Morex Int'l ATX-300C, worth a recap?

[Bungz]

Thought I would have a look inside this as I had a issue with a system freezing ( socket A Thunderbird system)

Thought it was a odd design, what do people make of it?

Edit - Nothing visual wrong, obviously may start with a recap but swapping the PSU solved the freezing issue.

I like older socket A systems so these older PSU that are weighted towards the 5V rail always catch my eye.

[PeteS in CA]

Looks like a fairly ordinary half bridge design, possibly with BJTs instead of MOSFETs. The core material for the output inductor is best for switch frequencies under 50KHz. The size of the heatsinks, main output transformer, and output inductor all seem more appropriate for around 200W rather than 300W.

[momaka]

The PSU looks acceptably OK for a socket A system, but isn't anything special really.

It's not built by Morex, though. Morex company's full name is Morex Information Enterprise Co. Ltd. This particular PSU is build by Channel Well (CWT), as is evident by the "CWT" markings on the transformers. It seems that CWT did this quite a bit in the old days - build a PSU and put some other manufacturer's logo/name/model on its label. Actually, I'm not sure if some middleman did that or CWT itself, but I've seen it with a CWT impersonating Macron Power (particularly a few MPT-301 units, like this one) and another brand I can't recall the name of ATM.

In any case, this CWT PSU isn't anything that special, as I mentioned above. Most of those old CWT PSUs used Fuhjyyu caps, so there's a good chance those are starting to go bad, and probably why replacing the PSU fixed the issue in your case. Though it could also have been marginal design, as these old CWT ISO series (and ISO series predecessors) are built a bit on the cheap side.

It's a half-bridge PSU, as PeteS mentioned, and with standard TO-220 BJTs on the primary side (likely 13007 or similar variants.) One thing that caught my eye is the lack of a "proper" BJT driver transformer. It seems the green toroidal inductor between the two transformers is likely what's taking on that duty... which honestly is quite a sketchy way to cut on costs. Apart from that and the Fuhjyyu caps, the rest of the PSU appears to be built OK - or at least for its times. A recap will likely fix the issues you were having with it. I don't see an output toroid for the 3.3V rail, so that is likely linearly regulated, which means the 3.3V rail should be pretty clean from noise. That said, I did run into issues eventually with ADT MPT-301 linked above with low ESR caps on the 3.3V rail, as it caused several systems I tried it in to boot-loop constantly or being unstable. So if you do recap it, don't go too low on the ESR. Something to the equivalent of Rubycon YXJ, Panasonic FC, UCC LXY/LXV/LXZ is probably as far as one should go on low ESR caps. If not, probably even GP 105c caps (from a reputable Japanese brand) should work OK too.

[Bungz]

Thank you for the detailed replies.

I recapped the secondary filter caps with FC series and it has made no difference.

Will possibly grab a multi meter and see if any of the rails drop but I would say it freezes the system faster now than before....

[Th3_uN1Qu3]

It has the driver transformer, it's just not visible under that daughterboard. The green inductor is the reactor for the 3.3V magamp supply. Fairly ordinary half bridge design indeed, but it does seem to have all of the input filtering components installed.

I would add a 5 watt 5V6 zener to the 5VSB rail, if it is the design I have in mind, it has a bad habit of delivering over 21 volts on 5VSB when certain components fail...

If the computer is still freezing I would suspect bad caps on CPU VRM, even if they appear ok.

[momaka]

Hey, good to see you back here again, uN1Qu3! Been a while!

Originally Posted by Th3_uN1Qu3

It has the driver transformer, it's just not visible under that daughterboard. The green inductor is the reactor for the 3.3V magamp supply.

Ah yes, the daughterboard - for some reason I completely ignored seeing it for whatever reason. Thanks for pointing it out to my eyes. Most likely the 5VSB circuit is there and the BJT driver TX should be that small yellow one on the main PCB.

That said, I still don't think the 3.3V on this PSU uses a magamp circuit, as I don't see an output filter toroid (or very long PI inductor, like some cheap Deer/L&C PSUs do sometimes.) You can also see a thick red wire with 1 or 2 turns wound around that green inductor in the middle and connected to the primary. Perhaps that's for some kind of OPP circuit. I've seen an old SPI/Sparkle and an old Topower PSUs with a similar arrangement like that - i.e. either a 4th small TX or an inductor placed closely next to the BJT driver TX.

Originally Posted by Th3_uN1Qu3

I would add a 5 watt 5V6 zener to the 5VSB rail, if it is the design I have in mind, it has a bad habit of delivering over 21 volts on 5VSB when certain components fail...

+1
And also recap all of the small caps related to the 5VSB circuit - both on primary and on the secondary side. That's usually a good way to avoid a 2-transistor self-oscillating 5VSB circuit from going insane and possibly going over-voltage.

Originally Posted by Th3_uN1Qu3

If the computer is still freezing I would suspect bad caps on CPU VRM, even if they appear ok.

True.
But if the system is no longer crashing with a different PSU and the O/P can confirm this by swapping the PSUs a few times, then the fault could still be with this PSU.

Originally Posted by Bungz

Thank you for the detailed replies.

I recapped the secondary filter caps with FC series and it has made no difference.

Did you replace the output caps on the 5VSB on that daughterboard too? If no, probably worth a try to replace those as well.

[Th3_uN1Qu3]

Originally Posted by momaka

Hey, good to see you back here again, uN1Qu3! Been a while!

Thanks. Always nice to receive greetings from old friends. Given that i was supposed to die 2 years ago, I must say that i'm doing quite well. Still fixing consumer crap despite swearing that I will quit the repair business.

Originally Posted by momaka

HeyAh yes, the daughterboard - for some reason I completely ignored seeing it for whatever reason. Thanks for pointing it out to my eyes. Most likely the 5VSB circuit is there and the BJT driver TX should be that small yellow one on the main PCB.

Plenty of ways to skin a cat. That being said, you're probably right.

Originally Posted by momaka

Perhaps that's for some kind of OPP circuit. I've seen an old SPI/Sparkle and an old Topower PSUs with a similar arrangement like that - i.e. either a 4th small TX or an inductor placed closely next to the BJT driver TX.

OPP in half bridge design is usually done with a diode and RC filter off the driver transformer into the DTC pin of the control circuit, but there are at least 5 other ways to do that so it's not out of the question.