Comp Power Supply - Rosewill 1200W - Blown Component - Need help Identifying

[lvshawnlv]

Yes it is an NTC inrush limiter, a few ohms at cold start and then they run hot (80-100C) and their resistance drops to a fraction of an ohm, they cheap ones do die quite often,

Go to element-14 or similar and buy one rated for the current at 110Vac, say 15 amps min, clean up the board and fit the new ntc and you're away...!

So I'm confused a little bit.. So when i'm being told to short something, where those two legs are sticking out, then it means fuse them together ? and solder them soild? I figured the one I bought was rated for up to 240 so it would be able to handle it better -- even though i dont think I am going to run it off 240, only 120 though...


So Element 14 ... Buy a new NTC from them and attach that to where the two legs are... ??? In the pictures, Those two gold legs where covered by the NTC when it was a whole... so now that its crumbled apart you see the legs now... the NTC that burned up i left by the two legs so that it would show where it was at one time..

Shawn

- - - Updated - - -

Can i run this unit without that NTC attached ? and if so will I experience any issues with it to cause a fire? If not ill just jump the two legs together and call it a day. When this unit is on ... it runs for about 1 minute before any load is pulled through it... the load isnt on it 100 percent from switch on to running...



Shawn

 

[Easy peasy]

yes you can short out this component and the psu will run, there is a slight risk of blowing the fuse...

 

[treez]

When this unit is on ... it runs for about 1 minute before any load is pulled through it... the load isnt on it 100 percent from switch on to running...

it doesn't make much difference in some ways if the load isn't on when plugged in, the capacitance at the input (or after the pfc output), will still take a large charging current spike, especially if you inadvertently plug it in at the instant of mains peak voltage, and as Easy Peasy says, you may blow the fuse...or worse, get a load of overvoltage ringing which blows something......maybe eventually the mov.

 

[c_mitra]

I believe all computer power supplies do not connect to the load till the output voltages stabilize - if I remember correctly, in the original IBM PS, the PG signal takes more than 100ms to come. The initial rush current can blow the fuse but most fuses will take overload for 5 ms (quarter cycle) and most circuit breakers are even slower. The risk of blowing the mains fuse is rather low.

 

[Easy peasy]

why is the ntc there in the first place then?

 

[c_mitra]

why is the ntc there in the first place then?

Good question. Possible answer is:

It is a good design practice.

However, as jonnyguru has pointed out in one of the reviews (see post #15), the NTC was shorted on the solder side with a small wire in the review unit he got. He did the tests with it (in the circuit) and without it (shorted out) and found only marginal loss of efficiency (**broken link removed**) with the NTC in place.

The last sentence reads: "But the best thing to do is just implement the NTC shorting relay so many other Gold designs use. It's less kludgey looking, at the very least, and it won't have techs like me discovering the bypass and then wondering what kind of logic installs a perfectly good NTC thermistor and then hardwires a bypass jumper for it so that it's never used"

Interesting are the ways the engineers' minds work.

 

[treez]

I've actually gotten though quite a few computer power supplies which clearly didn't have inrush limitation, or just didn't have enough of it...because you got a mini explosion every time you plug it in, then after a year or so, its a gonner and I have to go and buy another......I find I realise when the computer starts being slow because its running always off an insufficiently charged battery becuase the psu started struggling after the inrush problem.....inrush into a pfc output capacitor, through the pfc choke, without a diode across it to the output cap can ring very high in voltage, I bet they just turf them out and don't bother that they wont last long....one fails, sell another.

 

[c_mitra]

when the computer starts being slow because its running always off an insufficiently charged battery becuase the psu started struggling after the inrush problem

1. Computers do not run slow because of an insufficiently charged battery.

2. In fact, desktop computers do not have a battery to worry about (except the RTC battery)- they run off the mains (just like the current one we are talking about)

3. Laptops and other cousins run directly off the battery, and these batteries last at least 3 years (5 is more common).

4. The cheap SMPS cost around USD 10; the good ones around USD 50 and I think (personally) that going for an SMPS that lasts 10 years and costs USD 100 is simply a waste of money.

5. No modern computer SMPS starts under load (on the output side), the power is turned on only after power good is ready.

6. The supply side (power input) may be struggling with various problems but the delivery side (output connections to the MB) does not bother about these problems.

7. PFC is a necessary evil.

8. SMPS does die often; but they rarely take the MB with them. Anybody who can hold a screwdriver can replace a SMPS.

9. I have seen many SMPSs die a natural death; they invariably dies soon after the warranty (mostly one year) is over. This I call a design feature. What do you do with a 10 year young SMPS that is powering a 10 year old MB?

Computer problems are rarely (these days) due to SMPS problems. Thank China that we can buy a SMPS for USD 10!!

 

[lvshawnlv]

yes you can short out this component and the psu will run, there is a slight risk of blowing the fuse...

SO what am I going to do if I short and the PSU does not power on and, does not blow a fuse?

 

[treez]

yours is a 1400w output psu running off 115vac mains
So lets say your rms input current is 14.3 amps.
That's a heck of a job for an ntc, normally that would be handled by a power resistor and a relay to switch out the power resistor.......I mean, say you pick an ntc with an rmin of 0.05 ohms, then that would still dissipate 10w.
Can you see if theirs a relay in their, switching out that ntc...if not then it might be a problem in the psu.

 

[BradtheRad]

For a quick test in similar situations, I have thought it would be great to put a light bulb in place of a fuse (or power-carrying component). Immediately it would give me an idea how much current the equipment is trying to draw. The bulb becomes a fuse of sorts, so the equipment wouldn't blow up. Nor would the bulb blow up, being designed to carry full mains AC.

It would need to be low enough resistance so the equipment would start to operate. But such a light bulb would need to be several hundred watts. Not an ordinary lamp bulb.

If you can dig up some sort of instant current monitor, that might give you early warning, whether the power supply is trying to draw overmuch current, with risk of blowing up.

 

[FlapJack]

1288 Total Watts consumed if running full machine. But I'm running minus one hashboard so 1288 Minus 161 is the total wattage draw off that power supply.

What is a hashboard?