Tyco spi270f4b24 ATX supply repair tips (not starting)

[lpraderi]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Hi Dannyx in case you still hold this dead creature, received today a Tyco/Elo Touch POS/Cashier with the same PSU (SPI270F4B24).
Customer wants to upgrade CPU/Ram, and only reuse monitor/touchscreen.
Tested complete system before upgrade, PSU works but after a couple of minutes shutoff with that "relay like" sound.
Finally put (outside case) standard ATX PSU with borrowed CN19 connector piggybacked to ATX connector (without 24V lines) from original PSU.
So long everything works OK.
Let me know if you need some pictures/ measure from partial working PSU unit.

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Just in case someone is still following along at home, this project has been on hold ever since and I decided to look into it some more, not wanting to give up.

I did some more research on mag-amps like previously suggested and found resources ranging from scientific articles to more practical explanations like this one. At least now I kinda know what to look for. This opens up the possibility of the mag-amp not running and the output "overshooting", tripping OVP on one of the outputs which the PS223 senses and shuts off the supply. According to that article, it's the 3.3v output that's most likely to be mag-amp regulated from the 5v supply, while the 12 and 24v outputs are not directly regulated. If so, our mag amp would consist of op-amp LM339, those transistors on the riser board and the largest coil most likely, since there are no other components around.

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

I haven't actually replaced any ICs - just the PS224 with the PS223, which resulted in the aforementioned pop (which wasn't my idea to begin with ). Of course, I did some more digging in the meantime, removing the whole heatsink assembly on the secondary so I could get a better look down there, but again: nothing - just coils and jumpers, none of which appeared charred or burned. After all, the pop was rather small - not like what happens when something on the primary goes wrong for instance. It was more like a small spark with no smoke afterwards.

I also took this opportunity to measure the diodes on that heatsink out of circuit. All of them read good, with low Vf in one direction (forward biased) and no flow in the other direction. They're not identical - it's actually a cluster-f in there, with various types of diodes mixed together, so the results weren't consistent. One of them is actually a LM7912 regulator for the negative 12v bus, but that's ok too.

The pop was on the secondary side and it appeared to come from the area I pictured up close with the heatsink out. I thought that transistor or blue cap went, but that would mean no 5vSB which I AM getting, so those are fine. I also thought about those two TO-92 packages on the daughter board, but they show no sign of damage proportional to that spark (cracked case, unreadable serial number...). My only other idea is that one of those coils did it, like if the insulation on them gave in due to age and made contact with something it shouldn't...it might also explain why the PSU is now immediately shutting down. Just before the pop happened, the PSU came on and appeared to work fine, then there was the pop and now this (immediate shutdown) so my money is on a short somewhere in those coils.......but I can't find it. I doubt replacing the 224 will help, as it wasn't faulty to begin with....

The schematic is only for the daughter board. True, not that handy and it's a mess....doesn't really matter. It was more for proof of concept.

Mag-amps: I can indeed see some wires going THROUGH the big coil (the one with blue paint and reddish wire barely visible) and I believe it doesn't utilize series resistors for current sensing but instead these "mag-amp" things. There's only one actual shunt in there, on the 24v bus which I included in the schematic.

[momaka]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

No offense, but I think you're wasting time replacing all the ICs, especially on the secondary side. If there's anything that's going to be popped or bad, it would likely be on the primary side. After all, you had something pop. So there's definitely going to be a bad component in there somewhere now. Only way to find it is to dig deeper with the multimeter and soldering iron.

You still haven't reported back if you actually checked all the primary-side parts and the output rectifiers too. So no point drawing a schematic really. Only thing I can tell from your schematic is that the 3.3V and 5V rails are mag-amp regulated. If their parts check out alright, it won't be them that's causing the trouble.

That being said, with the 3.3V and 5V rails being mag-amp regulated, this means you need to get the 12V rail loaded up to a minimum of 1-2 Amps while testing. Otherwise, the PSU may act up and not want to start.

Last but not least:

It's not fixed yet though because he's tested it through a series lighbulb on the mains which stays on constantly because of the PFC booster - it's not a defect, I've seen it happen many times on PFC circuits. The added current limit prevents the circuit from going up to 400v which forces the "pump" circuit to stay on to try to compensate, thus drawing a lot of current, but it works fined if you attempt a startup straight on the mains.

