Power supply build quality pictorial. part 2

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Got this power supply from work. The power button failed.

It was the type with the red light to indicate that the psu is powered.

This psu was in storage and as it seems it was sitting on the button. That may have caused the problem.

The button would turn to ON but during computer operation it would suddenly loose contact and the psu turn off.

I replaced the button with an ordinary black one without a light, taken from a gutless psu.

I have posted this model again while I was serving in the army.

Chieftec GPS-500AB

Delta Electronics is the OEM.

Unit is heavily overbuilt and I believe it can do more on 12V compared to the label.

It is also true triple rail on 12V.

the fan is AFB1212H ball bearing Delta Electronics. I even found datasheet.

Posting some pics for you guys here.

mosfets for pfc and main psu switching: 4x 20N60C3

3845B PWM IC

UCC3818N PFC IC

TNY268PN 5VSB IC

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

I have a modern Huntkey condition : never used, I will try to post pictures of it sometime.

Thank you momaka for this detailed presentation. I would fix it and keep it.

[TechGeek]

Re: Power supply build quality pictorial. part 2

Seems like a pretty solidly built unit for 300W, just has cheap caps, a cheap fan controller, an oversight in heatsinking and airflow, and cheap caps. With some simple modifications, this could be a pretty strong 300W PSU.

[momaka]

BFG Tech BFGR550WGSPSU [PCB M/N: LW6550SE Rev. K]

Different day, different PSU…
For today’s consideration is a BFG Tech “550 Watt” ATX power supply, model BFGR550WGSPSU.





It was a relatively inexpensive eBay score. I actually got two identical units from the same seller, though one was in considerably worse shape (that one will be posted another time). Both were sold “for parts or repair”. The price didn’t have anything to do with why I got these PSUs. Rather it was the curiosity of what’s inside them. In the auction, I could barely make out from the pictures that one of the PSUs possibly had bad caps. And that’s the primary reason I went for both.

So the above… let’s call it unit #1.

It looks shiny, modern, and definitely capable of 550 Watts, yes?
Unfortunately, one or more of the above statements are not true. Which one(s)? Well, the PSU case does have a nice black mirror finish, I’ll give it that. And with the 140 mm fan, an EPS-12V connector, and two PCI-E 6-pin power connectors, one might conclude this is indeed a modern PSU. However, look at the first case picture above again. See that input voltage selector switch? That’s a dead giveaway we are dealing with a PSU with no Active PFC. In my book, that’s usually a good thing, as PSUs without APFC tend to have less problems with primary caps going bad (or in general, with anything on the primary, save for bad startup caps, of course.) However, PSUs without APFC also tend to use older and less-efficient topologies (though rare exceptions do exist.) Thus, we are probably not dealing with a very modern unit. In any case, the bigger concern is the fact that the label states the maximum output as “peak”. So what’s the continuous power rating then? I guess we will have to open the unit and see what’s inside to get an idea.

Tops off!


The above 2 pictures should give a good overall idea. First off, we have 3 transformers and two input caps with a big polypropylene film cap on the primary – that’s not very modern. In fact, that’s our long-time friend, the good old Half-bridge. Output-wise, we see two large output toroids and two smaller toroids, suggesting group-regulated design with a mag-amp –generated 3.3V rail.
Welp… there goes our statement about it being modern. It’s an ancient design… but a well-done one, it seems. Frankly, I personally don’t mind older topologies like this. After all, half-bridge PSUs can perform decently and be fairly efficient too, if done right.

What bugs me more is the 550 Watt “peak” rating on the label. Looking at the marks on the second picture above, the PCB says “350-450W”. I can’t quite tell if that means this PCB is capable of handling models rated from 350 to 450 Watts or if it means 350 Watts continuous and 450W “peak” and the 550W on the label being a complete lie.

According to a quick search on the PCB model number (LW6550SE Rev. K), this PSU was made by HuntKey. Going by that and looking at reviews for similar HuntKey-built PSUs suggests this PSU may be capable of 450 Watts continuous and really peak up to 550W… or just a 350W PSU capable of 450W peak for a minute or so (JG tested one a similar-looking HuntKey and it blew up after 1 minute of load at 450 Watts.) So what’s your guess on this one?

