Rhombu Tech is good for 250W continous, 300 peak. Nowhere close near 400.

PcBytes wrote on 2023-10-02, 11:11:

Rhombu Tech is good for 250W continous, 300 peak. Nowhere close near 400.

Yeh, this chinese c*ap is with at least twice increased power on the label (550 W), but in reality it will probably struggle to feed 250 W load continuously. The same applies for the sh**tiest PSU in this lot - LC Power 450 W maximum power on its label (really?! 😁 😁 😁) - maybe peak power for this will be no more than 170-180 W.

Not really, since the exact PCB design used for the Rhombus was actually rated for 250W, and the components look solid enough to do that. It's past 300W where things get ugly.

The LC Power is 0W, as not a single watt from it will be in spec, and should not be used to power anything, or to be powered on at all.

LC Power can be modified to AT-PSU and it will feed sucessfuly some old 386/486 system. I've even worse AT-PSU than this one for a 386SX system and it works fine from 1995 until now. The question here does it worth the effort? For me personally the answer is "it depends".

LC Power is not suitable for any PC. The components output too much ripple to be safe to use in any environment, to the point where I have seen AT PSUs fare much better than that crap.

The next two.
#1 (HP) feels heavy. Only drawabck is that it has only one cable with 2xMolex, one with 2xSata and the ATX plug is 24 pin and not 20+4
#2 (FSP) fells also quite heavy

PcBytes wrote on 2023-10-02, 14:20:

LC Power is not suitable for any PC. The components output too much ripple to be safe to use in any environment, to the point where I have seen AT PSUs fare much better than that crap.

Might work as a car battery desulphator 🤣

PcBytes wrote on 2023-10-02, 14:20:

LC Power is not suitable for any PC. The components output too much ripple to be safe to use in any environment, to the point where I have seen AT PSUs fare much better than that crap.

Agree to disagree. As I wrote, I run worse PSU (a totally no-name chinese fake 230 W sh*t, but within specs) than this LC Power one with 386SX and it just runs fine. Older systems are much more prone to ripple - you'll hardly find any low ESR electrolytic caps on cheap 386/486 mobos.

But anyway, this LC Power is the c*apiest of all the PSUs, next is Rhombu Tech.

HanSolo wrote on 2023-10-02, 14:39:

The next two.
#1 (HP) feels heavy. Only drawabck is that it has only one cable with 2xMolex, one with 2xSata and the ATX plug is 24 pin and not 20+4
#2 (FSP) fells also quite heavy

This HP-(re)branded PSU looks like massive build. FSP/Fortron is, lest say, OK (something like Power Man). Keep them both, if their output voltages are within specs.

Don't forget to remove the glue on the FSPs. It can get conductive and short out components.
Especially on the Powerman, that design is known to have glue that turns conductive.

HanSolo wrote on 2023-10-02, 14:39:

#1 (HP) feels heavy.
#2 (FSP) fells also quite heavy

passive pfc transformer is 1/3 the weight. At some point Chinese 🤣 brands were including one not connected to anything just to bump the weight 😀
https://electronics.stackexchange.com/questio … eally-connected
https://www.eevblog.com/forum/chat/fake-pfc-m … 5595/#msg105595

And the last 2.
#3 (LiteOn) feels heavy
#4 (Dell) I skip that one. The (short) power cable exits the PSU at the back of the underside. Who knows if it will fit any common board. And then the possibility of the non-standard plug. At least the tower gave me a SATA DVD drive and an AMD X2 4200. Better than nothing 😀

HanSolo wrote on 2023-10-02, 18:03:

#4 (Dell) I skip that one. The (short) power cable exits the PSU at the back of the underside. Who knows if it will fit any common board. And then the possibility of the non-standard plug.

At least the color coding for the Dell's wires at the connector looks like as of standard ATX-PSU. Check the output voltages with multimeter to confirm.

Lite-On - assume it's of same "OK quality" as FSP(Fortron) and Power Man.

Bonus 😀
Apparently I have two similar FSP/Fortron here. The second one (#9) feels somewhat lighter than the first (#2).
The FSP-sticker in my first post is from this lighter PSU shown here. (I didn't realize there are two and mixed that up.) The first and heavier one claims to deliver 30A on 5V (the rest is similar)

Reading through your comments it seems that except for the LC-Power all other seem to be at least 'ok', right?

That's about the size of it. I will cheerfully use a load of splitters and splices on Dell supplies though, as I have abused stray Dell PSUs at 120% of rated output before for long use periods. Though normally, just wanting 4 extra drives or something won't even get near the official output.

The "bonus" PSU is lighter than other rwo FSP/Fortron, 'couse it lacks passive PFC choke coil (i.e. no PFC) or it's (less likely) with active PFC schematic.

Yeah, the LC Power is the worst PSU in your lot, next is Rhombu tech - get rid of those. All others seems OK, but don't count just on our picture's judgement.

You must take proper measurements of the output voltages with some load before you decide to use them. Which PSUs aren't within the specs - give them for repair by someone who understands this type of switching PSUs (no offense, I figured out that you weren't much into electronics and repairs since you ask if it's easier to check voltages by readings through a motherboard's BIOS).

