Feeding load with two 220V AC power supplies to reach uninterrupted power supply.

[EEngineer93]

As shown in the diagram, I am trying to connect two power sources to load. The battery source is for emergency in the case of power outage of the wall source. I want the switching to be instantaneous with no interruption of power.

What is the best solution to this problem? Is there a board out there that I could use? There must be no interruption of power.

Thank you!

 

[KlausST]

Hi,

"No interruption" is not possible this way.
You have to define what is the maximum "off time" your load can handle.
Since it is AC there repeating are times with low or even no supplied energy (during each zero cross). Thus AC loads are uncritical with "OFF time" os some milliseconds.

For real "no interruption" you need the power to be continously supplied via the inverter. No direct mains connection to the load.

Klaus

 

[Easy peasy]

It seems unlikely your inverter will be in phase with the mains - the best you can hope for is the change over time of a relay, but then if you are switching, say, just at the conclusion of a pos half cycle of the mains and you happen to switch to a pos half cycle of the inverter - your down stream equipment has to be able to handle this - i.e. can't be a transformer ( as they don't handle 20mS of the same polarity very well ) ...

better to have the mains go thru a charger to the 24V battery - then, assuming the inverter is 100% reliable - you will have seamless power,

if you can synch the inv to the mains - you still have to decide what tolerance away from "good" mains - and for how long - before you let the relay change over - so you can not get "seamless" changeover - you have detect time & changeover time ...

 

[betwixt]

I think the only practical solution is to keep the inverter oscillator running continuously and locked in phase to the wall AC supply and with sufficient locking delay that it can outlive the outage. The output stage of the inverter can switch on almost instantly if the phase is already locked.
You would also have to ensure that the phase was re-locked when AC was restored before switching back.

Brian.

 

[KlausST]

Hi,

This only solves the "clean" switchover problem.
But not the "zero timing" requirement.

What still needs to be solved: how to detect a missing AC voltage within no time.

Klaus

 

[betwixt]

Interesting problem - if the phase is locked as in post #4 I suppose a comparator sensing a locked sine wave oscillator and a scaled version of the line AC would immediately notice any difference between them without having to monitor for the presence of line voltage. Just an idea ... I've never tried it.

Brian.

 

[KlausST]

Hi,

Misalignent in phase will give positive as well as negative comparator input, misalignent in voltage, too, distortion in the one or the other voltage, too.
Difficult to safely and fastly detect loss of mains.

Klaus

 

[Easy peasy]

this is a perennial problem for UPS designers - how far out is far out enough to trigger change-over to inverter - phase lock loops are commonly used to make a sine wave in sync with the mains ( but scaled down ) and as soon as the difference is "x" for "y" mS then changeover for 1 minute minimum - and keep monitoring the mains until it is "good" for more than a minute before you consider switching back.