Next, some specs about the components:
- Main transformer is 40 mm wide (i.e. “ERL40”.) 5VSB and driver transformers are standard 19 mm with proper height (not the cheapo short ones we see in gutless PSUs.)
- BJT’s on the primary side are a pair of 2SC3320 in TO-3P case.
- 5VSB is handled by a FSDM0265r PWM-FET IC – so no oldschool 2-transistor circuit here, at least.
- Input EMI/RFI filtering is adequate enough, as is the surge suppression (2x MOVs)
- Input caps are a pair of 200V, 1000 uF Teapo
- Heatsinks are pretty thick… though not very tall, due to the use of an overhead 140 mm fan. Not a whole lot of fins on them either, so they are not very good, IMO. I would have rather seen bigger/taller heatsinks with an 80 mm fan used instead.
- Bridge rectifier is a T15KB80, capable of 15 Amps @ 100C Tc with a heatsink… but only 3.2 Amps without @ 25C Ta. So it will probably run very hot at max PSU load and may not actually be adequate for the PSU at 120V AC line.

With that said, let’s move to the secondary side.

The secondary side is a bit cramped, IMO. However, the output rails have adequate filtering (2x 2200 uF caps for each major rail.) The main output toroid as well as the 3.3V toroid are large and use thick wire. In terms of design, this group regulated setup unfortunately does NOT use the newer method of generating the 12V rail, where a center-tapped 7V rail is superimposed on top of the rectified 5V rail (this reduces efficiency slightly, due to any current draw on the 12V rail also having to pass through the 5V rectifier, but it increases regulation and stability considerably.) Instead, this PSU uses the classic separate 3.3V/5V rail and 12V rail windings. On a similar note, the 12V rail uses two 20 Amp Schottky rectifiers in parallel. IMO, that’s not really ideal. If the PSU was built with a single 30 Amp rectifier or the two 20 Amp rectifiers were connected, such that the two diodes in each rectifier package are paralleled and handle only one tap of the 12V rail at a time, I think the 12V rail would have been able to handle more current (and is something I may try modding eventually.) Another item I would nick at this PSU is the -12V rail: although not really used anymore and not very important, I don’t like the fact that this PSU uses only a single 220 uF cap to filter it. Luckily, there is enough space around the -12V rail cap to fit an 8 mm cap (probably 470 to 1000 uF would be better.) This is another item I might mod.

Last but not least, I’m sure everyone has noticed the bad Teapo caps at this point. Probably every one of these units is destined with that fate. The other non-Teapo caps are FCon and KSC (though other units may come with similar dubious brands.) Moreover, it’s worth noting that the 140 mm cooling fan doesn’t have a sheet of plastic towards the back side to create better airflow inside the PSU. As such, airflow in the PSU is probably not optimal and likely contributes/accelerates the fate of the crap caps.

Now let’s move on to the solder/bottom side…

The soldering quality of the BGFR550WGSPSU is surprisingly good. I suppose that shouldn’t be too surprising, since Huntkey is also now a fairly big OEM PSU builder for major desktop brands like Lenovo, Acer, and etc. I didn’t find any large solder balls, long leads, or anything that could have caused a problem down the line.

And for the finish, we have the fan:

Yate Loon D14SM-12 (L-S04) sleeve bearing fan, rated for 0.7 Amps @ 12V. That’s quite powerful. Unfortunately, the speed controller of the PSU is not very good (haven’t traced it yet, though), as the fan sounds rather loud even when the PSU is not loaded. Something tells me I’ll have to mod that too.


Parts list………
ICs:
SG6105d (PWM controller, secondary side), AS339 (quad op-amp for handling some/all secondary-side protections), and DM0265r (5VSB PWM+FET combo, primary side)

Wiring:
* Input: 600V, 18 AWG
* Output: 300V, 18 AWG, except SATA and Molex HDD connectors after the first connector in every chain (20 AWG)
* Output connectors: 20+4-pin ATX, 4+4-pin EPS-12V, one 6+2 pin PCI-E, one –pin PCI-E, 4x SATA, 4x Molex drive connectors, 1x floppy connector.

Primary Side:
* Input Filtering: 1x 0.47 uF and 1x 0.1 uF X2-class caps, 2x 1 nF + 2x 2.2 nF Y2-class caps, 3x CM chokes
* Input protection: inrush current limiter, 2x MOVs for surge protection across primary caps., fuse
* T15KB80 bridge rectifier
* 2x Teapo LXK, 200V, 1000, 25x45 mm, 85°C caps
* 1x FCon KM, 50V, 47 uF, 6.3x11 mm as “startup” cap
* 2x 2SC3320 (TO-3P) BJTs for main PS
* 2x Teapo SC, 50V, 6.8 uF, 5x11 mm for BJT drive circuit

Secondary Side:
* 5VSB
*** 1x FCon GF, 16V, 1000 uF, 10x20 mm before PI coil
*** 1x Teapo SC, 10V, 2200 uF, 10x20 mm after PI coil
*** 3 Amp(?) schottky diode for rectification
*** PI coil: 7-turn, 20-AWG(?), 5 mm core
*** Load resistor: 1x 1.8-KOhm, 1/8 Watt, SMD

