DDR5 Thermal Testing & Analysis

[Cowboystrekk]

Nah 1.6 is ok 24/7 on DDR4 (B-Die)

As long as you cool it B-die if tuned loses stability between 40 and 50C. On ddr4 a few dies can handle a lot of heat and voltage like Micron re E and B and Hynix DJR. Many dies hate voltage and temp like Hynix AFR, Samsung C, E etc.

 

[mouacyk]

this is good science, thanks much

 

[Gigaherz]

Nah 1.6 is ok 24/7 on DDR4 (B-Die)

LOL ive been running 1.8 for over 2 years now. 1.5V Seems fine for DDR5.

 

[lexluthermiester]

Project Farm on YT tested a bunch and it didn't make a difference worth a damn.

I saw that video, this is true.

 

[Ruru]

Nah 1.6 is ok 24/7 on DDR4 (B-Die)

No experience of Bdie but in overall... I've ran my 2x16 HyperX-3200 kit at 3466 1.38V (they just don't OC any higher) for ages

LOL ive been running 1.8 for over 2 years now. 1.5V Seems fine for DDR5.

Damn, those are DDR2 voltages. Never ran even DDR3 at that high Vdimm

 

[Gigaherz]

Damn, those are DDR2 voltages. Never ran even DDR3 at that high Vdimm

DDR2 is 2v+ territory. Speaking of OC grade ics with active cooling tho. I wouldnt really run shit tier dimms that high just to get a decent speed. I mean you can buy DDR4 with 1.6v in xmp already, not too much of an extreme to go up to 1.8

 

[toooooot]

I got 2 sticks 16gb each of corsair dominator platinum ddr5 when I built a new PC in 2022.
Last year, I found a reasonably priced set of the same ram. I got it and plugged it into 2 remaining slots
on my Aero D. When I ran the memtest, after some time it would start giving errors.
To make sure it is or is not new set, I plugged the new set and ran memtest without my old ram.
It was working fine. After tweaking a few more things, I realized that ram gets extremely hot to touch.
My final conclusion came to this: when there are 4 sticks working together, there is no space between them,
therefore, temperature is boosted to 65c as it showed in AIDA64. I found out that with XMP turned off,
the temperature stays low enough (~60c) to not give any errors.
SO there I was, left with 2 options, sell the new ram set, or use all the ram without XMP. I decided to settle
without XMP. I also searched some forums, I was not alone with this inconvenience.
This is my main disappointment with DDR5. Perhaps, if I had a case with 10 fans, the air would
move fast enough for ram in my case to not overheat. But I do not intend to turn it into a vacuum cleaner.
Did anyone else experience this problem using DDR5?

 

[ir_cow]

Did anyone else experience this problem using DDR5?

While your Memory was "hot", most likely the problem was using 4 DIMMs. Without know exactly the platform and specs it's hard to say for certain, but it is a common problem because the CPU IMC can't handle 4x at higher frequencies.

 

[umeng2002]

4 dimms also restrict airflow more. I'd say do all of your RAM overclocking with a fan blowing on them. Then when you are stable, take the fan away and see if you get away without active cooling.

Overclocking without a fan can drive you in circles chasing down errors.

 

[toooooot]

4 dimms also restrict airflow more. I'd say do all of your RAM overclocking with a fan blowing on them. Then when you are stable, take the fan away and see if you get away without active cooling.

Overclocking without a fan can drive you in circles chasing down errors.

No overclocking, except for XMP on. My case is roomy and well ventilated. I do not intend to add more fans, that would be too much clutter.

While your Memory was "hot", most likely the problem was using 4 DIMMs. Without know exactly the platform and specs it's hard to say for certain, but it is a common problem because the CPU IMC can't handle 4x at higher frequencies.

It was not most likely, it was the main reason. I tested each stick alone in memtest. They all work with XMP on until I pluged all of them.
The thing about high temperature that I brought was that the ram was extremely hot. It was so hot that I did not doubt for a second
it simply cannot pass the test with four sticks plugged in.
I concluded that they are fairly hot on their own, but almost touching each other they just can't pass torture tests which was very disappointing to me.
I never had such problem with DDR4. And as I was troubleshooting, I found out other people who also experienced the same thing with 4 stick of DDR5.

One other thing I thought of was that this set is one of the early DDR5 ram sets. Perhaps they improved thermal
qualities of the newer sets. I would be interested in plugging in a fresh 4 stick set to test this theory, but I do not have one.