Raidmax RX-600 ringing noise and 0ms PG

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

Nice job!
The photo isn't bad, I can see everything okay.

But yeah, I was pretty sure that diode didn't look so good. Always a good idea to investigate anything that looks physically damaged.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Ok, PSU is working now. I replaced the burnt resistor with a thru-hole one, since I didn't have a proper SMD one on hand, and the problem persisted: PG was still 0.

I checked to see what voltage I get on that resistor and it was very low and unstable, bouncing between 1v and 2.5v. This led me to believe there was something wrong with the diode which comes before it. I measured it in circuit and the drop was very high - like 1.3v or more. I measured a similar one further up the board, also in circuit, for reference, and the drop was more what you'd expect, at around 0.6v, so my first diode seemed faulty at this point.

I removed it and I got the same high reading out of circuit too. Again, not having an SMD one readily available, I botched an 1n4148 in series with the resistor and tried it out. This time it worked - PG was now at 290ms Since this is for my own personal use, I dug no further to find proper SMD replacements and simply threw that black piece of heat-shrink over the solder joint there so it wouldn't come into contact with anything.

Picture turned out really bad due to 50Hz banding from the light of my illuminated magnifier I shot through. I can't seem to mitigate the effect on my SG S10+...the first picture I took a while back turned out crystal clear though...

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

LOL that would explain a lot.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

D'oh ! I think Danny just made a colossal screw-up with this whole story - too much work, not enough pay or lay
I had a look at the pinout of a 24p connector online, since it made no sense indeed and it just occurred to me that is NOT the PG wire there The PG wire should be gray and that's a brown one if I remember correctly I have no idea why I was so convinced PG is brown...brown is that 3.3v sense wire crimped together with one of the 3.3v ones, I guess for some sort of compensation to account for losses across the wires as close to the load as possible ? Come to think of it, this is most definitely a mag-amp regulator, since there's clearly no separate DC-DC regulator for the 3.3v rail, or ANY rail for that matter, since it's your run of the mill ATX.... With that out of the way, we can still analyze how this sense wire works, since we've gone that far

I'll definitely need to go back to the PS224, identify the gray wire and see what the deal is with that busted resistor...

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

The electrolytic could if it were open-circuit, or if the TL431 was shorted.
Or even if the transistor were shorted B-E the capacitor may charge too fast through it. But that seems unlikely.

This circuit makes no sense though. If the values of R314 and R312 are indeed 10 and 909 ohms respectively, that would produce a voltage divider that could never pull PG below 3.2v...

Unless the pad you've marked as GND is not GND, but actually some negative rail.

And also, why didn't they just use the PGO pin on the PS224?! It may be their PSU design isn't actually stable at the point where the PS224 outputs PGO so they had to make their own circuit with a longer delay?? Who knows....

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Here is a close-up of the side which has the PG wire and those two transistors. One of them is actually an AS341 IC (the one on the left), so we've got some regulator action going on ??

The device on the right is a PNP transistor, but I forgot the model right now - it's like B773 or something along those lines (has an UTC logo at the top of the can), but there's very little info on it. Still, I WAS able to find a pinout somewhere, though not an official datasheet...

There's also two diodes between the middle pin (collector) of the transistor and the GND plane....

It seems to me like they're using this setup to pull the PG wire low when the 3.3v rail has stabilized ???? Would this make that small electrolytic cap a culprit (red symbol) ??

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

It IS Raidmax, wouldn't surprise me...

Back in 2006 a couple friends of mine bought a few Raidmax supplies for themselves and family, all the same model PSU, and every single one had the primary capacitors go up in smoke within a month.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

There HAS to be some connection between that PG pin and the PG WIRE....otherwise why would the PG pin even be connected ??? I was thinking it may be a simple time-delay using those transistors and cap(s) and no "real" PG indication....will have to see.

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

It's almost always when I'm convinced that "this component can't possibly be faulty" that it turns out to be the problem. Murphy's law, or some variant thereof.

But yeah, draw up a schematic, this design sounds fascinatingly bizarre!

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

That diode looks like it's got a burn spot, too?

I measured it and seems fine - 0.6v drop across it in only one direction.
It could be a cap, though I've never seen small ceramic caps like this failing open, though I'm not saying it's not possible...
I shall do more digging, mostly for fun and because the supply seems decent enough to keep. I shall attempt a circuit diagram when I have the time.

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

That diode looks like it's got a burn spot, too?

If there was a short, it could have blown clear and you can't find it now. You mentioned a capacitor. Perhaps it works to provide a delay on PG line. Maybe the transistor circuit works as a buffer to provide a clean signal to the motherboard.

If that capacitor shorted, and blew, the diode/resistor could be damaged and now the capacitor is open-circuit, leading to the instant PG signal.


With regard to connecting the circuits, there could be top-side wire links etc hidden under other components or under blobs of glue?

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

According to that datasheet, your PS224 PGO pin is an open drain. So that burnt resistor and diode may be a pull up?