Good luck and... "May the force be with you" 😉

That bonus one is a APFC design of a FSP/Sparkle PSU.
The SPI markings on the transformers give it away as a FSP.
I would say it's the GLN design of FSP which I'd avoid at all costs.

PcBytes wrote on 2023-10-02, 14:20:

LC Power is not suitable for any PC. The components output too much ripple to be safe to use in any environment, to the point where I have seen AT PSUs fare much better than that crap.

How can you tell that there is a huge difference between the Rhombu Tech and LC Power? They look relatively similar to me.
Is it because of the smaller primary caps in the LC Power? Or the lack of caps on the secondary? Or something else?

I think I see a bad capacitor in the PowerMan (photo PSU_No08_PowerMan.jpg).
However, I would definitely keep that one and replace the capacitor because it has 30A on +5V, it provides 200W on +3.3V and +5V combined, and it has a -5V rail.
It's basically the ideal early ATX PSU.
I like FSP in general, they sometimes have bad caps, but none of them have failed after recapping in my experience.

analog_programmer wrote on 2023-10-03, 08:50:

The "bonus" PSU is lighter than other rwo FSP/Fortron, 'couse it lacks passive PFC choke coil (i.e. no PFC) or it's (less likely) with active PFC schematic.

Doesn't a single cap on the primary always indicate active PFC?

watson wrote on 2023-10-03, 09:52:

Doesn't a single cap on the primary always indicate active PFC?

I don't know it there's such a rule, but maybe this one really is with active PFC - now I compared its pictures to the sh**ty Rhombutech and seems that PcBytes is right. One cap or two in parallel - whats the difference, except that two in parallel seems to be more reliable (and maybe more expensive)?

watson wrote on 2023-10-03, 09:52:

How can you tell that there is a huge difference between the Rhombu Tech and LC Power? They look relatively similar to me.
Is it because of the smaller primary caps in the LC Power? Or the lack of caps on the secondary? Or something else?

I own both units (although the LC Power I have is the 12cm variant, but PCB is same, save for heatsinks being smaller to acommodate the top mounted fan) and the LC-Power has the worst materials used everywhere:

- PI coils on the secondary side are awfully thin (or even inexistent, though I did spot one through the wire harness)
- main switchers are GUARANTEED to blow past 200W
- capacitors on DC side are the lowest garbage known to mankind
- the only redeeming quality is the transformer
- unapproved disc capacitors on the primary side that can pose a fire hazard
- thin wiring on the common rails choke on the secondary side
- glass fuse
- awfully anorexic 330uF capacitors, nowhere you're gonna get 420W out of that, let alone 200W.

Rhombu-Tech isn't any much better on the AC side due to the small fuse (though it can be replaced with a better one at any time), but:

- thicker wiring on the common rail choke
- PI coils are decently thick, and are heatshrinked
- thermally controlled fan (which the LC-Power completely misses) including what I can assume is an external plug for motherboard reading of the fan
-680uF primary capacitors
- actual bridge rectifier instead of 4 puny diodes

So, this is why I would trash the LC-Power in favor of the Rhombu-Tech.
The Rhombu-Tech is made on a platform that although cheap, can have upgrade paths and can safely power up to 300-350W (the PFC choke actually states 350W so I'd say it could actually do 300W continuous provided the DC side has the capacitors replaced with japanese caps)

I have the 12cm version of the LC-Power and you can see how it looks even sadder than HanSolo's unit. The only things they were arsed enough to do was add a filtering AC PCB and different L shaped heatsinks.

I have done a lot of PSU experiments (replacing main transformers for bigger ones, recapping, upgrading rectifiers everywhere) but LC-Powers are totally awful units, at least the unknown OEM ones.
The GW(Great Wall) and Huntkey units are decent though - I have a GW-based LC6550H which looks much better. Picture is of the 6650H but the PCB on the 6550 is the same, save for lower rated secondary sillicon.

analog_programmer wrote on 2023-10-03, 10:01:

I don't know it there's such a rule, but maybe this one really is with active PFC - now I compared its pictures to the sh**ty Rhombutech and seems that PcBytes is right. One cap or two in parallel - whats the difference, except that two in parallel seems to be more reliable (and maybe more expensive)?

The two caps are not in parallel, they are in series, and they are typically rated at 200V.
This enables you to get a higher voltage out of 120 V line voltage. You rectify the positive half-wave on one cap, and the negative half-wave on the other.
This mode of operation is adjusted by the "115/230" switch. The resulting DC voltage is then ideally 120*sqrt(2)*2.
When operating in 230V mode, the DC voltage is simply 230*sqrt(2) stored across an effective "400V" capacitor.

Active PFC is basically a boost converter, meaning that it can theoretically boost any input voltage to (most often) slightly below 400V.
That's why you have a single 400V or 420V rated cap in PSUs with active PFC.
The only thing that I don't understand is why not all PSUs with active PFC have a full range input (110V-230V). Some are rated only for 230V. Maybe it is to save a few cents on some components?

Also, thank you PcBytes for the very detailed explanation. I am trying to learn as much as I can about PSUs and you are certainly a real expert.