LTT Power Supply Tier List 1.0

[turkey3_scratch]

Many of the Tier 5 PSUs end up going out under high load or have out-of-spec voltages or ripple or noise, which prebuilt computers do not experience high loads, but still not recommended. As for criteria, I sort of have one in my head, but don't have a strict mathematical table of numbers -> tiers. I have a good understanding of what is important, so I use that to determine the tiers. It really depends much on everything as a whole, though I mostly rank things on its worst critical quality.

[quan289]

1 hour ago, djdwosk97 said:

1. The G2 1000w and 1300w use chinese secondary (or is it tertiary?) capacitors while the rest don't (iirc).

The 1000/1300 version of the G2 uses CapXon on the modular board where it is under less stress, and it is placed after the caps that is filtering the 12v and minor rails on the secondary side which are Japanese Nippon Chemi Con electrolytes and polymers. Since he stated that "it'll be placed in a tier in accordance with its worst trait" and "The good don't balance out the bad.", he automatically drop it down to tier 3. The problem with that is the bad does not negate the good aspect of the PSU either, and because of that, there needs to be more thought put into it when valuated each unit (which can be subject to subjectivity).

 

Otherwise, you could have a top-performing, heavily over-engineered unit that is mostly uncompromising except for one thing that could have made it to Tier 1 to be drop to Tier 3. I mean something like the older Seasonic X KM based AX850 being put in the same league as the group-regulated Seasonic GB (S12II/M12II-B) based XFX Core Edition 450/550 seem quite odd, don't you think?

 

Anyways, I can't check atm, as I need to go, but the majority of the the tier list I have seen except for one, often make a comment that they had based their list off of JG, HWS, [H], TPU, Toms, etc. but never actually attempt to asked about the logistics of the grading methodology and the accuracy of the list in general with them. If you didn't, I suggest you take a step back and asked them about your list.

[PianoPlayer88Key]

Ahh, interesting.   (I almost am curious if I'm a bit pickier than some when it comes to PSU quality for a non-gaming build, for like home/productivity uses - for example less than 10 mV ripple on 12V, 0.75% voltage regulation at all loads including crossloads, high quality 105C capacitors throughout, clean pass for efficiency, etc.)

 

And what do you mean by "which prebuilt computers do not experience"?  Is that assuming their out-of-the-box configuration, or what?  (I would guess that if someone is going to add two Radeon R9 295X2s or 390X2s to a Dell Inspiron or similar PC and crossfire them, they'd best get a new PSU first, even if they had enough Molex->PCIe adapters, and even so maybe their motherboard might not support it anyway.)

 

Also what about PSUs in older computers, like this Pentium II, and this Pentium 4 or similar-vintage AMD (I forget exactly what it is though)?  They worked for several years back in the day, although neither of them have been powered on, I believe, for like 10 years now.  We still have the 2nd one sitting across the room, but I haven't attempted to power it up.  (Also I'd need a VGA to HDMI or DisplayPort adapter to hook it up to my monitor if I was going to attempt to start it.)

 

Also in my first build, c. February 2008, I used an OCZ StealthXStream 500-watt PSU in it.  I believe it doesn't have 80+ certification, but it does have Active PFC.  I still have the PSU, but the computer died in 2012 sometime.  At the time, upon asking online what might have happened and posting about the symptoms, I was told that the part of the motherboard (a Gigabyte GA-MA69G-S3H) that had the iGPU probably died.  Looking back now, 1 - I wonder if I should have gotten a discrete GPU to try to get video running again, and 2 - I wonder if it's possible my PSU wasn't all it was cracked up to be?  I still have that PSU, like I said - is there a way to test it to see if it still works?  Would hooking it up to my current build (stripping it down to the bare necessities in the meantime, like removing most of the hard drives, and temporarily removing the Corsair AX760) be a way to test it, or too risky?

 

Oh, also, I've heard that generally, PSUs that come included with cases are considered junk, right?  Are there any exceptions to that, though?  For example what about this Antec NSK3180?