Yes, but it's strange the brown PG wire is not connected there to make that a pull-up resistor. However....

Maybe they use 12v to pull up PGO and that transistor you find connected to PG wire is a level shifter circuit

This might make more sense, though I could not, at least so far, find a trace which connects the two "halves" of this PG circuit: the wire and the transistors are on one side, while the area in the picture is on the opposing side. Another thing I forgot to mention is that there IS what appears to be a pull-up resistor on the PG wire connected to the bunch of orange 3.3v wires !

Anyway, if the resistor burnt, something may be shorted on that PG line, or it was plugged into a faulty motherboard.

There's no immediate short that I could find, plus the supply starts up fine and outputs are perfect, but no MB is going to start with a 0ms PG signal. The possibility of an external factor killing the resistor remains though, which would be a best-case scenario since it would mean replacing it may solve the problem....or not....:|

[Agent24]

Re: Raidmax RX-600 ringing noise and 0ms PG

According to that datasheet, your PS224 PGO pin is an open drain. So that burnt resistor and diode may be a pull up? It should be to 5v though...

Maybe they use 12v to pull up PGO and that transistor you find connected to PG wire is a level shifter circuit. Why the heck they would do that though...?

Anyway, if the resistor burnt, something may be shorted on that PG line, or it was plugged into a faulty motherboard.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Well, just in case someone is still watching this thread, this is a follow-up to the last power supply I wrote about which has the same 0ms PG issue. Unfortunately this one was not solved after replacing the main cap. I tried tracing the brown PG wire to see where it comes from and it's tired with a couple of transistors, resistors and caps. I was expecting it to lead to the supervisor IC, a PS224, since it has a dedicated PG output on pin 16, but it doesn't seem to, at least after a brief inspection with my meter's continuity function. It plugs in on the far end of the board too !

I DID notice something else though: look at the pic. That SMD resistor is close to the PS224 IC, on the opposite side of the brown wire and has a burn in it. Sure enough its value is way off - it reads about 2kOhms, instead of the 10ohms the "100" marking on it implies - at least I THINK it's 100. I'm not quite able to grasp the function of this or why it burned in the first place: it's connected via a diode between the 12v output and the PG pin of the PS224, ironically enough ! There's also a cap to GND on that line further to the right, so it seems to me like they're using this setup as a slow-rise signal line, but this doesn't explain why the PG wire is not connected here, or heck, why it's connected to the PG pin of the IC, since that one SINKS current ! I haven't replaced the resistor yet, but I will soon just for fun. I imagine if the PG pin got "stuck", that gives us 12v/10ohms = 1.2A of current through that resistor !

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Sorry for bumping an old thread of mine, but I have another PSU with a similar issue and I'm very curious to see if it has the same fix: starts up fine, but the PC doesn't POST and although this joint doesn't have a power supply tester, I can pretty much guarantee it's related to PG again - I'll bring my own tester tomorrow just to confirm...either that, or there's too much noise on one of the rails which the MB is not liking....a scope would confirm this, but I don't have one so yeah.....it's a bit of gamble.

Speaking of which, does anyone know any "good" PSU testers ? All I can find are basically the same thing just with slightly different shapes and colours, all over Aliexpress and the rest of the internet, even local stores: blue screen in the middle, 24p connector at one end, 8p at opposite end.....but I think we can agree these are not the most "professional" things out there, which I agree, but they sadly look like the only option, unless maybe someone has designed their own. The lack of a load is the main issue I spot with these, though I imagine this could be improved by putting it in parallel with a constant-current load. After all, it's just a voltmeter with like 4 inputs, so the only truly meaningful readout is that PG value. The next step up would be to also connect a scope to see how bad the output ripple is. It just crossed my mind to get one of those scope kits off Ali and "integrate" it into the tester, though in this case "integrate" is too generous - "patch-in" would be more appropriate....not sure how great those are at picking up small amounts of noise. I mean, if the ripple is sufficiently bad to cause the MB to no start, would the scope reveal this, or would it still show a relatively flat line which to me would just be a false "good" ?

[momaka]

Re: Raidmax RX-600 ringing noise and 0ms PG

for cost/corner cutting reasons, obviously not! the bean counters dont want the extra unnecessary cost! they just want it to just outlast the warranty and blow up deliberately and unrepairably, so u will be forced to get another to keep them in business.

Maybe.
But now that you mention this, it is worth nothing that some of the high end PSUs out there have 10 year warranty and all kinds of bells and whistles. If they were trying to save on components, one extra PP cap certainly wouldn't have added that much more to the cost. So it makes me wonder now if a PP cap is not added there because it cannot stand the high ripple current??

