Power supply build quality pictorial. part 2

[c_hegge]

Re: Power supply build quality pictorial. part 2

It came from a Pentium II Machine btw. The owner brought it in last week saying he hadn't used it for a while and just needed his data saved off it.

[kaboom]

Re: Power supply build quality pictorial. part 2

You are trying me right?

First, I have never ever said anything about some VA meter, if someone is using such crap he is fool in the first palce.

Second, I am telling you it bloody matters if the unit has PFC or not. Most wall meters simply do not have fast enough sampling rate (which should in terms of MHz at least) to catch all the current harmonics of unit without PFC. As a result, they measure bullshit.

You speak bullshit.

Ranging from about -50 % of actual power to +50 % of actual power. Often it actually fluctuates between these values in real time.

You're really stuck on stupid, aren't you!! Of course it fluctuates! The computer you're measuring has ups-n-downs on the CPU, and other, loads.

Nice job contradicting yourself, too. "From about -50 % of actual power to +50 % of actual power." Excuse me! You're not even reading "actual," real, power unless it's a wattmeter, spinning disc W-Hr meter, TRMS DMM/scope/etc.

Have you ever even timed your POCO watthour meter, to see what you were using that very moment?

To measure that you need power analyzer which is built for such task in the first place.

Just like you need a "specific" device to measure currents below 100mA? You didn't think I'd forget about that litte tirade with Kiriakos, did you?

Again, all you need is either a wattmeter, such that the disc is acted upon by the [b]vector sum[b] of the flux from the V and I coils. Or, a TRMS DMM (or equivalent adapter for meter or scope) like the venerable Fluke 89. For "rougher" use, something likethe Fieldpiece HS36.

This:
http://www.candura.com/products/pqpro.html
is not used for measuring small-fry computer power supplies.

Don't try backpedalling, you've got your "sponsership" too far up your ass. Oh- wait, isn't it "maker madness" now?

You know, you almost seemed interested in one of my posts awhile back, where I mentioned scoping signal lines on a troublesome motherboard; you asked me to elaborate, and I did. Then, you never responded- as if you were put off (threatened?) by the detail I went into.

But what the hell do I know? I don't have a "website" where I "review" power supplies, therefore, according to you, I don't know two shits about watts, VARs, VA, and PF.

I guess I should just hack a website together, right? Then I can be god's gift to power supplies!!!


You're not weaseling your way out of this one.

Here are your power analyzers:
http://www.hialeahmeter.com/siphwame.html

The hell does PFC have to do with watt-hours recorded on a spinning disk meter?!?! Helll-lo! Anyone home? We don't have three-phase secondary-level distribution here, dammit, for typical resi services. Schools, offices, lite commercial (includes retail), anyplace with 480V services (or 480/277->208/120 stepdowns) or plain 3ph-4w 208/120 services (apartment "blocks," dormatories)- you will have triplens on the neutral from all the capacitor input supplies as the caps all charge at the peaks of the sinewave.

So the neutral is upsized! It's called a "super neutral!"

But for the common, single-phase pot, such as this one:

your theory is flawed.

The thousands of caps all "thumping" each half cycle to recharge is a real problem: flat-topping the sinewave. Scope the AC line with the correct divider probe and you'll see!

Don't come in here talking about how much bandwidth is supposedly required to measure anything. How much was there, for the last hundred years, in those old spinning disk meters? Say nothing about the newer "solid state" meters; bigger services have CTs and PTs to step the currents/voltages down to an acceptable level for metering. Applies, partly or completely, to anything from the 800A 208 service all the way to multi-feed primary-level, as for a modern hospital.

[kaboom]

Re: Power supply build quality pictorial. part 2

You really dont like Watt meters do you :L

Ignore that prick. There was a big fallout a few years ago, before you joined, over the same kind of issue.

It wast he absurd claim that APFC made the power supply so equipped more efficient. Misleading at best, patently false at middling, things I won't repeat at worst.

I ask you, why do the semiconductors in the PFC booster require heatsinks? Why does that boost inductor require quality cores wound with heavy wire? The answer, except for certain people, is self-evident.

The PFC booster, in itself, is a power supply, while uniquely configured, whose output feeds the single/double switch, half/full bridge "main" power supply. That's the one, in a computer, that give +/-12, +5, +3.3. +5VSB comes from part of the aux supply, always active with line voltage, independent of whether the booster is running.

There can be many power supply circuits in the box people call "power supply." The Dell Precision series (690, for instance) 1KW unit contains five power supplies!