[turkey3_scratch]

As I said, the G2 was probably the first I did, and was more harsh, so let it go. Obviously everyone goes right to the G2 because it's the first PSU that comes to everyone's mind. There are so many other PSUs.

 

I do not even look at the scoring page on Jonnyguru, the scores on that site should be completely ignored. I analyze the data myself, and do not listen to what they say at the end of the review. I judge the worst quality, and that's my unchanging methodology, so you can be assured quality is strictly quality in every tier.

 

Old P3 PSUs have so much load on the 3.3V and 5V rails and aren't suited for modern rigs with a GPU. And Jonny is too nice on scoring, crappy units get 8s all the time.

 

Also, henceforth I am writing in green why I placed a PSU in that tier.

[PianoPlayer88Key]

I'm usually pretty picky when I look at Jonny's reviews.  Anything that averages about less than a 9.5 in Performance, Value and Build Quality would for me be a Tier 5.   (I really want it to get a 10 in at least 2 of those categories, and no worse than a 9 in whatever is left.  Of course if I'm considering a particular PSU, I will look through the entire review, focusing on the load tables & charts.)

 

(Also I saw that you replied while I was editing my previous post.)

 

And I too look at the load tables in his reviews.  I do look at the scoring, too, at least in part to determine whether I should look deeper into the review.  (And often out of boredom, desire to be entertained, etc, I will have fun looking at his gutless wonder reviews.)  I do enjoy reading his reviews, it seems he sometimes has a fun sense of humor.

 

Also I do believe the two older rigs I pictured do have discrete GPUs, and my brother (whose computers they were) was a gamer, and still is.   I think they're so old that 1 - integrated GPUs hadn't been invented yet, and 2 - possibly even so old that even flagship discrete GPUs only needed a single slot and a simple heatsink, maybe a small fan at most.

[turkey3_scratch]

Updates and additions have been made, and people will be happy to know those two G2s have been moved into Tier 2.

[PianoPlayer88Key]

@turkey3_scratch, wouldn't one of those Deltas, as well as the Hairong, in Tier UN, be candidates for Tier 4 or 5?  Or does Jonnyguru's bargain basement roundup not give enough info?

 

[turkey3_scratch]

1 minute ago, PianoPlayer88Key said:

@turkey3_scratch, wouldn't one of those Deltas, as well as the Hairong, in Tier UN, be candidates for Tier 4 or 5?  Or does Jonnyguru's bargain basement roundup not give enough info?

 

I really don't like the way he tested it. Those PSUs are made for heavy load on the 3.3V and 5V rails, and light load on the 12V rail, and he completely ignored that and drew way too much current on the 12V rail, over its rated specification.

[quan289]

3 hours ago, turkey3_scratch said:

As I said, the G2 was probably the first I did, and was more harsh, so let it go. Obviously everyone goes right to the G2 because it's the first PSU that comes to everyone's mind. There are so many other PSUs.

Sorry, but the G2 isn't the first power supply that comes to mind for me. I was merely took the initiative of answering another person question regarding your own grading methodology. It was just so happened that it was the G2. I mean I could say the Antec Signature 650w is essentially a server grade PSU being sold in the consumer-enthusiast market from 2008 that is over-engineered and perform electrical better than some of the modern day units (under 1% v.reg and under 20mVpp ripple suppression on all rails in accordance to Wolf's review at JG) even when being overloaded probably should be in tier 1 to be something else if you want, as most people here probably won't think of that unit.

3 hours ago, turkey3_scratch said:

I do not even look at the scoring page on Jonnyguru, the scores on that site should be completely ignored. I analyze the data myself, and do not listen to what they say at the end of the review.

I don't either. Because scoring doesn't give you the whole picture; otherwise, it will be like saying the Seasonic KM3 based Cooler Master V700 is worse than the Seasonic S12II-B 520w because it yield a lower score (9.1 vs 9.7) when it's in actuality the other way around.

3 hours ago, turkey3_scratch said:

I judge the worst quality, and that's my unchanging methodology, so you can be assured quality is strictly quality in every tier.