* 3.3V Rail
*** 1x BH, 10V, 2200 uF, 10x20 mm before PI coil
*** 1x Teapo SC, 10V, 2200 uF, 10x20 mm after PI coil
*** 2x STPS2045ct schottky rectifiers in parallel
*** PI coil: 5.5-turn, 16-AWG, 4 mm core
*** Load resistor: 8x 100-Ohm parallel SMD resistors (12.8 Ohm load total)

* 5V Rail
*** 2x Teapo SC, 10V, 2200 uF, 10x20 mm with PI coil in between
*** 1x MOSPEC S40d40c schottky rectifier
*** PI coil: 5.5-turn, 16-AWG, 4 mm core
*** Load resistor: 7x 220-Ohm parallel SMD resistors (31.4 Ohm load total)

* 12V Rail
*** 2x Teapo SC, 16V, 2200 uF, 10x30 mm before PI coils
*** 2x STPS20s100 schottky rectifiers in parallel
*** PI coils (2x): 6.5-turn, 16-AWG, 5 mm core
*** Load resistor: 6x 1.2-KOhm parallel SMD resistors (200 Ohm load total)

* -12V Rail
*** 1x FCon GF, 16V, 100 uF, 6.3x11 mm after PI coil
*** 2x or 3x FR158 diodes for rectification
*** PI coil: multi-layer, multi-turn in shrinkwrap
*** Load resistor: 4x 330-Ohm + 1x 1.2-KOhm parallel SMD resistors (258.8 Ohm load total)

[momaka]

Re: Power supply build quality pictorial. part 2

I also got some other stuff for free that I have been struggling to find some free time to try to fix them, including a 32" 16:9 Sony CRT TV...

Nice! Those TVs are pretty nice, actually.
If it's not working and you're looking for a specific part for the power supply, let me know. I have saved several PCBs from Sony TVs from that era (none were widescreen, but many have similarly-designed power supplies and thus ICs on them), and chances are I might have it (like the PWM controller that likes to go bad, for example... I forget its part name now.)

Premier DR-8460BTX (MAX 450W)
...
The biggest letdown is that it doesn't have even a single Sata connector and of course no PCI-e or 8pin ATX 12V connector.

Yeah, I don't know why some cheapo PSUs continue to put so much old-style Molex connectors and just a few SATA (if any.) Granted this Premier PSU is old... but I'm pretty sure when it was sold, SATA was around for a while. So Deer should have put some on there.

As for 8-pin EPS 12V connector - I wouldn't put one on this PSU, lest someone happen to acquire it and think it can run a power-hungry 150+ Watt TDP CPU. A 4-pin 12V CPU connector and *maybe* a single 6-pin PCI-E is all I'd add to a unit like that... that is, if the 12V rail is even stable enough under any moderate load. Upgrading the 12V rectifier would surely help, but definitely consider replacing the main output toroid inductor too. The tiny one that is currently in there will likely start overheating with a high load.

Other than that, this PSU is one of the better Deer/Premiers, for sure.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

It will do 300W. The limiting factor is always the efficiency (or lack thereof). 3.3V and 5V diodes can be left alone since those rails do not see much use in modern computers. Use at least 60A diode for 12V and it should be fine.

The reason why you see such a huge overrating of diodes in brand name PSUs is that the higher the current rating of the diode, the lower its voltage drop. Lower voltage drop, higher efficiency. And with heatsinks as flimsy as that, you'll want to squeeze as much efficiency out of it as you can, so overrate your output rectifiers a lot, and also swap the output inductor if you have a larger one in your junk box. Upgrading the primary transistors will do nothing if the secondary is inefficient, because these transistors do not burn out because of overcurrent, the reason is always thermal runaway. If you can find one that fits, replacing the primary heatsink with a larger one is going to do more good than replacing the transistors with higher rated ones.

The reason these pieces of junk blow up at 250-300W output is the very low efficiency due to inadequate output filtering, high voltage drop in output rectifiers and thin wiring in the output inductor. Improving the efficiency of the secondary side makes the stress on the primary lower for the same output power.

Very interesting info here, thanks

200-250W max is what it will do. That is not a real 35 transformer, neither a 33 size. It's a 28 size transformer. Notice how it doesn't cover all the space where the transformer goes. 33 and real 35 transformers will cover the WHOLE space I've marked with red.

True, but it's taller than 33 size and thinner than 35 size.

[Dan81]

Re: Power supply build quality pictorial. part 2

200-250W max is what it will do. That is not a real 35 transformer, neither a 33 size. It's a 28 size transformer. Notice how it doesn't cover all the space where the transformer goes. 33 and real 35 transformers will cover the WHOLE space I've marked with red.

