Bestec ATX250-12E up and running!

[everell]

My project was a bad Bestec power supply with blown 5VSB. The buldging 220 Mfd capacitor C50 was starting point. That plus a couple of other parts, and the 5VSB line came back to life at 12 volts. I changed several other parts, but could not get that switcher to slow down. In process, it took out the filter caps, the PWM chip, and damaged the supervisory chip. So I decided that this PS had a poorly designed 5VSB circuit. I removed the parts associated with this 2 transistor switcher and replaced it with a Fairchild DM311 chip. Had to replace the PWM and supervisory chip, along with the newly damaged filter caps. But results were spectacular. The 5VSB is now rock solid at 5 volts. The main switcher works fine. 25 ohm resistor at pin 1 of optocoupler PC2 ran very hot. Now at 240 ohms and running cool. The 620 ohm resistor R23 in the switcher was also running hot - it was eliminated. So the 5VSB is running much cooler, and the DM311 chip gives overvoltage and overcurrent protection which wasn't there before. I will now attempt to show photos and my hand drawn schematics.

[KeriJane]

Re: Bestec ATX250-12E up and running!

Hi Everell
Welcome to the Forum!

Wow! I'm impressed. Re-designing a badly engineered PSU for stable and safe operation is a major undertaking.

Far too many would just throw it away and start over when with some thought and resourcefulness even a Bestec can be made useful.


Darn. those kinda schematic things make my head hurt. Kudos to you for being able to write them out, understand what is going on and for being able to express it.

Have Fun,
Keri

PS. I'm working on my own little project!

[Krankshaft]

Re: Bestec ATX250-12E up and running!

Nice job.

The 5 volt standby rocketing to 12 volts no wonder why Worsttec have earned their reputation here as mobo killers.

[HenryJ]

Re: Bestec ATX250-12E up and running!

you did change teapo suppressor caps right?


Added: Top Notch Job

[everell]

Re: Bestec ATX250-12E up and running!

This power supply was purchased CHEAPO as-is and was not working. When I took it apart, there was evidence of burn damage around the 5VSB transformer, a couple of buldging capacitors, and the 5VSB was dead. I tried fixing the 5VSB without success, then re-engineered successfully. The only capacitors replaced (so far) were those in the 5VSB circuit.

My take on the problem for this power supply - and many of the cheap power supplies like it - is a poorly engineered 5VSB circuit which goes overvoltage, or blows up and burns. Removing the cheap troublesome two transistor circuitry is not that hard, and the little add on board is not that hard to make.

If others are interested, I will attempt a write up on this conversion. It should be useful for many different brands and models.

[Wizard]

Re: Bestec ATX250-12E up and running!

Reason for this burn up were badcaps. Bad idea to make electrolytic cap part of regulation circuit with certain types of IC or components.

TEA2261 used two capacitors (2.2uF and oddly used 15uF both electrolytics, and they WERE 105C quality brown plastic sleeves, usually UCC). Both capacitors dry out and outputs gradually rises to the point where PSU shuts down (exceeds 200V on the B+ for flyback and SMPS shuts down when unit plugs in, or components fails or flyback or HV divider eventually fails due to short pulse of overvoltage during turn-on transtition from standby mode using hot side feedback and that two caps to run mode using cold-side feedback when customer turns tv on especially if SMPS didn't shut down on overvoltage (around 150V-170V does not trigger shut down in standby, the RCA CTC169 uses 143V B+ for both standby and run, the standby voltage is set with a pot resistor, the run regulation is on cold side with a pre-set pot via feedback transformer sealed during manufacture). Made lot of money on those.

Cheers, Wizard

[i4004]

Re: Bestec ATX250-12E up and running!

good work there.

>is a poorly engineered 5VSB circuit which goes overvoltage, or blows up and burns.

that's right. it's that 5v(going 12 or more) that blows mobos.
design you show at least has some sort of feedback circuit, i've seen self oscillating "marvels" without any feedback...huh...

if you absolutely-positively need to use such crap, yes, that's the way to make it much safer, but why even start if fortron makes psus with a good design for a rather low price?
offcourse having fun with it can be reason enough...

[kc8adu]

Re: Bestec ATX250-12E up and running!

watch out for this same unit rebadged awa.found in everex pc.

