Cheap PSU: Piece of junk or semi-decent?

[c_hegge]

Re: Cheap PSU: Piece of junk or semi-decent?

Broken fingers?! What kind of fans are you using?

4500RPM 120x38 Deltas? I have two of those on my load tester, and I'm sure one of those would quite easily break your finger. I've run the load tester near openings and windows before, and one of the fans ingested a small stick with some leaves that fell off a tree, and turned it into sawdust.

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

.... congrats on getting the PSU working .

Thanks for your very informative and enlightening advice Momaka .

If there ever was an award for spending the most time and effort to improve a cheap PSU, I think you would easily win .

I concur :

That's crazy! This is the first time I'm seeing this. Normally, those resistors have the same resistance value. No idea why Deer/Allied did that.

Either they knew something that we don't ... or they simply screwed up and put a resistor with the wrong Ω .



BTW, forgot to mention I washed the PCB of this PSU in the sink with abundant water while it was devoid of caps. Then 'centrifuged by hand', hair-driered, and left it 24 hours placed just over an electric heater to make sure it dried completely. And no afterwards.


Extra touches :

[momaka]

Re: Cheap PSU: Piece of junk or semi-decent?

Thanks for your very informative and enlightening advice Momaka .

You're welcome.

Either they knew something that we don't ... or they simply screwed up and put a resistor with the wrong Ω .

I think it's the latter, but since the PSU has worked for so many years up until it got the bad caps, then it's probably nothing to worry about.

BTW, forgot to mention I washed the PCB of this PSU in the sink with abundant water

I've done that with PC motherboards a few times. As long as they are left to dry properly, it's not a problem.

That is very cool . Topcat should really put this on the front page of BCN .
I'm saving the picture to my hard drive (if you don't mind, of course... I have a small picture collection of kludges and fixes from thereifixedit.com and other websites, and I think this definitely deserves to be saved ).

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

... I'm saving the picture to my hard drive (if you don't mind, of course...

Not at all, help yourself .



I'm trying to understand how MOV varistors work, and how to select their specs according to application ...

... There are two empty parking slots near them marked 'Z1' and 'Z2', could these be for MOV varistors? If so, which MOV specs (volts, joules, ns) would be adequate for this PSU?

Z1 is in parallel with one big cap, Z2 is in parallel with the other bigcap.

... If they are on the potential between the two caps, than it would be something like 115 Vrms/150 V DC ...

That 150V DC is maximum continuous operating voltage, right? So something like this would do?:

So I can learn, would you mind explaining why a MOV in parallel with a 200V rated bigcap, in a 230V PSU, must be 150V DC?


What if we add a single MOV to the PSU as per this scheme:

RV1 = MOV

Perhaps a ~271K MOV would be adequate for a 230V PSU in that position?

[Th3_uN1Qu3]

Re: Cheap PSU: Piece of junk or semi-decent?

So I can learn, would you mind explaining why a MOV in parallel with a 200V rated bigcap, in a 230V PSU, must be 150V DC?

It must NOT be 150V DC, else it'll blow first time you plug it in. It should be rated high enough to allow for the worst case - a 10% tolerance of mains voltage at zero load, which would be 253v. Multiplied by 1.414 (peak voltage of sine wave) this gives 178v per cap. So the MOV should be 180 volts or 200 volts.

If you use a single MOV in a half-bridge PSU, you will not get full protection. The two big caps in a half bridge work as a voltage divider, so if you only clamp the top to the bottom you could still get voltage imbalance. But given that MOVs are intended to absorb short-term transients which are over much too fast to cause any imbalance between the two big caps, you should be okay with a single one. Use a 360 or 400 volt MOV.

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

You'll often see a 350~375V AC MOV on the input to a power supply.

It seems a lot higher than you'd expect but it's because a MOV isn't there for brief transients. If it were to attempt to absorb them, it would be quickly damaged... there are hundreds of them per day in the UK... I wouldn't know about other countries with less reliable electrical systems. A MOV is for the "really bad" transients (such as phase faults, nearby lightning, etc.) that might start a fire or severely damage the rest of the power supply.

So, it should be rated at much more than the continuous operating voltage of the PSU. I would guess that the big caps can withstand a higher operating voltage for some short amount of time, and the two primary switchers are usually 13007, or two 400V switches in series. In the majority of the PSUs I've seen with 115V/230V switches flicked (which is similar to a really bad surge), the only thing that goes bang is the fuse.

[Behemot]

Re: Cheap PSU: Piece of junk or semi-decent?

I'm trying to understand how MOV varistors work, and how to select their specs according to application ...

That 150V DC is maximum continuous operating voltage, right? So something like this would do?:

…

Read my later reply, I suggested something VE09M00131K cause 150V continuous is really slightly too low.

tom66: I disagree, big caps could withstand like maybe 20-30% higher-than-nominal voltage spikes. But it is still decreases their life. I am trying to select as low varistors as possible above operating voltage cause it means longer input caps life. It's both cheaper and simplier to replace varistor than input cap.