[Th3_uN1Qu3]

Re: Power supply build quality pictorial. part 2

So, with a full recap to upgrade output filters to 2x 2200uF and coil, replacing primary caps with 2x 680uF, primary transistors replacement with 13009, and secondary rectifiers of at least 30A for each output this could do 300W max.

It will do 300W. The limiting factor is always the efficiency (or lack thereof). 3.3V and 5V diodes can be left alone since those rails do not see much use in modern computers. Use at least 60A diode for 12V and it should be fine.

The reason why you see such a huge overrating of diodes in brand name PSUs is that the higher the current rating of the diode, the lower its voltage drop. Lower voltage drop, higher efficiency. And with heatsinks as flimsy as that, you'll want to squeeze as much efficiency out of it as you can, so overrate your output rectifiers a lot, and also swap the output inductor if you have a larger one in your junk box. Upgrading the primary transistors will do nothing if the secondary is inefficient, because these transistors do not burn out because of overcurrent, the reason is always thermal runaway. If you can find one that fits, replacing the primary heatsink with a larger one is going to do more good than replacing the transistors with higher rated ones.

The reason these pieces of junk blow up at 250-300W output is the very low efficiency due to inadequate output filtering, high voltage drop in output rectifiers and thin wiring in the output inductor. Improving the efficiency of the secondary side makes the stress on the primary lower for the same output power.

The biggest letdown is that it doesn't have even a single Sata connector and of course no PCI-e or 8pin ATX 12V connector.

Solder in extra cables.

[Dan81]

Re: Power supply build quality pictorial. part 2

It might actually do 250W max. That transformer is actually a ERL28. I upgraded one just like that with an real 35 transformer (from another discarded Allied) and after desoldering the old transformer to replace it, I saw the "ERL-28 XP" text underneath.

Transistor-wise, I can't stop laughing at how big pads they stuck on the main transistors (13007s?), even though they probably don't put out that much heat. Can be considered a plus somewhat, since you don't need replacement pads if you go with 13009s in TO-3P package.

I might post my modded AL-8400BTX (12cm variant) here, since it might be worth it.

LOL
When the "highest end" Codegen is compared in decency to a Bestec 250-12Z and deemed "on par"... now that's sad.

Not that there is anything wrong with the Bestec ATX-250-12Z. I actually like those and they are very reliable with good caps. It's just that they are... very average PSUs in terms of 12V rail capability nowadays. Of course, for a regular office machine they are absolutely fine. And even for an entry-level budget gaming PC, they might handle something like a low-end modern i5 and a GTX 1050 or similar (i.e. a system with no more than 150W draw from the 12V rail.) So they are not worthless or useless by any means, and I still save them. The only thing they lack is SATA connectors.

I remember that the one I saw for sale was actually a pretty sturdy one.
It had thick heatsinks similar to Bestec (but half the height since it was a 12cm fan unit), a heavy PFC choke, and a complete line filter. I don't remember how much it was rated for (I think somewhere in the 400-450W range) but it was actually well built. The only thing it needed was a recap. (even my FSP Bluestorm II needed one before I put it into service so that's not something new)

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Cool!

I thought you'd have given up on fixing PSUs anymore, perhaps due to being busy with other things in life. For me, it's still a hobby that I enjoy doing. Glad to see you still have your "mojo" for it.

Well, I have been quite busy for a couple of months and the only PSU that I fixed was a Xiaomi 5V 1A USB charger (bad caps ). Not sure if it counts

I also got some other stuff for free that I have been struggling to find some free time to try to fix them, including a 32" 16:9 Sony CRT TV, some micro Hi-Fi systems, a Sony AVR and more. I will create a new thread about each one of them soon.

Just curious, does it have the same output filtering number and size of caps as the other Codegen you posted? (i.e. only 2x 1000 uF on the 3.3V and 5V, and a single 1000 uF cap on the 12V rail?) If so, then it's probably more realistic to say it's a 150W PSU, due to output ripple likely to become an issue at higher loads.

Yes, they are identical.

Another concern I've been having lately with some of my older/cheaper PSUs is regulation. In particular, the better group-regulated PSUs employ a technique, where the 12V rail is derived by superimposing a "7V rail" on top of the rectified 5V rail. This hurts the efficiency a bit, as any current drawn on the 12V rail also has to go through the 5V rail rectifier. But on the other hand, loading the 12V rail also loads the 5V rail rectifier, and so voltage regulation is much more stable.

Most older "gutless" Deer/Allied/Solytech and many other cheapo half-bridge PSUs don't use this technique and thus tend to suffer in voltage regulation when loaded with 12V-heavy or 5V-heavy PCs.

