Should I worry about the PSU I ordered?

[NunoLava1998]

I'm switching to a small form factor build and therefore need a new power supply, so I got a beQuiet! SFX Power 2 400W.
My system is a Ryzen 5 1500X, B450 motherboard, 24GB RAM, RX 570 4GB.
I only learned after purchasing it that the power supply in itself is group regulated. That sucks, but I want to know if I should be worried about my system. The only review for it (an Eteknix review) shows the 5V, 5Vsb and 12V voltages actually being at reasonable voltage levels (0.5-1%) with the 3.3V being at average of 1-2%. This isn't crossload, but it certainly is better than a lot of group regulated power supplies; plus everything else in the PSU seems to be good. It's literally rated for 50ºC lol
So, my question is if I should worry about the power supply I ordered. The eteknix review in question (has internals): https://www.eteknix.com/quiet-sfx-po...supply-review/
Thank you

[Jon Gerow]

I wouldn't worry about it being group regulated. It's only 300W and only Bronze efficiency. So it's a cheap way to build an SFX PSU. That doesn't make it a BAD PSU.

[Tator Tot]

It's overall a decent unit, and a slight upgrade over the Silverstone ST45SF v1.0 (PC Perspective Review).

In your kind of system you shouldn't have much to worry about besides the noise output of the unit. It's going to get loud under load.

[NunoLava1998]

It's overall a decent unit, and a slight upgrade over the Silverstone ST45SF v1.0 (PC Perspective Review).

In your kind of system you shouldn't have much to worry about besides the noise output of the unit. It's going to get loud under load.

I'm going to get the 400W and my system only consumes about 300W. So it should be quiet.

[NunoLava1998]

Thank you all though! I was just wondering if it would be dangerous or reduce my system's longevity. I didn't think so, but it being a group regulated PSU I just wanted to check.
It should be delivered to me within a few hours; DHL says it's arrived at the delivery facility

[Orion]

In the PCPerspective test linked by Tator, the Silverstone (using the same platform) has pretty bad load regulation: the 12V rail dropped by 0.43V (3.5%), and it dropped down to only 11.53V by the full load test. And keep in mind they put unrealistically high load on the minor rails (7A on 3.3V and 7A on 5V in last test), which does group regulated units a favour.

In your PC, which won't load the minor rails anywhere close to 7A on each, I'm afraid this voltage and load regulation should be even worse. If the Silverstone performs so poorly with such high load on minor rails, then it will be worse with almost no load on them.
You can see this in their crossload test, where the 12V rail drops down to 11.42V when there's only 1A on the minor rails - that's only 0.02V more than the ATX spec allows.

Unfortunately we don't know how it handles transient loads. It uses a double-forward primary converter, so it could cause problems with top end graphics cards.
That said, your system isn't really high end and RX570 doesn't exhibit high transient power spikes (in addition to not needing much power in the first place), so it shouldn't cause big problems.

That voltage regulation is the biggest issue here, and PCPer's review shows this platform to have really bad regulation even as far as group regulated power supplies go (even Cooler Master Masterwatt Lite has better regulation), but it shouldn't reduce lifetime of any part significantly. It's just, kind of a shame to put a low end group regulated unit in a new PC, even if it's a mid-end one like yours.

The only review for it (an Eteknix review) shows the 5V, 5Vsb and 12V voltages actually being at reasonable voltage levels (0.5-1%) with the 3.3V being at average of 1-2%.

Unfortunately Eteknix don't know how to use their equipment, and they have no idea how to test voltage regulation. They use a completely wrong formula to calculate their average deviance - they sum up all individual deviances of a given voltage and then average them, which completely misses the point of the test and makes no sense whatsoever.
By their logic, a power supply that outputs 12.6, 12.3, 12.0, 11.70 and 11.4, would achieve a score of 0% - perfect regulation, when in reality such score is as terrible as you can reasonably get. They also don't say how much load they put on any individual rail, which invalidates their results further, and if they can't even get that right, then I don't think they know the proper procedure of testing ripple either.

To properly judge the Be Quiet, we'd have to know the same things as with any other power supply - load regulation, voltage regulation, ripple, transient response, crossload performance, temperature resistance, hold-up time, and then when it comes to build quality, we'd have to know the layout, capacitor series, the parameters of bridge rectifiers, diodes, and FETs (current ratings and temperature tolerance)... and of course if the protections work as they should.

