Old Fortron FSP250-60GTM

[Behemot]

I've bought this PSU as a "working, just the fan does not spin up" for 1 CZK with 60 CZK shipping, coverted into USD it's around $3,3 alltogether

First of all, it has only 3,6 V on +5 V SB rail. After connecting the green wire on ground, nothing happened and no voltage anywhere. It was supposed to work connected to motherboard, so I tried. It boots up but the PSU whistles and it succesfully entered the BIOS on some 8. try (usually rebooted).

After having a look through case holes, most caps I've been able to see are popped. Bit unusual as most of the Fortrons I've seen had bad caps but visualy OK. I had in plan to recap it anway, I am just waiting for one shop in Czech republic to tell me how to order things from them (they do not have e-shop yet) because my low capacity caps supplies have run out and I have one another Fortron (FSP300GT) for recapping as well

The parameters: passive PFC, 250 W max (105 W on +3,3 and 5 V rails together), 14 A@+3,3 V, 18 A@+5 V, 13 A@+12 V, 2 A@+5 V SB, 0,8 A@-12 V

Stay tuned for photos

[Behemot]

Re: Old Fortron FSP250-60GTM

So I looked into, it was too dark for taking photos, so maybe tomorow...but it looks like the fan stopped working due to all the dust and most of caps overheated and blistered, half of them popped. There are mostly Teapo ones butworse is there is only 2x330 uF input, it's too low but I think I have some with higher capacity from shitty PSU's. Crap brands, but still higher capacity I guess

There are also 3 inductions replaced only with wire connection (we all know cost savings). As I never replaced inductors before, do I have to follow some specifications e.g. number of convolutes or I can just take coils from one PSU and place them here?

[Behemot]

Re: Old Fortron FSP250-60GTM

Here are the pictures. It had to be pretty overheated, look at all those color changes on transformers and fixing glue everywhere. The insulator spaghetti on the two resistors in down left corner is even totaly baked. I peeled it off together with some of the paint, I hope beside that the resistors are good

The wires are pretty thin for me, I consider whether to replace them or not. Even some old dead 250W EuroCase I have here somewhere has thicker wires

[c_hegge]

Re: Old Fortron FSP250-60GTM

Some serious heat damage there. Let's just hope that the caps are the only things cooked. I would be replacing those resistors, since they can suffer damage from heat too.

[Behemot]

Re: Old Fortron FSP250-60GTM

As I wrote, I already tested it connected to motherboard. It POSTed, just was very unstable and had difficulties even with entering BIOS. I blame mainly the caps. Maybe the dummy load resistors are damaged too, but recap is 1st thing to try I think

I am interested now in information about replacing the wire interconnects with actuall inductors which are supposed to be there acording to the PCB marking. Do I have to folow some rules or can I just take them from dead PSU's and place into this one directly?

ADD// I already have unsoldered the fan, cleaned it and lubricated. It's running fine now I also cuted-out the fan grill so I can slow the fan with 10-20 ohm resistors for the same airflow but lower noise

One other thing is there is no foil under the PCB which is unacceptable for my my paranoia, too. So I have to place there something, just for case, you know

[momaka]

Re: Old Fortron FSP250-60GTM

I am interested now in information about replacing the wire interconnects with actuall inductors which are supposed to be there acording to the PCB marking. Do I have to folow some rules or can I just take them from dead PSU's and place into this one directly?

Should be able to just take them from another power supply and place them directly. Many people here have done this.

I also cuted-out the fan grill so I can slow the fan with 10-20 ohm resistors for the same airflow but lower noise

Is the fan directly connected to the 12v rail or is there a speed controller?
If the fan is just connected to 12v, you can make it run at a lower voltage without the need to use resistors. Just connect the positive wire on 12v and the negative on either the 5v (for 7v) or 3.3v rail (8.3v). This should make it more quiet but still move plenty of air.

One other thing is there is no foil under the PCB which is unacceptable for my my paranoia, too. So I have to place there something, just for case, you know

Good idea. I've seen a few cheap PSUs die because of arcing between high-voltage/primary side to the case, usually due to lack of this plastic sheet.

By the way, what brand and series of capacitors are you getting for this PSU?

Lastly, once you get this PSU fixed and running, check the voltage on 5vsb (use a multimeter, not motherboard monitor). If it's above 5.25v, you'll need to replace one of the small caps in the 5vsb section (it's usually 10, 22, or 47uF).

[Behemot]

Re: Old Fortron FSP250-60GTM

Should be able to just take them from another power supply and place them directly. Many people here have done this.