Interesting




Here is another one:

Premier DR-8460BTX (MAX 450W)

I like this one and I plan not use it for parts but maybe someday recap it and upgrade it.

It even has metal oxide varistos and fan controller on a separate board.

It also has 2 caps and 1 pi coil for every output voltage, except -12V.

The input filter is complete, no parts missing and there is also a proper bridge rectifier.

IC is chip of the year 2005.

secondary rectifiers:
sb3045st @ 5V
stps2045ct @ 3.3V
SBL2060CT @ 12V

So, with a full recap to upgrade output filters to 2x 2200uF and coil, replacing primary caps with 2x 680uF, primary transistors replacement with 13009, and secondary rectifiers of at least 30A for each output this could do 300W max.

The biggest letdown is that it doesn't have even a single Sata connector and of course no PCI-e or 8pin ATX 12V connector.

[momaka]

Re: Power supply build quality pictorial. part 2

Posted a recap job of the above Delta DPS-300AB-24b here:
https://www.badcaps.net/forum/showthread.php?t=84121

Interesting that the multi-output output inductor is mounted on a base. That might improve its stability under vibration.

I think all decent manufacturers do that anyways.
Not only does it make the toroids more stable / able to endure vibration and shock better, but I think it's also done to help with manufacturing, as the inductor leads are spaced according to the base. Otherwise, I think a human would need to install the inductor manually, and that can increase costs for the large PSU manufacturers. Of course, for the cheapo PSUs that probably use child labor or very cheap labor in general, I'm sure they have someone installing those inductors by hand. And a lot of times, you can tell - the build quality simply isn't that great.

I got more than 20 power supplies for free, but most of them belong to the gutless psu thread.

Cool!

I thought you'd have given up on fixing PSUs anymore, perhaps due to being busy with other things in life. For me, it's still a hobby that I enjoy doing. Glad to see you still have your "mojo" for it.

EC 350X
...
So, I wouldn't call this a gutless power supply, just a half decent old half bridge 250 watter.

Yeah, not a bad PSU if honestly rated.


Just curious, does it have the same output filtering number and size of caps as the other Codegen you posted? (i.e. only 2x 1000 uF on the 3.3V and 5V, and a single 1000 uF cap on the 12V rail?) If so, then it's probably more realistic to say it's a 150W PSU, due to output ripple likely to become an issue at higher loads.

Another concern I've been having lately with some of my older/cheaper PSUs is regulation. In particular, the better group-regulated PSUs employ a technique, where the 12V rail is derived by superimposing a "7V rail" on top of the rectified 5V rail. This hurts the efficiency a bit, as any current drawn on the 12V rail also has to go through the 5V rail rectifier. But on the other hand, loading the 12V rail also loads the 5V rail rectifier, and so voltage regulation is much more stable.

Most older "gutless" Deer/Allied/Solytech and many other cheapo half-bridge PSUs don't use this technique and thus tend to suffer in voltage regulation when loaded with 12V-heavy or 5V-heavy PCs.

One last thing that I liked, is that the 5v stand by rail is handled with a 5H0165R Power Switch from Fairchild. It is very good to see something like this on a power supply that old instead of the typical 2 transistor 5vsb circuit.

Nice!

Their highest end (Codegen's) PSUs actually have same PCB. Just the heatsinks are MUCH thicker (on par with Bestec 250-12Z) and a 35 size main transformer. Of course, the heatsinks are low profile to acommodate tbe 12cm fan.

LOL
When the "highest end" Codegen is compared in decency to a Bestec 250-12Z and deemed "on par"... now that's sad.

Not that there is anything wrong with the Bestec ATX-250-12Z. I actually like those and they are very reliable with good caps. It's just that they are... very average PSUs in terms of 12V rail capability nowadays. Of course, for a regular office machine they are absolutely fine. And even for an entry-level budget gaming PC, they might handle something like a low-end modern i5 and a GTX 1050 or similar (i.e. a system with no more than 150W draw from the 12V rail.) So they are not worthless or useless by any means, and I still save them. The only thing they lack is SATA connectors.

[Dan81]

Re: Power supply build quality pictorial. part 2

Their highest end (Codegen's) PSUs actually have same PCB. Just the heatsinks are MUCH thicker (on par with Bestec 250-12Z) and a 35 size main transformer. Of course, the heatsinks are low profile to acommodate tbe 12cm fan.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

I got more than 20 power supplies for free, but most of them belong to the gutless psu thread.

I will post the first here, as I feel that it belongs better to this thread.

EC 350X

We have seen this model again here:
https://www.badcaps.net/forum/showthread.php?t=14822

but this one is still working (the other one had its fan seized and it burnt itself to death).

We have nice full input filter, and good output filtering for every output voltage rail, with coils.