Eteknix' entire review is worthless, so we have only PCPer, where they tested the same-platform Silverstone. The Be Quiet could perform better, but we have no way of knowing that. Silverstone showed really poor regulation with 12V rail dropping so much down, very bad crossload and good ripple. All the other things are still unknown, so it's kinda like trying to guess a graphics card's performance only from its clock rate... except the clock rate is already poor.

[Stefan Payne]

I'm switching to a small form factor build and therefore need a new power supply, so I got a beQuiet! SFX Power 2 400W.
My system is a Ryzen 5 1500X, B450 motherboard, 24GB RAM, RX 570 4GB.

Can you change your order?
If so the cheapest, decent Possibility is a 450W Silverstone SFX Series ST45SF V3.0 , wich you can distinguish from its predecessor by the bigger, 92mm fan.
What you should stay away from is something like the ENermax Revoultion SFX and Chieftec Compact Series because really awful fan controller (fan starts every other second. And no, I'M NOT joking!)

A little more (not much) is the new be quiet SFX-L Power and FSP Dagger...

Be quiet revisited the SFX-L Power and gave it a new internal NUmber (BN238) and replaced the older (BN214) models with it...
But I don't know what the Changes are.

[Orion]

I second the recommendation of ST45SF V3.0. In my shops (don't know about yours) it costs just 4% more than Be Quiet SFX Power 2, but it's already so much better - it has DC-DC converters, very good load and voltage regulation, great components (Infineon fets), great crossload, and ok transient and ripple.

[Stefan Payne]

And keep in mind they put unrealistically high load on the minor rails (7A on 3.3V and 7A on 5V in last test), which does group regulated units a favour.

We don't know.
It might be that the Unit is more designed towards 12V than 5V or that there are other things influencing the voltage drop. Sadly they didn't test that for whatever reason.
The 7A are NOT totally unrealistic though, if the Corsair integrated stuff in my HX750i is somewhat accurate, the 7A can acutally be archieved on 5V on modern AM4 Systems.
Didn't measure the ASROCK X570 Steel Legend though. Should have done it but the HX750 was in a different systen and I didn't think about it when I had the Steel Legend Running. Would have been interesting to see how much that uses on minor rails...

That voltage regulation is the biggest issue here, and PCPer's review shows this platform to have really bad regulation even as far as group regulated power supplies go (even Cooler Master Masterwatt Lite has better regulation)

Well, yeah, obviously the Master Watt Lite is better. But its also around 3 times the Size of the SFX Power. They are just THAT Small...

Unfortunately Eteknix don't know how to use their equipment, and they have no idea how to test voltage regulation.

Well, that's not necessarily the worst part, that is that:
Stingray DS1M12 USB Oscilloscope

It has only a bandwith of 250kHz, when 20MHz are required.
Someone calculated the time that the unit doesn't sample but it was somewhat significant for measurement and thus not viable.

And if someone asks about the Price: Every, known good, working, analogue Scope would have been better. Hell I'd even prefer an older CRT based Scope over earlier Digital ones (such as the Rigol 1052E) because the screens are just THAT bad. Only the recent generation seems to have a reasonable Display, though most of them without anti-glare film on the screen...

To properly judge the Be Quiet, we'd have to know the same things as with any other power supply - load regulation, voltage regulation, ripple, transient response, crossload performance, temperature resistance, hold-up time, and then when it comes to build quality, we'd have to know the layout, capacitor series, the parameters of bridge rectifiers, diodes, and FETs (current ratings and temperature tolerance)... and of course if the protections work as they should.

Build Quality is something we can't judge because we lack Knowledge and Training.
To be blunt: You need to be an Electrical Engineer, who works with Power Electronics all day and has specialized himself in that category to be able to judge the "build quality", all we can do is look for obvious faults such as bad solderjoints and judge it with that.

The Be Quiet could perform better

I'd assume that as be quiet is not known for Badge Engineering and instead custom modifications (even if its "just" the fan controller)...

But since its more of a lower volume part, I'd assume that the Changes are rather small, if at all...

[Orion]

The 7A are NOT totally unrealistic though, if the Corsair integrated stuff in my HX750i is somewhat accurate, the 7A can acutally be archieved on 5V on modern AM4 Systems.

Nice. What parts do you have in your system? I wonder how much of those 7A is taken by things other than drives.

Well, that's not necessarily the worst part, that is that:
Stingray DS1M12 USB Oscilloscope

It has only a bandwith of 250kHz, when 20MHz are required.

I didn't even notice, that's hilarious! That site is a disaster. (Or at least its PSU reviews.)