Thanks

Is the fan directly connected to the 12v rail or is there a speed controller?
If the fan is just connected to 12v, you can make it run at a lower voltage without the need to use resistors. Just connect the positive wire on 12v and the negative on either the 5v (for 7v) or 3.3v rail (8.3v). This should make it more quiet but still move plenty of air.

Yeah it's probably the last model without speed controller. There is even no OCP protection, just holes and marks for it.

I know about this possibility for slowing down fans, but I've heard about cases when the fan burned and than PSU died because the power ran from 12V to 5V rail. And I bought 150 pcs of 10ohm resistors some time ago, I need to use them somehow

By the way, what brand and series of capacitors are you getting for this PSU?

Lastly, once you get this PSU fixed and running, check the voltage on 5vsb (use a multimeter, not motherboard monitor). If it's above 5.25v, you'll need to replace one of the small caps in the 5vsb section (it's usually 10, 22, or 47uF).

I have Samxon RS 3300 uF/16 V for 12V rail and Nichicons HM 3300 uF/6,3 V for other rails. I also have in plan to recap the whole PSU even the small caps. These ones mostly with RL/RS Samxon's or with Chemi-Cons KMG which I'd like to order (some small-capacity Samxons already run out ).

The two input caps (330 uF CapXon's) are replaced with bigger ones (560 uF HEC) now. Better than nothing I think I'll leave it that way, input caps are usually OK and due to the price it's not worth the change I guess

[Scenic]

Re: Old Fortron FSP250-60GTM

please check the datecode on those nichicon HM's..
they had a mess up at the production affecting all HM / HN caps up to late (?) 2004.

if the datecode is newer than 04 it's probably OK..
(i've used HMs with late 06 datecode for ages now and they're still fine)

Datecode marking is something like "M0345" which means year 2003, week 45

[Behemot]

Re: Old Fortron FSP250-60GTM

Thanks for the info, I thought only HN were problematic

These are H0143, so year 2001. I have them from some motherboard and already used one or two, not sure where right now, I think in one PSU and it looks like they're running fine.

So should I use different ones or these are so old they were not affected with the problems yet?

[c_hegge]

Re: Old Fortron FSP250-60GTM

They are affected. I say anything between 2000 and 2005 is affected, since I have known caps with those date cades to fail.

[Behemot]

Re: Old Fortron FSP250-60GTM

OK I'll move them to shitty caps bag. I should have some Chemi-Cons or Sanyo's with similar specs as well, I'll have a look later.

But this is some fail, the were screwing it up for 5 years and nobody noticed? omg

[goodpsusearch]

Re: Old Fortron FSP250-60GTM

OK I'll move them to shitty caps bag. I should have some Chemi-Cons or Sanyo's with similar specs as well, I'll have a look later.

But this is some fail, the were screwing it up for 5 years and nobody noticed? omg

A possible explanation: They started to fail some years after production, so it took them some time to realize there is a problem with them.

[Behemot]

Re: Old Fortron FSP250-60GTM

A possible explanation: They started to fail some years after production, so it took them some time to realize there is a problem with them.

I'd expect from such a company to at least once a half-year do some serious stress tests behind limits the caps are supposed to handle. And they should still handle it, at least for some time. It will tell them something is not how it should be. Even if they had the caps under normal stress conditions, they should find it far sooner than after 5 years! This is total FAIL in my eyes and it tells me they were trying to save costs in the wrong place. Hope they don't do so since 2005.

As for this PSU, I'll use Chemi-Cons KZG and/or Rubycons MBZ, both 2200 uF/6,3 V instead of Nichicons. I think 2200 uF is enough for this PSU (the MoBo which will be powered by the Fortron is recaped anyway), I better save the 3300uF ones for some PSU's with higher wattage (like the 300W Fortron I have ).

[Behemot]

Re: Old Fortron FSP250-60GTM

So I've replaced all electrolytic caps and swapped wire connections for inductors. I also relaced input wire interconnect with actuall thermistor and did the fly mod for dummy load resistors (this is great thing indeed) since at least 4 of them are pretty hot and I don't want them to heat my Chemi-Cons

With dummy load resistors inside of the PSU and with the fan, some voltages seems good (+12 V - 11,44; +3,3 - 3,38, but others are far away from perfect - +5 V SB has 5,22 (almost over specs), -12 V has only 10,06 V and +5 V has 5,28 V on it. The -12 V rail does not matter much, but the +5 V is over specs. And it still whistles heavily.

