SMPS blowing fuses help!!

[2dye4]

We buy a SMPS at around 500 W capacity.

Driven by 120 VAC.
Line and Neut both have a 10A 3ag fuse with i^2*t=137
The round trip DC resistance of the input circuit from the line terminals to the AC input to the rectifier is 0.073 ohms.

There is a 140 VAC MOV after the fuses.

THere is a negative temp coef resistor that limits inrush but is switched out by a relay after cap charging.

The bulk filter caps are 3000 uf of low ESR Electrolytic.

The fuse failures occur where conditions cannot be monitored.

I think that with 3000 uF of capacitance and only .073 ohms of resistance the circuit is suceptable to low impedance line transients but i cannot figure a way to prove it.

Do you agree or have another idea??

 

[VEBill]

Is there anything in the environment making life difficult for the fuses? Vibration, extremely high temperatures, over-voltage events?

Those fuses are fast blow I think (?). Maybe switch to slow blow fuses.

Another old trick is to bypass the fuses, stand well back, and then wait to see what explodes. ;-)

 

[2dye4]

Unfortunately they are field failures and we cannot see the failure during testing.

The environment should not contain vibration or high temps.

The 137 A^2*T is suppose to be a slow blow in this series.

I suspect nothing would explode if they were bypassed.

I wonder if charging the large capacitance 3000 uF (low ESR ) from transients could do it??

 

[2dye4]

Forgot to add something that is very important.

this is a custom designed supply for us and has never previously been field tested.

Thanks

 

[IRstuff]

Did you get the failed units back?

What limits the load surge? Your 500W supply draws 4A in steady state. Many systems surge 10x the steady state draw on power up.

TTFN

FAQ731-376

 

[2dye4]

Ir

We got the units. They worked fine after fuse replacement.

There is a neg temp coef resistor in the circuit during power up. It gets bypassed by a relay after the bulk caps are charged.

My main question is whether the heavy filter capacitor with only minimal impedance to the AC line (0.073 ohms resistive) leaves it vulnerable to transients. Transients must charge that large capacitance through the fuses and it may be possible that relatively minor transients, say 200 additional volts for 100 us with low source impedance could trip these fuses.

I want the manufacturer to include something to limit the surge current that must charge the capacitors, 200 uH inductor or 1/4 ohm resistor.

Of course they will have none of this. However i an not really sure i am right, need second and third and fourth .....opinions.

 

[IRstuff]

You didn't bother answering the question. How do you know it's not the load side? Which is way more likely than input transients.

TTFN

FAQ731-376

 

[2dye4]

I don't know what limits the load surge. I would assume the supply has an internal current limiting circuit.

Based on the load we apply i believe I can rule out the load causing the issue.

And unfortunately we didn't spec any transient requirements.
Can anybody name a spec that we should have required ??

 

[BobM3]

Could you try a higher rated (voltage) MOV? Maybe the transients are producing a high enough voltage that your current MOV is conducting too much.

 

[VEBill]

"...negative temp coef resistor that limits inrush but is switched out by a relay after cap charging."

Does this circuit work with short power interruptions?

The reason I ask is that typically they don't. The designer makes the assumption that the power is off then on. And they often fail to account for the on-off-on sequence.

 

[2dye4]

VE1BLL
This NTC switch out was a problem. It has been changed to place it back in circuit as soon as power loss is detected.

Does anyone think this circuit is suceptable to blowing the fuses due to line transients, as they must charge the 3.0 mf caps??

Is there any basis to fault the design for not having an intentional full time current limiting impedance in the charge path of the caps. Inductor or low ohm resistor.

 

[zappedagain]

I just stumbled across this current limiting article in my 'library' so I thought I'd post it. A NTC seems much simpler...

http://www.edn.com/article/CA6486029.html?spacedesc=designideas&industryid=44217

John D

 

[westom]

The NTC (thermistor) is installed so that you do not have a massive power on current surge. Why is it not doing its job?

Why do you want to add resistors or an inductor to do what the neg coeffi thermal resistor (NTC) is supposed to be doing?

 

[2dye4]

Westom

The NTC is switched out of the circuit to avoid the loss after the bulk caps have charged up.

My concern is that the bulk caps are charged through the fuses can draw enough current in a power line voltage spike to blow the fuse.

 

[westom]

This sentence still applies and is unanswered:
The NTC (thermistor) is installed so that you do not have a massive power on current surge. Why is it not doing its job?

Bulk caps are charged through the NTC so that fuses never blow and so that the power supply is even more robust.

Power line spike long after power up? These spikes are how big and how long? Put those numbers into the chart. Maybe one hundred consecutive spikes would occur before the fuse even considered blowing.

Now, using an oscilloscope, what exactly is the current draw of that power supply as power is applied and during constant operation? What exactly is the current of these 'assumed' spikes? And what are in-line fuse parameters? For better answers, provide specific numbers from test equipment. If some spike exists, then you already have those numbers. Where are those numbers?