You must log in or register to comment.
I’m so sorry 😔
If you are still under warranty, complain. If not, try to re-seat the RAM.
Check the warranty status! Some modules have a lifetime warranty for hardware defects. Not sure if this applies to you, but it could be an easy replacement without having to pay today’s ridiculous RAM prices.
Lifetime warranty until they just flat say “we won’t replace this cuz money, go fuck yourself”
As an Aussie, I’m not surprised by Umart (MSY)’s actions - they’ve ALWAYS been dodgy and have been fined multiple times by our consumer protection body (ACCC).
The customer in the article will be made whole (eventually), and the retailer (possibly supplier and vendor as well) may all face fines too.
ACCC is one of the best things about Australia, shame it’s underfunded. Seriously, pump bucks into the ACCC, fine the fuck out of dodgy cunts all over the place - government revenue stream (betcha it’s profitable) and happy voters. But lobbying, sigh.
“oopsie, how did 3 of your stores and corporate hq burn down last night? what a coincidence!”
Lifetime warranty until they just flat say “we won’t replace this cuz money, go fuck yourself”
From that link "Despite qualifying for warranty coverage, Umart is only willing to issue a refund for the original purchase. "
That sounds like an extended service plan denial, not the manufacturer’s warranty. I believe @[email protected] is referring to manufacturer’s warranty.
So Australian consumer Law is one of the strongest in the world, it requires goods to be warranted for “a reasonable lifespan” which is independent of the manufacturers warranty, from what it says in the article is say that umart are calling foul of multiple posts of Australian consumer Law, reasonable lifespan and right to choice of repair, replacement or refund.
As to their statement of it being a minor fault I would say in response that it causes the PC to be unstable which would be a major fault in my books.
Here is the link for the ACCC page, it you’re Australian I strongly recommend you check it out.
I had a pair of OCZ sticks die at the same time like seven years after I bought them. I submitted an RMA and they approved it and didn’t even want a receipt. They sent me two new sticks!
This was lien 15 years ago, so YMMV nowadays.
Dont be shy to RMA RAM that’s faulty. Even if its old - they will replace it. Yes, RAM is far more expensive now for them to RMA - but they are making absolute bank.
Samsung as an entire business reported profits up 800% in just the last 3 months. Memory was 90% of their earnings. That’s insane when you look at how many market sectors they’re in.
SK Hynix is up 500% last 3 months. Their operating margin is 72% which is also insane. Its essentially an indicator of markup against costs - true profits. Apple is known as offering famously overpriced “you pay for the name but you get pretty good service” products… And their operating margin is just 27%. Any common huge grocery store you pick as comparison will be sitting at about 3-5% operating margin.
Don’t believe any of them if they say they ‘can’t afford’ to RMA you. Memory companies are currently the most valuable businesses in the world due to the AI market stupidity, and absolutely swimming in cash.
It won’t hurt asking but if you think just because their profits are up they’ll be fine replacing that RAM when the alternative is to simply have the customer buy new, now expensive RAM (because what else would they do?), then I am not sure what to say.
You could just say ‘I sadly live in a country with lax consumer protections’ and move on. I’d empathize.
If the product is faulty and within the mandated warranty or reasonable lifespan of the product, the manufacturer or supplier legally must remedy it in most countries.
I’m going to contact them today. G.skill says they have a lifetime warranty on their RAM, and these 2 sticks were bought in December 2024 (for $65…)
Remove the RAM, clean the contacts, run it again. It probably won’t change anything, but it might.
If the ram is borderline then lowering the clock speeds/relaxing timing might keep it going. Same with giving it a bit more voltage.
LOL Twins over here. I even have G.Skill memory, too.
On the plus side, G.Skill seems to be chill about RMAs. I got an RMA approval for my memory and sent it in late last week. Should be getting a replacement soon.
Check your RAM timings. Test RAM at default speeds before going up to 3600. Im betting it works fine at 2133 (or sometimes 2400 is the default)
This. My PC needs slight voltage tweaks to be stable at full speed as well; just a +0.1V on the northbridge, and decreasing the RAM voltage from 1.33V default down to 1.2V. Been stable for years that way.
If the ram fault is persistently just at some locations, you can set a kernel parameter not to use them. That’s more for starting up to ensure your backups are up-to-date. If your ram is legitimately dying, then don’t use that pc for anything you’d care about losing.
Oh no. 😥
F
My heart hurts…
I just had some G Skillz DDR4 3200CL16 2x8GB go bad. Sticks had caused problems before. I lowered the clocks to 2133 and then it passed when I had instabilities before, but when they came back nothing changed it. Degredation I guess? Def odd for RAM, but maybe bad batch. Mine were made 2018 Sep.
Do clean the contacts and reseat to be sure.
But about a month ago I did the RMA form, had to email them to remind them to send me the info, got the RMA info, sent it in, and a week or so later they sent me back newly manufactured identical RAM sticks new in box.
There’s absolutely no info when you send it in till when you get it back, but they did send the replacement. You will be without RAM for a bit. They did not want an invoice since its lifetime babyyy
Why are all but one failure shown using the same CPU core? There is a failure on core 14 as well, so it’s not that it only uses one core. Are they more spread out if you scroll?
Idk if that means anything, but it would definitely have me double checking things like timing. I’d expect true failures to be less tied to specific CPU cores.
There’s only 11 out of 1894 errors being displayed. Each core starts the test at a different address, so it’s probably core 6 was the last to hit a bad area.
That makes sense, thx
😭
Oof I’m so sorry
