I can I have some high end OCZ RAM and I wouldn't buy it again plus after reading all of the bad stories about their SSD drives only supports my experience with them.
And I've heard OCZ have bad PSU's but decent RAM. 99% of bad reviews I've heard about their SSD's have been written by people that admittedly either don't know how to follow directions to update firmware and/or treat them like a traditional hard disk. I havent had issues with either the vertex or the vertex2, so my bias makes me believe it. I guess we all get our news from different places!
The SSD complaints I'm speaking of had nothing to do with morons it had to do with failure rates and at just over a year of use and at worst 37% of wear the SSD's were failing. This was a very common theme in the threads I researched. As for SSD's I'll probably buy Corsair or maybe even Intel even though they are hi-priced but will not buy OCZ even if they are supposed to be faster. I won't buy their RAM again or at least for some time I'll buy Crucial before I'll buy OCZ. My RAM still works but it performs poorly compared to other RAM in its category so if I'm going to pay good money for RAM it will be with a better product. Their RAM sure looks pretty though and that is a major factor in today’s favor, iPod's anyone?
Mr-Movies, This Crucial seems to be about the most reliable of all the major brands, and sells very well. http://www.newegg.com/Product/Product.aspx?Item=N82E16820148348 Works with either Sata II or Sata III. At 140MB/s it doesn't have the fastest Write speed, but they seem to be very reliable compared to other brands. Russ
I've always had good luck with Crucial even though it is less expensive and not considered a brand name, although I think it is becoming that more and more. I do agree with you... Stevo
The GameXStream was meant to come with a fan controller, but in the PSUs that were shipped for at least the entire first year it was on sale, they were non-functional. Presumably they've since fixed that issue. The main issue to watch out for with PSUs is thermal sensitivity, and fan gradient with load. Many PSUs will run fairly quietly and cool by themselves, but as soon as they pick up the smallest amount of ambient case heat start to become noisy, and plenty more PSUs that are very quiet at minimal load suddenly become quite noisy with moderate load. For example, the corsair CX 400W is 19dB in an anechoic chamber (compare to 11dB for the Nexus units I use) at idle, but in a hot-box setup (i.e. a PC where the PSU is at the top, picking up case heat), it reaches a considerable 26dB by 150W, and a very noisy 32dB at 200W. When running in isolation on the test bench, it's only 19dB at 150W, and 24dB at 200W, so it does make a sizeable difference. Things have certainly changed in the last few years as far as component reliability goes. OCZ have been bad for memory for quite some time now, and that hasn't changed. G-Skill, who have a good reputation have produced a few bad sticks, and although their average failure rate across the entire range is quite low, the most unreliable individual memory products mostly come from corsair. This isn't obscure high-spec dominator GT stuff, this is commonly bought pairs, like the 8500C5 XMS2 dominators and the 6400C5 XMS2. The most reliable brand (perhaps surprisingly) is Kingston, with Crucial and Corsair not far behind, but with all these brands' failure rates less than 1 in 70, none of them are really a risky buy. With OCZ's failure rate more than 1 in 15, they're best avoided. OCZ's PSU's used to be quite poor back in the early days, they improved slightly with the GameXStream but were still relatively mediocre implementations of quite mediocre OEMs (Fortron). They weren't going to go pop like the cheap units, but for the price tag, there was better stuff out there. Corsair used to be the gold standard for PSUs from the introduction of the excellent HX520/620 units several years ago, up until the end of 2010. Since then, things have gone seriously downhill. The CX Builder series are very cheap, and it shows, they're proving quite unreliable. There have been a fair few problems with the flagship AX units too, despite their excellent credentials. Worst of all though, previously solid units, the HX620, HX1000, and a quite commonly bought newcomer the TX950, take silver, bronze and gold respectively for the most unreliable power supplies sold in france, with the HX620/TX950 failure rates more than twice any other PSU. For large mechanical drives, with the exception of the WD2001FASS which is faring quite poorly (1 in 10), the only real risky buy is the Hitachi Deskstar, the least reliable large capacity drive (5-6% for the 1TB, 7% for the 2TB). On the whole WD Green drives score quite well (especially the earlier EADS drives, which are top of each size class for reliability), and by contrast, the WD1001FALS scores very well too (1 in 74). For 1TB drives, if you stick with WD you can't really go too wrong, and even the Samsung F3 and Seagate LP are reasonably solid buys, just avoid the 7200rpm seagates (especially the 7200.11), the Samsung F1, and anything by Hitachi. With the exception of the WD20EADS way out in front, none of the 2TB drives score that well, with second place (Ecogreen F3) scoring worse than all bar the Hitachis at 1TB, but the failure rates of the main competitors, barracuda LP, Ecogreen F3 and WD20EARS, are all fairly similar, at 4-5% each. As before, just stay away from hitachi (oh, and the WD2001FASS) For SSDs, there is simply no contest for reliability. The failure rate for most SSD brands falls between 1 in 46 for Corsair, and 1 in 34 for OCZ. For intel, it's a staggering 1 in 170. Intel know how to make SSDs, it seems. By comparison, the average failure rate for CPUs? 1 in 550. Why can't all hardware be so reliable?
