random YT page crashes when browsing and other anomalies) and I am wondering if they are indeed B-Die, what is a safe 24/7 voltage to run them at? I have heard anything up to 1.55v 24/7 is fine for well cooled B-Die IC's, I'm currently at 4000mhz and 1.47v though I feel I need another 0.1-2v to be fully stable, so if they are indeed B-Die IC's is 1.49-1.5v ok for 24/7 operation? if they aren't B-Die and turn out to be C-Die, is 1.47-1.5v going to harm them?
 
C-die is allergic to voltage (performs best between 1.325v - 1.375v) and I don't think it could do even 3400 C14. You potentially have B-die but it's hard to tell without pulling off heatspreaders (not recommended unless you have experience or are able to live with it being broken).

There is a wide difference in performance between B-die, and that is indeed a result of different bins, but also can just be sample variance within the same bin (not all 4400 C19 B-die is created equal for example). I say do as much testing as you are able, find out its limits in your configuration, and then you have a relative idea of where your kit stands vs. others online.

Voltage: B-die is fine till 1.55v as you said. I'd try 1.5v and probably not budge from there, unless you have good airflow. The concern is less so the longevity, but really the instability at higher temperatures caused by higher voltages.
 
B die is fine 24/7 up to 1.75v if you keep it cool.

There's people who've run it that or higher for years with 0 issues.

I run 1.55/1.6v daily, used to do 1.65 for 4000/14 - 16gb but i'm on dual rank sticks now that won't go above 3833 on a 5950x.
 
I am ok with 1.55-1.6, I haven’t really needed more.. except that one time.
 
Well, too much voltage will certainly cause black marks to appear. :p

There is a point, mentioned, where things become unstable, but wont blow up, how much higher than that before it goes boom, IDK.
 
Ok, so assuming that they are indeed B-Die and I have ran them at 3400/c14 1.45v for 2+ years without any problem, then me worrying about going to 1.48-1.5v is likely nothing to actually worry about and anything up to around 1.55-1.6v if cool should be ok for B-Die 24/7, I mean I paid £65 for them 2-3 years ago and could likely grab a confirmed 16GB kit of B-Die for around £75 today if the worst happens so it's not a major issue, more of an inconvenience, pretty sure if they were C-Die they wouldn't even hit 4000 and 1.5v anyway?
 
DoLlyBirD said:
I have some TeamGroup UD4 DDR4 3000 8GBx2 (K4A8G085W[B/C/D]-BCPB) that I believe may be B-Die though could well be C-Die, I have ran them with a 1600AF for 2+ years at 3400 CL14/1.45v with no issues, now I can push them to 4000, though 1.45v doesn't seem to be cutting it (random YT page crashes when browsing and other anomalies) and I am wondering if they are indeed B-Die, what is a safe 24/7 voltage to run them at? I have heard anything up to 1.55v 24/7 is fine for well cooled B-Die IC's, I'm currently at 4000mhz and 1.47v though I feel I need another 0.1-2v to be fully stable, so if they are indeed B-Die IC's is 1.49-1.5v ok for 24/7 operation? if they aren't B-Die and turn out to be C-Die, is 1.47-1.5v going to harm them?
Click to expand...

Below 1.5V we don't need to have this discussion. B-die is a tank.

1.45-1.5V is pretty standard for 4000 16-16-16 on kits that aren't the top bins (4000CL14, 4266CL16, 4400CL16 XMP). Nothing alarming there, the average Viper Steel CL19 kit lies in that range. 1.55V is a decent place to be for a daily, both my 3800CL14 and 4333CL16 are around that mark.

It doesn't really care as long as you give it the cooling it needs, iirc cooling needs increase the tighter tRFC you run (eg. 140ns may be fine up to 50C but 130ns may start destabilizing at 45C). North of 1.55 you may or may not find you need to start reducing VTTDDR to boot/be stable, but that's not limited to B-die.

DoLlyBirD said:
if they were C-Die they wouldn't even hit 4000 and 1.5v anyway?
Click to expand...

If they were 8Gb C-die, they might not even hit 3600CL16. It's not so much whether they'd survive 1.5V (doubt), but there are only a few ICs out there that scale at all with voltage.
 