The bridge rectifier consists os 4 discreet diodes rated for 5A, I would prefer to see a proper bridge rectifier, but this one should also work for AC 230V.

2x D13007 NPN transistors limit it to 250W max, same as the 2x470uF primary capacitors and the 33 sized transformer.

The heatsinks are not anything impressive, but should be good for 250W, combined with the 120mm fan. Which btw is thermally controlled.

So, I wouldn't call this a gutless power supply, just a half decent old half bridge 250 watter.

One last thing that I liked, is that the 5v stand by rail is handled with a 5H0165R Power Switch from Fairchild. It is very good to see something like this on a power supply that old instead of the typical 2 transistor 5vsb circuit.

[PeteS in CA]

Re: Power supply build quality pictorial. part 2

Interesting that the multi-output output inductor is mounted on a base. That might improve its stability under vibration.

[momaka]

Delta DPS-300AB-24b [PCB M/N: DPS-300AP-24]

Hmmm… I thought I posted this already, but I guess not.
Today's menu has a Delta DPS-300AB-24b. It's a 300W PSU used in many OEM systems, just like the HiPro HP-P3017F3 and it has essentially the same connector arrangement too (24-pin ATX main, 1x 4-pin 12V CPU power, 4x SATA, 1x floppy.)




On the outside, it looks very much like the 350W DPS-350VB I posted above. The lack of 6-pin PCI-E power connector and minor differences in the label are two ways to tell them apart. Also, as you can see on the label, someone wrote “BAD” on it. Foreshadowing of what the issue might be??

Even on the inside, the DPS-300AB is very similar to the DPS-350VB.


One notable difference, of course, is the lack of a PPFC coil (though the PCB has provision for it, and some 300W models probably do come with PPFC.) The 350VB also has a daughterboard on the secondary for the ICs, whereas the 300AB has everything on the main PCB. Other than that, it's the same solid build quality throughout, with good-sized heatsinks and perfectly adequate input and output filtering.

Next, a picture of the primary side:
https://www.badcaps.net/forum/attach...1&d=1587610857
Speaking of input filtering, look at all of those Y2-class caps there! Delta is NOT cutting corners here. All of the input filtering is present and adequate. The two big input electrolytic caps are Samxon HP, 200V, 560 uF, 22x42 mm, 105°C. The only thing I'm not too excited about is the bridge rectifier lying sideways on the PCB like that, as it probably impedes its cooling capacity. Of course, I can understand Delta probably did this for the models with PPFC coil, as the bridge rectifier would likely get in its way otherwise. Also, one major difference between the DPS-300AB and DPS-350VB is that the 350VB uses double-forward, whereas this 300AB here uses a single-transistor forward topology, as seen by the single MOSFET on the primary heatsink.

Now it's time for the secondary side.
https://www.badcaps.net/forum/attach...1&d=1587610857
https://www.badcaps.net/forum/attach...1&d=1587610857
https://www.badcaps.net/forum/attach...1&d=1587610857
It's a forest of mostly 10 mm and some 8 mm caps for the output. The output design is a classic group-regulated setup with the 3.3V rail generated by a mag-amp circuit, as seen by the two output toroid inductors. The main toroid may not look very big on this picture, but it is not small either – same as the 350VB, actually, with a light-green core (I believe those are the higher frequency capable ones, IIRC.)

And of course, now you can truly see why someone wrote “BAD” on the PSU label: yup, bad capacitors. In particular, it's the first 5VSB output cap and the second cap on the 5V rail. Needless to say, this power supply has been fully recapped and put back into a working PC. I'll save the recap info for another thread/post, though.

Next, a PCB shot, as usual… or half of it anyways (didn't feel like removing the foam to show the full board when I was cleaning the unit and taking pictures.)

It's really just the secondary. Clean Delta soldering, like always.

Last but not least, a fan shot:
https://www.badcaps.net/forum/attach...1&d=1587610865
This time, it's not a Delta fan. Instead, Delta used a Sunon KD1208PTS1 80 mm fan, rated for 2.6W (about 0.217 Amps) @ 12V. Still a pretty high quality fan.