Two aux supplies, one isolated (5VSB, supervisory opamps), and a line-connected one for driving the power MOSFETs in the booster and step-downs.

One 12V stepdown converter, for all the +/-12 load in the computer; over half the capacity is right here. It may be as much a 2/3 total capacity, split over several rails, with shunts to monitor current.

One 5/3.3 stepdown for raw voltage for +5 and +3.3. Each of those hits a magamp for truly independent regulation.

And finally, the PFC-booster.

[Per Hansson]

Re: Power supply build quality pictorial. part 2

Replying ad-hoc to some of the posts here, in the case of my shitty Etech PM-300 power meter the problem is I think that it can be fooled simply by a capacitor between phase and neutral.

Agreed, I have a Etech PM-300 power meter and just a few weeks ago figured out it's completely useless.
It was showing 32w consumption from a PSU with nothing connected to it when it was off, and sure the 5VSB can be inefficient on old PSU's but not THAT bad
I bought a Voltcraft Energy Check 3000 instead ofter checking some reviews, it shows a more sensible 3.7w under the same conditions
Turns out my original meter does not handle VA at all, yet the display shows "watt"
Yea right, watt if you load it with a purely resistive load like an incandescent light bulb, but I think I can read the rating on those myself so don't need a meter for that

Well, it is weird, because I have the same Kill-a-Watt as coreAngel, but I haven't seen it get tricked by non PFC PSUs yet. Actually, only APFC units tend to trick it. In particular, I have one PSU that shows to be drawing something like 8 or 9 Watts with nothing on the 5 VSB. Compare that to less than 2 Watts for my PPFC HiPro PSUs and about 2 to 7 Watts for my other standard 2-transistor designs (all without load and in "soft-off" mode).

I forgot to mention one thing, yes it shows 32w with an Antec TruePower 2.0 PFC PSU in standby (i.e. 5vsb active but nothing connected and PSU "off")
But if I turn off the rocker switch on the PSU the meter shows 16w, that's how I figured it out.
I.e. X and Y caps between phase/neutral and phase/neutral and ground respectively is enough to fool it!
The new power meter I bought shows 0w with the rocker switch off and 3.7w with 5vsb active but nothing connected...

[TELVM]

Re: Power supply build quality pictorial. part 2

This thing:

Velleman netbsem5

Works surprisingly well, it can read accurately from 0.1W AC from a mini-ITX system in standby, to ~1500W AC from a washing machine for a brief moment when it starts centrifuging, and from Oh Deers to modern PSUs.

190~250VAC only, Schuko plug.

[Behemot]

Re: Power supply build quality pictorial. part 2

…bs…

Let the true speak for itself.
https://www.hwworld.cz/files/SilverC...334-FR_PFC.rar
Unless you come up with some wild theory like we should not believe our own eyes and you are at least half the person you are trying to imply you are, I expect an apology.

For those who can talk like adults and do not need wet rag on their head first, this is my SilverCrest meter and undisclosed 550W PSU prototype runing roughly on maximum power.

Now I had that wonderfull Voltcraft Energy Monitor 3000 and it was better, but even that one was easily fooled at some moment unless the unit had power factor correction. That one died so I opened it and know basic principle of its operation which is common for all these wall meters. So I am fucking sure what I am talking about, unles Mr. Clever here. These cheap meters are intended just for simple resistive, capacitative or inductive load (or combination of both), not SMPS with harmonic distortion on frequencies starting at 60 kHz. So when your wall meter shows crap, it is crap as it is just unable to measure what flows through that.

TELVM: yeah, those are the specifications. Do you still believe them after what I showed you? Why else do you think most power analyzers start at 500 USD yet they often still cannot even measure THD for that price? I have selected one chinese analyzer which is and I am working my ass hard to get it so I can finally make really accurate readings.

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

Something brand new:

[kc8adu]

Re: Power supply build quality pictorial. part 2

brand new and i still see the damm brown glue!

Something brand new:

[momaka]

Re: Power supply build quality pictorial. part 2

But if I turn off the rocker switch on the PSU the meter shows 16w, that's how I figured it out.
I.e. X and Y caps between phase/neutral and phase/neutral and ground respectively is enough to fool it!

Interesting.
I tested my P3 Kill-a-Watt with a PSU like that (i.e. one that has an X cap between L-N and 2 Y caps between L-G-N... all of this before the rocker switch).
Result: 0 Watts with rocker switch in the OFF position. So I guess my meter isn't fooled at least by that .