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

Well, I've only repaired a few 115/230V jobs, and the primary caps haven't blown on them. Always the fuse. Sienoki has done a lot more... only had one which had more substantial damage.

The caps basically behave like short circuits at those voltages due to dielectric breakdown. The heat from this would quickly damage it but it is so quick it doesn't usually do any harm.

Perhaps it reduces the life of the capacitors but this just moves the problem to the MOVs. So you end up changing more of them.

[Behemot]

Re: Cheap PSU: Piece of junk or semi-decent?

Yeah but degrading electrolyte is what kills the caps in the first place…

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

The cap just breaks down internally -- it isn't a chemical reaction (so doesn't affect the electrolyte), but simple arcing across the dielectric. The dielectric is really thin, and decides the breakdown voltage and working voltage of the capacitor.

If it gets too bad, it creates local hot-spots, which damage the capacitor by decreasing capacitance. But this usually takes a few hundred ms.

If it gets REALLY bad, it creates a short circuit, the cap vents violently from the heat (it's got a few ohms of resistance so the fuse doesn't pop immediately), and then the fuse goes pop.

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

Thanks! So if I get it right the formula is:

Max continuous operating voltage of MOV = 230 * 1.1 * 1.414 = ~360V

For two MOVs in parallel with bigcaps (half-bridge topology), halve that value (~180V each MOV).


Dang, just scavenged six MOVs from an old broken UPS and none of them suits :

· 561K -> 560V
· 471K -> 470V
· 2x 320K -> 320V
· 271K -> 270V
· 101K -> 100V

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

... If you use a single MOV in a half-bridge PSU, you will not get full protection. The two big caps in a half bridge work as a voltage divider, so if you only clamp the top to the bottom you could still get voltage imbalance ...

And what if instead of two 180V MOVs, one in parallel to each bigcap ...

... I lay just one 360V MOV, in parallel to both bigcaps?:

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

Yes essentially.

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

... R1 is 330K (orange/orange/yellow/gold), R2 is 470K (yellow/violet/yellow/gold) ... Either they knew something that we don't ... or they simply screwed up and put a resistor with the wrong Ω .

Seems they knew what they were doing after all. While on educational patrol I found this thread, and grew uneasy on the prospect of one of my brand new big Pannys eating excess voltage due to the weird assimetry in the bleed resistors .

So I checked on the live with DMM, and found with relief that both big caps read 158V each .

The Thing has been working like a charm for two or three dozen hours flight time now.

[momaka]

Re: Cheap PSU: Piece of junk or semi-decent?

Seems they knew what they were doing after all. While on educational patrol I found this thread, and grew uneasy on the prospect of one of my brand new big Pannys eating excess voltage due to the weird assimetry in the bleed resistors .

Funny, I remember reading that thread but I don't remember those 2 particular posts in it. I didn't expect those bleeder resistors to make that much of a difference. Seems like they do, though. Time to check some of my PSUs that have bulgy primary caps.

So I checked on the live with DMM, and found with relief that both big caps read 158V each .

That's good. I'm still puzzled by the mismatched resistors though.

[c_hegge]

Re: Cheap PSU: Piece of junk or semi-decent?

Maybe Deer ran out of 330K resistors half way through assembling it?

[Behemot]

Re: Cheap PSU: Piece of junk or semi-decent?

Is the bigger one accross first, or second cap? If second, isnt there higher voltage (from first AND second cap), so they maybe placed it there to limit current to +- same value as from first cap? Just a wild guess

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

Is the bigger one accross first, or second cap? If second, isnt there higher voltage (from first AND second cap), so they maybe placed it there to limit current to +- same value as from first cap? Just a wild guess

As in the pic, R2 is in parallel with left cap, R1 with right cap.

[Behemot]

Re: Cheap PSU: Piece of junk or semi-decent?

I am hungry and cant focus Which has what resistance?

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

I'm thinking about a moderate recap of the 'TOOQ' PSU from the beginning of this thread, just the output filtering and +5VSB critical caps:

1,2 (+5V) · 1000uF · 10V · 105C · ASIA
3 (+5VSB) · 470· 16V· 105C · ASIA
4 (-12V) · 470· 16V· 105C · ASIA
5,6 (+12V) · 1000uF · 16V· 105C · ChengX
7,8 (+3.3V) · 1000uF · 10V · 105C · ASIA

12 (+5VSB) · 1000uF · 16V· 105C · ChengX
13 (+5VSB) · 47uF · 50V · ChengX

This crappy PSU doesn't have any PI coil in the output , so I wonder if it could be a good idea replacing all the output caps with higher capacitance ones, say 1500~2200uF replacing the 1000s, and perhaps ~1000uF replacing the 470s. Or something like that.

I have four spare Panasonic FJ 1500uF 6.3V 105C caps scrounged from an old mobo, apparently in good health. What do you think about using them for 1-2 (+5V) and 7-8 (+3.3V)?