Dell Power Supply NPS-250KB B

[Pilm2013]

For the above mentioned power supply, I've had issues where the power supply will not restart after a power outage. Both times I removed the power supply and examined it looking for any damaged parts like a blown fuse, or caps, etc, but found nothing. Then did some live probing, had AC where expected and rectified high voltage (doubled) DC in some places, but w/o a schematic that's about all I could do, nothing thorough. Anyway, despite having some AC and the high voltage DC, it had none of the lower voltages (+12, +/-5, etc). So I'm wondering if there's something that isn't getting RESET properly after an unexpected loss of power that's inhibiting the lower voltages. I see there's a PS_ON signal, I think it comes from the motherboard, to put the PS into Standby mode, wondering if that might not reset to 0V after power failure? Or maybe something in the protection circuit isn't resetting? Anyway, what I've found is if I leave the unit unplugged overnight, it'll restart in the morning, at least it has the 2 times this has happened. Kind of seems like something on the board is still energized and inhibiting turn on. I did by the way discharged the 2 input filter caps, but perhaps there's more stored energy elsewhere that's keeping the thing from immediately restarting after a power failure. Any ideas appreciated, thanks.

[ben7]

Re: Dell Power Supply NPS-250KB B

Is it also hard to turn on after you unplug it and let it cool down?

Most likely a small electrolytic capacitor has failed.

[Pilm2013]

Re: Dell Power Supply NPS-250KB B

Thanks ben7. If I turn off my PC normally (through Windows), it always turns back on without a problem, however twice I've had the power fail momentarily and when I went to turn it back on it did nothing, even after sitting for 10 minutes or so. So I suspect when I do a normal power down the power supply has some sequence of events it follows, perhaps discharging or powering down in some fashion, that isn't getting done when I suddenly lose power. So I guess the issue here is it works fine in normal operation, but has trouble rebooting after a power failure, and my guess is there may be a design issue with the PSU that Dell didn't account for sudden loss of power hanging the thing up. The main reason I suspect some kind of improper discharging is because if I leave the thing sit overnight after losing power, it'll start right up the next morning and work as expected.

[ben7]

Re: Dell Power Supply NPS-250KB B

Might have to do with the PG (Power Good) sensing.

Not sure what else to look at...

[Wester547]

Re: Dell Power Supply NPS-250KB B

What happens if you remove AC power after powering it down normally, through Windows (unplugging it from the wall after that)? Does it turn on immediately after giving it AC power again (plugging it back in) and pressing the power button, after a few minutes, or does it exhibit the same behavior as it would in the duration of a power outage (not coming back on for a long time)?

[Pilm2013]

Re: Dell Power Supply NPS-250KB B

What happens if you remove AC power after powering it down normally, through Windows (unplugging it from the wall after that)? Does it turn on immediately after giving it AC power again (plugging it back in) and pressing the power button, after a few minutes, or does it exhibit the same behavior as it would in the duration of a power outage (not coming back on for a long time)?

I gave that a try and it worked just fine. I guess that means it had a chance to power down normally and do whatever it is it needs to do to prepare for the next power up. At this point my only guess is that during a power failure, the standby +5V part of the power supply, which should always be on when AC is applied, is, for some reason, not turning back on when AC is reapplied. I believe it's the +5V standby that powers the PC ON/OFF power button, so without that the PC has no idea the power button was pressed.

[JayArr_BC]

Re: Dell Power Supply NPS-250KB B

Correct, and it's the circuit that logs the most "on" time since it produces that 5VSB whenever plugged in so it has the electrolytic caps that fail first, Just replace all of the small caps around the small standby circuit and see if it solves.

[Wester547]

Re: Dell Power Supply NPS-250KB B

I gave that a try and it worked just fine. I guess that means it had a chance to power down normally and do whatever it is it needs to do to prepare for the next power up. At this point my only guess is that during a power failure, the standby +5V part of the power supply, which should always be on when AC is applied, is, for some reason, not turning back on when AC is reapplied. I believe it's the +5V standby that powers the PC ON/OFF power button, so without that the PC has no idea the power button was pressed.

