$20 "500W" PSU From MicroCenter, how bad can it be?

[dmill89]

Here we have the Solid Gear (MicroCenter House Brand) Basix SDGR-500BX "500W" PSU, one of their cheapest ATX power supplies selling for a whopping $19.99,


lets take a look at how good/bad it is:

Box, pretty basic packaging:


It claims 26A on the 12V rail and 22A ea. on the 3.3V and 5V rail.



The unit itself is a pretty basic grey box as expected at this price point:


No APFC (not surprising at this price point):



Label claims the same as the box (at least the lies are consistent).




Lets take a look inside. Oh boy, this isn't good:




Primary has 2X "Rubycon" PX 220uf 250v caps (these caps look really old and dirty so they may be "recycled"), a pair of 13007 (8A 400V) transistors for the primary and a C3150 (50W 800V) on the standby. The input "filtering" (if you could call it that) consists of 1 x-cap and 2 y-caps directly on the mains connector. It is also using the 4 diodes treatment for the bridge rectifier.



On the secondary side we have a Mospec F16C20C (16A 200V) ultra-fast diode on the 12V rail (hmm didn't the label claim 26A??), and MCC MBR2045CT (20A 45V) Schottky diodes for the 5V and 3.3V rails (and the label claimed 22A for these).


Secondary Caps as follows:

3x Jayi 1000uf 16v
2x ZTLCon 1000uf 10v
2x Jwco 470uf 16v


Bottom of the board, not terrible for a cheap unit:


Lets see how it handles some load:

Starting with the 137.6W load from my Seasonic power supply tester

Note: the oscilloscope is set to the 100mV per division scale

12V:




Voltage at 12.586V which is in spec, but the 250mV ripple isn't, so this thing fails even with a fairly light load.

5V:




5.0335V, but 225 mV ripple again out of spec.


3.3V



3.1099V and 200 mV ripple out of spec on both counts here.

Power draw:


178W for a 77% efficiency.

Now lets add an additional 10A on the 12V rail to bump it up to 257.6W:

12V:



12.363V and over 300 mV ripple (and look at how ugly the ripple in the +/-100mV range is ).

5V:



5.082V with around 275 mV ripple


3.3V



3.1206V with around 225 mV ripple

Power draw:


338W for a 76% efficiency.


Well, it didn't blow up, though it did start to smell "bad" (like hot melt glue) in the second test so I doubt it would last very long under that load. The ripple is downright ugly though even under the lighter load so I wouldn't use this thing on any PC I cared about (or any PC for that matter).

[eccerr0r]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Keep on seeing that silly skinny extruded T heatsink... powmax?

looks like 200W... if that much...

KA7500/TL494 or what custom chip does it use?

[dmill89]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Keep on seeing that silly skinny extruded T heatsink... powmax?

looks like 200W... if that much...

KA7500/TL494 or what custom chip does it use?

UTC7520D (looks like a knock off of a Weltrend WT7520):

[cashkennedy]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

I don't blame them with all the idiots I hear still insisting on 750-850w power supplies for systems with midrange processors (ryzen 3600x), and upper mid video cards (gtx 1660). Those idiots (which seem to be the majority of people building computers) usually think 500w wouldn't be enough for any system with a discrete video card.

[dmill89]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

looks like 200W... if that much...

It couldn't even do the lower 137W in spec (granted that is an ancient PSU tester so it was very 3.3V/5V heavy, but that is also an ancient PSU design so it should actually handle that better than a 12V heavy load of a modern system), so as far as meeting ATX specs it looks like a <100W unit, granted it could probably handle 200W without blowing up, but it would be feeding your components a ton of ripple that would slowly destroy them.

At the higher 257W load (the built in 137W from the PSU tester + 2 60W electronic loads on the 12V rail) it really did feel like it was on the verge of blowing up with the smell it was making and the amount of hot air coming off the exhaust.

[PeteS in CA]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Keep on seeing that silly skinny extruded T heatsink... powmax?

looks like 200W... if that much...

KA7500/TL494 or what custom chip does it use?

The heatsinks, 13007 BJTs, and main transformer look like ~200W, but those 220uF input lytics are more like 150W. I'm surprised it lasted long enough at 257W to get all the readings and w/f's without glowing and launching.