Off to a detailed part list summary…

ICs:
UC3843b (PWM controller), DWA108 (secondary-side supervisor), LM339 (secondary side protections or fan control?), TNY266PN (5VSB PWM+FET combo), and 7912 linear regulator (-12V rail)

Wiring:
* 600V, 18 AWG input wiring
* 300V, 18 AWG output wiring, except ATX connector (mostly 20 AWG)
* Output connectors: 24-pin ATX, 4-pin 12V CPU, 4x SATA, 1x floppy

Primary Side:
* Input Filtering: two X2-class caps, seven Y2-class caps, two CM chokes + one SM choke
* Input protection: inrush current limiter, three MOVS for surge prot., fuse
* 8 or 10 Amp (?) bridge rectifier (can't read info due to being glued)
* 2x Samxon HP, 200V, 560, 22x42 mm, 105°C caps
* 1x Ltec LZG, 35V, 100 uF, 8x13 mm cap as “startup” cap

Secondary Side:
* 5VSB
*** 1x Ltec LZG, 10V, 2200 uF, 10x30 mm before PI coil
*** 1x Taicon PW, 10V, 680 uF, 8x15 mm after PI coil (note: spot for this cap has 5 mm lead spacing and can accommodate a 10 mm diameter cap, if needed)
*** 3A (?) schottky diode for rectification

* 3.3V Rail
*** 2x Ltec LZG, 6.3V, 2200 uF, 10x25 mm with PI coil in between
*** one STPS20c45 schottky rectifier

* 5V Rail
*** 2x Ltec LZG, 10V, 2200 uF, 10x30 mm with PI coil in between
*** two STPS20c45 schottky rectifiers in parallel

* 12V Rail
*** 1x United Chemicon KZH, 16V, 1000 uF, 8x20 mm (note: spot for this cap has 5 mm lead spacing and can accommodate a 10 mm diameter cap, if needed)
*** 1x Ltec LTG, 16V, 1500 uF, 10x20 mm
*** PI coil: NONE. Rail has current shunt only
*** one STPS20s100 schottky rectifier

* -12V Rail
*** 1x Taicon PW, 25V, 220 uF, 8x13 mm before 7912 linear regulator
*** 1x Taicon PW, 25V, 100 uF, 6.3x11 mm after 7912 linear regulator
*** PI coil: NONE
*** 1.5 or 2 Amp diode as rectifier

[TechGeek]

Re: Power supply build quality pictorial. part 2

Sounds about right. I acquired this PSU in mid-2017. I believe the PSU was used in the PC it was in until then. So 9 years of life, more or less - not bad!
But now it should do even more with the new caps.


You're welcome!
And yes, I meant Ltec, not Lelon. Not sure why I put down Lelon as the brand of the caps in my notes. I dug through my bin of bad caps (yes I keep those... I don't know why! ) and found the caps I pulled out of this PSU. They were all Ltec, indeed.
Ah well, same difference whether Ltec or Lelon. They are still both crap brands.

Blow 'em up on mains AC and make a video of it!

[momaka]

Re: Power supply build quality pictorial. part 2

I see that Delta P/S was made in late 2008. So 8-11 years of life would be pretty decent.

Sounds about right. I acquired this PSU in mid-2017. I believe the PSU was used in the PC it was in until then. So 9 years of life, more or less - not bad!
But now it should do even more with the new caps.

Thanks for that great review momaka! I suppose you meant to say Ltec LTG and not Lelon LTG?

You're welcome!
And yes, I meant Ltec, not Lelon. Not sure why I put down Lelon as the brand of the caps in my notes. I dug through my bin of bad caps (yes I keep those... I don't know why! ) and found the caps I pulled out of this PSU. They were all Ltec, indeed.
Ah well, same difference whether Ltec or Lelon. They are still both crap brands.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Thanks for that great review momaka! I suppose you meant to say Ltec LTG and not Lelon LTG?

[PeteS in CA]

Re: Power supply build quality pictorial. part 2

I see that Delta P/S was made in late 2008. So 8-11 years of life would be pretty decent. But 4 or 5 would not, unless the customer kept the computer in a confined space with poor cooling.

[momaka]

Delta DPS-350VB-1 A [PCB M/N: DPS-350VP-1

Ready for another solid 350W ATX PSU? Here we go then: I present you the Delta DPS-350VB-1 A.



Yup, that's another boring-looking gray box. Nice clamshell case design, though. And according to the label, we get two 12V rails that offer a combined maximum of 300 Watts – not bad! It fits the output wiring, which in terms of connectors is the same as the 350W Chicony / HiPro I posted previously, consisting of a 24-pin ATX connector, a 4-pin 12V CPU connector, four SATA connectors, single floppy connector, and a single 6-pin PCI-E power connector…all with the same wire gauges as the HiPro.

OK, let's take that clamshell apart.



Ugh, yeah, I'm sure everyone noticed right away: bad caps galore.
That's actually the reason I even have this PSU. It came from a customer's computer back when I worked as a tech in PC retailer here. Since that place doesn't do any component level repairs and the customer expressed no interest to pickup the old PSU after it was replaced with a new one, it meant this one was destined for the “recycle” bin (i.e. cruncher.)
- Ha! A solid Delta PSU getting trashed?! Not on my watch!
Actually, it was all thanks to one of my colleagues there, as he knew I liked to fix stuff. So he just walked by my bench with it and said, “here, I got a gift for you.”
- Thanks bud! This PSU has been absolutely solid after I recapped it and has been powering one of my computers for a few years now. I'll leave the details of the recap for another thread, though. (and it turns out I never took pictures of the recap, nor jotted down too many details… so need to do that some day when I open the PSU again.)

