Why are switching supplies over voltage protection circuits so high

[kc3ase]

I've noticed that switching power supply circuits generally use a ~140% over voltage protection circuit. If you have a +/- 5% supply why is the over voltage protection so high?

I've run into a case where I have to add circuitry to protect against this case because it will ruin our day why isn't it something like 10-15%?

Cheers,
John

[Kleinstein]

Pure switched mode regulators are prone to overshoot when subkect to more problematic loads (e.g.  low ESR capacitor in the 1-10 mF range). So an overvoltage protection set to low might trip during turn on or when load is connected / disconnected. Depending on the ciruit you may need to turn off the supply or even replace a fuse - so tripping to often is not so good.

The better supplies use a linear regulatior after the switched mode and can thus provide better dynamic regulation. Some of them also provide an individual adjustably overvoltage protection. So the user can choose the level - 5.1 V for a senstive circuit or 10 V for just a fan.

[T3sl4co1l]

Overvoltage where?  What part are you looking at?

If you mean output control, I wasn't aware most if any even considered it.  If the output voltage rises above setpoint, PWM just shuts off entirely, and the output is able to float.  It's not in the controller's design nor ability to do anything more than that.

If you mean input, they're generally made to work up to rated voltage, and usually somewhat beyond that, but not by much; so you'd want to operate with a steady-state maximum of perhaps 70% of ratings, with a TVS to clamp transient maximum inputs to within 100% of rating.

If you're looking at a type actually with UVLO and OVLO, I don't know, all the load controllers I've seen with that type of functionality do it accurately against an internal reference (low-%).  Example: automotive/telecoms power input/load controllers (polarity protection, undervoltage/startup control, overvoltage (load dump) limiting/shutdown, and often, current limiting as well).

Tim

[DanielS]

I've run into a case where I have to add circuitry to protect against this case because it will ruin our day why isn't it something like 10-15%?

If you need more precise voltage limits, use a point-of-load regulator tuned to your specific requirements which takes into account the ~40% tolerance on the bulk supply's OVP circuit.

You do not want your main supply to shut down on every minor event that might cause the output to momentarily go out of specs and the main supply needs to cope with 0-200% (brownout to surge/swell) input voltage variation, 0-100% output load variation, switching noise, ripple from output load, etc. Then you have the nominal tolerance on the output voltage plus the accuracy tolerance on the OVP threshold as well.

The OVP is there to cut off power in case of output voltage regulation failure to prevent the loads from turning into fire hazards through sustained over-voltage.

[saturation]

Many secondary PSU use the ITIC or CBEMA curve as a design guide, which accounts for a lot worse conditions than +/- 5% variation from mains, it also includes worse case frequency variations.  In addition, the curve shows how long surge, sag, and transient protection should last.  Red zone breakdown are need dependent, if your needs are tighter, then you need OVP must be set lower than 140%.



Close up of the steady state secondary output conditions:



Primary line variations can vary more than +/-5% and still be ANSI standard power in the USA.

http://www.powerqualityworld.com/2011/04/ansi-c84-1-voltage-ratings-60-hertz.html

Range A nominal, Range B under high demand conditions.