PSU tier list 2.0

Page 196 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.


A Raspberry Pi cannot be built by just anybody - you need at the least someone to make a PCB (too complex for home-brew), and a reflow oven or very very good soldering skills.

PSUs need very specific transformers that are likely custom-built, or you'll be at <70% efficiency.

Plus, you're playing around with mains. Big legal issues and it's hazardous as hell.

Stick with extra low voltage for DIY, please.
 


I was thinking along these lines the other day. With all the expertise Tom's has on hand, there's no reason that the community couldn't spec out a power supply (or other PC components) and approach an OEM for manufacture. Tomshardware branded components would sell like hotcakes.
 
RealHardX has no ratings though, although I'd say some level of 80+ means it's not tier-5 (assuming it's not fake); 80+ Gold is probably at least tier-3, but I wouldn't be surprised if there's an exception, like Coolermaster or Chokemax that might still only be 4.
 
Yes, there are PLENTY of units out there, Raidmax, Cooler Master, Thermaltake, Rosewill, Cougar, Coolmax, Enermax, Lepa, NZXT, Topower, Startech, Xion and many others who have units with various levels from Bronze to Gold that I wouldn't put above a tier 3 and many of them are much lower quality than that. A good many of them, despite 80plus ratings, are really worthless and probably tier 5 or just plain "crappy" if you want to go that route and don't like tiers, like some people.

So there are easily a bucket full of units out there with 80plus Bronze or higher ratings that in my opinion are not worthy of even being a tier 4 or "decent" rated unit. The Thermaltake TR2 and Cougar A series are good examples.

 
The issue I have with a 2 tier approach is its very simplicity. There's no room for mediocrity. You'll take most of those 'ok' psus and label them the same as 'excellent' psus, and the rest of the mediocre and label them junk. This just makes for a butt load of reading, trying to chase down reviews etc just to differentiate between a corsair CX and a SuperFlower leadex.

It's not hard to understand SR-71 liking the 2 tiers, he's already extremely well versed in the psu arena, so it's really not much more than a reference tool when he can't be bothered thinking (no disrespect intended) but for ppl with less education, they'll see the CX as 'good' so think it's great for the price. Having 5 tiers at least gives some definition to great-good-ok-bad-fugly.
 
I think best tier list would be

1. Green light. Always try to buy these.
2. Yellow light. Buy these if budget and availability is issue on the green ones.
3. Red lights. Avoid. Do not buy
 

This works, as does Inkiad's virtually identical three; basically taking the Go/Nogo idea that JohnnyLucky uses and adding budget/availability considerations. After all, why would anyone buy Coolermaster if Seasonic is available? Who chooses Raidmax if they can get Antec? If a goal is to simplify the selection process, there's no need to include so many superfluous choices, or superfluous tiers.
 
On the five-tier lists, given that Tier-5 units are [potentially] system-destroying PSU-shaped objects, a PSU that does have an honest 80+ [Any] label on it might still be low-quality and deserve tier-4, having been tested at 100% (although too cool), it is probably not going to kill your system merely because it was connected to A/C power. In the absence of all other information, I still maintain that 80+ is at least minimal assurance of some minimum but safe level of performance.
 
The only problem with that is we know for a fact that many companies have no qualms about outright LYING on their labels and many units that claim 80+ can't even come close. Many of them can't even operate under normal conditions.
 
Ok, I'm assuming (it's not hard to check, unlike claimed wattage) that the 80+ part of the label is honest. Then you can apply some de-rating for the difference between 23C and ~40C to get an idea of minimum actual wattage. One hundred percent of the claimed wattage was tested, for half an hour, and the unit not only did not croak, but it maintained efficiency so it wasn't on the ragged edge.
 
The two lists I have are not "tiers". It is just one list of recommended power supplies and one list of psu's to avoid. Both are based on competent technical reviews. Links are provided to the technical reviews. That's all. It is up to the consumer to do research and make an informed purchasing decision.

I specifically mentioned at the top of the psu recommendation list:

"NOTA BENE - Many well known brands and models are not listed because they were not reviewed by competent hardware review sites. Other brands and models are not listed due to average or below average performance during stringent testing. Still others are not listed because they failed to pass rigorous testing. To see which power supplies to avoid please check the PSU Lemon List."

I specifically mentioned this at the top of the psu lemon list:

"NOTA BENE - Only brands and models that were reviewed and tested by competent hardware review sites are listed. There may be additional brands and models that would qualify for the psu lemon list but they are not listed because they were not reviewed or subjected to rigorous testing."

Years ago when I inherited the recommended list from Proximon there were 7 review sites that published competent technical reviews. The number has been reduced and the number of new power supplies being released has also been reduced. That is understandable since sales of desktop pc's have been declining for years. Currently more small. portable, mobile pc's are being sold than desktop pc's.

My original thinking was to keep things fairly simply by eliminating "tiers". The problem with tiers is that there is no international standard that can be applied. When the discussion comes up in forums it usually boils down to personal opinions.




 
I don't get the problem with tiers, although I do understand that many of you don't like them. To me, a tier by any other name......


I see it like it's basically being said that:


Tier 1: There is no doubt these are great units.

Tier 2: We think, based on the reviews and available data, that these are pretty damn good units.

Tier 3: These units are ok and probably won't blow up or catch on fire. They might not last as long as some others and there are situations where they're probably not the best choice.

Tier 4: You could use one if you had to, but honestly it's a crapshoot and you're taking your chances that you won't soon be replacing it with something else.

Tier 5: If this has to be explained after knowing the criteria for tier 4, you probably won't believe or understand how bad they are anyhow. If you buy one, after being told they are dangerously bad units, you deserve what happens to your system when it fails.


Really, the tiers are just an expanded version of good/bad that makes allowances for more criteria. Great, good, ok, bad, really really bad. I don't see the issue. Since we largely recommend only tiers 1 and 2, it's very unlikely anybody that buys one of those units is ever going to regret it. For the most part, buying a unit from tiers 3, 4 or 5 are accurately going to be either a waste of money that could have netted you a much better unit for the same price or an outright hardware failure situation, so it seems to me that overall culling the separation of units by tiers, groupings, intended uses or any other categorical methodology boils down quite simply to semantics.

Good vs bad is still fundamentally a two tier list, regardless whether you call them tiers or not.
 
We are putting Corsair AX and EVGA G2 in the same tier. How is it justified? AX series is definitely better unit.

Putting EVGA B2, Seasonic G and XFX XTR in the same tier? IMO it is not doing justice to the better units.
 
Status
Not open for further replies.