To get around testing APFC PSUs with a "dim bulb" tester circuit, you just need a more powerful light bulb. 200-400 Watts halogen or incandescent should do the trick. Maybe even 500W, if it's still acting up. Even with such a powerful bulb, the current will still be limited to about 2.2 Amps @ 230V AC line (note: this is not running current but rather short-circuit current if the PSU shorts somewhere. The running current will be limited to whatever makes the line voltage drop below 90V AC, as that's usually the lowest most APFC PSUs can handle).

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

v3.0 - replaced physical LM339 op-amp with individual internal op-amps

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

v2.0 Corrections, tidied up. This might happen later on too, so sorry for posting so often

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Here's the schematic for the riser board with the PS224 IC, I'm just gonna leave it here for people to analyze, but it's provided "as-is", so it may not be correct and it's also rather crudely drawn in MS Paint.

There's a couple of unknown SOT-23 components on there, but I ASSUMED they're equivalents of KIA431, because there's another one of these controlling the feedback optocoupler for the 5VSB bus and they bear the same marking, A98F, so by analysing its pins and comparing them to the ones on the riser, I concluded they're the same: anode at GND, cathode to whatever's being controlled and the REF pin to a resistor divider usually.

The values on the resistors are the actual markings on them to help me follow along easier as I'd get lost otherwise. I don't think there's anything wrong with this circuit, but there WAS a short there which I now removed - it was between pin 3 of PS224 and GND, causing the LED in the optocoupler to turn on as soon as power was applied regardless of what the PS224 was doing. Don't know why this caused something to pop though and I sure hope it's the PS223 currently on there that popped, since it will be replaced anyway with the correct 224 part.

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Schematic is almost complete and it's a good thing I spend all that time with a M/M probing each and every single component because I did spot a f-up: there was a short created by solder blob between a ground plane and a components lead which was not supposed to go to GND (I think the FGO pin, which also explains why the supply fired up by itself for no reason back there)...

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

I also don't imagine this thing running with a 223 instead of a 224 because of the slight differences on pin 9 - this is why I began drawing a schematic (a software better suited for this purpose might've been better than MS Paint, but hey....). I'm now back to square one where there's some voltage coming on briefly at startup, but then shuts down (this with a series lightbulb, so it could be normal). Still, something popped there for sure but I can't find it...might have to start removing components, since something could be buried towards the board and I can't see it. Why don't people have DVRs in their eyes ?

[stj]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

it probably turned on because of that fuckery with the opto-isolator!

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Just discovered another hurdle: there's a dedicated 6P connector (beside the 24p one) where the brown wires are 24v outputs, which are obviously not found on any ol' crappy ATX supply, so adapting one is not possible, even if it were to physically fit in the enclosure. I realised this only when looking at the top cover again that it said "24v (BROWN)", because I was seriously considering replacing the whole unit at this point. What threw me off all this time is that the point on the board where the bundle of brown wires is soldered and crimped is marked as "+3.3VSN", so I just assumed it's another 3.3v rail and I could adapt that from the regular 3.3v line on a standard ATX, but no...don't know which to trust: board or cover

I actually started doodling a schematic of the riser board in MS Paint...it's going slow with all those SMDs, but I'm doing it...

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Well that didn't work so well and we broke the damn thing: a startup on the mains was somewhat successful, in that the fan started to spin, but here's the weird part: the green wire wasn't jumped to GND, yet the thing turned on...it worked for a few seconds then something popped in the secondary (I saw a little spark in the general area of the heatsink), but even inspect it for hours today I COULD NOT see what the hell popped there to cause that light I saw...there aren't even any components there !

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Follow-up: today, my co worker messed around with this supply some more after I put it aside for a while. I'm not entirely sure what he did to it, but it's a step in the right direction, annoyingly enough

From what I gather (the old man himself forgot what he did to it ), he pulled off the CM6800 and replaced it with another one from a different supply, along with a small SMD cap on its VCC pin and also the PC223 on the secondary. This I believe is the second/third time he's replaced these ICs and the PS223 was actually a PS224 originally. Today he noticed a difference on pin 9 between the two. One is called Pext and the other OTP...funnily enough the original 224 used Pext, whereas the board actually has an NTC on there which would make it more suited for the 223 which has NTC "support" (or it could very well be for fan control and not involved in any temperature sensing)...he put back the original 224 and somehow, all of these combined lead to 14v passing through the optocoupler which originally caused so much trouble...my money is on the PS224 - there was nothing wrong with the primary to begin with.