Well, that's why it is flawed. Since "it'll be placed in a tier in accordance with its worst trait" and "The good don't balance out the bad", you made it sound like you would negate any positive aspect of a particular unit. Tier 1 build quality but iffy ripple suppression on the minor rails may be group in tier 4 with units of similar ripple suppression but tier 3-4 build quality. That would give the impression that the two unit are of identical build.

 

Seeing how had you stated:

"As for criteria, I sort of have one in my head, but don't have a strict mathematical table of numbers -> tiers. I have a good understanding of what is important, so I use that to determine the tiers. It really depends much on everything as a whole, though I mostly rank things on its worst critical quality."

in one of your post that I had missed, it seem you do put some thought into it sometimes, but that "sometimes" can yield inconsistency, especially if you "sort of have" a grading system "in your head" - not a set, controlled one.

 

Anyways, the main reason I posted was mostly to say is this, whether the list make any sense to me or not "the majority of the the tier list I have seen except for one, often make a comment that they had based their list off of JG, HWS, [H], TPU, Toms, etc. but never actually attempt to asked about the logistics of the grading methodology and the accuracy of the list in general with them. If you didn't, I suggest you take a step back and asked them about your list." After all, I don't think they would appreciate being credit for something they may not necessarily agree with.

 

ATM, you seem like every other tier list maker who feel that they can make a better list but their background is pretty unknown. All I know about you is that, your knowledge about PSUs seem to be self-taught based on what you had said a little over two months ago at Jonnyguru when you were still in the process of learning and about the importance of load regulation three weeks ago. Not to mentioned, your list apparently is getting some feedback on their forum.

 

Anyways, I'm out.

 

 

[turkey3_scratch]

Yes, if you don't like my methodology, then Abschied.  You can make your own list, if you so desire. I'm just saying, I see you using Jonnyguru's score as a standpoint of judgement, which I do not even look at because of the mistakes in the scoring.

 

And Jonnyguru's old reviews checked line regulation, so what was told to me two months ago was ironic because they used to calculate line, not load. I respect your opinions, and respect your further absence.

 

Note: my question on Jonnyguru was before I started this list.

Edited February 9, 2016 by turkey3_scratch

[quan289]

1 hour ago, turkey3_scratch said:

Yes, if you don't like my methodology, then Abschied.  You can make your own list, if you so desire. I'm just saying, I see you using Jonnyguru's score as a standpoint of judgement, which I do not even look at because of the mistakes in the scoring.

 

And Jonnyguru's old reviews checked line regulation, so what was told to me two months ago was ironic because they used to calculate line, not load. I respect your opinions, and respect your further absence.

 

Note: my question on Jonnyguru was before I started this list.

A tier list is meant to be informative list for your readers to used, as well as expected to be accurate. If it's based on a flawed methodology, then the units in the respected tiers is questionable, and as such, it is disrespectful for the readers, as they are putting their trust in your list that could very well be flawed. If it wasn't flawed, then you wouldn't getting these inputs on your list at JG: http://www.jonnyguru.com/forums/showthread.php?t=13048

Also, saying you "respect your further absence" and "Abschied" to me shows that you don't wished to improved on your list. It also undermining your claim of wanting to learn more about PSUs.

 

I do not use "Jonnyguru's score as a standpoint of judgement". I mean I even replied "I don't either." to your "I do not even look at the scoring page on Jonnyguru". Meaning I too read the reviews.

 

Technically, Jonnyguru's reviews does have line regulation in it (deviations from the nominal values: 12V, 5V, 3.3V), but they don't put it into percentages. If you want to see it in percentage, you will do (([measured value] - nominal)/nominal)*100. If you see 12.3V on the 12v rail on one test as the highest value, that's a +2.5% line regulation ( ((12.3-12)/12)*100 ). While the lowest value of another test you see it to be 11.8, that's a -1.67% line regulation. Load regulation would be a 4.17% or a 0.5V voltage drop.

 

While Jonnyguru's older test did not offer the percentage, he does offer the voltage drop on the 12v rail as shown here: "the 12V rails dropped as much as .46V" or 3.83% load regulation.