The NPS-250KBs I've seen have 4 optocouplers. I believe one of them is a crowbar protection circuit of sorts that precludes the user from simply restarting the computer right away if something causes the power supply to shut down. Of course, this might only apply if +5VSB is still being supplied by the power supply (which it wouldn't be in the case of a power outage), but I wonder if that fourth optocoupler is configured such that it prevents you from turning on the PSU for a while in the case of a power loss... but replacing the startup capacitor - possibly an 100uF, 25V Ltec TK or maybe Taicon VT - and the +5VSB capacitors - either Rubycon YXG, Chemi-con KY, Nichicon (possibly Taicon as well) PW, or Ltec LZG, may help too - there should be a 2200uF, 10V one and a 470uF, 10V one and a PI filter coil in between. I don't think you'd have to replace them if the brands are Rubycon, Nichicon, or Chemi-con. I listed the above mentioned brands and series because they were typical choices for Newton/Delta at the time (if it has a 2002 date code - 2003 and later date codes seemed to have mostly Ltec capacitors).

As for +5VSB capacitors being the first to fail, it doesn't help that there's no air movement and it's usually a linear regulated flyback topology used (along with only a single diode) - very inefficient (not to mention some two transitor +5VSB circuits).

[mockingbird]

Re: Dell Power Supply NPS-250KB B

I have my NPS-250KB B (Date Code 0218) out of the chassis sitting on my lap right now. I inititally planned to re-cap all the Taicon in it eventhough it has not even been tested, but changed my mind and instead relegated it to the pile of 20-Pin PSUs I will probably never use.

All the secondary filtering caps are Rubycon YXG and Nichicon PW, save for one Taicon PW 16V 470uF cap in the 12V pi filter. The 5VSB pi filter is composed (If I understand correctly) of the Rubycon 10V 2200uF, the small winding, and a Nichicon 10V 470uF cap. Check if these are also Japanese on your particular unit.

Regarding the startup caps, there are only a total of 4 5mm electrolytics in the primary section. One of which is 50V 1uF, one 50V 100uF, one 25V 47uF, and one 50V 0.47uF. All are Taicon "VT" series. I can't say with utmost certainty that the second 5mm cap is 50V 100uF because eventhough I can make out the voltage on the cap just fine, the capacitance is covered with the selastic and the bottom part of the sleeve ripped off witht he selastic as I bended it to get a good look so I can only make out the 1 and 0 on the top. I *think* I can just barely see a seond zero on the top part of the sleeve though, so coupled with your suggestion I'm going to conclude it is 100uF.

Also, I desoldered the daughterboard in the primary section hoping to find a standby chip for the 5VSB. As far as I can remember, there was no standby chip on it.

[Wester547]

Re: Dell Power Supply NPS-250KB B

I believe those PSUs use a Delta-made chip known as DNA1001DL or DNA1001D, which I believe combines a chip that powers +5VSB and PWM (like UC3843) into one. 470uF, 16V sounds very small for the +12V rail. It could have been on the -12V rail. I believe the +12V rail uses two 1500uF, 16V capacitors in those PSUs. I have not been able to locate the PI filter for the +12V rail for whats it's worth, but I do see a silver shunt on that rail, though I don't think resistors pass ripple and that it's probably for OCP (but I have seen a bronze shunt in some Hipro PSU's 12V rails in place of a PI filter as well, so...).

[mockingbird]

Re: Dell Power Supply NPS-250KB B

There's a DNA1002E on daughterboard all the way to the edge of of the secondary side. Is this what you're refering to? The little daughterboard near the primary caps what looks like a SOIC-8 SMT IC. There is a single 16V 1500uF Nichicon PW with the aforemtnioned 470uF Taicon in the 12V area, and two Rubycon 6.3V 2200uF caps right next to them. There is a single blue wire connected further down the PCB somewhat near 3 Rubycon caps. One 25V 470uF and two more 6.3V 2200uF caps...

As far as I can tell, those are the only three 16V caps in this PSU so I assumed the Taicon 16V 470uF and the Nichicon 16V 1500 compose the single +12V pi filter... Maybe the -12V rail only has that single 25V 470uF 8mm Rubycon YXG as a smoothing cap.

