PSU tier list 2.0

[Nuckles_56]

Jonnyguru updated the scoring system for voltage regulation now. Anything above or equal to 0.5% now gets a deduction. So the EVGA 1000W T2 got deducted because regulation on the 12V rail was 0.5% exactly. Anything below is considered "mystic" instead of excellent now.

I'm actually surprised that he hadn't done it sooner, considering how many PSUs are close to that now

 

[turkey3_scratch]

Yeah, I think it's a good idea.

 

[Onus]

This brings up a point that a PSU with a mediocre score of "7" today might have earned a "9" only a few years ago, and is probably not a bad PSU, even though there are better ones.

 

[tiny voices]

Johnny's scoring system is all screwed up. When looking at his reviews, avoid the scores, and just read what is important.

 

[turkey3_scratch]

Is it just me, or is there a miscalculation on Jonnyguru for the 850W GS voltage regulation on the 12V rail. I did the math and it's definitely not 0.4%, it's 1.1%. http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=414

Also, great deal here:

PCPartPicker part list / Price breakdown by merchant

Power Supply: EVGA 1000W 80+ Gold Certified Semi-Modular ATX Power Supply ($119.99 @ Amazon)
Total: $119.99
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2016-01-13 18:17 EST-0500

 

[tiny voices]

You misread him.

Read here: "But what's even sweeter is the improved voltage stability over the cold tests. Now, I see 0.6%, 0.4%, and 0.4% - averaging 0.47%."

Those numbers are for the cold tests not the hot tests (chart above).

 

[turkey3_scratch]

The cold test he says 0.8% regulation for the 12V rail. I am looking at the hot test. The numbers are:

12.16
12.15
12.14
12.12
12.11
CL 12.18
CL 12.09

(12.16+12.15+12.14+12.12+12.11+12.18+12.09)/(12*7) - 1 = 0.0113095238095 = 1.1%

Jonny says: "Now, I see 0.6%, 0.4%, and 0.4%", which goes left to right, so he's saying it's 0.4% on the 12V rail. Unless you're saying Jonny is talking about a 0.4% improvement over the cold tests, which is also not true.

 

[tiny voices]

Huh that's confusing. Doesn't matter. Wither way it is a good PSU. I'd forget about it.

 

[turkey3_scratch]

I also just tested his cold test, and I get 0.95% rather than 0.8%. I think I'm going to go on the Jonnyguru forums and ask OklahomaWolf about this.

 

[Darkbreeze]

Had us going for a minute there.

 

[turkey3_scratch]

Okay, so apparently how it works is they take the highest voltage minus the lowest voltage and divide it by 12 to get their voltage regulation. Apparently this is load regulation (relative voltage sag from the initial to the final load), whereas I was calculating is closer to line regulation.

So basically, the +=5% is not the basis of the calculations on Jonny. You could have 12.55V as highest and 12.5V as lowest (obviously would be a really poor PSU) and it would be considered .4% according to their methodology.

That doesn't seem right to me. They should still be calculating offset in percentage from 12.0V along with the sag difference / 12.

 

[Darkbreeze]

That's pretty strange actually, considering we don't even typically condemn a PSU as "done haddit" until it's at like 11.6v, Jonny's site even says lower at 11.4 for discard.

http://www.jonnyguru.com/modules.php?name=NDFAQs&op=FAQ_Question&ndfaq_id=28


I've seen brand new Seasonic and Super Flower units that were only doing a consistent 12.1-12.3 on my Fluke.

 

[turkey3_scratch]

ATX spec says it needs to be within 11.4V-12.6V, so anything past that is obviously a definite "no", so Jonny follows the spec. But then again, quality in power supplies these days is quite good, I personally wouldn't recommend anything with voltages outside of 11.7V-12.3V. Most quality units that people recommend anyway are within 11.8V-12.2V or even quite better. It's a good thing that in general PSUs are higher quality than they used to be.

I just can't come to grasp with their methodology, though. They have strict scoring parameters for load regulation, but I don't see any strict parameters based on line regulation in their scoring process.