I'll try it with MoBo, maybe it needs higher load, but IMHO the PSU should run even without any external load.

[Behemot]

Re: Old Fortron FSP250-60GTM

Voltages are better with PSU plugged into motherboard. It also whistles less. I suspect the main shared choke or the huge coil near heatsink of being the source. I'll try to paster it with hot glue

But I've been surprised with efficiency. Measured consumption against Rexpower PL-350 and the Fortron sucks 8,5 W more electricity (50 to 58,5 W) ! I see two possible explanations - either it's due to the dummy loads (so it'll be better with higer load) or the wattmeter measures badly (because of the Rexpower has no PFC, it could make probles with power factor).

[Behemot]

Re: Old Fortron FSP250-60GTM

The main choke inductor is now burried under the hot glue but no change.

I think it could be main transformer, something probably happened inside due to the overheating and now it resonates with the switching frequency...is there some way how to make it silent?

Could some mod merge the last posts please :-) Longer time for editing would help...

[c_hegge]

Re: Old Fortron FSP250-60GTM

As for this PSU, I'll use Chemi-Cons KZG and/or Rubycons MBZ

Chemi-con KZG are also unreliable Unlike HM, they always have been problematic and still are to this day. I see just as many of those failed as HM and HN

The only brands of caps that NEVER had problems with ANY series are Rubycon, Sanyo and Panasonic.

[momaka]

Re: Old Fortron FSP250-60GTM

You should remove the hot glue from under the big coil. Hot glue expands quite a lot when heated up and if applied under hot components, it can crack their solder joints.

As far as the whine, is it there only when the PSU is on, or can you also hear it when the PSU is plugged in but turned off? The second case would suggest it's coming from the 5vsb. It could also be that the Nichicon HM and Chemicon KZG caps have too low ESR and are causing the coils on the output circuit to oscillate.

5.22v for 5vsb also sounds a bit high. Are you getting the same reading when PSU is connected to a load (i.e. motherboard)? If yes, then did you replace those small caps in the 5vsb I mentioned?

[Behemot]

Re: Old Fortron FSP250-60GTM

You should remove the hot glue from under the big coil. Hot glue expands quite a lot when heated up and if applied under hot components, it can crack their solder joints.

As far as the whine, is it there only when the PSU is on, or can you also hear it when the PSU is plugged in but turned off? The second case would suggest it's coming from the 5vsb. It could also be that the Nichicon HM and Chemicon KZG caps have too low ESR and are causing the coils on the output circuit to oscillate.

5.22v for 5vsb also sounds a bit high. Are you getting the same reading when PSU is connected to a load (i.e. motherboard)? If yes, then did you replace those small caps in the 5vsb I mentioned?

I applied the glue on top of the coil, I won't be able to get it under it, it's impossible

The sound is coming from the PSU only when it's turned on and it varies according to load. That's what leads me to main transformer. I desoldered the main transformer, heated it with hot air gun and soldered back, I also tried to replace the inductors with wires again (there could be oscillation between them) and nothing. Than I've put the inductors back and suddenly no whistling with no load. When powering mainboard, it whistles again

I get exactly the same 5.22 volts even when the PSU powers the motherboard. And yes, as I wrote, I replaced every single one electrolytic capacitor.

As for the caps, I believe ESR is not the problem - e.g. Samxons RS have much lower ESR and they are actually meant for power supplies. Are not caps in PSU's almost only low-ESR ones anyway? I've been quiet surprised when I saw recommendations of caps which usually are in VRM's on motherboards as a replace for failed caps in PSU's.

As the test with inductors implied, I think it's not problem with oscillation between some of the components, but probably the wires inside transformer resonate with the switching frequency. Remember the PSU was overheated, there are color changes even on the main transformer so there could be some extra room made inside where the wires are moving now and it makes the sound I am considering now whether to try to put the transformer into mineral oil to fill the room?

On the top of all, I'll let the KMG's there and see...there are definitelly better than bulged OST's and Teapo's now, aren't they? I'll have a look on them from time to time and report...once again, what was the reason for them to fail, they did not handle the heat or even the ripple?

Well, maybe I won't use the PSU to one of my normal computers...maybe I'll keep it for running in oil. I plan it for quite some time now

[c_hegge]

Re: Old Fortron FSP250-60GTM

Did you use KMG or KZG? KMG are reliable, but they are general purpose caps, not low-ESR, but any cap which hasn't failed would be better than a busted one.