[Wester547]

Re: Dell Power Supply NPS-250KB B

DNA1200E would be for secondary side overcurrent protection (I don't think that's causing the OP's problem, but you never know), etc. 1500uF + 470uF for the +12V rail... that's it? That seems really low coming from a PSU brand that knows how to make even underpar capacitors last ages. I have a Newton Power (NPS-250GB, Rev 01 B, datecode 0142) that also has a 470uF, 25V Rubycon YXG for -12V filtering (as well as L7912CV mounted to the other side of the heatsink and Panjit ER802 without a heatsink as its rectifier, I think), but all the other output capacitors are Ltec LZG and unfortunately the small capacitors are Ltec TK... voltage doubler has Nichicon LU though... it probably still works after 20,000 hours and 5,400 cycles of use because the Sunon sleeve bearing fan it has blows out a stunning amount of air and the heatsinks are very good and thick. It never really spits out hot or even warm air. It probably has the same configuration as yours (2200uF, 6.3V x2 for +3.3V and +5V, 2200uF + 470uF for +5VSB). Mine only has 3 optocouplers, though, and the only other PCB in the PSU is the input filtering PCB.

I find the 1500uF + 470uF configuration confusing because even older (albeit proprietary) Dell Newton 200W power supplies have more than that on the +12V rail, 2200uF, 16V Chemi-con LXJ + 330uF, 16V Rubycon YXF, though there's no PI filter or shunt to be found on that rail. Do you remember what the rectifiers were on the secondary side of the PSU, as well as the main switcher, if you checked? Also, I don't mean to derail this thread away from the OP's problem, so I'm suspect that either the 4th optocoupler is configured unfittingly or one of the small startup capacitors - a Ltec TK or Taicon VT - failed.

[mockingbird]

Re: Dell Power Supply NPS-250KB B

Don't forget these come out of Dell systems that probably never went over 150W... Yes, two 200V 560uF Nichicon LU Primaries. I've got the UG10DCT for what I assume is -12V rectification, like yours, without the heatsink. All of my small caps are Taicon I think.

The other rectifiers on the secondary are:
STPS2045CT (2 of them one on each side)
L7912CV
STPS3045CW
B20100
MBR3045CT

[Wester547]

Re: Dell Power Supply NPS-250KB B

Whoa... wait a second.... STPS2045CT x2 for the +12V rail!? That's 40A of current (28A in forward topology, but still)!!! Okay, forget the lack of output filtering, that's so overspec'd for the touted 14A (16A as of ATX12V 1.3) on the +12V rail... makes the single STPS20H100CT for the +12V rail on my out of commission Dell Hipro 250W look like nothing, though it had two 3300uF 16V Asiacon LEs and a PI filter for +12V filtering so it was far from lacking. What's the B20100 for? Is the main switcher a Toshiba part, IE 2SK2611? I also wonder if it's possible that one of the power outages did damage to one of the capacitors in the OP's PSU's voltage doubler.

[mockingbird]

Re: Dell Power Supply NPS-250KB B

The B20100's middle leg connects to ont of the legs of the 8 pin large ring coil. My guess is it has something to do with the +12V rail. Main switcher is a Fairchild FQA7N80C.

[2007blueprius]

Re: Dell Power Supply NPS-250KB B

DNA1200E would be for secondary side overcurrent protection (I don't think that's causing the OP's problem, but you never know), etc. 1500uF + 470uF for the +12V rail... that's it? That seems really low coming from a PSU brand that knows how to make even underpar capacitors last ages. I have a Newton Power (NPS-250GB, Rev 01 B, datecode 0142) that also has a 470uF, 25V Rubycon YXG for -12V filtering (as well as L7912CV mounted to the other side of the heatsink and Panjit ER802 without a heatsink as its rectifier, I think), but all the other output capacitors are Ltec LZG and unfortunately the small capacitors are Ltec TK... voltage doubler has Nichicon LU though... it probably still works after 20,000 hours and 5,400 cycles of use because the Sunon sleeve bearing fan it has blows out a stunning amount of air and the heatsinks are very good and thick. It never really spits out hot or even warm air. It probably has the same configuration as yours (2200uF, 6.3V x2 for +3.3V and +5V, 2200uF + 470uF for +5VSB). Mine only has 3 optocouplers, though, and the only other PCB in the PSU is the input filtering PCB.

