RAM needs less voltage after cleaning the contacts?

[GerKNG]

so i have a weird kit of G.Skill RAM that fits roughly in the time frame of the "Hynix sold worse than usual ICs scandal".
it needs around 1.38V to run with XMP (on all of my platforms, from mid tier Z390 Gigabyte boards over a Z590 Strix board with a lot of CPUs tested, and on all my AMD boards. it just needs 1.38V under load (not in bios, in HWInfo, it needs around 1.4V set in Bios)

now i just inspected the RAM again after around 6 months and i just thought, why not cleaning the contacts on both sticks with ipa and see what happens. (both sticks were perfectly clean with zero residue or corrosion. just clean golden pads.)

well after cleaning them properly they run totally fine at 1.33V with the identical clocks and timings.
i ran Prime95, Memtest (over night to around 700%) with zero issues.
any ideas what happened?

 

[eidairaman1]

Less resistance, better continuity between contacts on ram and motherboard

 

[GerKNG]

Less resistance, better continuity between contacts on ram and motherboard

but almost 50mv?

 

[eidairaman1]

but almost 50mv?

Yup

Id be thankful that happened to you

 

[joemama]

Contamination on the gold pads could take up some part of your voltage input but your voltage difference is a little big though.

 

[tabascosauz]

I'm not a fan of G.skill CJR, mine always behaved really weirdly as well. Before the 3600CL14 Trident Z kit I was using a 2x16 dual rank 3600CL17 Trident Z kit.

It always did my 3600 16-19-19 profile at 1.37V, but once in a blue moon it would crap out and fail to POST. I'd clear CMOS, put in my settings again, and it was back to normal. I never managed to find errors at 1.37V or 1.38V in any combination of HCI, TM5, Prime95 Large etc. Happened on two different boards.

With tRFC as well, I ran 450 @ 3600 for more than a year. All of a sudden it became unstable and I had to come back down to 471 @ 3600, no hardware changes. I never ran more than 1.38v.

Since the CJR kit went in fall of 2019, I never managed to solve my chronic frame skipping/stuttering problem in MW2019. After years of trying to solve it with drivers, reinstall, reinstall windows, MSI-mode, CPUs, boards, GPUs......it turns out that the CJR was the culprit. I've never seen an instance of stuttering with my dual rank B-die, ever.

My 4Gb E-die, A0 single rank B-die, A2 single rank B-die, and dual rank B-die combined have never given me nearly as much headache as the CJR.

 

[Bill_Bright]

I'm not really buying this - at least not completely.

The contacts in today's RAM slots are specifically designed to keep dirt, crud, moisture, oxygen (perhaps most important), and corrosion from getting in between the mating surfaces of the slot contacts and RAM contacts. So if they (slots and RAM) were completely clean, in excellent condition, and properly inserted in the first place, continuity/resistance between the contact points will not change - especially after a mere 6 months.

Remember ZIF slots? Those "zero insertion force" slots made it easy to install and remove the sticks, but they lacked the necessary pressures to keep crud out. So they did away with them (at least for RAM). Yes, ZIF connectors are more expensive so that does play a role, but they still use ZIF on other things - like CPUs - where insertion forces might cause damage.

Today's RAM slots, with their spring loaded contacts that require considerable force (and strategic placement of tongue) to properly insert, have two important tasks. (1) They scrape clean the contacts during insertion and (2) they effectively create a seal around the mating surfaces (where the contacts touch) to prevent crud and air from getting in between those mating surfaces.

The point is, "IF" the RAM sticks and slots are clean of crud and corrosion in the first place, and "IF" the slots are in good repair, and "IF" the sticks are properly inserted, there is no way for air (oxygen) or dirt to get in between the mating surfaces to form corrosion or degrade continuity.

So if you are claiming you tried 1.33V right before cleaning and it didn't work, then immediately after cleaning it did, I contend the sticks (and/or slot contacts) were not really as clean 6 months ago as they initially appeared. Perhaps there was some machine oils, or skin oils that were not visible. I don't know. What I am certain of is there are other unknown variables at play here.

 

[Andy Shiekh]

I believe that the springy contacts in the RAM slots are of phosphor bronze and not gold plated (I could be wrong),
so removing and replacing your RAM scraped the oxide layer from the slot contacts; gold plating one side would be
to avoid galvanic corrosion.

It is hard to avoid air getting in; for this reason I tend to use silicone oil (sparingly) to seal the contacts from air, but
I really don't want to get into an argument about this (been there, done that).

 

[Bill_Bright]

I believe that the springy contacts in the RAM slots are of phosphor bronze and not gold plated (I could be wrong);

No, you are right!

Gold is rarely used. It simply costs too much, plus it is too soft. Some contacts may use alloys that contain gold to help prevent corrosion. But if gold is used, it really is a very tiny amount due to costs and hardness issues.

Copper is actually a better electrical conductor than gold. And silver is even better yet. But they do corrode (tarnish) easily. So yeah, bronze is commonly used because it is mostly copper, thus an excellent conductor, but bronze typically is an alloy with tin. And tin is very good at resisting corrosion and increases strength and wear resistance too.

Silicone (not to be confused with silicon) is great for sealing a connector to keep air and contaminates out. This is why it is commonly used in applications that are exposed to the elements, like "under-the-hood" automotive connectors. But care should be taken to ensure none gets in between the mating surfaces since it is a semiconductor and can act as good insulator.

But I am not so sure one can avoid air getting in

If you pinch your thumb and forefinger together with some decent pressure and you maintain that pressure, is air going to get in between? Note the contacts on a RAM stick frequently are called "fingers".

If this were a problem - especially indoors - it would be needed with essentially all low-voltage DC, high-speed digital connections - like CPU sockets, PCIe connectors, SATA, USB and network connectors. But that is just not done. It's not needed.

The bigger issue I have frequently encountered is with RAM sticks that have sat on the shelf, out of their original packaging for some length of time. For them, I carefully take a clean pencil erasure to the contacts, being careful not to scratch or drop them, or zap them with ESD. Then I blast the RAM slots with a good Electrical Contact Cleaner before carefully inserting the RAM. I have never had to "re-seat" RAM after that.

 

[Andy Shiekh]

Metal contacts are not pliable like human fingers; without a pliable gasket, two metal surfaces are likely to leak.

How about the hard drive internal contacts?

• Hard Drive REPAIR - Western Digital 500GB Green - cleaning contacts - Bing video
• Here’s a Surprising Trick Repairing Hard Disk (HDD) | Electronics Repair And Technology News (jestineyong.com)

where one can see the issue.

I also worry over tin whiskers
Basic Info on Tin Whiskers (nasa.gov)

 

[Bill_Bright]

Metal contacts are not pliable like human fingers; without a pliable gasket, two metal surfaces are likely to leak.

That's true. But in this case, those contacts in the slots are spring loaded. And with pretty stiff springs. So in that respect, there is pliability to them. The metal surfaces are not simply laying against each other. So I totally disagree they will leak - unless you bounce the computer off the floor causing a gap to appear.

As far as your hard drive - there are always exceptions. Exceptions don't make the rule. We don't know the condition of those connectors or the drive prior to the original connection. I personally have worked with 100s (probably 1000s) of hard drives over the last 30 plus years. I have never had to clean those contacts.

 

[ThrashZone]

Hi,
Always good to clean dimms but even then sometimes sticks just don't install right same goes for cpu /... power connections too always good to reinstall once in a while.

 

[Bill_Bright]

always good to reinstall once in a while.

Nah! If it ain't broke, don't fix it! Chances are you will break it then.