[EOL] PSU Tier List rev. 14.8

[Juular]

Just now, jonnyGURU said:

So you've never read any of my reviews back in the day?  That kills me. 

If you're talking about your own reviews, i don't reread them every day, especially since 90% of them are down. And i don't understand what do you mean by 'low to high voltage drop' either, sorry, i don't think i've seen this terminology in other reviews.

[Guest]

8 minutes ago, Juular said:

If you're talking about your own reviews, i don't reread them every day, especially since 90% of them are down. And i don't understand what do you mean by 'low to high voltage drop' either, sorry, i don't think i've seen this terminology in other reviews.

Literally everybody.  It's like... sometimes the ONLY thing the reviewer touches on.

 

Example:  https://www.tomshardware.com/uk/reviews/how-we-test-psu,4042-3.html

 

Low load to high load voltage drop.

 

You take the PSU's highest voltage output.  And the lowest output.  You find the percentage of drop.  This is "voltage regulation".   The voltage should stay as close to nominal as possible.  This one is a yo yo.

[Juular]

1 hour ago, jonnyGURU said:

Literally everybody.  It's like... sometimes the ONLY thing the reviewer touches on.

 

Example:  https://www.tomshardware.com/uk/reviews/how-we-test-psu,4042-3.html

 

Low load to high load voltage drop.

 

You take the PSU's highest voltage output.  And the lowest output.  You find the percentage of drop.  This is "voltage regulation".   The voltage should stay as close to nominal as possible.  This one is a yo yo.

I see what you mean now. But isn't calling it 'voltage drop' a little bit weird ? It is drop if it drops from a nominal value. Like a voltage drop from the nominal right at the PSU modular receptacles to the end of the cables.

The difference between extreme values isn't a 'drop', it's for the lack of a simple term, min-to-max voltage regulation range or smth. Also, isn't it rather irrelevant what the difference between the extremes is ? If min-to-max voltage regulation is 5% but the median value is right at the nominal, then the deviation is 2.5%, which is okay ? ATX specs are defined as deviation from nominal in 5%, right ? So min-to-max is 10% assuming the median value is, again, right at the nominal. So say, in two cases with one having -5% nominal minimum value and 0% (right at the nominal) maximum, and the one having -2.5% minimum and +2.5% maximum, there would be 5% min-to-max in both cases, but clearly, the first case is way worse.

I prefer defining 'voltage regulation' as a maximum deviation from nominal value, whether it is lower or higher. It makes more sense this way IMO.

So what were actual voltage regulation figures for EVGA GA ?

[Guest]

21 minutes ago, Juular said:

I see what you mean now. But isn't calling it 'voltage drop' a little bit weird ? It is drop if it drops from a nominal value. Like a voltage drop from the nominal right at the PSU modular receptacles to the end of the cables.

No.  We've had this discussion to death on the JG forums.  Including what math to use.  Like, mix/max*nominal or max-min/max, etc.  Even the way it's measured on a Chroma is max - min.

[electropical]

5 hours ago, jonnyGURU said:

Even the way it's measured on a Chroma is max - min.

Is it measured like that because it's better (and easier) to maintain highly stable voltages within spec than it is to keep voltages around nominal?

Basically, is A better than B?

[Guest]

42 minutes ago, electropical said:

Is it measured like that because it's better (and easier) to maintain highly stable voltages within spec than it is to keep voltages around nominal?

Basically, is A better than B?

A would be better than B.  While the voltages are well within spec, the loads are static.  So in rear world use, you have to consider how much the voltages are bouncing during dynamic loads.  Therefore, nominal voltage is a moot point.

 

[Juular]

3 hours ago, jonnyGURU said:

No.  We've had this discussion to death on the JG forums.  Including what math to use.  Like, mix/max*nominal or max-min/max, etc.  Even the way it's measured on a Chroma is max - min.

Then it seems like i missed it.