I find the 1500uF + 470uF configuration confusing because even older (albeit proprietary) Dell Newton 200W power supplies have more than that on the +12V rail, 2200uF, 16V Chemi-con LXJ + 330uF, 16V Rubycon YXF, though there's no PI filter or shunt to be found on that rail. Do you remember what the rectifiers were on the secondary side of the PSU, as well as the main switcher, if you checked? Also, I don't mean to derail this thread away from the OP's problem, so I'm suspect that either the 4th optocoupler is configured unfittingly or one of the small startup capacitors - a Ltec TK or Taicon VT - failed.

I am tinkering with the exact model and quite honestly I don't know exactly what I am looking at, but I am trying to get the voltages slightly adjusted, preferably the amp output too, problem is I am kinof clueless, haven't been able to find a diagram I am guessing,
up untill I cam across this post I thought the dna1002e on the sister/daughter board might have something to go with voltage regulation, I see there are lines comming from all 3 outputs to it, the 3.3v goes through a resistor, the 5 a jumper, and 12 goes straight into pin 2 I believe it was, led me to believe that was the feedback circuit, I am realy not too good at this stuff.
either way you all semm to know a bit, can any suggest as to how I could go about adjusting the voltages and current output?
I am aiming for 3.65v at 10 amp preferably, idealy I would like to step down the 5v output to well below spec, I am looking to build a single cell lithioum charger for testing individual cells, charging and discharging, I could charge 2 cells at once to split the amps, but that would defy the whole point I am afraid, I would like to be able to monitor these bateries individually,
This may not have been my best choice for a mod but it was reman for $5 so it will do.
any input much apreciated, also, what is the small white trimpot for? tinkered with it a bit did not do anything

[Wester547]

Re: Dell Power Supply NPS-250KB B

Main switcher is a Fairchild FQA7N80C.

Mine has 2SK2746 from Toshiba as its main switcher, but like yours is rated for 7A/800V@25*C (same switcher to be found in those Dell 200W Newtons), so it probably cross references to it. A bit on the inefficient side but it does seem good for 250W so I won't complain (I have seen similar grade FETs in Bestec 250Ws, like STW7NC90Z), especially given the heatsink it's soldered to and the amount of air the Sunon sleeve bearing fan pushes. Also happens to have the same rectifier for the -12V rail as yours, I had mistakenly said Panjit because I was looking at the bridge rectifier at the time (GBU 8K from Panjit Semiconductor without a heatsink). I think B20100 has more to do with the +5V rail. The +12V rail schottky wise is too overspec'd to need anything more. I wouldn't be surprised if B20100 was freewheeling for +5V and MBR3045CT was rectifying.

what looks like a SOIC-8 SMT IC

Is that UC3843B?

This may not have been my best choice for a mod but it was reman for $5 so it will do.
any input much apreciated

Do you mean that it's working but you're trying to improve upon the voltage regulation? I wouldn't be worried about the voltage regulation on any of the rails unless it's out of spec. Unless you need to because you're using the PSU for other means (if I read correctly). The amp output should be as the label says - 18A on +3.3V, 22A on +5V, 14A on +12V, 1A on -12V, 2A on +5VSB, 170W for +3.3V and +5V combined maximum, 250W maximum.

[2007blueprius]

Re: Dell Power Supply NPS-250KB B

Do you mean that it's working but you're trying to improve upon the voltage regulation? I wouldn't be worried about the voltage regulation on any of the rails unless it's out of spec. Unless you need to because you're using the PSU for other means (if I read correctly). The amp output should be as the label says - 18A on +3.3V, 22A on +5V, 14A on +12V, 1A on -12V, 2A on +5VSB, 170W for +3.3V and +5V combined maximum, 250W maximum.[/QUOTE]

not necesarrely improve, more like change, and yess you guessed right, the supply was working perfectly fine when I got it, still does, well my knowledge is limited but it works fine best I can tell, however I did buy if for different purposes, I have these lithium batterypacks with some issues, I need to weed some bad cells, in order to do that I need to charge and discharge them individualy to see how they behave,
longstorryshort without going too far off topic, I need 3.65v at 10amps, now I know batteries are a bit funny and draw whatever they need so I am a bit in the dark there but for now I would be happy if I figured out how to tamper with the feedback, and lie to it thet the 3.3v is low so that it will give me 3.65 in return or step down the 5v ( the dummy load is on that so not sure how that would interact, not in a desktop unit), I traced the back of the board and I see that there is a line from the 3.3v output going through a resistor to the sisterboard, figured a trimpot might do the trick as I have measured the voltage across that resistor, 3.35 goes in, 1.xx comes out, I have seen some other mod examples for atx but they have different chips, can't find a diagram on this one nor a datasheet for most these chips.

