FSP400-60GLN worth fixing?

[Behemot]

Re: FSP400-60GLN worth fixing?

For sure 1A didoes are fine. First, 1N4007 are 1A and it mostly works fine with them. Becase the question here is whether the diodes overheat so much they kill the caps, or the caps die first and the extra switching makes diode overheat so much. I think that 1N4007 may be almost fine for normal operation but when caps start going down, it burns itself to dust. Faster diode is better because it produces less heat already and with quality caps it should never fail again.

Second, because of available space/diode (and electrode) size you won't squeeze more than maybe 2 A anyway. That 75ns time is OK, ordinary diode is around 1 us (and they usually do not even state the time). One of the diodes needs to be 1 kV, other may be just 100-200 V, that will work OK.

As for resistors, use ordinary SMD resistor. I think all of them are from the same material but it should not play any role - the thing is the actual resistive area is so small they will burn anyway (remember that great part of the resistor is just package). Unlike ordinary resistors which can handle much higher power before they overheat and burn hence why you seen fireworks before it affected the resistor itself.

[stj]

Re: FSP400-60GLN worth fixing?

replace the term 1N4007 with UF4007 - they are much faster.

[Behemot]

Re: FSP400-60GLN worth fixing?

What?

[stj]

Re: FSP400-60GLN worth fixing?

whenever you intend to buy or use a 1N4007, use a UF4007 instead.

[Behemot]

Re: FSP400-60GLN worth fixing?

Anybody here wanted to buy 1N4007?

[budm]

Re: FSP400-60GLN worth fixing?

http://www.voltagemultipliers.com/html/appendix_b.html

high frequency diode applications, reverse recovery losses can significantly contribute to the power dissipated in the diodes. Reverse recovery losses occur during the transition from forward current to reverse voltage. When reverse voltage is applied to a diode, it will conduct in the reverse direction for a short time (the reverse recovery time). While the diode is conducting in the reverse direction, the power dissipated is equal to the reverse recovery current multiplied by the reverse voltage.

[Behemot]

Re: FSP400-60GLN worth fixing?

Yes. Anybody said anything else?

[budm]

Re: FSP400-60GLN worth fixing?

Nothing wrong with understanding as to why.

[Per Hansson]

Re: FSP400-60GLN worth fixing?

Found out what was wrong now, last time I tried it I said 5vsb was dead.
I did some poking now with my multimeter and found that Zener diode Z42 was shorted in both directions, you can see it in this old photo.
It shunts the power going to the feedback pin of the DM311 to ground if the breakdown voltage of the Zener diode is passed.
It should be a 18v Zener diode according to ipo3nk but I could not find one, seeing as it's only for protection I simply removed it and fired the PSU up, and it was working fine!
5vsb is rock stable and I'm bench testing it now with some old decommissioned hard drives for load
I also replaced diodes D41 & D42 with the Ultrafasts I mentioned that Elfa had: 70-024-80.
And the SMD resistor R4A I replaced with a regular resistor because I could not find the original now over a year later

The following schematics where really useful in troubleshooting:
https://www.badcaps.net/forum/showpo...20&postcount=5
https://www.badcaps.net/forum/showpo...7&postcount=13

[Behemot]

Re: FSP400-60GLN worth fixing?

Suggest buying replacement one, if I am not mistaken it is parallel to that 100uF/25V (usually) cap. If it goes too high the cap will blast. Also get the fusible resistor, if there would be any problem, ordinary resistor would not burn and as a result the whole circuit may explode again.

[Per Hansson]

Re: FSP400-60GLN worth fixing?

Yes I agree Behemot, problem is I was not able to locate a fusible resistor at Elfa.
And I can't warrant paying much if anything for it, this PSU has already been too expensive to fix honestly...
From what I can see there is only a tiny SMD cap on the feedback circuit: C44
I have already replaced all electrolytic caps, even the small ones so it should be quite happy like this.
I will of course buy the Zener diode next time I need anything, but for now I think it works fine like this.

I also tried to find fusible SMD resistors on Elfa but was not able to, but maybe I just don't know the eight search terms...
If you remember I incorrectly replaced the fusible resistor 2R0 at location R44 with a regular resistor which caused some nice fireworks when the DM311 chip went up in flames

[Behemot]

Re: FSP400-60GLN worth fixing?

Ordinary SMD, they will burn if too much power flows through. https://www.elfa.se/elfa3~se_en/elfa...2-74&toc=19543

[Per Hansson]

Re: FSP400-60GLN worth fixing?

Not sure I follow you now, are we talking about resistor R44?
Because the one I put in is a quite normal SMD resistor too.
It was this one: RK73H2BTTD2R00F.

Will the reduction of 0.25w > 0.2w power rating of the one you linked help that much?
EDIT: One more thing, the "Dielectric strength" of the one you linked is 150v, the one I used is rated for 200v
But as that circuit is connected to the main bulk capacitor it sees close to 400vdc, don't I need a resistor rated for that then?
Or in other words why hasn't the one I put it caught fire (yet?).

Ahh ok, now I understand.
So the original fusible resistor had already failed open "saving" the DM311.
Then I put in a _regular_ 2Ω resistor: RK73H2BTTD2R00F
And it blows up more spectacularly since it's not a fusible resistor.
And the DM311 was already shorted but was now allown to explode since the resistor did not burn up fast enough.
Thanks for setting me on the right track

[Behemot]

Re: FSP400-60GLN worth fixing?