It's not fixed yet though because he's tested it through a series lighbulb on the mains which stays on constantly because of the PFC booster - it's not a defect, I've seen it happen many times on PFC circuits. The added current limit prevents the circuit from going up to 400v which forces the "pump" circuit to stay on to try to compensate, thus drawing a lot of current, but it works fined if you attempt a startup straight on the mains. No major components were popped if you think about it.

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Turns out I was wrong: there's only ONE IC on the primary - the other one is on the secondary.
The one on the primary is a CM6800 so VCC is on pin 13 which is indeed connected to the other side of the photocouple. My buddy suggested jumping across the optocoupler, but I'm not sure that's a good idea...

Also, the secondary DOES seem to go up for a brief moment - like 1-2v, so it looks like it's shutting down for some reason...the IC on the secondary is a PS223 and it's supposed to "sink" the cathode of the LED in the photocoupler to GND when it's issued a "PS_ON" command. This works, because I removed the "output" leg on the primary side of the optocoupler and there was 14v going through when the green wire is shorted to GND and it disappeared when it was removed. Still, when I put it back in, it's being shut off by the secondary I THINK, so my buddy's assumption was flawed all along and I rest my case: I've NEVER seen a bad optocoupler causing issues in a PSU...

[Stefan Payne]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

It's an FSP according to the UL code E190414.

A look at the Label would already have told you that
As they didn't even change that, just put their name instead of FSP on it.

And there is the SPI (Sparkle Power Incorporated??) Number on the PSU as well, wich points also towards FSP...

[PeteS in CA]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

https://en.wikipedia.org/wiki/FSP_Group

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Oh so it's a brand

Tomorrow it's back to work on this PSU, so I'll start by measuring for shorts just for good measure and also check the ICs and post back.

[eccerr0r]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

FSP seems to make acceptable PSUs...

Wish it were true that you can make assumptions of previous attempts, but it rarely turns out to be the case, and have to do all of the analysis over again

[Dannyx]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Silly question: what does FSP stand for ?

I didn't try any of the "major" stuff personally (like measuring the output for shorts or checking large components etc) because I PRESUMED any of the guys before with the least bit of electronics knowledge would've done that first thing....of course you never know. I delved straight into the nitty-gritty of the PSU without checking major stuff first, which I shall just for good measure. One thing is certain: nothing was burnt to a crisp from the get-go, it just wouldn't start up.

About the optoisolator: I mentioned in my original post, but since I tend to write a lot people often give up and miss certain details: it's soldered like that solely for testing so we could measure the resistance across the internal transistor (the legs sticking out in the air) when the LED lights up when the PS_ON command is sent (green wire to GND). The LED does light up, because standby is handled by a small 8 pin IC and transformer (you can see them towards the edge of the board).

PFC most likely doesn't start because there's no VCC AUX being sent to the ICs (what actually turns them on). I also said I'd track down and measure the actual VCC pin of the IC rather than assuming the optocoupler is connected straight to the VCC pin. There's most likely going to be a resistor in there too.

Might not start without a load: that's another thing that needs to be tested.

I shall post more info about the ICs when I can get to them.

[momaka]

Re: Tyco spi270f4b24 ATX supply repair tips (not starting)

Ironically, in one of the repair shops where I worked a while back, I had a PSU very similar to this to fix. I don't remember if it was made by SPI/Forton specifically, but it was for a POS/cash register system too. It had a blown APFC side and fuse. I didn't know how to fix it at the time, so we just replaced it with a regular ATX PSU wired externally with long wires (ghetto solution, but owner was okay with that).

Anyways... regarding your unit, I see an optocoupler removed and soldered on the bottom. Make sure that opto is soldered the right way - not just diode to diode side and output to output side, but that the pins for the diode and output are also wired correctly. If the diode, for example, is reverse-biased, of course the output side won't pass current. So that's #1 thing on the list.

Next, check the voltage across the primary cap when you jumper PS_ON to ground (make sure you got some load on the PSU too). Most likely, it should go to at least 360V. Any less means it's probably not boosting.

#3: can you tell us what ICs are in this PSU? APFC chip? PWM chip? (I suspect UC384x family, if not CM6800 PWM-APFC combo). Supervisor?

Yes, it's a good challenge, but when it comes to electronics, I sometimes really hate working on other people's half-assed jobs, lol.

*EDIT*
Seems like I'm repeating a lot of what Kaboom has said. Should have refreshed the page before typing my post.