[everell]

Re: Bestec ATX250-12E up and running!

My question is - how can you know if the power supply you are purchasing has a decently engineered 5VSB circuit? Like the old saying, a bird in hand is better than two in the bush! I feel much more confident running this power supply with modified 5VSB than purchasing an unknown and hoping for the best.

I have put several days running time on it since modification - so far it is doing great.

Here is a drawing showing the eleven components which need to be removed. And photos of the added pc board with the new circuit - only six components.

[PCBONEZ]

Re: Bestec ATX250-12E up and running!

Very cool project everell!

[i4004]

Re: Bestec ATX250-12E up and running!

>My question is - how can you know if the power supply you are purchasing has a decently engineered 5VSB circuit?

actually easy...offcourse you need to see its inside, but that's not too hard to find...this forum has many psu's pix, many on the web too...

once you see it, it's easy to see if it's crappy few-transistors-design, or one that you used(ie one with chip, that all mfrs we label "reliable" are using)

>I feel much more confident running this power supply with modified 5VSB than purchasing an unknown and hoping for the best.

well yeah, if you purchase something you don't know anything about, and it's as cheap as these, then you have good chances to have crappy 5vstby circuit...
otoh, these are the psus we don't recommend and that we usually dump, because they're not reliable...
your mod improves it, but i seriously doubt you'll make an industry of modding cheap psus, reason being that decent psus are not too expensive and they have such 5vstb circuits...
fortron, delta, corsair, to name few....

offcourse, your details on the mod are still welcomed...
i personally had the idea to go one better than this design, and to put a simple linear supply for 5vstby(mount it on the outside of the psu), but there's no real need to be such a purist...and 1 chip 5vstby designs are pretty reliable too...

[linuxguru]

Re: Bestec ATX250-12E up and running!

Actually, the two transistor +5Vsb circuits (either BJT or MOS switch) are reliable - but one electrolytic (usually 22uF/50v or 47uF/50v) used on the primary side of the flyback is critical. It eventually dries out, causing an auxiliary secondary-side over-voltage on both the controller power rail, as well as +5Vsb. If it doesn't take out the mobo first, it will take out the PWM controller circuitry and then the mobo.

Bestec is the one most cited, but the flaw exists on numerous designs, including better-rated brands like HiPro, AcBel, Fortron/Sparkle, Enhance, etc.

I replace these caps with 2x the capacitance in a long-life series, like NCC LXA/LXF or similar non-aqueous only. The extra capacitance only affects the startup of +5Vsb, and has not been a problem with the discrete switchers so far - but if you're using a TOPSwitch or similar integrated switcher, it's better to stay exactly with the values recommended on the reference designs.

[Wizard]

Re: Bestec ATX250-12E up and running!

Better yet, pick a IC that doesn't use electrolytic for standby. This have been done.

Cheers, Wizard

[i4004]

Re: Bestec ATX250-12E up and running!

>Actually, the two transistor +5Vsb circuits (either BJT or MOS switch) are reliable - but one electrolytic (usually 22uF/50v or 47uF/50v) used on the primary side of the flyback is critical. It eventually dries out, causing an auxiliary secondary-side over-voltage on both the controller power rail, as well as +5Vsb. If it doesn't take out the mobo first, it will take out the PWM controller circuitry and then the mobo.

well yeah, it would be better circuit if they didn't use lytics there at all...
it would still be less reliable than ic design just because it has more components...
to that extent it's better to mod the psu like everell did, than to rely on better caps...if that crap is getting hot even better caps will get hurt...ic design also has lytic, but i doubt it would do any damage as ic usually has many types of protection...

what i'm wondering is what's the feedback circuit doing(as this crappy original design had it, we can see it from pdf in the first post) when voltage goes overboard...in this case it did nothing if there was 12v on output...
so it indeed is a crappy design, given the lytic and the fact that feedback is not protecting anything...may as well be that feedback is not working because of lytiuc(and it's probable), but again that's bad design...
no doubts there.

[everell]

Re: Bestec ATX250-12E up and running!

The feedback circuit in the original design was a big question mark for me. Take a look at the series resistor between the +5VSB and the optocoupler PC2 pin 1. Actual measured voltage across this 25 ohm resistor taken on a different up and running Bestec, same model, was 0.46 volts. This means the resistor and the LED in the optocoupler were pulling 20 ma. It is no wonder the resistor was running hot, and when voltage rose to 12 volts began smoking. Makes me wonder if the optoisolator was also damaged. So I replaced it!