190~250VAC only, Schuko plug.

Yay, Schuko plug! .
Wish I had a few in my room, along with 230/240 V AC.

[ReeceyBurger123]

Re: Power supply build quality pictorial. part 2

Ignore that prick. There was a big fallout a few years ago, before you joined, over the same kind of issue.

It wast he absurd claim that APFC made the power supply so equipped more efficient. Misleading at best, patently false at middling, things I won't repeat at worst.

I ask you, why do the semiconductors in the PFC booster require heatsinks? Why does that boost inductor require quality cores wound with heavy wire? The answer, except for certain people, is self-evident.

The PFC booster, in itself, is a power supply, while uniquely configured, whose output feeds the single/double switch, half/full bridge "main" power supply. That's the one, in a computer, that give +/-12, +5, +3.3. +5VSB comes from part of the aux supply, always active with line voltage, independent of whether the booster is running.

There can be many power supply circuits in the box people call "power supply." The Dell Precision series (690, for instance) 1KW unit contains five power supplies!

Two aux supplies, one isolated (5VSB, supervisory opamps), and a line-connected one for driving the power MOSFETs in the booster and step-downs.

One 12V stepdown converter, for all the +/-12 load in the computer; over half the capacity is right here. It may be as much a 2/3 total capacity, split over several rails, with shunts to monitor current.

One 5/3.3 stepdown for raw voltage for +5 and +3.3. Each of those hits a magamp for truly independent regulation.

And finally, the PFC-booster.

I know right, thanks for the info

[TELVM]

Re: Power supply build quality pictorial. part 2

... Yay, Schuko plug! .
Wish I had a few in my room, along with 230/240 V AC.

I wish our Schukos had unambiguous 'polarity' like your sockets do.

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

Here the rest of that unit I posted yesterday:


I think I don't need to say anything, do I?
Well, except for:
Thanks to Caseking.de for that unit

I wish our Schukos had

Well, depends..
Sometimes it's better to have the other...

Anyway: Schuko = Schutz Kontakt -> protected contact.
Another one for the idiotic names category (like every BUS there was or is in your IBM compatible PC X-D like Peripheral Component Interconnect...)

[Behemot]

Re: Power supply build quality pictorial. part 2

Interesting.
I tested my P3 Kill-a-Watt with a PSU like that (i.e. one that has an X cap between L-N and 2 Y caps between L-G-N... all of this before the rocker switch).
Result: 0 Watts with rocker switch in the OFF position. So I guess my meter isn't fooled at least by that .

They have internal power draw as long as you don't power them by batteries. My current meter always states at least 0.5 W because of that.

Stefan Payne: HEC? Interesting, I would have guessed some chinese maker…

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

Well, it's the same PCB as be quiets L8 series and some others...
EVGA also has some similar designs.

Just in a 350W version with tiny heatsinks...

Interesting thing by the way: With 360W primary load, that one is more silent than a Seasonic G-Series or Enermax Triathlor unit...

[Behemot]

Re: Power supply build quality pictorial. part 2

That does not mean anything…look at CM G550M, it is so silent…because the fan is hardly spinning and it was the first unit to overheat with sweater.

When you put slow fan in there it will be silent, but the consequences…

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

To use a PSU inside a sweater is not specified, so it is not something a PSU has to endure...

Besides: The 0,5A fan inside the Cougar STX350 isn't that slow...
According to my calculation that fan should be something around 3000rpm @ 12V or so (-500rpm)...

[Behemot]

Re: Power supply build quality pictorial. part 2

Well the maximum may be fixed by the fan controller…

It is not specified but good manufacturers have their units tested if they can work or at least safely die (without harming connected equipement) under special conditions like stuck fan. And some users, well…the sweater is nice simulation of situation between these two

[Agent24]

Re: Power supply build quality pictorial. part 2

Well for proper testing you need to see what the maximum operating temperature the PSU is rated for and then test at that temperature to see if it can meet the claims.

A sweater may help with this, but you need to do measurements too

[momaka]

Re: Power supply build quality pictorial. part 2

Most PSUs are rated at 25C ambient. Lite-ON, on the other hand, usually rate their PSUs to meet specs up to 50C.

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

Yeah, many Lower End Units are specced at 25 or 35°C Room Temperature, more higher end units are specced for 40-50°C, where 40°C room temperature becomes more and more the standard than 50°C...