junky Jetway P4MDPT board

[zunasthegreat]

Re: junky Jetway P4MDPT board

Read an old article from goodpsusearch talking about a DR-8600BTX that I just have around with different sticker (FORCE). Internals look good nice heatsink/filters/bridge diode even PPFC it seems. All filter caps (Y? smthing brand) are either busted or have super high capacity. Did a quick recap with Lelon RGA's and some KY 2200uf 10V / Nichicon HD/HV and some nichicon 1000uf PW for 5VSB. It's among the best gutless wonders Deer/LC/Premier I've seen so far and it even uses 18 AWG Cables

[goodpsusearch]

Re: junky Jetway P4MDPT board

Please NEVER connect Powertech/Hantol/Approx/Aqprox power supplies to any valuable equipment.

If you search this forum you will find some of those that I have posted pics of he internals. They even use fake Passive PFC. I have seen myself many of those burn up and take with them motherboards and hdds. They are cheap, cost 11-20 euros here and they claim 450-600W but the ripple they deliver is out of spec even at 50W. And they explode at 150-200W

[momaka]

Re: junky Jetway P4MDPT board

Do you think Panasonic FJ 3300uf 6.3V will make things better for the 3.3V line ?

Don't know.
FJ is motherboard grade low-ESR, so quite a bit lower than KY. Most likely the PSU will work fine with it, given that this is an FSP design, so I don't expect it to oscillate or do anything crazy. But the ripple may still end up the same due to increased noise from output filter ringing. The only way to know for sure is to load the PSU and measure the output with an o-scope.

I also think that it uses 6 SBR's (2 for each rail maybe?)

I wouldn't be surprised. Actually, that's pretty normal for better-built PSUs that tout higher efficiency (especially anything 80+ that isn't synchronous rectification.)

Those SBR's are TO-247 and I think It was U30D60 something, not 100% sure.

MBR-, SBL-, SBR-, STPR-.... there are a number of different parts out there... but yeah, first number is typically the maximum average current, and the second is the breakdown voltage... and 30 Amp, 60V would be pretty standard there. On the half-bridge PSUs, though, 60V rectifiers on the 12V rail may not be enough sometimes, since with half-bridge, the reverse voltage spikes will be 2x higher than with forward topologies. So for H-bridge PSUs, aim for 100V parts to be on the safe side, unless the PSU came with 60V parts to begin with.

13007 transistors : Check
One Cap per output wonder : Check
Full nudity : Check
1.5mm metal for heatsink : Check
No 3.3V toroid : Check
No PI Coils : Check
Awe-sum quality cap only : Check
AWG20 quality cabling : Check
Classy Coolbox Fan : Check
... we have another winner

All perfect, except for one missing item: a POWMAX label.

[zunasthegreat]

Re: junky Jetway P4MDPT board

That's good anyways, since KY series tends to be a good replacement for Teapo SC. The lower capacity will just mean you won't be able to pull as much power as the label says... but that's rarely a concern if you just want a basic system build anyways.

Do you think Panasonic FJ 3300uf 6.3V will make things better for the 3.3V line ?
Had nothing to do due to weather and was about to even check some GP Nichicon VZ there for lols. I also think that it uses 6 SBR's (2 for each rail maybe?) and 12V Rail had 2X 30A SBR's, Its so crowded area and pretty tough to read the type. Those SBR's are TO-247 and I think It was U30D60 something, not 100% sure.

[zunasthegreat]

Re: junky Jetway P4MDPT board

13007 transistors : Check
One Cap per output wonder : Check
Full nudity : Check
1.5mm metal for heatsink : Check
No 3.3V toroid : Check
No PI Coils : Check
Awe-sum quality cap only : Check
AWG20 quality cabling : Check
Classy Coolbox Fan : Check
... we have another winner

Its the perfect psu for LoLs and having fun with bad-caps or maybe the perfect April's fool gift.

[momaka]

Re: junky Jetway P4MDPT board

I think 3.3V line uses MBR2045CT, hard to say for sure cause its well hidden. To improve this psu's nudity I've installed some X Y Caps. All filter caps are 1500uf LOWESR from YHC? with space for one cap for each as usual. I'll put Lelon RGA 2200/3300uf there and will let it be. The 3.3V line on DMM went up to ~3.15V on the Jetway P4 board and shows around ~3.4V on a C2Q P31 board.