Ok, maybe I'll just jump right in at 1.5v and see how they get on, FYI, I am running the XMP profile for 3000 which defaults to 16.15.15 35.93 1T and tRFC is really high at 4000, but I'm fine to get it stable first and tweak timings afterwards, I am happy to throw 1.55/1.6v at these as I said I bought them cheap years ago so a loss wouldn't be that bad afaic though would be nice to get the best out of them rather than having to buy another set if I can
 
Run thaiphoonburner to find out what ics they are. Pretty sure they are b-die tho.
 
maxfly said:
Run thaiphoonburner to find out what ics they are. Pretty sure they are b-die tho.
Click to expand...
Thaiphoon is bad at identifying B vs C die, MFR vs AFR, etc. in my experience. It will likely say "Samsung B/C/D" die for the readout - or mislabel C-die as B-die (have had this happen)
 
maxfly said:
Run thaiphoonburner to find out what ics they are. Pretty sure they are b-die tho.
Click to expand...
I posted the part number as reported by Thaiphoon : K4A8G085W[B/C/D]-BCPB they could be B or C hence my hesitation though the likelihood is pointing towards B-Die, in my mind at least, as @MachineLearning confirmed, Thaiphoon isn't always so clear cut when it comes to IC revision
 
In that case I'd definitely lean towards b-die. I've never seen or heard of c-die clocking as well as yours seem to.
 
Any advice for SOC voltage on Ryzen 5000 series? I always had it at 1.15v on my 1600AF, currently at a.1175v though have heard you can't set it as high with Ryzen 3000/5000 CPU's
 
DoLlyBirD said:
Any advice for SOC voltage on Ryzen 5000 series? I always had it at 1.15v on my 1600AF, currently at a.1175v though have heard you can't set it as high with Ryzen 3000/5000 CPU's
Click to expand...

5500 follows APU rules, it's not related to 5600X.

For 4000 you can probably make do with 1.1V since no iGPU. Most likely even less, unless the 5500s are all just very badly binned for some reason. Since Renoir I've only needed about 1.05V or less to hit 4000 when no iGPU load.

There's no VDDGs for APUs, just VDDP which never really matters (just make sure it's 0.9V or something not bizarrely stupid). VSOC is be-all-end-all.
 
DoLlyBirD said:
what do you mean by this?
Click to expand...

I vaguely recall you had a 5500......maybe I remembered wrong?

For Fabric stability you only have to be concerned with VSOC on 5500. On 5600X and the like you sometimes also have to worry about the 2 VDDGs at higher speeds.
 
tabascosauz said:
I vaguely recall you had a 5500......maybe I remembered wrong?

For Fabric stability you only have to be concerned with VSOC on 5500. On 5600X and the like you sometimes also have to worry about the 2 VDDGs at higher speeds.
Click to expand...
I do have a 5500, so what should my VSOC look like?
 
DoLlyBirD said:
I do have a 5500, so what should my VSOC look like?
Click to expand...

1.1 is a good place to start

If everything's stable, I'd just leave it there and work on RAM and other stuff. APU SOC draws very little power (guessing 5W worst case at 1.1), so there's much less need to shave off excess VSOC.
 
Hi,
My trident-z royal 4000c16 2x16 state 1.4v with 16-16-16-36 timings so there's one exception to typical 1.45v

Lower cl just drives the dimm voltage up in other words to do 4000c14 I'd just need more dimm voltage where cooling matters on how much is to much
Think 1.5v is a good limit though if you can't afford to replace it :cool:
 
ThrashZone said:
My trident-z royal 4000c16 2x16 state 1.4v with 16-16-16-36 timings so there's one exception to typical 1.45v
Click to expand...
My 3200C14 Royals do that speed with the same voltage, noice! I think my Black and Whites need a bit closer to 1.5v, though I can’t remember off hand.
 
This is the first thread I have ever read on the net that says 1.6 is ok for daily driver voltages. If you are not sure, I wouldnt go above 1.5 for daily same advice as @ThrashZone.

Also the dimm refresh intervals (tRFC and tREFI) are affected by temperature for stability, so more voltage means either making that more lenient or applying proper cooling.

Annoyingly my ram has no temp sensors.
 
freeagent said:
My 3200C14 Royals do that speed with the same voltage, nice! I think my Black and Whites need a bit closer to 1.5v, though I can’t remember off hand.
Click to expand...
Hi,
Think all memory manufactures understate what auto max actually is used under stress.
DDR5 especially 1.1v yeah right :laugh:
 
(K4A8G085W[B/C/D]-BCPB) thaiphoon reports my Team B die sticks, so yours are 99% B die.
 
Downbin Samsung B Die is a thing. This is one of those kits : Corsair kit version 4.31.
 

Attachments

  • Downbin B die.png
    Downbin B die.png
    36.9 KB · Views: 331
Intervention said:
Downbin Samsung B Die is a thing. This is one of those kits : Corsair kit version 4.31.
Click to expand...

Ran one of those Corsair 3000 MHz B-die downbins with my old 5950X build, although mine were Dominator Platinum sticks. It did this at 1.375V. Not too bad, and my 5950X probably had a very good IOD too... the VSOC I ran for a 4x16 DR was insane. The timings were not particularly tight, but I never saw the point in pushing it too far, as it performed as well as a DDR4 kit ought to anyhow.

I'll probably reuse them in a future side build someday.

1685211426706.png
 
You must log in or register to reply here.
Back
Top