I see normal resistor there, no SMD

[Per Hansson]

Re: FSP400-60GLN worth fixing?

Referring to my picture that budm annoted again.
The positive from the bulk filtering capacitor goes through R44 that is supposed to be a fusible resistor.
This one I have replaced with a regular SMD resistor at 2Ω namely this one: RK73H2BTTD2R00F.
And due to my original DM311 being shorted I know how it burns up too; not very pretty.
After this resistor we come to R4A which is a 1MΩ resistor connected to the DM311's Vstr pin, this is the one I have replaced with a regular resistor.
Actually I found the original SMD resistor, I had nicely put it in a plastic bag next to the PSU together with it's screws but it was so small that I did not see it

I did some more searching and realized that there are two revisions of the DM311, the DM311A that budm attached a datasheet for when this thread was started.
And the end of life DM311 (As of 23-Sep-2010) that was in my PSU originally.
They seem to basically be the same but there is a difference in the datasheet, the DM311 datasheet has an implementation example for a 5vsb PC circuit
You can identify them because DM311 should be printed on the first line of a DM311.
And DM311A should be printed on a DM311A on the second line.
I bought my replacement on eBay and they actually have the DM311 marking on the second line so not sure what's up with that.

But back to the implementation example, in it they use a 91V 1W Zener Diode instead of a 1MΩ resistor at location R4A.
And they also list the Zener diode for the feedback pin in their example which is ZD4 as optional, sadly no recommended values are shown for it.
When measuring the capacitor yesterday I had 386VDC on it, the 1MΩ resistor drops this down to 330v for the DM311.
Using the implementation examples 91V Zener diode instead I guess would drop the voltage to 295V

[Elitist]

Re: FSP400-60GLN worth fixing?

Greetings, friends! So long since I was here last that the birthday greetings stopped arriving. But have recommended many.
Hope this is the right location for another FSP400-60GLN problem? If not, please can moderator re-locate - thank you.
Had this beast for ~2yrs during which flawless and silent performance. Suddenly it stopped (whist ON). Hard reset runs for about 20s. before everything cuts out, ie during main OS loading and eqpmnt detection. Off load, all voltages OK. Can load 5V and 12V lines with 10ohm, but the 3.3V goes down and stays down with anything less than ~30ohms. Recovers after off for ~1min.
Clean as a whistle inside (at least, it is now!). No bad caps apparent, nothing burnt, nothing splattered around. Looking for assistance, please.
Although I've no formal electronics training, have fixed many PSUs for blown fuses, bad caps, blown varistors, blown bridge rectifiers, a very few blown MOSFETs and their associated components. It's these partial faults that really show up my lack of detailed knowledge. Can trace circuits but, whilst appreciating the hazards, limited ability with live PWM-switched mode devices! No longer possess signal generator or 'scope.
Thank you for any suggestions.

[Marianelforum]

Re: FSP400-60GLN worth fixing?

Greetings.

You said that the other 2 outputs seam to work fine under load, so knowing that all 3 main outputs draw their power from the same transformer, logic dictates that the problem could ( or rather should ) be located at the 3.3V output. It cannot be a lack of power because that would reflect on all the main outputs, so, one can conclude that it could be the under voltage protection that comes in, for some reason the voltage on the 3.3V output drops under load too much, so the superviser detects that and steps in. Now about the reason the voltage drops under load, that could be down to a failed capacitor ( or more than one ), and you should know that failed caps do not always have visible damage, sometimes the only way to determine their state is to measure them, check both the capacitance and the esr, you could find one or both of theese parameters out of order, and if so, than change it.

Another reason could be down to a problem on some other component of the mag-amp for the 3.3V, but i think that should be quite vissible, so the caps remain the most probable culprit. My advice is to change all the electrolytic capacitors ( and i do mean all of them ) with new ones, that is in my book always the first and main step on reconditioning a psu.

PS: You could always check for any cold solder, you might get lucky that way

[everell]

Re: FSP400-60GLN worth fixing?

Post #35 in this thread has a schematic drawing of the 3.3 volt circuit. Component values may vary from model to model, but this is the basic circuit.

[Elitist]

Re: FSP400-60GLN worth fixing?

Thanks, guys. Will work on it this week as time permits. Cap changing I can do blindfolded! [Can't we all!]
Later: - as you were. Closer inspection of the underside with a lens,
1) confirms D35 missing
2) very bad soldering around fan lead-thro's with solder bridging
3) serious burning of R92, presently metered as 57ohm but marking erased, of D90, polarisation line erased but giving static diode action on simple meter, and of R93, reading ~17kmarked 2102, all in-circuit, all smc.
Adjacent D92, R9F & R9G somewhat discoloured but giving reasonable static responses, also in-circuit.
Anyone got the circuit diag, please and/or confirm correct value of R92?
O.P. caps mainly TEAPO.

[Elitist]

Re: FSP400-60GLN worth fixing?

Here's what I've done so far:
-tidied up the soldering
-installed D35
-installed C99 (fan shunt, originally missing).
-swapped 3v CapXon O.P. cap for Teapo
and the result:
-3.3v now stays on with 10ohm load but with 2.3v reduced output!
They say, in research, if you can change something, you're working in the right area! {but only supposed to change one parameter at a time}. The burn marks tells it's the right area, but limited knowledge prevents further progress. It would help to know what value R92 is supposed to be, along with more suggestions, please.