Ah, I think I know exactly what's going on now - you have one of those "super-group regulation" PSUs, like this POS Cyberlink PSU here:
https://www.badcaps.net/forum/showthread.php?t=39222

^ In essence, that PSU doesn't have a mag-amp or linear regulator circuit for the 3.3V rail. Instead, it just derives the 3.3V rail from its own separate transformer tap just like the 5V and 12V rails... and so the 3.3V will vary quite wild-ly with any load. Cheap PSUs like this are some of the worst garbage you can get. I wouldn't put one of these in *ANY* PC.

The picture of yours here confirms you have the same design in that PSU. IMO, you shouldn't waste any time on PSUs like that, unless you really have nothing else around and need to put a system together... but even then, I'd say that's a questionable endeavor. (And if you've read my posts long enough on BCN, you will find out that I am the type of person to try to reuse/rebuild just about any PSU... but some of the cheapo PSUs like this, even I don't bother with, lol. )

Another nasty discovery upon cleaning a chieftec aps-500 (10y old PSU). One Teapo SC is bloated and others are starting to raise in capacity (pics bellow).

That's a Teapo-cal failure mode for the SC series.
10 years is about the right age too, if not a pretty good run, I must say. Would like to see Cheng/Chang/Chong/X garbage make it that long, but I'm pretty sure they won't.

The very crammed area around 3.3V forced me to swap that cap for a 2200uf KY 10V as my other GP caps are Nichicon VZ 3300uf 6.3 and lelon GP.

That's good anyways, since KY series tends to be a good replacement for Teapo SC. The lower capacity will just mean you won't be able to pull as much power as the label says... but that's rarely a concern if you just want a basic system build anyways.

PSU is Group Regulated with proper equipment, nothing fancy.

Nothing fancy, sure... but at least 100000 times better than that crappy Powertech supergroup-regulated POS.

Replaced all SC caps with the following ->

3.3V -> 2X KY 2200uf 10V (capacity downgrade from 6.3V 3300uf)
5V -> 2X KY 2200uf 10V
12V -> 2X Nichicon HD 2200uf 16V
5VSB -> 2X Nichicon PW 1000uf 10V

Looks good to me.

in any case getting 10mm diameter caps is really hard ...

Wait until you run into PSU with custom-order 8 mm caps... or even 6.3 mm skinny tall caps. I haven't, but I've heard terrible stories.

For 10 mm, and depending on what e-tailers you have/use/deliver to you, look for these"
United Chemicon KYB series for 16V, 2200 uF in 10 mm dia.
Rubycon ZLQ series for 6.3V, 3900 uF in 10 mm dia. (these can be used on 3.3V or 5V rails.)

^ At least those have been my go-to caps in 10 mm dia and higher capacity (particularly when needing to replace 3300 uF ones.) I still haven't found a good source of 3300 uF 16V caps in 10 mm. Badcaps.net used to have Samxon RS, but not sure if available anymore. I think only member behemot on this forum has some skinny custom-ordered caps that can fit newer PSUs.

[RJARRRPCGP]

Re: junky Jetway P4MDPT board

That cap looks like an e-tailer-scammer special.

[zunasthegreat]

Re: junky Jetway P4MDPT board

The other two caps (5V+3.3V) show super high capacity and it seems their about to meet their doom.

Replaced all SC caps with the following ->

3.3V -> 2X KY 2200uf 10V (capacity downgrade from 6.3V 3300uf)
5V -> 2X KY 2200uf 10V
12V -> 2X Nichicon HD 2200uf 16V
5VSB -> 2X Nichicon PW 1000uf 10V

in any case getting 10mm diameter caps is really hard ...

[RJARRRPCGP]

Re: junky Jetway P4MDPT board

what the...?! sanyo cap with a plus vent?! totally screams fake to me...

Yikes! QFT!

[zunasthegreat]

Re: junky Jetway P4MDPT board

Another nasty discovery upon cleaning a chieftec aps-500 (10y old PSU). One Teapo SC is bloated and others are starting to raise in capacity (pics bellow). The bloated one (3300uf) is filtering the 3.3V line, 5V and 12V lines use 2200uf caps and psu has 2 for each line. The very crammed area around 3.3V forced me to swap that cap for a 2200uf KY 10V as my other GP caps are Nichicon VZ 3300uf 6.3 and lelon GP. I choose some nichicon PW 10v 1000uf for 5VSB as the space is OK there. 5V line got a KY 2200uf 10V and 12V a Nichicon HD 16V 2200uf (12mm was hard to fit).