Honestly i don't care which way it's measured by Chroma, we can just look at actual voltage numbers, not just voltage regulation value it spits out, can we ?

I understand that more stable voltages are better if they're slightly off nominal (to the higher side at least), that wasn't my point. My point is just min-max voltage regulation figures are misleading without context. 5% min max could be -5% to 0% or -2% to +3%.

Also, unless we're specifically talking about dynamic (transient response) voltage regulation, static voltage regulation doesn't really represent PSU's ability to keep voltages in specifications in real world usage, i mean, if it's bad in static then it would be even worse in dynamic, but if it's good in static that doesn't mean that it would be good in dynamic too. For dynamic voltage regulation, min-max value makes more sense because it's indeed better when it's tighter there, because large instantaneous voltage changes would actually affect stability, while for static voltage regulation it's a useless metric, changes are too slow to make any difference, unless it's actually goes out of specifications (and even there, a good VRM would be perfectly capable to cope with slightly off-spec voltages i presume if it's still relatively stable there).

[Guest]

4 hours ago, Juular said:

Then it seems like i missed it.

Honestly i don't care which way it's measured by Chroma, we can just look at actual voltage numbers, not just voltage regulation value it spits out, can we ?

I understand that more stable voltages are better if they're slightly off nominal (to the higher side at least), that wasn't my point. My point is just min-max voltage regulation figures are misleading without context. 5% min max could be -5% to 0% or -2% to +3%.

Also, unless we're specifically talking about dynamic (transient response) voltage regulation, static voltage regulation doesn't really represent PSU's ability to keep voltages in specifications in real world usage, i mean, if it's bad in static then it would be even worse in dynamic, but if it's good in static that doesn't mean that it would be good in dynamic too. For dynamic voltage regulation, min-max value makes more sense because it's indeed better when it's tighter there, because large instantaneous voltage changes would actually affect stability, while for static voltage regulation it's a useless metric, changes are too slow to make any difference, unless it's actually goes out of specifications (and even there, a good VRM would be perfectly capable to cope with slightly off-spec voltages i presume if it's still relatively stable there).

No.. That's just it.  It's NOT stable.  If the output voltage swings from a high to low value, even under static loads, it will do the same or worse under dynamic loads.  It doesn't matter what the nominal voltage is as long as the voltage is within spec.

 

Transient load testing is different.  Typically from off to full load, +5VSB to full load, etc..  You don't want to see any dips on the way up from that no load to full load rise.  The curve on the scope needs to be as smooth as possible.

  

[Juular]

2 hours ago, jonnyGURU said:

No.. That's just it.  It's NOT stable.  If the output voltage swings from a high to low value, even under static loads, it will do the same or worse under dynamic loads.  It doesn't matter what the nominal voltage is as long as the voltage is within spec.

Yeah, i agree, it's just that min-max value is not very informative regardless. But whatever, so you're sating that EVGA GA can't hold it's 12V even under static load ? What are actual voltage figures ? I assume it drops to out of spec values from about the nominal ? Is it consistent among samples, i mean, how many of them you've tested ?

[LukeSavenije]

On 7/1/2021 at 12:08 AM, jonnyGURU said:

Just as an FYI:  I've tested both the EVGA GT and GA.  GT is "not bad".  It's Gold, has about 5% voltage regulation.  Not as good as RM, but passable.  The GA.... is just terrible.  LOL!  Failed efficiency and load regulation tests.  It sure looks pretty, though. 

 

EDIT:  I will say that I ran the tests at both room temperature and 50°C and it didn't blow up.  So there's that.

 

I wouldn't mind getting the full numbers/report if you have them noted somewhere

[IIIIIIIIII]

On 6/29/2021 at 1:34 AM, IIIIIIIIII said:

Lian Li SP750 review

 

Cybenetics report

SP750 TPU review

[IIIIIIIIII]

Thermaltake Toughpower TF1 1550W cybenetics report

 

Yeah, It is CWT CTT.

