The 2014 El-cheapo Power Supply Roundup

[c_hegge]

http://hardwareinsights.com/the-2014 ... y-roundup/

[LongRunner]

On the page for the Aywun:

The primary capacitors are rated at 680µF and look to be from Nippon Chemi-Con (KMG series to be exact). However, we are sure that these parts are fake. They have short, thick snap-in pins, while real KMG caps have long leads and the KMG series datasheet doesn’t list a 250V 680µF part.

Just for the record, there was a snap-in KMG (as well as SMG) series in the distant past, although I don't have the specifications for either. But even then, the font as well as the colour (real 105°C Chemi-cons have brown sleeves, except for LXY/LXZ/LZA which are dark blue and older KZEs which were green) give the game away.

Also, two minor corrections regarding the Auriga/Macron:

Unfortunately, though, the critical capacitor is made by GoldLink, which is not one of the high quality manufacturers.

The two 10µF capacitors you see on the primary side are both in the part of the circuit that drives the main switchers — neither has anything to do with the standby supply. (If there was one in the two-transistor circuit, there would be a total of 3 small caps on the primary side.)

The fan is made by Superred. It’s a 1Ball/1 Sleeve bearing part, which are arguably the worst kinds of bearings available – they are noisier than traditional sleeve bearing fans, but are not inherently more reliable.

In the case of Superred fans, the B you saw on the label refers to the speed (A, B, C, etc.) — not to there being a ball bearing. I have seen 1B/S fans in those units but under Superred's model number system, such a fan would be the CHA0812BBS. So far I haven't seen one with a 2BB fan…

And why does the score section only take 2 points off for the (lack of) capacitor quality (I thought we had agreed on 3 points off)?

But anyway, I love how you listed:

Fireworks display was entertaining

…as a "Pro" for the Numan.

[c_hegge]

I'll fix those few things up at home when I'm back on my computer. As for the caps, it depend on just how low quality they are. If I see fuhjyyu, then 3 points do come off. For brands which I don't see failed quite as often, then only one or two points come off.

[Wester547]

Just for the record, there was a snap-in KMG (as well as SMG) series in the distant past, although I don't have the specifications for either. But even then, the font as well as the colour (real 105°C Chemi-cons have brown sleeves, except for LXY/LXZ/LZA which are dark blue and older KZEs which were green) give the game away.

KZJs have black sleeves, but they might as well be fake in way of their quality despite the fact that they've been out of production since 2007 or 2008...

I don't get why this Huntkey exploded at 350W and the other one with the 80mm fan exploded at 420W last year (maybe the switchers aren't being properly driven in this unit...). This one has a 120mm fan rated for 1350RPM, 47 CFM, and 29 DBA, and the other one had an 80mm fan rated for 2000RPM, 25 CFM, and 26 DBA... (both are wired directly to +12V)... the other unit did have lower VF rectifiers on the secondary, though. That might have something to do with it. This unit does have fairly good voltage regulation minus the -12V rail which takes me aback because usually half bridge units (as well as Huntkeys) excel at that even without a linear regulator but I suppose it depends on the rest of the design (the other Huntkey had this issue too). What's even sadder is the fact that the exhaust temperature wasn't all that high when it did finally explode. If it performs this poorly with an ambient temperature of 22*C... would not want to know how it fares in warmer weather. The FCON capacitors may also be a reason for the poor ripple suppression as I wouldn't be surprised if they aged drastically on the shelf, being a Chinese brand. And I noticed a small Teapo SC is right next to the +5VSB diode by the flyback transformer. I don't know what that Teapo is for but that can't be good...

As to why the unit exploded, I think you need proper protection on the primary side to properly avert that (assuming you haven't exceeded the individual current limit on each output). Similarly, I think you need protection on the output as far as +5VSB goes in order to have true protection. Bad output capacitors could still cause an overvoltage, I think, but with a recap this becomes a non problem. And I find it bewildering that the A-power's +3.3V output wasn't cleaner since it used linear regulation (that might be because the capacitors failed silently, though). Its voltage regulation is shockingly inept given that as well.

And as for its fan not having a proper sealing, I have noticed this issue with older Sunon sleeve bearing fans which use a cheap orange plug that comes off very easily. I oiled a Sunon sleeve bearing that came out of a Newton 250W earlier this year and noticed that the back of the sticker (which also came off very easily and didn't want to go back on. read: cheap) was wet (not "wet wet", but you could tell that oil had leaked out even after the sticker dried) but there was no lubricant left in the bearing (the fan was just starting to slow down and oiling it restored its speed and significantly quieted the fan). This suggests that Sunon fans have the same issue because I've seen new Sunon fans and the shaft is wet yet the sticker is not... and I noticed that their plugs are significantly thinner than say, ADDA's, Delta's, AVC's, and Yate Loon's, etc...

[LongRunner]

Also not indicated in the review is that +5V in the MPT-301 has 2 SBL3040PTs in parallel.

[c_hegge]

I have fixed those few issues up with the Auriga.

[LongRunner]

A few more goofs:

In the test 4 table for the Powercase, you slipped the test 5 efficiency result into the +5VSB load cell. And on the Aywun's page:

Not only is there visible damage to the washers and transistors, but it appears that they god hot enough to start de-soldering themselves!

Also, on the summary table for the Auriga/Macron, you accidentally left the regulation numbers from the Tsunami. They should be 4.8%, 2.0%, and 0.6% for the +12V, +5V, and +3.3V respectively.

[c_hegge]

Fixed those up too. The regulation numbers do not include the overload test, so the numbers on the Auriga are 3.25%, 1.6% and 0.6%.

[LongRunner]

For the +12V, yes, but the +5V is listed as 5.1V in the test 3 table, which is 2% high.

Also, on the A-Power's page:

…with the 80m fan version…

[c_hegge]

Haha. Fixed.