Original Cap spec
3.3V -> 2X Teapo SC 3300uf 6.3V
5V -> 2X Teapo SC 2200uf 10V
12V -> 2X Teapo SC 2200uf 16V
5VSB -> 2X Teapo SC 1000 10V

TEMP SOLUTION
3.3V -> 1X Teapo / 1X KY 2200uf 10V
5V -> 1X Teapo / 1X KY 2200uf 10V
12 -> 1X Teapo / 1X Nichicon HD 2200uf 16V
5VSB -> 2X Nichicon PW 1000uf 10V

PSU is Group Regulated with proper equipment, nothing fancy.

[zunasthegreat]

Re: junky Jetway P4MDPT board

Powertech got the Polymer care for the 5VSB it needed. Plus I've installed :
5V -> 2200uf KY 10V
12V -> Lelon RGA 3300uf 16V
3.3 -> Lelon RGA 2200uf 6.3V
5VSB -> 470uf 16V KY + OCRZ 1000uf 16V (polymer)

Just left chengx masterpieces for -12V and an 50V 47uf near 5VSB cause they had decent Esr. The 47uf cap has like ~0.2Ω ESR

PSU has very weak 3.3V (14A) / 5V (10A) btw, won't be a big deal for a Xeon 771 modded on a P31 with a HD 5450. System pulls most juice from 12V anyway

[zunasthegreat]

Re: junky Jetway P4MDPT board

I think 3.3V line uses MBR2045CT, hard to say for sure cause its well hidden. To improve this psu's nudity I've installed some X Y Caps. All filter caps are 1500uf LOWESR from YHC? with space for one cap for each as usual. I'll put Lelon RGA 2200/3300uf there and will let it be. The 3.3V line on DMM went up to ~3.15V on the Jetway P4 board and shows around ~3.4V on a C2Q P31 board. Btw it uses 13007T transistors and I doubt that it will do over 200W

[momaka]

Re: junky Jetway P4MDPT board

Another half bridge beast with special scissor marks on the pcb. The R37 680Kohm resistor is cut and has that scissor icon, any clue whats the deal with it ? PSU is a powertech brand without even the 3.3V coil plus next to that resistor there is marking with "OPP". I guess its an overpower protection ...

Yup, if that resistor was cut out from a section labeled OPP, probably they did that to either raise the maximum allowable power or lower it (depending on circuit design), probably according to the parts they installed in the PSU.

In any case, I'm always a bit skeptical if the OPP would work properly on these cheapo PSUs. On some, it does. But just don't count on it.

Regarding the missing 2nd 3.3V coil... PSU probably uses linear regulation for the 3.3V rail then. See what part on the heatsink connects to the 3.3V rail. If it's a MOSFET, 3.3V rail is surely a linear job (and likely taking power from the 5V rail.) This isn't a bad design, per se, but just a little more inefficient. On the plus side, the 3.3V rail does end up with very little ripple, even with crap undersized caps.

Upon using it on a testbed with a Soltek P4 board the 3.3V reading was ... 3.05V

:\
Curious what they done goofed up to do that. I imagine this reading was with a multimeter and not taken from mobo monitor? If with mobo monitor, I'd disregard it.

Usual business with one F16C20C Mospec SBR for 12V.

Well, you got somewhat lucky. A lot of cheaper PSUs come with a F12C20C - just barely enough to run a low-end CPU with an entry-level GPU without blowing up... like something along the lines of an i3 or low-end i5 and a GTX1030. It's all the same trash though, - these cheap PSU manufacturers haven't really changed their ways too much over the years.

On the other hand, I still find it amazing that 40-year-old designs (namely, the half-bridge) is still able to run modern PCs.

what the...?! sanyo cap with a plus vent?! totally screams fake to me...

LOL, good catch!

I didn't even see the Sanyo text. The moment my peripheral vision caught the crappy stamp on the vent, I didn't even bother to look at what the cap was and just assumed it to be a crap brand... or I guess in this case, a counterfeit!

[zunasthegreat]

Re: junky Jetway P4MDPT board

what the...?! sanyo cap with a plus vent?! totally screams fake to me...

Haha yeap ChengX - ChongX or Sanyo sometimes. Those have the same maze diagram between the cap legs. I see those very often on many boards.

[ChaosLegionnaire]

Re: junky Jetway P4MDPT board

what the...?! sanyo cap with a plus vent?! totally screams fake to me...

