Hacker News Reader

Comment by pling | Hacker News Reader

And this is why I always buy Crucial. Cruial and Micron are the same company and micron actually make the chips rather than just assemble them. They have control over the entire process end to end and no motivation to pick up chips anywhere else as they already have the best deal.

Over the last 10-12 years I've bought about 2000-3000 sticks of RAM from them. Not one failure or problem ever. I used to buy vendor provided RAM (mainly IBM and Sun back then) and just got ripped off every time. Zero failures on Sun stuff but the ECC stuff hurts. IBM sold us crap lots of times.

Only pain is you can't get an 8Mb EDO SIMM anymore for embedded stuff so its off to a dealer to be shafted repetitively for stuff like that.

View on HN

> They have control over the entire process end to end

How is that supposed to protect you from a completely intentional shady business tactic? Because that's what this is. "We looked the other way and our supplier screwed us over" is a flimsy excuse at the best of times (quality control and penalties should be built into the contract in the first place), but in the Kingston/PNY example it isn't even that.

PNT's other choice is to pay more for better quality NAND or controllers and then pass the cost along to the consumer.

Presumably because semiconductor manufacturing has massive setup costs, so changing anything in the middle of a production run is likely to be uneconomical. But an assembler can simply switch parts or suppliers.

Well the supplier and the manufacturer are the same for Micron. They wouldn't screw themselves over. That's illogical.

the issue at hand was using cheaper fabrication for lesser quality parts.

micron/crucial could still do this.

this is all about saving money once the reputation of a model has been established.

I avoid Crucial SSDs like the plague. They shipped the M4 and similar models with a firmware bug that causes the drive to stop working after 5000 hours of power on time:

http://www.tomshardware.com/news/Crucial-m4-Firmware-BSOD,14...

Not cool.

Crucial, OCZ and Kingston were the favorite failure complaints on the email list I was on at Microsoft ("Hardware Junkies"). The wisdom a few years ago was to stick with Intel and Samsung.

I see no reason to change that. I'd rather spend fifty or a hundred bucks more and get a drive that I'm pretty sure won't fall flat on its face.

OCZ is the worst - I avoid them like the plague now. I've had 2 of 3 of their SSDs fail on me.

They actually found the bug and released a fimware update fixing it before anyone's SSD was 5000 hours old, and SSD owners (at least back in those days) are generally expected to install firmware updates. It was quite well-publicised, too; they handled it really well, except for the actual release.

That wasn't even the only firmware problem with those drives. There was a later bug that would cause the drive to not be recognized after a hard power down, and it wouldn't show up again until you left the computer sitting at the BIOS screen for at least 20 minutes.

It was fixed quickly and that's been the only major complaint I've seen. Further, the bug would cause reboots, but wasn't outright destructive. This is about the same time Sandforce based drives were failing left and right.

tl;dr - Crucial is one of a handful of brands I do trust, and I've owned a number of the M4 series.

SSDs are complicated beasts. We don't use them on anything other than disposable workstations. Every problem with SSDs is down to software rather then hardware.

Your M4s I wouldn't touch. There's nothing wrong with the hardware at all but the firmware sucks.

As far as you can go with SSDs is buy the ones with the least shit firmware, which is Samsung 840 Pro.

If your kit is mission critical, SAS is still the right choice.

Crucial/Micron RAM has no firmware!

As an ex EE this sort if crap is why I trust the hardware guys more than the software ones in life. We still haven't figured out software engineering yet.

I think the main issue is that hardware companies in general really suck at writing software.

ALL companies suck at writing software. Our tools are in the dark ages still. There is nothing that combines the wisdom of the past and applies it to the problems of today.

Wouldn't you say the same was true of hardware? What EE tools are available that help people avoid making stupid choices?

I'm pretty new to actual engineering (I will be getting a BSEE in one year), but I've worked for a couple of years at a hardware company. From what I have seen I think EE and other types of engineering are quite a bit more rigorous than Software Engineering is currently.

Also keep in mind that for firmware often times an EE will be the one doing the programming and much of the time they picked programming as another tool on the way rather than as another separate field of work.

I think when you start picking apart a lot of electronic designs you see that there are a lot of bad EEs out there too.

Software engineering and EE are different in a lot of ways though. It's a lot harder to modify a EE design after it's shipped. You're also often designing around a particular manufacturing process. I think that makes you focus on getting things working correctly before it gets out the door.

Software engineering has a long way to go, but I can also see the EE world does too. A lot of the EE design tools suck, lack standards, and often don't (or aren't) integrate well with revision control. The design process therefore, tends to be a bit messy during development.

The fundamentals of EE work do seem to been clearer than software engineering though. I think this is partly due to EE having developed a mature set of design patterns, but also because it relies on a standard set of components which are really based on physical properties and as such strict and concrete in their definition as opposed to most concepts in software engineering.

Reply on news.ycombinator.com