Looks like they replaced the fan with Hong sheng → HongHua

[IIIIIIIIII]

On 7/2/2021 at 2:15 PM, Juular said:

Yeah, i agree, it's just that min-max value is not very informative regardless. But whatever, so you're sating that EVGA GA can't hold it's 12V even under static load ? What are actual voltage figures ? I assume it drops to out of spec values from about the nominal ? Is it consistent among samples, i mean, how many of them you've tested ?

In this review, At least, 12V and efficiency seem to be fine in a better environment than 50℃....

[seon123]

9 minutes ago, Nacht said:

Where is the Seasonic Prime GX-1300 PSU on tier list, i may buy new PSU soon

It is listed as the GX. It uses a crappy fan, lacks multi rail OCP, and is a glorified Focus. 

[Juular]

41 minutes ago, seon123 said:

It is listed as the GX. It uses a crappy fan, lacks multi rail OCP, and is a glorified Focus. 

Well, 1.3kW version is actually Prime, it's 550/650W versions that are based on Focus platform, but still, Seasonic Prime is usually not the best option neither value nor performance or components choice wise.

42 minutes ago, Nacht said:

Any better alternatives you may recommend that are least 1200w nearly same size or not ?

be quiet! Straight Power 11, Corsair HX.

[ldmendes7]

lol I've been using a Corsair | VS 2012 [orange-label] (Tier E • Potentially dangerous in multiple scenarios) for over 5 years now. My system is worth over $2000 with an RTX 2070 SUPER. 

I think I should probably update it xD

[ldmendes7]

Just now, Nacht said:

 

I would not want to lose my system due PSU that kills it all honestly in this current economy etc

Yep, gonna buy a new one now - have you got any recommendations? 

Something good quality that will last 650/700 watts range

[IIIIIIIIII]

30 minutes ago, ldmendes7 said:

Yep, gonna buy a new one now - have you got any recommendations? 

Something good quality that will last 650/700 watts range

Please create a new thread with a link to the store where you will buy the PSU.

[GoldenLag]

3 hours ago, Nacht said:

 

What about be quiet! Dark Power Pro 12 1200W ?

create your own thread, but the DPP12 would be a step above the HX or HXi

[Juular]

3 hours ago, Nacht said:

What about be quiet! Dark Power Pro 12 1200W ?

It's very good, just be aware that Dark Power Pro 12 and Dark Power 12 are very different. The latter is still good but not worth the money generally, it's just a bit fancied up Straight Power 11. But honestly even though DPP12 is basically one of the best PSUs on the market, i don't think it worth it over Corsair HX seeing how it's often like 150$ more expensive.

[Juular]

@jonnyGURUCorsair CX-M refresh ? CWT GPN ? Can we expect some reviews ?

PS: Also, damn, someone at Corsair loves this new grille design ...

[Guest]

49 minutes ago, Juular said:

@jonnyGURUCorsair CX-M refresh ? CWT GPN ? Can we expect some reviews ?

PS: Also, damn, someone at Corsair loves this new grille design ...

[IIIIIIIIII]

55 minutes ago, Juular said:

PS: Also, damn, someone at Corsair loves this new grille design ...

RM and the new RM White did this as well.

@jonnyGURU Is the fan curve of the new design RM the same as before?

[Guest]

1 minute ago, IIIIIIIIII said:

RM and the new RM White did this as well.

@jonnyGURU Is the fan curve of the new design RM the same as before?

[Juular]

@jonnyGURUAlso, could you guys please add some discerning suffix or smth, at least on the label to these all new refreshes ? Now when Corsair seems to use all the same design for new refreshes it's getting hard to distinguish them, there were differences in label color before. Something like Corsair CX-M v3 or smth, or just literally a year besides model (on the label). Yes, i know there's SKU code (or whatever you guys call it), like CP-9020*, but we can't just write a bunch of codes for the entire wattage range here...