[zunasthegreat]

Re: junky Jetway P4MDPT board

Another half bridge beast with special scissor marks on the pcb. The R37 680Kohm resistor is cut and has that scissor icon, any clue whats the deal with it ? PSU is a powertech brand without even the 3.3V coil plus next to that resistor there is marking with "OPP". I guess its an overpower protection ...


Upon using it on a testbed with a Soltek P4 board the 3.3V reading was ... 3.05V
Usual business with one F16C20C Mospec SBR for 12V. The laber markings are quite funny : 3.3V - 14A / 5V - 10A and 12V 29A (14+15)

[zunasthegreat]

Re: junky Jetway P4MDPT board

I think I might have one and can salvage some more MBR20200CT's for the 12V rail. Those are salvagable from dead PSU's, mighty Deer/codegen/allied/l+c PSU's offer a ton of spare parts
For the scope I think you have to put some caps to isolate ground, it's on some intel slide.

[momaka]

Re: junky Jetway P4MDPT board

One funny thing that still bugs me is the Nitrox PSU that switches off when I play with DVD-ROM drives. I had two drives connected and upon pressing eject on both it switched off. I have to power cycle the PSU to boot again. Will have to wait for the scope to check for ripple.

I guess the low-ESR Nichicon HV caps may just have too low of an ESR for this PSU. Run it with the VY's only, then. When it comes to these older half-bridge designs, you can always count on capacity as being "king" in terms of ripple filtering.

Would definitely be interested to see what your scope is showing. Make sure you do bypass your scope properly, of course (IIRC, there were a few articles for the one or two caps needed to do that so your scope picks up the noise from the PSU correctly.)

This PSU is like the one you fix'ed on another post with I think 2X F12C202 SBR's for 12V. Can it really provide 18-20A on 12V rail btw ?

You mean the KDMPower MIPC MI-X8775CD?
Yeah, when I was looking at the pictures of your power supply, it indeed reminded me of my KDMPower PSU as it looks very similar in terms of component layout in some areas on the PCB.

That one actually had 2x 20A 100V rectifiers... though I removed one.

It still did 15 Amps on the 12V rail, and that's with a low-line voltage input. So I think 18 Amps may actually be feasible... though I imagine the ripple on the 12V is going to be quite high (given that your PSU has a spot only for one cap there), if even in spec at all. I certainly don't suggest peaking higher than 20 Amps for a few seconds, if even that much.

[zunasthegreat]

Re: junky Jetway P4MDPT board

One funny thing that still bugs me is the Nitrox PSU that switches off when I play with DVD-ROM drives. I had two drives connected and upon pressing eject on both it switched off. I have to power cycle the PSU to boot again. Will have to wait for the scope to check for ripple. This PSU is like the one you fix'ed on another post with I think 2X F12C202 SBR's for 12V. Can it really provide 18-20A on 12V rail btw ?

[momaka]

Re: junky Jetway P4MDPT board

Here is the FSP 300W I talked about earlier, all caps are from Teapo and two of them are inside heat-shrink tube. It also seems that each 12V output V1/V2 has 2 16V 1000V caps. I don't think it needs any improvements atm.

Yup, looks pretty good... though at 5-8 years old, you might want to check back on it again - that's roughly how long it takes for some of the Teapo caps to become problematic. Not all of them will, though. I also see this PSU has mostly black Teapo SY series, which IMO tend to be a little more stable than their green/gold SC and SM series. The worst ones are the blue/gold -colored SZ series, which I see one (6.3V 2200 uF) on the secondary side. Most likely, that will be the first one to fail... and then probably the 5VSB output filter after that (or around the same time.)

Primary input cap is also Teapo... which can be a bit risky with APFC PSUs (which this one is) after a few years. At least it's rated for 420V... though capacity is a bit low at 180 uF for a "300W" PSU. I'm sure the PSU will do its rated output on the label, but that primary cap is just borderline cutting it in with the capacitance. 270 uF would be a little better to see here.

But other than the cap choices, that's a pretty decent PSU that actually will do the rated power numbers on its label and not blow up.

Another non-visible GSC c[r]ap from a P4 board. Dimm detection is dodgy so this might be an issue from the aging caps.

Yup, I got a few failed Evercon caps like that too - not bulging and high ESR with high vloss. GSC/Evercon/Sacon really are some of the worst caps, right down with Rubycon-copycats, like Rulycon and Rudoycon.