[ChaosLegionnaire]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

ah well... i guess i'll just use it as a 100w bench testing power supply for testing non-pc related electronic components and hope it doesnt launch and release the magic smoke from those non-pc parts too...!

[momaka]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Before anything else, I want to say, THANK YOU! (For posting this PSU here.)

I've been curious how bad these SolidGear PSUs (more like SolidWaste, I suspect ) would be ever since I saw them hit the shelves of MicroCenter. When I was working there as a tech a few years back, we had store discount savings. But for some reason, we didn't get much or any at all for the store-brand or very cheap items. Either way, I've been thinking about buying one of these cheapo SolidGear PSUs to open and post here for a long time, but just never did. Most of this was because I didn't want to waste $20, and I kind of knew what to expect would be inside.

Or did I?

Seeing the pictures of the (lack of) build quality, I'd say I didn't expect such junk at all. Good you created a separate thread for this and didn't post it in the gutless PSU thread. I think this Solid Gear PSU is so bad that it can make even some of the crappiest PSUs in the gutless thread appear well-made compared to this.

Right off the bat, as soon as I saw the "500W" rating on the label, yet not even a single PCI-E power connector, I knew we'd be dealing with a crap PSU. Just how crap surprised me even more after you opened it.

I think the worst thing about this SG PSU is the small output toroid inductor. Not only that, but notice there is only one output toroid inductor instead of two, like you would expect on most group-regulated PSUs. This suggests the 3.3V rail should then be linearly regulated. BUT! That would require a MOSFET or other linear device on the secondary heatsink, and there is none!! So what regulates the 3.3V rail down to 3.3V, then?
Answer: NOTHING!
This means the 3.3V rail is derived the same way as the 5V and 12V rails - i.e. through its own winding from the main transformer. This mean there is absolutely no addition regulation for the 3.3V rail, and it can vary as the other rails do.

In other words, the 3.3V rail has an absolute garbage design. I've only seen this on an extremely gutless CWT-built ISO PSU posted here:
https://www.badcaps.net/forum/showthread.php?t=39222

I never tested that ISO PSU in a PC, as I feared the regulation and output ripple would be terrible (even in the junk PC I found it in I didn't test it with that PSU.) Your load test results, particularly those showing the 3.3V rail, confirmed my suspicion that it would have terrible regulation.

Although an old design based on half-bridge topology, I think this SG PSU would actually be more dangerous to use on an old PC than a new one - particularly one with SDRAM. Reason for that is because many SDRAM motherboards do not regulate power going into the SDRAM slots. Instead, it is usually fed directly off the 3.3V rail. So if the PSU outputs garbage on the 3.3V rail, the RAM will take that garbage and produce garbage results itself (unstable system if nothing else.)

Also, with the Seasonic load test @ 137 Watts and no additional loading on the 12V rail, I notice the 12V rail is quite high. I suspect SG or whoever manufactured this PSU for SG, added a few more windings on the 12V rain taps of the main transformer to get a higher voltage output on the 12V rail - most likely to account for 12V rail sag when loaded with a more modern 12V-based PC, like all PCs are nowadays. That's why it looks like both the 12V rail and 3.3V rail regulation improved quite a bit when an additional 10 Amps (120 Watts) load was added on the 12V rail.

That noted, I am also very surprised that this PSU did not blow up when it was loaded past 200 Watts. The smell was probably a good indicator that it likely would go bang if left loaded like that for an extended period of time. I suspect the smell was coming from the output toroid and maybe parts getting cooked near the heatsinks (and it sure would have been interesting to add a thermocouple to the heatsinks to see how hot they got. )

Lastly, I'm also surprised with the ripple and noise on the output with that load. Sure 200-300 mV is absolutely terrible. But with no more than one to two 1000 uF caps per rail and no PI coils, along with a tiny output inductor.... that's actually a decent result. However, even with all output filter cap spots populated and bigger capacity, I still don't think the ripple would be in spec past 150 Watts - mostly due to the lack of PI coils and small output toroid.

Anyways, thanks again for posting this, dmill. You took an item off of my crap PSU bucket list.
It's been a pleasure reading.