On my Enermax 525W PSU that had bad (open) primary Panny caps, there was also one 0.1 uF ceramic cap in parallel with them, which is what I think saved the APFC section from blowing up ultimately (along with the NTC thermistor volunteering to take the fall and save everything else - thanks buddy! )

With that said, maybe a high-value ceramic cap would be better/more suitable? Of course, finding a 1 uF ceramic cap with a voltage rating of 450+ VDC is going to be quite a bit more expensive than a PP of the same voltage and capacitance ratings (I'm seeing prices on Digikey indicating $2+ for a single 1 uF cap... vs. 50-70 cents for a PP cap of twice the capacity.)

So maybe it is price and application reasons they are not using them there. In which case, we will have to see how my "experiment" with my Antec EarthWatts 430 fares over time. Perhaps the cap will explode and cause major damage? We will see. I do need to put it in a PC, first, though. It's in the garage currently, and one of my last spare "high-power" PSUs.

[ChaosLegionnaire]

Re: Raidmax RX-600 ringing noise and 0ms PG

makes me wonder why they're not doing this out of the box already....

for cost/corner cutting reasons, obviously not! the bean counters dont want the extra unnecessary cost! they just want it to just outlast the warranty and blow up deliberately and unrepairably, so u will be forced to get another to keep them in business.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Sorry for replying so late to this. I must've been really busy and just read the notification in my inbox, but forgot to check back on the whole thing.

This.....

As of lately, I've started (as an experiment) putting polypropylene metal film caps in parallel with the primary caps of these APFC PSUs - typically about 1 to 3.3 uF. I figure the PP cap should help with absorbing some of the high frequency ripple current (due to its lower ESR) and relieve some of the stress from the primary electro cap. But we will see if this (experiment) works or not. At the very least, I expect the PP cap to act as a "fail-safe" reservoir: that is, if the primary electrolytic cap fails, the PP, even with its tiny capacitance, should provide enough storage to keep the APFC circuit voltage from spiking too high and killing silicone parts. But again, that's all a theory in my head.

Interesting stuff. Definitely interested in hearing more about this, as it could very well increase the life of all sorts of PSUs that I might come across in the future....makes me wonder why they're not doing this out of the box already....

By the way, side story: remember that Tyco PSU I kept yapping about ? Here's some more recent pics of its current status (scroll down to see an older one for comparison...it's like one of those "spot the difference" games)

[momaka]

Re: Raidmax RX-600 ringing noise and 0ms PG

It had absolutely no signs of bulging at the top, that's why I completely disregarded it at first, so hats off to the person who suggested it might be the culprit. Good thing to keep in mind in the future when the issue appears to be on the secondary

You're welcome!

Okay, okay, I don't want to give myself too much credit... but yeah, I kinda did mention it in my first reply to the thread that replacing the primary cap could be behind at least one of the issues you were having. And it wasn't just a wild guess, either.

Now you know the drill. If you see PSU with APFC, always check the primary cap. This actually applies regardless if the cap is a known good Japanese brand or not. APFC PSUs simply put too much stress on them, especially if they are undersized (which is typically the case with the more "budget" models.) As of lately, I've started (as an experiment) putting polypropylene metal film caps in parallel with the primary caps of these APFC PSUs - typically about 1 to 3.3 uF. I figure the PP cap should help with absorbing some of the high frequency ripple current (due to its lower ESR) and relieve some of the stress from the primary electro cap. But we will see if this (experiment) works or not. At the very least, I expect the PP cap to act as a "fail-safe" reservoir: that is, if the primary electrolytic cap fails, the PP, even with its tiny capacitance, should provide enough storage to keep the APFC circuit voltage from spiking too high and killing silicone parts. But again, that's all a theory in my head.

simply wiring the fan to the 12v. output is the lazy way of fixing this, rather than bothering with figuring out why it's not PWM-ing

It would also make the PSU too loud and collect too much dust.
Connecting the fan to run on 7V (positive wire on 12V and negative on 5V) is equally lazy, but should at least make the fan a lot more quiet while still pushing a good amount of air. Only con with doing that is, if the PSU runs very hot under high load and expects the fan to run at 12V to keep it cool, you'll have a problem. So it's possible that the PSU may no longer be able to provide "full power" with the 7V fan mod. Of course, if the PSU will never be used under 100% full load, then this mod will be good enough. I have a few PSUs of my own wired like that, because the PCs they are in will never put a load large enough on them to require the fans to speed up more than the 7V mod I am running them on.

[Dannyx]

Re: Raidmax RX-600 ringing noise and 0ms PG

Nice! I'm kinda surprised the PG issue was caused by the bad primary capacitor.

Some PSUs really do leave the fan off until things warm up. You can always try heating the sensor with hot air and see if the fan kicks in.

It's also interesting to note that there's also a red latching push button on the back, next to the mains inlet, which upon first glance you might think is the main power switch, but it's not: it's already got the regular rocker main power switch next to that, so what's it for ? Looking inside, it's got some thin wires running in parallel (I THINK) to that NTC thermistor, so my guess is that the button shunts the thermistor to toggle between "full on" and "auto" operation of the fan, but nothing is happening regardless of whether it's pushed in or not, so it's likely the small circuit at fault....no idea how it operates, since I haven't dug deeper...