Another failure

[Skuzzy]

By the way, Corsair uses three different companies to make their supplies.  Seasonic, Channel Well, and Flextronics.  The problem is they are interspersed within the same family line (i.e. TX, GS...).

I can track down who made each model.  It does change from time to time.  You really never know what you are getting with a Corsair supply, if you do not check first.

[Bizman]

Here's the list of brands and their manufacturers from 2013 - although already three years old, many models still are available or at least in use: http://www.tomshardware.com/reviews/power-supply-oem-manufacturer,2913.html

More on the subject for those who like reading:
 
http://www.tomshardware.com/reviews/power-supplies-101,4193.html
http://www.tomshardware.com/reviews/psu-buying-guide,2916.html
http://www.tomshardware.com/reviews/best-psus,4229.html

[Condor]

Is anyone familiar with FSP PSUs? I have a 700W Aurum series 80+Gold. The original review several years ago was quite good. I looked at voltages in the BIOS as Skuzzy suggested. Just to be sure I checked the right values they were for Vcore, DDR15V, +12V, Vcc3, Vcc, and Vtt. I watched them for 5-10 min and could see that they updated (flashed) several times but the values never changed, They were the same out to 3 decimal places for the entire time. I assume that's good.

I suspect many of us don't give much thought to power supplies. Obviously that's a mistake. What's a reasonable life span for a good PSU? I've always thou ht of them as lasting forever.

[Bizman]

FSP does quite good PSU's. They both design and build them, as well as build for other brands. Way back in history they were known for having the smallest RMA percentage in the lists of French insurance companies (Take the source with a grain of salt, my memory may fail since it was a decade ago). A friend who's an electrician used to say they don't usually fail and when they do, they don't break anything else.

A good PSU should last as long as the entire computer. Then again PSU's are what I have to replace most often because of failure. But that's not because of the nature of the beast, it's because many cheap computers have poor power supplies. I also have a pile of fully working used PSU's salvaged from long abandoned (read XP) computers.

You're right in that people should give more thought to PSU's. An extra $30 or so shouldn't break the budget, at least not in a dedicated build.

[Skuzzy]

Power supplies have gotten more critical due to the shrinking of the silicon which has occurred.  The smaller the transistor, the more susceptible to transient/noise they become.

Many high end motherboards incorporate additional power filtering to deal with the lower quality power supplies available on the market, but filtering does not help bad regulation.  It only helps remove noise from the power.

Think if it this way.  If we took a transistor from 15 years ago, and scaled it to the size of a diesel rig.  Take a transient spike the transistor could deal with and scaled it using the same value.  It would be about the size of a Smart car.  Well, a Smart car hitting a diesel rig will be not more than a shudder to the diesel rig.

Now, take that scale value and apply it to the transistors available today.  Suddenly that diesel rig becomes a roller skate.  Now the Smart car transient which would not be noticed by the older transistor, takes a major hit.  Possibly damaging the transistor.

Hope that puts things in perspective and helps people to understand why power supplies have gotten so critical.

[MADe]

Do PSU's do any ac conditioning? or are they at the whim of basic power delivered.

If the voltage levels are changing, at all (for any of the voltages), then you have a cheap power supply that will end up taking out sensitive components, or giving you grief in other ways (i.e. lockups, overheating, noise...).


Skuzzy, how much to you consider fluctuation? 100th of volt or 10ths or whole numbers.


http://whirlpool.net.au/wiki/psu_manufacturers

[Skuzzy]

Ideally, the outputs never change, once they are up.

That is something I should mention.  If a power supply output is overloaded, it could cause drift.

No power supply, I am aware of, does input AC line conditioning.

[Chalenge]

I think you should do line conditioning from the power line to your house, and again just before the UPS. It can get rather crazy how much equipment you use in the end, but it's the best way to save your PC. Once you've seen how much should go into protecting your equipment you will look at the 'plug-and-play' computers in a completely different way.

[MADe]

being an apt dweller, I can only condition what comes out me walls. I know for a fact that it fluctuates. The pisser was every one of my UPS's died, like clock work, all having the same appx life span. My elec bill dropped $80/mth. So to clean the power was costly, the UPS's are costly, I turn pc off when not home so I am trying to convince myself not to go there, but.

[LCADolby]

Anyway got the SSD back from the rapair shop and it's a total loss. They couldn't recover anything. So I have lost all my unreleased training films and the scenario guncam film I was doing. I've lost 9 months of work... 

[Bizman]

Ouch! 

That's a good reminder to all of us to do some backing up. My photos are mainly only on the computer, only the oldest are duplicated on an external disk.

[Skuzzy]

I find the complete failure of the SSD to be odd and intriguing.  Theoretically, as long as the storage chips are not physically damaged, they should retain their data.

[MADe]

Anyway got the SSD back from the rapair shop and it's a total loss. They couldn't recover anything. So I have lost all my unreleased training films and the scenario guncam film I was doing. I've lost 9 months of work... 

that bites.
NAS is the wtg.

[Bizman]

I find the complete failure of the SSD to be odd and intriguing.  Theoretically, as long as the storage chips are not physically damaged, they should retain their data.

Supposedly, if the failure is in the circuit board, not many repair shops have the equipment to detach the storage chips and attach them to a working board. It might be even more difficult than swapping a hard disk platter.

[Skuzzy]

Well that is unfortunate, as it is not really that difficult to remove chips from an SMT board. A decent heat gun, an Exacto knife and BINGO! Pop those chips right off.

Use a board with sockets for the chips and get busy reading the data.  Or put them through another solder wave bath and get the data that way.