 

Yes, I kind of figure this list was made after you had asked your question. My point was "is a person who was learning about what's a diode, bridge rectifier, PFC, etc are a little over two months ago qualify to be making a "tier" list".

 

But since it seem like you are sticking with that abschied, I guess?

 

[turkey3_scratch]

I have the right to make a Tier list. It is legal, so you must adhere to my free will. If you don't like it, make your own. That's what they told me before, which is why I did.

[quan289]

2 minutes ago, turkey3_scratch said:

I have the right to make a Tier list. It is legal, so you must adhere to my free will. If you don't like it, make your own.

Your right doesn't invalidate my right to comment on it. If you don't like the comment, feel free to ignore it.

[quan289]

But yes, since it's a waste of time, I'll take my leave.

[turkey3_scratch]

2 minutes ago, quan289 said:

Your right doesn't invalidate my right to comment on it.

I never made such assertions.

[ConiferousJelly]

What tier would the Corsair CX 600M Be put in? Tell me if i'm blind and its on the list but I didn't see it

[turkey3_scratch]

Funny you ask, I'm just updating the list with a bunch of CX/CX V2/CX M units.

 

It is a Tier 4 unit.

[Dive5885]

What Tier would the Cougar GX 1050 V3 fall under?

[turkey3_scratch]

5 minutes ago, Dive5885 said:

What Tier would the Cougar GX 1050 V3 fall under?

I'll check on that and add it as quick as possible!

[Dive5885]

3 minutes ago, turkey3_scratch said:

I'll check on that and add it as quick as possible!

Okay cool!

 

I had a EVGA 650G1 before now I have the Cougar PSU so I'm wondering if I should upgrade to something better than the Cougar unit or the Cougar PSU should be fine.

 

 

[turkey3_scratch]

9 minutes ago, Dive5885 said:

Okay cool!

 

I had a EVGA 650G1 before now I have the Cougar PSU so I'm wondering if I should upgrade to something better than the Cougar unit or the Cougar PSU should be fine.

 

 

The Cougar GX 1050 V3 will go in Tier UN. There is no review on it. There is one review on a site called nikktech.com but they have a horrible incorrect testing methodology of collecting voltage readings using computer software. Yuck, yuck, and yuck. They obviously can't get ripple data, either, so it's pointless. What I do know is that is uses Chemi-Con capacitors mostly, with some Teapo, which is fine. There are a lot of reviews of the non-V3, but none of the V3.

 

It's probably better than the 650G1, those are just poor.

[Dive5885]

1 minute ago, turkey3_scratch said:

The Cougar GX 1050 V3 will go in Tier UN. There is no review on it. There is one review on a site called nikktech.com but they have a horrible incorrect testing methodology of collecting voltage readings using computer software. Yuck, yuck, and yuck. They obviously can't get ripple data, either, so it's pointless. What I do know is that is uses Chemi-Con capacitors mostly, with some Teapo, which is fine. There are a lot of reviews of the non-V3, but none of the V3.

 

It's probably better than the 650G1, those are just poor.

Yeah that's the only review I found on V3.

 

I thought the 650G1 was good until people said not to use it for no more than 3 months..

 

Had me thinking I should go ahead and put back my V3 lol

[turkey3_scratch]

Most people talk about the NEX 650G poorly for its capacitors, but I think of the bad ripple and noise suppression as its largest downfall.

[Dive5885]

1 minute ago, turkey3_scratch said:

Most people talk about the NEX 650G poorly for its capacitors, but I think of the bad ripple and noise suppression as its largest downfall.

It was also a B-Stock Unit so I guess that was also a red flag.

 

Paid about $40 USD for it.

[79wjd]

1 hour ago, Dive5885 said:

Yeah that's the only review I found on V3.

 

I thought the 650G1 was good until people said not to use it for no more than 3 months..

 

Had me thinking I should go ahead and put back my V3 lol

 

1 hour ago, turkey3_scratch said:

Most people talk about the NEX 650G poorly for its capacitors, but I think of the bad ripple and noise suppression as its largest downfall.

The 650G1 might not be a great unit; but I definitely wouldn't be afraid of it though.