Consideration for robust AC input protection?

[Red_Micro]

The circuit attached is a simple linear power supply: mains->MOV->resistor->cap->transformer->bridge

Let's say the mains is the typical 120VAC. What are good design practices for choosing components for robust protection at the input side? For example, I have seen cases where resistor R1 goes open, maybe because not handling surges or MOV being over dimensioned? I read that metal film resistors are not good handling surges. Is this correct?

[T3sl4co1l]

Nothing is necessary; the transformer's series impedance (leakage), saturation if applicable (not really for mains transformers), and output capacitors, clamp IEC 61000-4-5 level spikes.

You can put an MOV on if you like, after the fuse of course, to protect against freakishly large surges.  Say if you're in a mountaintop installation frequently subject to direct or nearby hits?  (It should be a big one then, to give reasonable lifetime under such service.)

Tim

[Konkedout]

I would agree that the transformer provides a lot of protection, and you may not need any more unless your need for protection is extreme.

If you do require an inrush limiter for some reason, go with a bulk resistance device (large resistor mass) such as an NTC inrush limiter (better) or a ceramic composition resistor.  If you do have high inrush, then the higher thermal mass of a bulk resistance (instead of a film type resistor) can quickly absorb more energy than a film resistor without burning out.  Some wirewound resistors are also specified for high peak surge current. 

If you have a large transformer with high quality steel laminations and minimum winding turns and winding resistance, you can encounter core saturation at startup (or maybe as a result of other mains disturbances) and an inrush limiter of some sort might be indicated.  I have seen this happen a couple of times in....45 years of power supply engineering.  Most recently with a 1 KVA toroid mains transformer.  I fixed that problem with an NTC inrush limiter which was bypassed by a time delay relay after power supply startup.  This was a one-off application for a test circuit.

[bdunham7]

I think it would be rare application that would require--or even benefit from-- a MOV and I'm not sure what the capacitor is going to do.  I'd add an appropriate fuse and use R1 to limit inrush and fault current as needed.  Knowing the size of this PSU would be helpful, as well as whether or not efficiency is important.  If you want a very robust surge protection, use bridge rectifier rated well above the operating voltage and then some sort of crowbar after that, like a TVS.  Between the TVS and the filter caps, nothing should get by before the fuse blows.

[T3sl4co1l]

If you have a large transformer with high quality steel laminations and minimum winding turns and winding resistance, you can encounter core saturation at startup (or maybe as a result of other mains disturbances) and an inrush limiter of some sort might be indicated.  I have seen this happen a couple of times in....45 years of power supply engineering.  Most recently with a 1 KVA toroid mains transformer.  I fixed that problem with an NTC inrush limiter which was bypassed by a time delay relay after power supply startup.  This was a one-off application for a test circuit.

Yes, it's particularly pronounced for toroids, which saturate quite sharply.

Transformer impedance can be high enough that inrush isn't much of a problem ("impedance limited" and most bank-wound / split-bobbin types), but mind that these also come with the steady-state consequence, i.e. high output impedance (poor regulation).  If you're powering a switching regulator, this doesn't matter much (just design for the full range of Vin between max line + min load vs. min line + max load), but this is particularly troublesome for linear supplies where all that overhead must be just burned off.

If you want a very robust surge protection, use bridge rectifier rated well above the operating voltage and then some sort of crowbar after that, like a TVS.  Between the TVS and the filter caps, nothing should get by before the fuse blows.

Meh, rectifiers are quite robust -- a 6A or so rated part is adequate to handle IEC surge, and a like one of those quick-connect FWBs (say GBPC3508) will handle even slow-blow fuses; maybe not at industrial level (100kA+ fault current), but I've had 3/4ths of one survive faulting a residential 240V 50A circuit for example.  That scenario was probably beyond the guaranteed fusing point and the three survivors were luck, but to say they can handle regular fuses, heck, sure.


Voltage doesn't matter at all, as long as surge current * cap ESR doesn't blow things up before the crowbar fires.

Crowbarring that, actually, is a strong challenge in itself, as that filter cap needs to be discharged on top of the surge (and mains itself until the fuse opens).  Unless you add another FWB to separate crowbar from cap, which, if you need this much protection I guess you could.

It's also the kind of thing that, if you really need this much protection, can you just sacrifice the PSU itself, as long as the precious (sensitive, bespoke, unserviceable, irreplaceable..) stuff being powered, doesn't get overvolted?  Secondary side crowbar is a far sight easier!

Tim