[dmill89]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Seeing the pictures of the (lack of) build quality, I'd say I didn't expect such junk at all.

Same, here I wasn't expecting a great power supply for $20, but I wasn't expecting it to be that bad either (especially since it is a store brand unit so there shouldn't be much in the way of markup between the manufacture and the retailer, but it looks like they cut pretty much every corner possible to ensure the highest possible profit margin rather than building an "ok" but basic unit). I've seen plenty of "free with case" units from L&C, HEC, CWT, etc. that were better than this thing. Even this $15.57 no-name unit from Amazon I looked at back in 2018 was vastly better: https://www.badcaps.net/forum/showth...ghlight=Amazon

[RukyCon]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

My Raidmax PSU has a single toroid for all three major rails, i wonder if they played a part in the making of this thing. :/

[momaka]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Same, here I wasn't expecting a great power supply for $20, but I wasn't expecting it to be that bad either (especially since it is a store brand unit so there shouldn't be much in the way of markup between the manufacture and the retailer, but it looks like they cut pretty much every corner possible to ensure the highest possible profit margin rather than building an "ok" but basic unit).

Not sure if SG is a MicroCenter exclussive house brand. I think other places sell them as well. Inland, on the other hand, is an MC house brand for sure. Not that this makes them much better either, I suspect (probably the same quality more or less.)

Either way, the bottom line is you're getting a pure garbage PSU for $20. However, if you do shell out twice that (about $40 or more), I think the PowerSpec 430 Watt PSU is actually pretty decent. I haven't had the chance to open one or looked any reviews, but I remember looking through the fan grills on an open package one, and it seemed pretty much similar to what Corsair and EVGA offered in that price range.

My Raidmax PSU has a single toroid for all three major rails

Single toroid doesn't necessarily mean poor design, though.
In the case of your Raidmax (and mine - I have a nearly identical but slightly more gutless version), there is a 40N03 MOSFET that provides linear regulation for the 3.3V rail.

As for whether Sun Pro made that SG PSU above... I dunno , though something tells me the SG is not based on a PowMax or Sun Pro platform.

[Per Hansson]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

Well at least it has those genuine Rubycon primary side caps going for it.
In all likelyhood scavanged by children at a dump somewhere in China.
Someone should do a investigation on a unit like this.
I imagine it is very similar to child labor for some expensive western clothes brands.

Can you get a picture of the back of the primary side caps, so we can see any date codes?
Found it on Mouser, price for 2 would be over $2 which is just impossible in the BOM cost for a unit like this.
https://www.mouser.se/ProductDetail/...%252BXMA%3D%3D

[stj]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

it actually looks like a board designed for a 250w psu in the 32bit days is just still being used in a different case with bigger lies on the label.

btw, i have a faulty supply coming maybe at the weekend that will make for great photo's

[RJARRRPCGP]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

it actually looks like a board designed for a 250w psu in the 32bit days is just still being used in a different case with bigger lies on the label.

QFT! Looks more like a typical PSU of the socket 462 days and it would be considered cheapo, for even back then!

[Per Hansson]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

A question, the writing on the PCB does it say this?
WARNING: For protection against fire we recommend that you install a fuse on PCB location "F1" that we have bypassed.

[stj]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

it's up-to-code.
isnt that what weller said when eevblog found their u.s. model soldering station had no fuse?

[PeteS in CA]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

to momaka for catching that really small output inductor core! An additional worrisome thing about it is that if enough power is drawn for the core temperature to get much above 90C or 100C, it's going to cause deterioration of the core's magnetic properties, leading to ever worsening ripple.

[lti]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

A question, the writing on the PCB does it say this?
WARNING: For protection against fire we recommend that you install a fuse on PCB location "F1" that we have bypassed.

I think there is one of those weird tiny fuses above the word "warning." It took me a while to find it because it's so small that it's hidden behind the wires.

[dmill89]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

I think there is one of those weird tiny fuses above the word "warning." It took me a while to find it because it's so small that it's hidden behind the wires.

Yep, there is one of those little soldered in glass fuses:

[stj]

Re: $20 "500W" PSU From MicroCenter, how bad can it be?

well that will be fun when it fails!!
not because it's small, but because it's not ceramic.