once I figure that out, I would like to limit the output to 10 amps, preferably make it adjustable, thou all this is well beyound my skill, therefore I need some help

[2007blueprius]

Re: Dell Power Supply NPS-250KB B

bump.

I know there is a wealth of knowledge among your circles, and what I am trying to accomplish is a bit unconventional to what you boys do for fun, I could use some insight here before I do something silly and blow this thing up, I took some pics and poster them here http://s1365.photobucket.com/user/20...?sort=3&page=1
I am not sure what the chip on the main board does but I measured some voltages and this is what I got:
voltage readings, starting at pin #1 top left, they changed a bit by the time I wrote them all down, I am sure some of my outputs did as well, I shall doublecheck on that.
1 = .52v
2 = 11.85
3 = 11.84
if I know better I would say this is where the 12v is beeing regulated
4 = 17.24 sounds like v+
5 = 2.49
6 = .52
7 = 16.23
I would say that is simmilar to the adjustable voltage circuit I seen on some of the atx chips
8 = .5
9 = 4.95
10 = 4.94
this looks like the 5v regulating circuit
11 nothing, must be ground, I'll doublecheck
12 = 3.27
13 = 3.28
14 = .6
it is labeled lm324n, found this for a datasheet not even sure if it is the right thing https://cdn.badcaps-static.com/pdfs/...c4e67ba06d.pdf
for a minute I thought it was controlling the output voltages, now I look on the sister board I have the dna1002e it has all sorts of voltages going to it and I see some paths on the main board from the outputs to it I am assuming this is the feedback circuit, goes to show how much I know,
Honestly I am a gearhead, electronics for the most part I regarded as too complicated and disposable, I do know the basics, thou I never had much luck with them.
I know it can be done as I have seen some examples on youtube:
http://www.youtube.com/watch?v=rpUk9hpFQL4 this guy done a few but he knows his stuff and had schematics, I have neither
also: http://www.youtube.com/watch?v=rkApRg5VWnc another good example
now both aimed at the 12v circuit, I need 3.65v, no need for a range from 2v to 30v, one thing I am unclear of:
I read that some of these power supplies have individual feedback for each output, the examples I seen had 1 pot controlling all outputs, also not sure if the output amperage can be adjusted, I am sure that there is a circuit that regulates it, and the mods are probl as simple as replacing 2 resistors with 2 trim pots, 1 for voltage one for amps, but which ones? and what size 2k seems to be the norm.
I also read about the opto couplers, I read somewhere this unit has 4, I know what they are, read about it, no clue what they look like, would help identifying the feedback circuit and the resistor to be replaced with a pot, I'll have a look again when I get home.
if any of you know more than I do ( and you must since I don't know much ), a schematic, advice pointers, I'll take anything, by all means please enlighten me, I read the theory, cant really place it on the board.
also there is a small white pot, it was glued fix, I tinkered with it, it did noy change anything that I was monitoring, what does it do?
all thoughts are appreciated.
George

[2007blueprius]

Re: Dell Power Supply NPS-250KB B

oh and I touched some of those surface mount resisters on the daughter board with a screwdriver, the draw from the extra metal was large enough to cause a voltage drop, to 8v on the 12v output, unlike other silly experiments, it did not trip the psu, as soon as the screwdriver was removed voltage went right back up. therefore I feel I am on the right track but again this is not science, pocking things and see what happens.
there is also a meanwell supply mod that I read about that inspired me to tinker with this
http://endless-sphere.com/forums/vie...l+mods#p653187
different animal but same principal, if only I knew which resister regulates the amp output on this newton unit, should be able to accomplish the same.
there are plenty examples converting these units to a bench supply but not so much about making the output variable.