Continuing along with the build quality discussion seen in the above pictures… aside from the bad caps, this Delta PSU is build very well: good solid heatsinks, a massive passive PFC choke (lol, that sounded funny ), good size input caps (680 uF, IIRC), and overall just a clean and solid design. Look! Even the AC plug is held with screws, so the entire PCB assembly can be separated from the metal case without desoldering anything. - Neat!

Speaking of design, this PSU uses a double-forward for the main PS and a TNY266 PWM-FET combo IC for the 5VSB.

I also like that Delta put all of the X2-class caps and input common-mode chokes on the PCB instead of the AC plug. On that note, this PSU has two X2 caps, two large common-mode chokes, a single-mode choke, a film cap before the bridge rectifier, and seven (or eight?) Y2-class caps. Proper EMI/RFI filtering guaranteed here.

Next, we look at the secondary side.

The main output inductor toroid is not as big as on the 305W and 350W Chicony / HiPro units, but it is not undersized in any way at all either. On the other hand, this PSU uses a mag-amp circuit for the 3.3V rail, as is visible by the second (3.3V rail) output inductor toroid (which is quite big, I should add.) In terms of output filtering caps, there is definitely enough: two large caps for all of the main rails. It's just a shame that many of them are crap brands and have bulged – mostly Ltec with a few CapXon in there. But look! There are also two United Chemicon caps as well. These two are actually for the 12V rail output. I couldn't see what series they were, but most likely either KY or KZE… or possibly KZH. So even though all the other rails may have questionable quality caps, at least Delta didn't mess around when it came to the 12V rail, as they knew that one would be under the most load. – Smart! They put good parts where it matters the most. Too bad they didn't do that to the 5VSB as well. Oh well, I fixed that for them now.
Also notice the two current shunts in the corner – this PSU indeed has two 12V rails. But as the combined power on the label suggests, they both draw power from one single 12V rail. So not quite truly independent 12V rails, but the label is not lying about anything (it's a Delta, after all – why would they lie? )

Next a PCB shot… or half of a PCB shot anyways (I didn't feel like fully ungluing that foam pad.)

Nice and clean, as one would expect from Delta.

And here is the daughterboard:
https://www.badcaps.net/forum/attach...1&d=1582995032
It has a bit of discoloration on its back, which I believe is from the transistor used for the fan speed control.

The PPFC choke:
https://www.badcaps.net/forum/attach...1&d=1582995046
You're not gonna find a chunk of concrete with a looped wire in that one, that's for sure! It's a real PPFC choke with lots of copper inside - nice and heavy!

And last but not least, the fan:
https://www.badcaps.net/forum/attach...1&d=1582995046
A Delta fan in a Delta PSU… who could've guessed that one?
Model DSB0812VH, rated for 0.30 Amps at 12V – that's quite powerful for a PSU fan, actually. No wonder that control transistor on the daughterboard has been getting hot.

Unfortunately, I don't have a detailed part summary for this PSU like I do with my other ones, due to being in a rush at the time to recap it and use it. Thus, I only noted down the output cap values and proceeded from there to fix up the PSU. So the output cap values are the only thing I will include here.

* 5VSB
*** 1x Lelon LTG, 10V, 2200 uF, 10x30 mm before PI coil
*** 1x Lelon LTG, 10V, 1500 uF, 8x18 mm after PI coil (PCB has 5 mm LS for a 10 mm dia. cap)

* 3.3V Rail
*** 2x CapXon GL, 6.3V, 3300 uF, 10x25 mm, with PI coil in between
*** PI coil: 7-turn, 16-AWG, 5 mm core

* 5V Rail
*** 2x Lelon LTG, 10V, 2200 uF, 10x30 mm, with PI coil in between
*** PI coil: 7-turn, 16-AWG, 5 mm core

* 12V Rail
*** 2x United Chemicon KY or KZE, 16V, 1500 uF, 10x20 mm, no PI coil

* -12V Rail
*** 1x Lelon LZG, 25V, 330 uF, 8x18 mm before 7912 linear regulator
*** 1x Lelon LTG, 25V, 100 uF, 6.3x11 mm after 7912 linear regulator

That is all for this one. I'll make a cap diagram of where caps go later… and will see if I post it here on in a separate thread to show the recap. Stay tuned for more PSUs – I still have a ton of units I have not posted on BCN yet.