Anyway, I think the Tier list still needs a major overhaul. I encounter tons of units people have on this forum that are not present on the tier list. It can use a lot more power supplies added to it.

I'm thinking of a PSU Tier List 3.0. No offense at all to dottorrent, but I know he's very busy so perhaps it'd be better if someone else had ownership of the Tier List to update it more frequently? And get work done on an updated tier list. I encounter more units not on the list than I do ones that are on it.

 

[Darkbreeze]

That's because there are tons of people with discontinued, regionally unavailable and low end units. Not everything can be tiered. Some things simply don't merit the effort. If you have 90% of the most popular enthusiast and mainstream models covered, which I think we do, I don't think you can ask for a lot more than that. Plus, a lot of those units you see don't have any reliable/reputable in depth reviews to base tiering off of.

 

[turkey3_scratch]

I think some of these mainstream units need updated, though. I took a look at all the Antec High Current Pro models on Jonnyguru, and they definitely vary in quality. For instance, the 900W is not near the 1300W in quality. The 900W should be a Tier 2. It has 50mv ripple on the 12V rail, which back then was fine, but seeing that Superflower and Seasonic have units in 2016 now that have under 20mv on the 12V rail, they should be a tier higher. It also lacks any MOV in its transient filter stage, and line voltage regulation (I'll be more specific about this this time) probably more worthy of Tier 2. It also only has two PCIe cables for a 900W unit.

I don't know. I feel like the older units from 2010 and prior should be rechecked, because what was excellent and worthy of Tier 1 back then is not now.

Edit: Well, it seems that the Tier 1 high current are the platinums, so that makes sense.

 

[logainofhades]

Thoughts on the Evga GQ series? Apparently it is a new platform, from FSP.

 

[tiny voices]

It's meh in my opinion. Definitely solid units but I doubt I would buy one. I'm not a big EVGA PSU fan.

 

[logainofhades]

I am a fan of their B2 and G2/GS units. I was thinking they are probably a bit better than G1, but still not as good as G2/GS. The 650w GQ is going for about $70 right now. For those wanting modular, but don't need a 750w, the GQ could be a good buy.

 

[Darkbreeze]

Yeah, they're in between G1 and G2. For the price, I don't see where they serve a needed market. If you can get a better unit for somewhere between the same price and 10 bucks more, from the same company, seems pointless.

 

[okcnaline]

What is so bad about the CX series? Consider the first version. No 80+, but hangs around the 80% mark. Pretty good for something cheap.
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=214

And the CXM series? That's actually what one would call poor workmanship. Stuff ranging from problems with debris on the mainboard to some serious ripple spikes.
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story5&reid=416

The CX revision 2 is actually much better than the original 2.
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story4&reid=239

I think all that's sufficient to say that the CX unit isn't one with fireworks.

 

[tiny voices]

Who said the CX were fireworks? They are simply not great and there is always something better/ more sensible for the price. They are not a good buy.

 

[okcnaline]

Well, people gve me various reasons to dislike CX. They aren't good, obviously, if you put it against the EVGA SuperNOVA T2 1000. But people do that, so I don't get what the big fuss is about these CX units.

 

[tiny voices]

Because people use things like a CX750 to power an i7 and r9 390 in a $1500 PC. That makes absolutely no sense at all.

If you are spending the money to buy a PC that needs 600+ watts of power, there you should NOT be cheaping out and buying a Corsair CX. Plain and simple.

 

[Onus]

The issue with the CX is that it tests well when new, but is made with inferior capacitors that degrade rapidly, so they just don't last. With good parts, it would be a solid tier-3 (or even 2A) unit; as built it is not a tier-5 system-killer, but belongs on tier-4 of stuff to avoid.

 

[logainofhades]

Yeah, they're in between G1 and G2. For the price, I don't see where they serve a needed market. If you can get a better unit for somewhere between the same price and 10 bucks more, from the same company, seems pointless.

With it being in between G1 and G2, should we consider these Tier 2 level PSU's? G2 and GS cost more, @ that wattage. The GS is almost $20 more and the G2 is $30 more. The GQ 750 is a potential winner, as it is a bit cheaper, than B2, right now.