Because CPUs are load tested as part of their manufacturing process aren't they? If a chip is bad it never becomes a complete product anyway.
Eh what can I say never had a bad CPU I didn't break myself. I first went AM2 with my X2 4400+ because I broke a pin off my X2 3800+ XD
I remember when the pins were on the CPU... Thankfully that was before I started dabbling with hardware too much
Sam, I discovered a minor flaw with the Coolit, in that sensor control is just too slow. That's why when the fan controlled by the bios barely starts to ramp up when it slows right back down again the minute cooling starts to take effect. The sensors are just too sluggish to be effective as it's always a day late and a dollar short. That's why I run both the Push and Pull fans at full speed all the time. At 1800 rpm for the Push and 1200 rpm for the Pull, it maintains lower temperatures across the entire load range. This is something that Coolit neds to address! It's not that it's very loud, but they need to come up with load sensing/thermal solution so that the fans can be controlled more efficiently. Thankfully for now, both the stock fan and the scythe are able to control the temps to the point of never exceeding 53C under full load, all 6 cores at 100%! Which brings me to the OCZ 550FTY. The fan in it has both load sensitive and thermal sensitive controllers, so in effect you get instant linear cooling the instant a load is applied to the PSU, while the thermal sensor, by it's more sluggish nature both up and down, controls the de-acceleration of the fan, when the load is removed, better removing the vast majority of the heat! From all the reviews I read, it seems to work quite well. Russ
With a 120mm fan at 1800rpm all the time, it's no wonder you think the OCZ PSU is quiet, most PSUs even those rated quite poorly by noise-sensitive sites would be inaudible behind such a noisy fan. Noise is of course relative, for most people noise wouldn't get too irritating until multiple 2000rpm 120s were in use, but for me, just one 1800, even a very good one, is quite loud, at least at idle. There is no resonance with the SFF21F in my CPU cooler (the plastic fan holders are surprisingly good for vibrations), yet I can still clearly hear it when it's at 100% speed (1550rpm in this case), and while it's understandable and livable at full load, at idle it'd really annoy me, and I'm glad it drops down to 1000rpm or less at idle where I can't really hear it over the 120hz hard disk noise. You can't have two means of controlling the same fan, otherwise they'd conflict. You could have two inputs into a logic-controlled fan speed (which seems a very complex and expensive method) and take the highest value, but still, you'd get the same results with the load-sensitive method. What most companies do that want more sensitive (read:noisier) fan control, is place the sensor right on top of a fundamental component like a power transistor, instead of just having one somewhere in the unit to take an ambient temperature. Such units can be frustrating though, as spikes in power draw from a few seconds' CPU usage (opening a program for example) are enough to raise the PSU fan speed. The fan speed changes with ambient-sense PSUs are slow enough not to annoy you with the fan speed going up and down all the time, but usually come with idle fan speeds high enough for cooling to not be an issue, i.e. they're quite noisy even at idle (e.g. Corsair HX 1000). The ideal sort of PSU is one that has a flat idle line for a considerable load before ramping, so you only ever see raised fan speed when the whole system is under heavy load. The 850W Zalman I use for example, runs at idle fan speed all the way up to 700W DC when run with a separate air supply (such as in the HAF 932 case), which is ideal. The only time the fan speed would ever rise is if I opened OCCT or furmark with my old 4870X2s installed (which actually ran the unit over 100% load), and you'd never hear a 1400rpm 140mm fan beneath two 6000rpm graphics fans [which always run at 100% in OCCT due to the 99ºC GPU temps even at that speed)
This couldn't be more correct and although Russ's pressurized case is the best way to run and keep temps down it is way too noisy. I've pressurized old amplifiers in the old days to keep them from clipping under intense load and it works very well, but I don't want my PC screaming at me under these same conditions. From the manufactures specs green drive only save about 1.5w under load and 0.3w idle. Seek times are 12-13ms verses 8ms for a good drive and then there is that 13 seconds for the drive to ready on green drives. Green drives do not play music or video well on a media server as you get hiccups’ from time-to-time. Green drives from the annual failure percentages are around 1% compared to 0.3% for a good drive that equates to 3x more failures but really 1% is even really good if not excellent. So is a Green drive GREEN? And is a Green drive the way to go? I think not! Good critique Sam, very good! Stevo
Green drives are no good for media use ?, i have to disagree (quite strongly), no hiccups here. And the servers we run at work with green drives (as the main drives, plus servers with greens as data drives) will have to disagree too
I'm not 'believing' anything, in fact i have no idea what the manufacturers support (other than vague recollections that they're not any good in certain NAS devices), am just going by my own experiences, on work servers and all my home machines. 'fraid we'll have to agree to disagree.
I like my green drives The heads unload after a few seconds, but I don't allow the platters to spin down. So my wait time for readiness is trivial. They'll probably run this way for about a year, and then I'll upgrade to 3Tb drives, and the 2000EARS will be shelved. The transfer speeds are equal to my WD1001FALS drives
It's common knowledge that they're not ideal for everything, as in certain NAS applications as i mentioned, but for my uses, in media use at home plus obviously shunting large amounts of data around between drives, and in servers at work.