I was looking at an application note for these optocouplers and they showed a series resistor of 200 to 330 ohms. The original circuit would not work at all with these higher values. In my circuit modification, I experimented and changed this resistor value from 220 ohms to 270 ohms - all of which work well.

Perhaps one of our engineering gurus can explain this one!

[i4004]

Re: Bestec ATX250-12E up and running!

nobody can explain chinese cheap designs...hehe...
not with reason anyway...with price cuts..perhaps...

they probably take older "western" designs and then mod them with cheaper and cheaper compoinents untill it barely works...then they sell it...
we were discussing this here
https://www.badcaps.net/forum/showth...ght=bad+design
there were some mistakes and misunderstandings there...

there you can see pix of 5vstb design with no feedback(no optocouplers)

something else: i dunno if 12v on 5vstby can destroy the complete system...ie 5vstb problems alone don't explain why hdds and optical drives got toasted...
or do they? perhaps that 5vstby(now 12v) finds it's way thru mobo onto data cables and hdd and optical drives....huh...

because main pwm ic usually takes high span of supply voltage, so 5 or 12v wouldn't hurt it's operation too much....
may as well be that that faulty 5vstb is the sole reason for destroying whole systems...not its interaction with main pwm ic(ie the fact that it supplies main ic)...
can't say for sure, i wasn't measuring all voltages when it happened...

[everell]

Re: Bestec ATX250-12E up and running!

I have really enjoyed the discussion on this one. For anyone else who wants to try this modification I will post the remaining drawings I made. They are specific to the Bestec ATX250-12E but probably a similar pattern for other "cheap" power supplies. This one was cheap enough that I didn't have to worry about failing and losing any money. Best of all, I learned a lot - both from the project and from the discussions on this forum. Thanks guys.

Tom

[everell]

Re: Bestec ATX250-12E up and running!

I would like to respond to i4004 comment - "because main pwm is usually takes high span of supply voltage, so 5 or 12v wouldn't hurt its operation to much"

Once the original 5VSB circuit was repaired and functional - but running at about 12 volts - did it damage the pwm circuit. While working on this PSU I did not have it connected to a mobo. Once I put the modified circuit in place and had 5VSB measuring 5 volts, I was ready to look at the remaining problems - mainly - nothing else was working. I began by looking at the pwm chip, a UC3842. It had 17 volts of power on pin 7, but the VREF which should have been 5V was zero. I replaced this chip and the VREF returned to its proper 5 volts. The PSU tried to start, but shut down. So I moved to the supervisory chip which is a TPS3510. I found that the FPO voltage on pin 3 was running a little more than 3 volts. I replaced this chip and the power supply came up running fine. The FPO voltage on pin 3 was now 0.7 volt. From this it can bee seen that problems in the 5VSB circuit can damage the PWM chip AND the supervisory chip which controls the main switcher.

I went ahead and replaced all of the caps on the output side - they were all made by JPCE-TUR. I don't remember reading anything about this brand. Any comments?

[PCBONEZ]

Re: Bestec ATX250-12E up and running!

JPCE-TUR is known junk but not 'replace on site' like Fuhjyyu.
JPCE is the Brand Logo. - TUR is the series.

They are made by CEC International Holdings out of Hong Kong and China.
CEC's website is useless but here it is:
http://www.ceccoils.com/CECWEB/Produ...oID=0000000008

No.
I dunno how they get from "CEC International Holdings" to "JPCE".
.

[everell]

Re: Bestec ATX250-12E up and running!

I just finished putting together and testing my new Blue ESR meter (by Bob Parker), so decided to put it to use on this project. These are the measurements of the seven JPCE-TUR capacitors which I removed from this PSU

2200 uF 16 V .02
1000 uF 16 V .04
1000 uF 16 V .04
2200 uF 10 V .03
1000 uF 10 V .06
1000 uF 10 V .04
1000 uF 10 V .04

I interpret these readings as reasonably good. Therefore the heart of the problem with this PSU was the bad 5VSB circuit, including the one bad bulging capacitor which was trashed many days ago.